User login
New targeted treatments are major advances for HER2-positive breast cancer
Before 2001, HER2/neu-positive breast cancer (HER2+) was one of the most dreaded diagnoses a woman could face, as treatment was largely ineffective. The discovery of trastuzumab changed that dramatically.
Over the next 20 years, two additional HER2-targeted therapies – lapatinib and trastuzumab emtansine (TDM-1) – earned approval from the Food and Drug Administration for selected patients with early and late HER2+ breast cancer.
Since 2019, four additional HER2-targeted therapies have been approved by the FDA for HER2+ metastatic breast cancer (MBC), changing the treatment paradigm for those patients substantially.
The new agents are especially useful in certain patient populations. The agents offer the promise of improved survival for patients with recurrent metastatic disease and the potential for further reductions in relapse rates in earlier settings.
Trastuzumab deruxtecan
Trastuzumab deruxtecan is an antibody-drug conjugate that links three components: an anti-HER2 monoclonal antibody, a highly potent topoisomerase I inhibitor payload, and a tetrapeptide-based cleavable linker.
Trastuzumab deruxtecan has a high drug-to-antibody ratio. A membrane-permeable payload offers the potential for activity against adjacent HER2-negative cells in heterogeneous tumors. It has a long half-life (6 days).
Trastuzumab deruxtecan received accelerated approval from the FDA in December 2019 to treat patients with HER2+ MBC who have received two or more prior HER2-targeted regimens, based on the results of the DESTINY-Breast 01 trial.
DESTINY-Breast 01 trial
In the phase 2 DESTINY-Breast 01 trial, 184 patients with a median of six previous treatments received trastuzumab deruxtecan (5.4 mg/kg) intravenously every 21 days. There were 24 patients with treated, asymptomatic brain metastases who participated. Patients with untreated or symptomatic brain metastases were excluded.
Overall, a response to therapy was reported in 112 patients (60.9%), with 6.0% complete and 54.9% partial responses. Most of the patients for whom both baseline and postbaseline data were available had a reduction in tumor size.
The median time until response was 1.6 months, an interval that corresponded to the time until the first scheduled imaging. Three patients (1.6%) had progressive disease, and two patients (1.1%) could not be evaluated.
The median duration of follow-up was 11.1 months, and the median response duration was 14.8 months.
The median progression-free survival (PFS) was 16.4 months, and the median overall survival (OS) was not reached. The median PFS in the patients with brain involvement was 18.1 months.
The most common adverse events of grade 3 or higher were a decreased neutrophil count (20.7%), anemia (8.7%), and nausea (7.6%). Most concerning was that trastuzumab deruxtecan was associated with interstitial lung disease in 13.6% of patients.
Tucatinib
Tucatinib is an oral, highly selective HER2 tyrosine kinase inhibitor (TKI). In April 2020, it was approved by the FDA, in combination with trastuzumab and capecitabine, for adult patients with advanced unresectable or metastatic HER2+ breast cancer who have received one or more prior anti-HER2–based regimens for MBC. The approval included patients with brain metastases.
The recommended tucatinib dose is 300 mg orally twice a day in combination with trastuzumab (at the standard dose) and capecitabine (1,000 mg/m2 given orally twice daily on days 1-14) on a 21-day cycle, until disease progression or unacceptable toxicity.
HER2CLIMB trial
The study that led to the approval of tucatinib was the HER2CLIMB trial. The trial enrolled 612 HER2+ MBC patients who had prior treatment with trastuzumab, pertuzumab, and T-DM1. Patients had received a median of 4 (range, 2-17) prior lines of HER2-targeted therapy.
The patients were randomized 2:1 to receive trastuzumab plus capecitabine and either tucatinib or an identical placebo twice daily.
The primary endpoint was PFS, evaluated in the initial 480 randomized patients. The median PFS was 7.8 months in the tucatinib arm and 5.6 months in the control arm (hazard ratio, 0.54; 95% confidence interval, 0.42-0.71; P < .001).
The confirmed overall response rate for patients with measurable disease was 40.6% in the tucatinib arm and 22.8% in the control arm (P = .001). The proportion of patients still in response at 12 months was 33.1% and 12.3%, respectively.
The median OS was 21.9 months in the tucatinib arm and 17.4 months in the placebo arm (HR, 0.66; 95% CI, 0.50-0.88; P = .005). At 24 months, 44.9% and 26.6% of patients, respectively, were still alive.
The most common grade 3 or higher adverse events (in the tucatinib and placebo arms, respectively) were palmar-plantar erythrodysesthesia syndrome (13.1% vs. 9.1%), diarrhea (12.9% vs. 8.6%), elevations in ALT and AST (approximately 5% vs. 0.5% for each), and fatigue (4.7% vs. 4.1%).
Tucatinib in patients with brain involvement
A unique feature of the HER2CLIMB study was that patients with MBC and untreated, symptomatic brain metastases were eligible. Patients with active, untreated central nervous system disease are excluded from virtually all other trials, especially drug-approval trials.
There were 291 patients with brain metastases in HER2CLIMB, 198 (48%) in the tucatinib arm and 93 (46%) in the control arm.
The risk of intracranial progression or death was reduced by 68% in the tucatinib arm (HR, 0.32; 95% CI, 0.22 to 0.48; P < .0001).
The 1-year CNS-PFS rate was 40.2% in the tucatinib arm and 0% in the placebo arm. The median duration of CNS-PFS was 9.9 months and 4.2 months, respectively.
The risk of death was reduced by 42% in the tucatinib arm (HR, 0.58; 95% CI, 0.40-0.85; P = .005). The median OS was 18.1 months and 12.0 months, respectively.
There were more objective responses in the brain with tucatinib (47.3%) than with placebo (20.0%; P = .03). The median duration of response was 6.8 months and 3.0 months, respectively.
Particularly because of its CNS activity and lack of serious, long-term toxicity, tucatinib combination therapy represents an attractive new option for patients with HER2+ MBC.
Neratinib
Neratinib is an irreversible pan-HER TKI that was approved by the FDA in July 2017 for extended adjuvant therapy in patients with early-stage HER2+ breast cancer, following the use of trastuzumab-based therapy.
Long-term results of the ExteNet study led to the approval for use as extended adjuvant therapy.
In February 2020, neratinib was FDA approved in combination with capecitabine for patients with HER2+ MBC after two or more prior anti-HER2–based regimens. The more recent FDA approval was based on results of the NALA trial.
NALA trial
The phase 3 NALA trial included 621 patients with HER2+ MBC who had received at least two prior anti-HER2 based regimens.
Patients were randomized 1:1 to receive neratinib at 240 mg orally once daily on days 1-21 with capecitabine at 750 mg/m2 orally twice daily on days 1-14 or lapatinib at 1,250 mg orally once daily on days 1-21 with capecitabine at 1,000 mg/m2 orally twice daily on days 1-14 for each 21-day cycle. Patients were treated until disease progression or unacceptable toxicity.
The primary endpoints were PFS and OS by blinded, independent, central review.
The median PFS was 5.6 months in the neratinib arm and 5.5 months in the lapatinib arm (HR, 0.76; 95% CI, 0.63-0.93; P = .0059). The PFS rate at 12 months was 28.8% and 14.8%, respectively.
The median OS was 21.0 months in the neratinib arm and 18.7 months in the lapatinib arm (HR, 0.88; 95% CI, 0.72-1.07; P = .2086). The ORR was 32.8% and 26.7%, respectively. The median response duration was 8.5 months and 5.6 months, respectively.
Fewer interventions for CNS disease were required in the neratinib arm than in the lapatinib arm (cumulative incidence, 22.8% vs. 29.2%; P = .043).
The most frequently reported grade 3-4 adverse reactions for the neratinib combination were diarrhea, nausea, vomiting, fatigue, and decreased appetite.
Grade 3 diarrhea occurred in 24.4% of those in the neratinib arm and 12.5% of those in the lapatinib arm. Antidiarrheal medication was used by 98.3% of patients receiving neratinib and 62.1% of patients receiving lapatinib.
Margetuximab-cmkb
Margetuximab is a chimeric Fc-engineered anti-HER2 monoclonal antibody that targets the same epitope as trastuzumab and exerts similar antiproliferative effects.
Compared with trastuzumab, margetuximab has higher affinity for both 158V (high-binding) and 158F (low-binding) alleles of the activating Fc receptor, CD16A. As a result, margetuximab enhances innate immunity, including CD16A-mediated antibody-dependent cellular cytotoxicity, more effectively than trastuzumab. Margetuximab also potentiates adaptive immunity, including enhanced clonality of the T-cell repertoire and induction of HER2-specific T- and B-cell responses.
In December 2020, margetuximab, in combination with chemotherapy, was approved by the FDA for patients with HER2+ MBC after two or more prior anti-HER2 regimens, at least one of which was for metastatic disease. The approved dose is 15 mg/kg IV every 3 weeks.
The study that led to margetuximab’s approval was the phase 3 SOPHIA trial.
SOPHIA trial
SOPHIA was a randomized trial of 536 patients with HER2+ MBC who had received prior treatment with other anti-HER2 therapies, including one to three lines of therapy for MBC.
Patients were randomly assigned 1:1 to receive margetuximab plus chemotherapy or trastuzumab plus chemotherapy. Assignment was stratified by chemotherapy choice (capecitabine, eribulin, gemcitabine, or vinorelbine), the number of previous lines of therapy for MBC, and disease extent.
Co–primary outcome measures were PFS by blinded, independent, central review and OS.
At the second interim analysis, the median PFS was 5.8 months in the margetuximab arm and 4.9 months in the trastuzumab arm (HR, 0.76; 95% CI, 0.59-0.98; P = .033). Results were more impressive in patients with CD16A genotypes containing a 158F allele. In this group, the median PFS was 6.9 months with margetuximab and 5.1 months with trastuzumab (HR, 0.68, 95% CI, 0.52-0.90; P = .005).
At the second interim analysis, the median OS was 21.6 months in the margetuximab arm and 19.8 months in the trastuzumab arm (HR, 0.89; 95% CI, 0.69-1.13; P = .33).
Subgroup data showed no differences in OS between the two arms for any subgroup except HER2+ MBC patients with an IHC score of 2 or higher. This is consistent with the postulated mechanism of action of margetuximab.
The confirmed ORR was 25% in the margetuximab arm and 14% in the trastuzumab arm, with similar durations of response between the study arms.
The most common adverse events in both arms (≥20%), regardless of causality, were fatigue, nausea, diarrhea, and neutropenia. Vomiting was common in the margetuximab arm, and anemia was common in the trastuzumab arm.
Grade 3 or higher adverse events occurred in 53.8% of patients receiving margetuximab and 52.6% of those receiving trastuzumab.
In view of margetuximab’s modest benefits in the SOPHIA trial, the ultimate role for margetuximab in HER2+ MBC may be restricted to patients with the CD16A-158F allele. A neoadjuvant trial is planned in that population.
Take-home messages
There are legitimate arguments regarding whether curing MBC is within reach for certain patient subsets, but there is no argument about whether the outlook for patients with HER2+ MBC has improved dramatically in recent years; it has.
The approval of four unique, new agents for the treatment of women with HER2+ MBC in relapse provides further improvements in outcome for these patients and distinctly different opportunities for tailoring treatment to the special circumstances of each patient (e.g., whether brain metastases are present, desire for oral therapy, comorbidities, experience with prior chemotherapy, etc).
When considered along with the potential for incorporating these drugs in earlier settings in well-designed clinical trials, these new drugs offer great promise to a group of patients who faced a dismal outcome just 2 decades ago.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Before 2001, HER2/neu-positive breast cancer (HER2+) was one of the most dreaded diagnoses a woman could face, as treatment was largely ineffective. The discovery of trastuzumab changed that dramatically.
Over the next 20 years, two additional HER2-targeted therapies – lapatinib and trastuzumab emtansine (TDM-1) – earned approval from the Food and Drug Administration for selected patients with early and late HER2+ breast cancer.
Since 2019, four additional HER2-targeted therapies have been approved by the FDA for HER2+ metastatic breast cancer (MBC), changing the treatment paradigm for those patients substantially.
The new agents are especially useful in certain patient populations. The agents offer the promise of improved survival for patients with recurrent metastatic disease and the potential for further reductions in relapse rates in earlier settings.
Trastuzumab deruxtecan
Trastuzumab deruxtecan is an antibody-drug conjugate that links three components: an anti-HER2 monoclonal antibody, a highly potent topoisomerase I inhibitor payload, and a tetrapeptide-based cleavable linker.
Trastuzumab deruxtecan has a high drug-to-antibody ratio. A membrane-permeable payload offers the potential for activity against adjacent HER2-negative cells in heterogeneous tumors. It has a long half-life (6 days).
Trastuzumab deruxtecan received accelerated approval from the FDA in December 2019 to treat patients with HER2+ MBC who have received two or more prior HER2-targeted regimens, based on the results of the DESTINY-Breast 01 trial.
DESTINY-Breast 01 trial
In the phase 2 DESTINY-Breast 01 trial, 184 patients with a median of six previous treatments received trastuzumab deruxtecan (5.4 mg/kg) intravenously every 21 days. There were 24 patients with treated, asymptomatic brain metastases who participated. Patients with untreated or symptomatic brain metastases were excluded.
Overall, a response to therapy was reported in 112 patients (60.9%), with 6.0% complete and 54.9% partial responses. Most of the patients for whom both baseline and postbaseline data were available had a reduction in tumor size.
The median time until response was 1.6 months, an interval that corresponded to the time until the first scheduled imaging. Three patients (1.6%) had progressive disease, and two patients (1.1%) could not be evaluated.
The median duration of follow-up was 11.1 months, and the median response duration was 14.8 months.
The median progression-free survival (PFS) was 16.4 months, and the median overall survival (OS) was not reached. The median PFS in the patients with brain involvement was 18.1 months.
The most common adverse events of grade 3 or higher were a decreased neutrophil count (20.7%), anemia (8.7%), and nausea (7.6%). Most concerning was that trastuzumab deruxtecan was associated with interstitial lung disease in 13.6% of patients.
Tucatinib
Tucatinib is an oral, highly selective HER2 tyrosine kinase inhibitor (TKI). In April 2020, it was approved by the FDA, in combination with trastuzumab and capecitabine, for adult patients with advanced unresectable or metastatic HER2+ breast cancer who have received one or more prior anti-HER2–based regimens for MBC. The approval included patients with brain metastases.
The recommended tucatinib dose is 300 mg orally twice a day in combination with trastuzumab (at the standard dose) and capecitabine (1,000 mg/m2 given orally twice daily on days 1-14) on a 21-day cycle, until disease progression or unacceptable toxicity.
HER2CLIMB trial
The study that led to the approval of tucatinib was the HER2CLIMB trial. The trial enrolled 612 HER2+ MBC patients who had prior treatment with trastuzumab, pertuzumab, and T-DM1. Patients had received a median of 4 (range, 2-17) prior lines of HER2-targeted therapy.
The patients were randomized 2:1 to receive trastuzumab plus capecitabine and either tucatinib or an identical placebo twice daily.
The primary endpoint was PFS, evaluated in the initial 480 randomized patients. The median PFS was 7.8 months in the tucatinib arm and 5.6 months in the control arm (hazard ratio, 0.54; 95% confidence interval, 0.42-0.71; P < .001).
The confirmed overall response rate for patients with measurable disease was 40.6% in the tucatinib arm and 22.8% in the control arm (P = .001). The proportion of patients still in response at 12 months was 33.1% and 12.3%, respectively.
The median OS was 21.9 months in the tucatinib arm and 17.4 months in the placebo arm (HR, 0.66; 95% CI, 0.50-0.88; P = .005). At 24 months, 44.9% and 26.6% of patients, respectively, were still alive.
The most common grade 3 or higher adverse events (in the tucatinib and placebo arms, respectively) were palmar-plantar erythrodysesthesia syndrome (13.1% vs. 9.1%), diarrhea (12.9% vs. 8.6%), elevations in ALT and AST (approximately 5% vs. 0.5% for each), and fatigue (4.7% vs. 4.1%).
Tucatinib in patients with brain involvement
A unique feature of the HER2CLIMB study was that patients with MBC and untreated, symptomatic brain metastases were eligible. Patients with active, untreated central nervous system disease are excluded from virtually all other trials, especially drug-approval trials.
There were 291 patients with brain metastases in HER2CLIMB, 198 (48%) in the tucatinib arm and 93 (46%) in the control arm.
The risk of intracranial progression or death was reduced by 68% in the tucatinib arm (HR, 0.32; 95% CI, 0.22 to 0.48; P < .0001).
The 1-year CNS-PFS rate was 40.2% in the tucatinib arm and 0% in the placebo arm. The median duration of CNS-PFS was 9.9 months and 4.2 months, respectively.
The risk of death was reduced by 42% in the tucatinib arm (HR, 0.58; 95% CI, 0.40-0.85; P = .005). The median OS was 18.1 months and 12.0 months, respectively.
There were more objective responses in the brain with tucatinib (47.3%) than with placebo (20.0%; P = .03). The median duration of response was 6.8 months and 3.0 months, respectively.
Particularly because of its CNS activity and lack of serious, long-term toxicity, tucatinib combination therapy represents an attractive new option for patients with HER2+ MBC.
Neratinib
Neratinib is an irreversible pan-HER TKI that was approved by the FDA in July 2017 for extended adjuvant therapy in patients with early-stage HER2+ breast cancer, following the use of trastuzumab-based therapy.
Long-term results of the ExteNet study led to the approval for use as extended adjuvant therapy.
In February 2020, neratinib was FDA approved in combination with capecitabine for patients with HER2+ MBC after two or more prior anti-HER2–based regimens. The more recent FDA approval was based on results of the NALA trial.
NALA trial
The phase 3 NALA trial included 621 patients with HER2+ MBC who had received at least two prior anti-HER2 based regimens.
Patients were randomized 1:1 to receive neratinib at 240 mg orally once daily on days 1-21 with capecitabine at 750 mg/m2 orally twice daily on days 1-14 or lapatinib at 1,250 mg orally once daily on days 1-21 with capecitabine at 1,000 mg/m2 orally twice daily on days 1-14 for each 21-day cycle. Patients were treated until disease progression or unacceptable toxicity.
The primary endpoints were PFS and OS by blinded, independent, central review.
The median PFS was 5.6 months in the neratinib arm and 5.5 months in the lapatinib arm (HR, 0.76; 95% CI, 0.63-0.93; P = .0059). The PFS rate at 12 months was 28.8% and 14.8%, respectively.
The median OS was 21.0 months in the neratinib arm and 18.7 months in the lapatinib arm (HR, 0.88; 95% CI, 0.72-1.07; P = .2086). The ORR was 32.8% and 26.7%, respectively. The median response duration was 8.5 months and 5.6 months, respectively.
Fewer interventions for CNS disease were required in the neratinib arm than in the lapatinib arm (cumulative incidence, 22.8% vs. 29.2%; P = .043).
The most frequently reported grade 3-4 adverse reactions for the neratinib combination were diarrhea, nausea, vomiting, fatigue, and decreased appetite.
Grade 3 diarrhea occurred in 24.4% of those in the neratinib arm and 12.5% of those in the lapatinib arm. Antidiarrheal medication was used by 98.3% of patients receiving neratinib and 62.1% of patients receiving lapatinib.
Margetuximab-cmkb
Margetuximab is a chimeric Fc-engineered anti-HER2 monoclonal antibody that targets the same epitope as trastuzumab and exerts similar antiproliferative effects.
Compared with trastuzumab, margetuximab has higher affinity for both 158V (high-binding) and 158F (low-binding) alleles of the activating Fc receptor, CD16A. As a result, margetuximab enhances innate immunity, including CD16A-mediated antibody-dependent cellular cytotoxicity, more effectively than trastuzumab. Margetuximab also potentiates adaptive immunity, including enhanced clonality of the T-cell repertoire and induction of HER2-specific T- and B-cell responses.
In December 2020, margetuximab, in combination with chemotherapy, was approved by the FDA for patients with HER2+ MBC after two or more prior anti-HER2 regimens, at least one of which was for metastatic disease. The approved dose is 15 mg/kg IV every 3 weeks.
The study that led to margetuximab’s approval was the phase 3 SOPHIA trial.
SOPHIA trial
SOPHIA was a randomized trial of 536 patients with HER2+ MBC who had received prior treatment with other anti-HER2 therapies, including one to three lines of therapy for MBC.
Patients were randomly assigned 1:1 to receive margetuximab plus chemotherapy or trastuzumab plus chemotherapy. Assignment was stratified by chemotherapy choice (capecitabine, eribulin, gemcitabine, or vinorelbine), the number of previous lines of therapy for MBC, and disease extent.
Co–primary outcome measures were PFS by blinded, independent, central review and OS.
At the second interim analysis, the median PFS was 5.8 months in the margetuximab arm and 4.9 months in the trastuzumab arm (HR, 0.76; 95% CI, 0.59-0.98; P = .033). Results were more impressive in patients with CD16A genotypes containing a 158F allele. In this group, the median PFS was 6.9 months with margetuximab and 5.1 months with trastuzumab (HR, 0.68, 95% CI, 0.52-0.90; P = .005).
At the second interim analysis, the median OS was 21.6 months in the margetuximab arm and 19.8 months in the trastuzumab arm (HR, 0.89; 95% CI, 0.69-1.13; P = .33).
Subgroup data showed no differences in OS between the two arms for any subgroup except HER2+ MBC patients with an IHC score of 2 or higher. This is consistent with the postulated mechanism of action of margetuximab.
The confirmed ORR was 25% in the margetuximab arm and 14% in the trastuzumab arm, with similar durations of response between the study arms.
The most common adverse events in both arms (≥20%), regardless of causality, were fatigue, nausea, diarrhea, and neutropenia. Vomiting was common in the margetuximab arm, and anemia was common in the trastuzumab arm.
Grade 3 or higher adverse events occurred in 53.8% of patients receiving margetuximab and 52.6% of those receiving trastuzumab.
In view of margetuximab’s modest benefits in the SOPHIA trial, the ultimate role for margetuximab in HER2+ MBC may be restricted to patients with the CD16A-158F allele. A neoadjuvant trial is planned in that population.
Take-home messages
There are legitimate arguments regarding whether curing MBC is within reach for certain patient subsets, but there is no argument about whether the outlook for patients with HER2+ MBC has improved dramatically in recent years; it has.
The approval of four unique, new agents for the treatment of women with HER2+ MBC in relapse provides further improvements in outcome for these patients and distinctly different opportunities for tailoring treatment to the special circumstances of each patient (e.g., whether brain metastases are present, desire for oral therapy, comorbidities, experience with prior chemotherapy, etc).
When considered along with the potential for incorporating these drugs in earlier settings in well-designed clinical trials, these new drugs offer great promise to a group of patients who faced a dismal outcome just 2 decades ago.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Before 2001, HER2/neu-positive breast cancer (HER2+) was one of the most dreaded diagnoses a woman could face, as treatment was largely ineffective. The discovery of trastuzumab changed that dramatically.
Over the next 20 years, two additional HER2-targeted therapies – lapatinib and trastuzumab emtansine (TDM-1) – earned approval from the Food and Drug Administration for selected patients with early and late HER2+ breast cancer.
Since 2019, four additional HER2-targeted therapies have been approved by the FDA for HER2+ metastatic breast cancer (MBC), changing the treatment paradigm for those patients substantially.
The new agents are especially useful in certain patient populations. The agents offer the promise of improved survival for patients with recurrent metastatic disease and the potential for further reductions in relapse rates in earlier settings.
Trastuzumab deruxtecan
Trastuzumab deruxtecan is an antibody-drug conjugate that links three components: an anti-HER2 monoclonal antibody, a highly potent topoisomerase I inhibitor payload, and a tetrapeptide-based cleavable linker.
Trastuzumab deruxtecan has a high drug-to-antibody ratio. A membrane-permeable payload offers the potential for activity against adjacent HER2-negative cells in heterogeneous tumors. It has a long half-life (6 days).
Trastuzumab deruxtecan received accelerated approval from the FDA in December 2019 to treat patients with HER2+ MBC who have received two or more prior HER2-targeted regimens, based on the results of the DESTINY-Breast 01 trial.
DESTINY-Breast 01 trial
In the phase 2 DESTINY-Breast 01 trial, 184 patients with a median of six previous treatments received trastuzumab deruxtecan (5.4 mg/kg) intravenously every 21 days. There were 24 patients with treated, asymptomatic brain metastases who participated. Patients with untreated or symptomatic brain metastases were excluded.
Overall, a response to therapy was reported in 112 patients (60.9%), with 6.0% complete and 54.9% partial responses. Most of the patients for whom both baseline and postbaseline data were available had a reduction in tumor size.
The median time until response was 1.6 months, an interval that corresponded to the time until the first scheduled imaging. Three patients (1.6%) had progressive disease, and two patients (1.1%) could not be evaluated.
The median duration of follow-up was 11.1 months, and the median response duration was 14.8 months.
The median progression-free survival (PFS) was 16.4 months, and the median overall survival (OS) was not reached. The median PFS in the patients with brain involvement was 18.1 months.
The most common adverse events of grade 3 or higher were a decreased neutrophil count (20.7%), anemia (8.7%), and nausea (7.6%). Most concerning was that trastuzumab deruxtecan was associated with interstitial lung disease in 13.6% of patients.
Tucatinib
Tucatinib is an oral, highly selective HER2 tyrosine kinase inhibitor (TKI). In April 2020, it was approved by the FDA, in combination with trastuzumab and capecitabine, for adult patients with advanced unresectable or metastatic HER2+ breast cancer who have received one or more prior anti-HER2–based regimens for MBC. The approval included patients with brain metastases.
The recommended tucatinib dose is 300 mg orally twice a day in combination with trastuzumab (at the standard dose) and capecitabine (1,000 mg/m2 given orally twice daily on days 1-14) on a 21-day cycle, until disease progression or unacceptable toxicity.
HER2CLIMB trial
The study that led to the approval of tucatinib was the HER2CLIMB trial. The trial enrolled 612 HER2+ MBC patients who had prior treatment with trastuzumab, pertuzumab, and T-DM1. Patients had received a median of 4 (range, 2-17) prior lines of HER2-targeted therapy.
The patients were randomized 2:1 to receive trastuzumab plus capecitabine and either tucatinib or an identical placebo twice daily.
The primary endpoint was PFS, evaluated in the initial 480 randomized patients. The median PFS was 7.8 months in the tucatinib arm and 5.6 months in the control arm (hazard ratio, 0.54; 95% confidence interval, 0.42-0.71; P < .001).
The confirmed overall response rate for patients with measurable disease was 40.6% in the tucatinib arm and 22.8% in the control arm (P = .001). The proportion of patients still in response at 12 months was 33.1% and 12.3%, respectively.
The median OS was 21.9 months in the tucatinib arm and 17.4 months in the placebo arm (HR, 0.66; 95% CI, 0.50-0.88; P = .005). At 24 months, 44.9% and 26.6% of patients, respectively, were still alive.
The most common grade 3 or higher adverse events (in the tucatinib and placebo arms, respectively) were palmar-plantar erythrodysesthesia syndrome (13.1% vs. 9.1%), diarrhea (12.9% vs. 8.6%), elevations in ALT and AST (approximately 5% vs. 0.5% for each), and fatigue (4.7% vs. 4.1%).
Tucatinib in patients with brain involvement
A unique feature of the HER2CLIMB study was that patients with MBC and untreated, symptomatic brain metastases were eligible. Patients with active, untreated central nervous system disease are excluded from virtually all other trials, especially drug-approval trials.
There were 291 patients with brain metastases in HER2CLIMB, 198 (48%) in the tucatinib arm and 93 (46%) in the control arm.
The risk of intracranial progression or death was reduced by 68% in the tucatinib arm (HR, 0.32; 95% CI, 0.22 to 0.48; P < .0001).
The 1-year CNS-PFS rate was 40.2% in the tucatinib arm and 0% in the placebo arm. The median duration of CNS-PFS was 9.9 months and 4.2 months, respectively.
The risk of death was reduced by 42% in the tucatinib arm (HR, 0.58; 95% CI, 0.40-0.85; P = .005). The median OS was 18.1 months and 12.0 months, respectively.
There were more objective responses in the brain with tucatinib (47.3%) than with placebo (20.0%; P = .03). The median duration of response was 6.8 months and 3.0 months, respectively.
Particularly because of its CNS activity and lack of serious, long-term toxicity, tucatinib combination therapy represents an attractive new option for patients with HER2+ MBC.
Neratinib
Neratinib is an irreversible pan-HER TKI that was approved by the FDA in July 2017 for extended adjuvant therapy in patients with early-stage HER2+ breast cancer, following the use of trastuzumab-based therapy.
Long-term results of the ExteNet study led to the approval for use as extended adjuvant therapy.
In February 2020, neratinib was FDA approved in combination with capecitabine for patients with HER2+ MBC after two or more prior anti-HER2–based regimens. The more recent FDA approval was based on results of the NALA trial.
NALA trial
The phase 3 NALA trial included 621 patients with HER2+ MBC who had received at least two prior anti-HER2 based regimens.
Patients were randomized 1:1 to receive neratinib at 240 mg orally once daily on days 1-21 with capecitabine at 750 mg/m2 orally twice daily on days 1-14 or lapatinib at 1,250 mg orally once daily on days 1-21 with capecitabine at 1,000 mg/m2 orally twice daily on days 1-14 for each 21-day cycle. Patients were treated until disease progression or unacceptable toxicity.
The primary endpoints were PFS and OS by blinded, independent, central review.
The median PFS was 5.6 months in the neratinib arm and 5.5 months in the lapatinib arm (HR, 0.76; 95% CI, 0.63-0.93; P = .0059). The PFS rate at 12 months was 28.8% and 14.8%, respectively.
The median OS was 21.0 months in the neratinib arm and 18.7 months in the lapatinib arm (HR, 0.88; 95% CI, 0.72-1.07; P = .2086). The ORR was 32.8% and 26.7%, respectively. The median response duration was 8.5 months and 5.6 months, respectively.
Fewer interventions for CNS disease were required in the neratinib arm than in the lapatinib arm (cumulative incidence, 22.8% vs. 29.2%; P = .043).
The most frequently reported grade 3-4 adverse reactions for the neratinib combination were diarrhea, nausea, vomiting, fatigue, and decreased appetite.
Grade 3 diarrhea occurred in 24.4% of those in the neratinib arm and 12.5% of those in the lapatinib arm. Antidiarrheal medication was used by 98.3% of patients receiving neratinib and 62.1% of patients receiving lapatinib.
Margetuximab-cmkb
Margetuximab is a chimeric Fc-engineered anti-HER2 monoclonal antibody that targets the same epitope as trastuzumab and exerts similar antiproliferative effects.
Compared with trastuzumab, margetuximab has higher affinity for both 158V (high-binding) and 158F (low-binding) alleles of the activating Fc receptor, CD16A. As a result, margetuximab enhances innate immunity, including CD16A-mediated antibody-dependent cellular cytotoxicity, more effectively than trastuzumab. Margetuximab also potentiates adaptive immunity, including enhanced clonality of the T-cell repertoire and induction of HER2-specific T- and B-cell responses.
In December 2020, margetuximab, in combination with chemotherapy, was approved by the FDA for patients with HER2+ MBC after two or more prior anti-HER2 regimens, at least one of which was for metastatic disease. The approved dose is 15 mg/kg IV every 3 weeks.
The study that led to margetuximab’s approval was the phase 3 SOPHIA trial.
SOPHIA trial
SOPHIA was a randomized trial of 536 patients with HER2+ MBC who had received prior treatment with other anti-HER2 therapies, including one to three lines of therapy for MBC.
Patients were randomly assigned 1:1 to receive margetuximab plus chemotherapy or trastuzumab plus chemotherapy. Assignment was stratified by chemotherapy choice (capecitabine, eribulin, gemcitabine, or vinorelbine), the number of previous lines of therapy for MBC, and disease extent.
Co–primary outcome measures were PFS by blinded, independent, central review and OS.
At the second interim analysis, the median PFS was 5.8 months in the margetuximab arm and 4.9 months in the trastuzumab arm (HR, 0.76; 95% CI, 0.59-0.98; P = .033). Results were more impressive in patients with CD16A genotypes containing a 158F allele. In this group, the median PFS was 6.9 months with margetuximab and 5.1 months with trastuzumab (HR, 0.68, 95% CI, 0.52-0.90; P = .005).
At the second interim analysis, the median OS was 21.6 months in the margetuximab arm and 19.8 months in the trastuzumab arm (HR, 0.89; 95% CI, 0.69-1.13; P = .33).
Subgroup data showed no differences in OS between the two arms for any subgroup except HER2+ MBC patients with an IHC score of 2 or higher. This is consistent with the postulated mechanism of action of margetuximab.
The confirmed ORR was 25% in the margetuximab arm and 14% in the trastuzumab arm, with similar durations of response between the study arms.
The most common adverse events in both arms (≥20%), regardless of causality, were fatigue, nausea, diarrhea, and neutropenia. Vomiting was common in the margetuximab arm, and anemia was common in the trastuzumab arm.
Grade 3 or higher adverse events occurred in 53.8% of patients receiving margetuximab and 52.6% of those receiving trastuzumab.
In view of margetuximab’s modest benefits in the SOPHIA trial, the ultimate role for margetuximab in HER2+ MBC may be restricted to patients with the CD16A-158F allele. A neoadjuvant trial is planned in that population.
Take-home messages
There are legitimate arguments regarding whether curing MBC is within reach for certain patient subsets, but there is no argument about whether the outlook for patients with HER2+ MBC has improved dramatically in recent years; it has.
The approval of four unique, new agents for the treatment of women with HER2+ MBC in relapse provides further improvements in outcome for these patients and distinctly different opportunities for tailoring treatment to the special circumstances of each patient (e.g., whether brain metastases are present, desire for oral therapy, comorbidities, experience with prior chemotherapy, etc).
When considered along with the potential for incorporating these drugs in earlier settings in well-designed clinical trials, these new drugs offer great promise to a group of patients who faced a dismal outcome just 2 decades ago.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
The power and promise of social media in oncology
Mark A. Lewis, MD, explained to the COSMO meeting audience how storytelling on social media can educate and engage patients, advocates, and professional colleagues – advancing knowledge, dispelling misinformation, and promoting clinical research.
Dr. Lewis, an oncologist at Intermountain Healthcare in Salt Lake City, reflected on the bifid roles of oncologists as scientists engaged in life-long learning and humanists who can internalize and appreciate the unique character and circumstances of their patients.
Patients who have serious illnesses are necessarily aggregated by statistics. However, in an essay published in 2011, Dr. Lewis noted that “each individual patient partakes in a unique, irreproducible experiment where n = 1” (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
Dr. Lewis highlighted the duality of individual data points on a survival curve as descriptors of common disease trajectories and treatment effects. However, those data points also conceal important narratives regarding the most highly valued aspects of the doctor-patient relationship and the impact of cancer treatment on patients’ lives.
In referring to the futuristic essay “Ars Brevis,” Dr. Lewis contrasted the humanism of oncology specialists in the present day with the fictional image of data-regurgitating robots programmed to maximize the efficiency of each patient encounter (J Clin Oncol. 2013 May 10;31[14]:1792-4).
Dr. Lewis reminded attendees that to practice medicine without using both “head and heart” undermines the inherent nature of medical care.
Unfortunately, that perspective may not match the public perception of oncologists. Dr. Lewis described his experience of typing “oncologists are” into an Internet search engine and seeing the auto-complete function prompt words such as “criminals,” “evil,” “murderers,” and “confused.”
Obviously, it is hard to establish a trusting patient-doctor relationship if that is the prima facie perception of the oncology specialty.
Dispelling myths and creating community via social media
A primary goal of consultation with a newly-diagnosed cancer patient is for the patient to feel that the oncologist will be there to take care of them, regardless of what the future holds.
Dr. Lewis has found that social media can potentially extend that feeling to a global community of patients, caregivers, and others seeking information relevant to a cancer diagnosis. He believes that oncologists have an opportunity to dispel myths and fears by being attentive to the real-life concerns of patients.
Dr. Lewis took advantage of this opportunity when he underwent a Whipple procedure (pancreaticoduodenectomy) for a pancreatic neuroendocrine tumor. He and the hospital’s media services staff “live-tweeted” his surgery and recovery.
With those tweets, Dr. Lewis demystified each step of a major surgical procedure. From messages he received on social media, Dr. Lewis knows he made the decision to have a Whipple procedure more acceptable to other patients.
His personal medical experience notwithstanding, Dr. Lewis acknowledged that every patient’s circumstances are unique.
Oncologists cannot possibly empathize with every circumstance. However, when they show sensitivity to personal elements of the cancer experience, they shed light on the complicated role they play in patient care and can facilitate good decision-making among patients across the globe.
Social media for professional development and patient care
The publication of his 2011 essay was gratifying for Dr. Lewis, but the finite number of comments he received thereafter illustrated the rather limited audience that traditional academic publications have and the laborious process for subsequent interaction (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
First as an observer and later as a participant on social media, Dr. Lewis appreciated that teaching points and publications can be amplified by global distribution and the potential for informal bidirectional communication.
Social media platforms enable physicians to connect with a larger audience through participative communication, in which users develop, share, and react to content (N Engl J Med. 2009 Aug 13;361[7]:649-51).
Dr. Lewis reflected on how oncologists are challenged to sort through the thousands of oncology-focused publications annually. Through social media, one can see the studies on which the experts are commenting and appreciate the nuances that contextualize the results. Focused interactions with renowned doctors, at regular intervals, require little formality.
Online journal clubs enable the sharing of ideas, opinions, multimedia resources, and references across institutional and international borders (J Gen Intern Med. 2014 Oct;29[10]:1317-8).
Social media in oncology: Accomplishments and promise
The development of broadband Internet, wireless connectivity, and social media for peer-to-peer and general communication are among the major technological advances that have transformed medical communication.
As an organization, COSMO aims to describe, understand, and improve the use of social media to increase the penetration of evidence-based guidelines and research insights into clinical practice (Future Oncol. 2017 Jun;13[15]:1281-5).
At the inaugural COSMO meeting, areas of progress since COSMO’s inception in 2015 were highlighted, including:
- The involvement of cancer professionals and advocates in multiple distinctive platforms.
- The development of hashtag libraries to aggregate interest groups and topics.
- The refinement of strategies for engaging advocates with attention to inclusiveness.
- A steady trajectory of growth in tweeting at scientific conferences.
An overarching theme of the COSMO meeting was “authenticity,” a virtue that is easy to admire but requires conscious, consistent effort to achieve.
Disclosure of conflicts of interest and avoiding using social media simply as a recruitment tool for clinical trials are basic components of accurate self-representation.
In addition, Dr. Lewis advocated for sharing personal experiences in a component of social media posts so oncologists can show humanity as a feature of their professional online identity and inherent nature.
Dr. Lewis disclosed consultancy with Medscape/WebMD, which are owned by the same parent company as MDedge. He also disclosed relationships with Foundation Medicine, Natera, Exelixis, QED, HalioDX, and Ipsen.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Mark A. Lewis, MD, explained to the COSMO meeting audience how storytelling on social media can educate and engage patients, advocates, and professional colleagues – advancing knowledge, dispelling misinformation, and promoting clinical research.
Dr. Lewis, an oncologist at Intermountain Healthcare in Salt Lake City, reflected on the bifid roles of oncologists as scientists engaged in life-long learning and humanists who can internalize and appreciate the unique character and circumstances of their patients.
Patients who have serious illnesses are necessarily aggregated by statistics. However, in an essay published in 2011, Dr. Lewis noted that “each individual patient partakes in a unique, irreproducible experiment where n = 1” (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
Dr. Lewis highlighted the duality of individual data points on a survival curve as descriptors of common disease trajectories and treatment effects. However, those data points also conceal important narratives regarding the most highly valued aspects of the doctor-patient relationship and the impact of cancer treatment on patients’ lives.
In referring to the futuristic essay “Ars Brevis,” Dr. Lewis contrasted the humanism of oncology specialists in the present day with the fictional image of data-regurgitating robots programmed to maximize the efficiency of each patient encounter (J Clin Oncol. 2013 May 10;31[14]:1792-4).
Dr. Lewis reminded attendees that to practice medicine without using both “head and heart” undermines the inherent nature of medical care.
Unfortunately, that perspective may not match the public perception of oncologists. Dr. Lewis described his experience of typing “oncologists are” into an Internet search engine and seeing the auto-complete function prompt words such as “criminals,” “evil,” “murderers,” and “confused.”
Obviously, it is hard to establish a trusting patient-doctor relationship if that is the prima facie perception of the oncology specialty.
Dispelling myths and creating community via social media
A primary goal of consultation with a newly-diagnosed cancer patient is for the patient to feel that the oncologist will be there to take care of them, regardless of what the future holds.
Dr. Lewis has found that social media can potentially extend that feeling to a global community of patients, caregivers, and others seeking information relevant to a cancer diagnosis. He believes that oncologists have an opportunity to dispel myths and fears by being attentive to the real-life concerns of patients.
Dr. Lewis took advantage of this opportunity when he underwent a Whipple procedure (pancreaticoduodenectomy) for a pancreatic neuroendocrine tumor. He and the hospital’s media services staff “live-tweeted” his surgery and recovery.
With those tweets, Dr. Lewis demystified each step of a major surgical procedure. From messages he received on social media, Dr. Lewis knows he made the decision to have a Whipple procedure more acceptable to other patients.
His personal medical experience notwithstanding, Dr. Lewis acknowledged that every patient’s circumstances are unique.
Oncologists cannot possibly empathize with every circumstance. However, when they show sensitivity to personal elements of the cancer experience, they shed light on the complicated role they play in patient care and can facilitate good decision-making among patients across the globe.
Social media for professional development and patient care
The publication of his 2011 essay was gratifying for Dr. Lewis, but the finite number of comments he received thereafter illustrated the rather limited audience that traditional academic publications have and the laborious process for subsequent interaction (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
First as an observer and later as a participant on social media, Dr. Lewis appreciated that teaching points and publications can be amplified by global distribution and the potential for informal bidirectional communication.
Social media platforms enable physicians to connect with a larger audience through participative communication, in which users develop, share, and react to content (N Engl J Med. 2009 Aug 13;361[7]:649-51).
Dr. Lewis reflected on how oncologists are challenged to sort through the thousands of oncology-focused publications annually. Through social media, one can see the studies on which the experts are commenting and appreciate the nuances that contextualize the results. Focused interactions with renowned doctors, at regular intervals, require little formality.
Online journal clubs enable the sharing of ideas, opinions, multimedia resources, and references across institutional and international borders (J Gen Intern Med. 2014 Oct;29[10]:1317-8).
Social media in oncology: Accomplishments and promise
The development of broadband Internet, wireless connectivity, and social media for peer-to-peer and general communication are among the major technological advances that have transformed medical communication.
As an organization, COSMO aims to describe, understand, and improve the use of social media to increase the penetration of evidence-based guidelines and research insights into clinical practice (Future Oncol. 2017 Jun;13[15]:1281-5).
At the inaugural COSMO meeting, areas of progress since COSMO’s inception in 2015 were highlighted, including:
- The involvement of cancer professionals and advocates in multiple distinctive platforms.
- The development of hashtag libraries to aggregate interest groups and topics.
- The refinement of strategies for engaging advocates with attention to inclusiveness.
- A steady trajectory of growth in tweeting at scientific conferences.
An overarching theme of the COSMO meeting was “authenticity,” a virtue that is easy to admire but requires conscious, consistent effort to achieve.
Disclosure of conflicts of interest and avoiding using social media simply as a recruitment tool for clinical trials are basic components of accurate self-representation.
In addition, Dr. Lewis advocated for sharing personal experiences in a component of social media posts so oncologists can show humanity as a feature of their professional online identity and inherent nature.
Dr. Lewis disclosed consultancy with Medscape/WebMD, which are owned by the same parent company as MDedge. He also disclosed relationships with Foundation Medicine, Natera, Exelixis, QED, HalioDX, and Ipsen.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Mark A. Lewis, MD, explained to the COSMO meeting audience how storytelling on social media can educate and engage patients, advocates, and professional colleagues – advancing knowledge, dispelling misinformation, and promoting clinical research.
Dr. Lewis, an oncologist at Intermountain Healthcare in Salt Lake City, reflected on the bifid roles of oncologists as scientists engaged in life-long learning and humanists who can internalize and appreciate the unique character and circumstances of their patients.
Patients who have serious illnesses are necessarily aggregated by statistics. However, in an essay published in 2011, Dr. Lewis noted that “each individual patient partakes in a unique, irreproducible experiment where n = 1” (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
Dr. Lewis highlighted the duality of individual data points on a survival curve as descriptors of common disease trajectories and treatment effects. However, those data points also conceal important narratives regarding the most highly valued aspects of the doctor-patient relationship and the impact of cancer treatment on patients’ lives.
In referring to the futuristic essay “Ars Brevis,” Dr. Lewis contrasted the humanism of oncology specialists in the present day with the fictional image of data-regurgitating robots programmed to maximize the efficiency of each patient encounter (J Clin Oncol. 2013 May 10;31[14]:1792-4).
Dr. Lewis reminded attendees that to practice medicine without using both “head and heart” undermines the inherent nature of medical care.
Unfortunately, that perspective may not match the public perception of oncologists. Dr. Lewis described his experience of typing “oncologists are” into an Internet search engine and seeing the auto-complete function prompt words such as “criminals,” “evil,” “murderers,” and “confused.”
Obviously, it is hard to establish a trusting patient-doctor relationship if that is the prima facie perception of the oncology specialty.
Dispelling myths and creating community via social media
A primary goal of consultation with a newly-diagnosed cancer patient is for the patient to feel that the oncologist will be there to take care of them, regardless of what the future holds.
Dr. Lewis has found that social media can potentially extend that feeling to a global community of patients, caregivers, and others seeking information relevant to a cancer diagnosis. He believes that oncologists have an opportunity to dispel myths and fears by being attentive to the real-life concerns of patients.
Dr. Lewis took advantage of this opportunity when he underwent a Whipple procedure (pancreaticoduodenectomy) for a pancreatic neuroendocrine tumor. He and the hospital’s media services staff “live-tweeted” his surgery and recovery.
With those tweets, Dr. Lewis demystified each step of a major surgical procedure. From messages he received on social media, Dr. Lewis knows he made the decision to have a Whipple procedure more acceptable to other patients.
His personal medical experience notwithstanding, Dr. Lewis acknowledged that every patient’s circumstances are unique.
Oncologists cannot possibly empathize with every circumstance. However, when they show sensitivity to personal elements of the cancer experience, they shed light on the complicated role they play in patient care and can facilitate good decision-making among patients across the globe.
Social media for professional development and patient care
The publication of his 2011 essay was gratifying for Dr. Lewis, but the finite number of comments he received thereafter illustrated the rather limited audience that traditional academic publications have and the laborious process for subsequent interaction (J Clin Oncol. 2011 Aug 1;29[22]:3103-4).
First as an observer and later as a participant on social media, Dr. Lewis appreciated that teaching points and publications can be amplified by global distribution and the potential for informal bidirectional communication.
Social media platforms enable physicians to connect with a larger audience through participative communication, in which users develop, share, and react to content (N Engl J Med. 2009 Aug 13;361[7]:649-51).
Dr. Lewis reflected on how oncologists are challenged to sort through the thousands of oncology-focused publications annually. Through social media, one can see the studies on which the experts are commenting and appreciate the nuances that contextualize the results. Focused interactions with renowned doctors, at regular intervals, require little formality.
Online journal clubs enable the sharing of ideas, opinions, multimedia resources, and references across institutional and international borders (J Gen Intern Med. 2014 Oct;29[10]:1317-8).
Social media in oncology: Accomplishments and promise
The development of broadband Internet, wireless connectivity, and social media for peer-to-peer and general communication are among the major technological advances that have transformed medical communication.
As an organization, COSMO aims to describe, understand, and improve the use of social media to increase the penetration of evidence-based guidelines and research insights into clinical practice (Future Oncol. 2017 Jun;13[15]:1281-5).
At the inaugural COSMO meeting, areas of progress since COSMO’s inception in 2015 were highlighted, including:
- The involvement of cancer professionals and advocates in multiple distinctive platforms.
- The development of hashtag libraries to aggregate interest groups and topics.
- The refinement of strategies for engaging advocates with attention to inclusiveness.
- A steady trajectory of growth in tweeting at scientific conferences.
An overarching theme of the COSMO meeting was “authenticity,” a virtue that is easy to admire but requires conscious, consistent effort to achieve.
Disclosure of conflicts of interest and avoiding using social media simply as a recruitment tool for clinical trials are basic components of accurate self-representation.
In addition, Dr. Lewis advocated for sharing personal experiences in a component of social media posts so oncologists can show humanity as a feature of their professional online identity and inherent nature.
Dr. Lewis disclosed consultancy with Medscape/WebMD, which are owned by the same parent company as MDedge. He also disclosed relationships with Foundation Medicine, Natera, Exelixis, QED, HalioDX, and Ipsen.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM COSMO 2021
IMvigor130: A treasure trove of data for urothelial carcinoma
A second interim overall survival (OS) analysis suggested that atezolizumab monotherapy provides a clinical benefit as first-line treatment for mUC patients with PD-L1–expressing immune cells representing at least 5% of the tumor area (IC2/3), including patients who are cisplatin ineligible.
The analysis also suggested that atezolizumab plus chemotherapy produces similar OS results as chemotherapy plus placebo, but patients receiving atezolizumab may do better with cisplatin-based chemotherapy than with carboplatin-based chemotherapy.
These results were reported in two presentations at the American Association for Cancer Research Annual Meeting 2021: Week 1.
Current guidelines from the National Comprehensive Cancer Network and the European Society for Medical Oncology recommend atezolizumab monotherapy for cisplatin-ineligible patients with mUC and PD-L1 IC2/3.
The ongoing phase 3 IMvigor130 trial was designed to compare atezolizumab plus gemcitabine/platinum chemotherapy, atezolizumab monotherapy, and placebo plus chemotherapy. Platinum-based chemotherapy included either cisplatin or carboplatin, per investigator choice.
Coprimary endpoints for IMvigor130 were progression-free survival (PFS) and OS for atezolizumab plus chemotherapy versus placebo plus chemotherapy. The hierarchical study design dictated that OS would only be assessed for the comparison of atezolizumab monotherapy versus placebo-chemotherapy in the overall and PD-L1 IC2/3 populations if there was statistical improvement in OS for the atezolizumab-chemotherapy arm over the placebo-chemotherapy arm.
Secondary endpoints were overall response rate (ORR; per RECIST 1.1), duration of response (DOR) for all patients, and PFS for the comparison between atezolizumab monotherapy and placebo-chemotherapy. Exploratory analyses were performed on cisplatin-ineligible patients by PD-L1 status.
At the time of the primary analysis, an OS benefit for atezolizumab-chemotherapy over placebo-chemotherapy was not observed. Therefore, the OS benefit of atezolizumab monotherapy versus placebo-chemotherapy was not assessed. However, a trend toward improved OS was noted with atezolizumab for PD-L1 IC2/3 patients, including cisplatin-ineligible patients.
Atezolizumab vs. placebo-chemo
Ian D. Davis, MBBS, PhD, of Monash University in Melbourne, presented the second interim analysis of OS with atezolizumab monotherapy versus placebo plus chemotherapy (Abstract CT040).
The median follow-up was 14.9 months for atezolizumab monotherapy (n = 360) and 11.8 months for placebo-chemotherapy (n = 359). The median OS was 15.2 months and 13.1 months, respectively (hazard ratio, 0.99; 95% confidence interval, 0.83-1.19). There was no apparent OS benefit of atezolizumab for any clinically selected subgroup.
The ORR was 23.4% for atezolizumab monotherapy and 44.1% for placebo-chemotherapy. The median DOR was more than 3.5 times longer for atezolizumab monotherapy than for placebo-chemotherapy – 29.6 months and 8.1 months, respectively.
Although there was no formal statistical comparison, exploratory subgroup analyses demonstrated that the median OS for the PD-L1 IC2/3 patients appeared higher in the atezolizumab monotherapy arm than in the placebo-chemotherapy arm – 27.5 months and 16.7 months, respectively.
Similarly, the median OS for cisplatin-ineligible PD-L1 IC2/3 patients appeared higher for atezolizumab monotherapy than for placebo-chemotherapy – 18.6 months and 10.0 months, respectively.
In terms of safety, atezolizumab monotherapy compared favorably with placebo plus chemotherapy. There were similar numbers of grade 3/4 adverse events and comparable adverse events leading to discontinuation of treatment in both arms.
The atezolizumab monotherapy arm had fewer adverse events leading to withdrawal from any treatment, when compared with the placebo-chemotherapy arm – 7% and 34%, respectively. Two patients in the atezolizumab arm and one in the placebo-chemotherapy died of treatment-related causes.
Atezolizumab-chemo vs. placebo-chemo
Matthew D. Galsky, MD, of Mount Sinai Health System and Icahn School of Medicine at Mount Sinai in New York, presented the second interim OS comparison of atezolizumab plus chemotherapy with placebo plus chemotherapy (Abstract CT042).
The primary analysis had shown a statistically significant improvement in PFS for patients on atezolizumab-chemotherapy, in comparison with placebo-chemotherapy, with encouraging OS improvement, but the boundary for declaring significance for the OS endpoint was not crossed (Lancet. 2020 May 16;395[10236]:1547-1557).
Because IMvigor130 included both patients who received cisplatin and patients who investigators deemed cisplatin ineligible, the second interim analysis included an exploratory analysis of whether there was a difference in outcome between patients who received or did not receive cisplatin.
At a median follow-up of 13.3 months, the median OS was not significantly different in the atezolizumab-chemotherapy arm (n = 451) and the placebo-chemotherapy arm (n = 400) – 16.1 months and 13.4 months, respectively (HR, 0.84; 95% CI, 0.71-1.00; P = .026).
There were no clinically or pathologically defined subgroups that experienced an OS benefit from atezolizumab-chemotherapy over placebo-chemotherapy.
As for subsequent nonprotocol therapy, 24% of the placebo-chemotherapy arm received an immune checkpoint inhibitor at progression, as did 7% of the atezolizumab-chemotherapy arm. There was no difference in receipt of an immune checkpoint inhibitor post progression among patients treated with cisplatin versus carboplatin.
The benefit of combining atezolizumab with chemotherapy appeared more substantial with cisplatin-based chemotherapy than with carboplatin-based treatment. With cisplatin, the median OS was 21.6 months for the atezolizumab-chemotherapy arm and 14.6 months for the placebo-chemotherapy arm. With carboplatin, the median OS was 14.3 months and 13.0 months, respectively.
PD-L1 status was prognostic for patients who received cisplatin, with lower OS being observed for patients with PD-L1 IC0/1 status and higher OS observed for patients with PD-L1 IC2/3 status. Atezolizumab plus cisplatin-based chemotherapy appeared superior to cisplatin-based chemotherapy alone in both PD-L1–low and –high groups.
Atezolizumab did not seem to benefit patients who were treated with carboplatin, and PD-L1 status did not seem to influence OS among the carboplatin-treated patients.
Although similar ORR results were seen with cisplatin and carboplatin, there appeared to be a longer median DOR among cisplatin-treated patients who received atezolizumab than among those who did not – 13.2 months and 8.3 months, respectively.
No such benefit from atezolizumab was seen in carboplatin-treated patients. The median DOR was 8.1 months among patients who received atezolizumab and 7.1 months among those who did not.
The overall safety profile for atezolizumab plus chemotherapy was consistent with prior reports of the combination. Treatment-related grade 3-5 adverse events were similar on the atezolizumab-chemotherapy arm and the placebo-chemotherapy arm.
The present and future
The investigators who presented the second interim analysis for OS of the IMvigor130 trial were appropriately modest in their conclusions. After all, the prespecified boundary for significant improvement in OS for the addition of atezolizumab to chemotherapy was not crossed. No change in guideline-based clinical practice would be appropriate at the present time.
The various exploratory analyses are hypothesis generating and invite potential mechanistic explanations. However, given the nonrandom allocation of patients to cisplatin- or carboplatin-based chemotherapy, unrecognized variables may have influenced any appearance of a difference in OS between the regimens.
In IMvigor130, treatment was given until unacceptable toxicity or disease progression. It is uncertain whether the current National Comprehensive Cancer Network category 1 recommendation of chemotherapy induction followed by immune checkpoint inhibitor maintenance therapy will prove superior to the IMvigor130 strategy.
Clearly – and concordant with current treatment guidelines – atezolizumab monotherapy can benefit some patients, though the response rate for atezolizumab monotherapy was lower than for chemotherapy (23.4% vs. 44.1%).
As noted by the session chair, Marina Chiara Garassino, MD, of the University of Chicago, the OS curves were initially superior for chemotherapy over atezolizumab. However, the apparent early OS benefit for chemotherapy dissipated over time and, among responders to atezolizumab, response duration was considerably longer than for chemotherapy.
IMvigor130 will ultimately have a final OS analysis to clarify the relative benefits of the various treatment strategies. Fortunately, this large phase 3 study will yield a treasure trove of data to inform future research and build on the advances of recent years for patients with advanced urothelial cancer.
IMvigor130 is sponsored by Hoffmann-La Roche. Dr. Davis, Dr. Galsky, and Dr. Garassino disclosed relationships with Hoffmann-La Roche and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
A second interim overall survival (OS) analysis suggested that atezolizumab monotherapy provides a clinical benefit as first-line treatment for mUC patients with PD-L1–expressing immune cells representing at least 5% of the tumor area (IC2/3), including patients who are cisplatin ineligible.
The analysis also suggested that atezolizumab plus chemotherapy produces similar OS results as chemotherapy plus placebo, but patients receiving atezolizumab may do better with cisplatin-based chemotherapy than with carboplatin-based chemotherapy.
These results were reported in two presentations at the American Association for Cancer Research Annual Meeting 2021: Week 1.
Current guidelines from the National Comprehensive Cancer Network and the European Society for Medical Oncology recommend atezolizumab monotherapy for cisplatin-ineligible patients with mUC and PD-L1 IC2/3.
The ongoing phase 3 IMvigor130 trial was designed to compare atezolizumab plus gemcitabine/platinum chemotherapy, atezolizumab monotherapy, and placebo plus chemotherapy. Platinum-based chemotherapy included either cisplatin or carboplatin, per investigator choice.
Coprimary endpoints for IMvigor130 were progression-free survival (PFS) and OS for atezolizumab plus chemotherapy versus placebo plus chemotherapy. The hierarchical study design dictated that OS would only be assessed for the comparison of atezolizumab monotherapy versus placebo-chemotherapy in the overall and PD-L1 IC2/3 populations if there was statistical improvement in OS for the atezolizumab-chemotherapy arm over the placebo-chemotherapy arm.
Secondary endpoints were overall response rate (ORR; per RECIST 1.1), duration of response (DOR) for all patients, and PFS for the comparison between atezolizumab monotherapy and placebo-chemotherapy. Exploratory analyses were performed on cisplatin-ineligible patients by PD-L1 status.
At the time of the primary analysis, an OS benefit for atezolizumab-chemotherapy over placebo-chemotherapy was not observed. Therefore, the OS benefit of atezolizumab monotherapy versus placebo-chemotherapy was not assessed. However, a trend toward improved OS was noted with atezolizumab for PD-L1 IC2/3 patients, including cisplatin-ineligible patients.
Atezolizumab vs. placebo-chemo
Ian D. Davis, MBBS, PhD, of Monash University in Melbourne, presented the second interim analysis of OS with atezolizumab monotherapy versus placebo plus chemotherapy (Abstract CT040).
The median follow-up was 14.9 months for atezolizumab monotherapy (n = 360) and 11.8 months for placebo-chemotherapy (n = 359). The median OS was 15.2 months and 13.1 months, respectively (hazard ratio, 0.99; 95% confidence interval, 0.83-1.19). There was no apparent OS benefit of atezolizumab for any clinically selected subgroup.
The ORR was 23.4% for atezolizumab monotherapy and 44.1% for placebo-chemotherapy. The median DOR was more than 3.5 times longer for atezolizumab monotherapy than for placebo-chemotherapy – 29.6 months and 8.1 months, respectively.
Although there was no formal statistical comparison, exploratory subgroup analyses demonstrated that the median OS for the PD-L1 IC2/3 patients appeared higher in the atezolizumab monotherapy arm than in the placebo-chemotherapy arm – 27.5 months and 16.7 months, respectively.
Similarly, the median OS for cisplatin-ineligible PD-L1 IC2/3 patients appeared higher for atezolizumab monotherapy than for placebo-chemotherapy – 18.6 months and 10.0 months, respectively.
In terms of safety, atezolizumab monotherapy compared favorably with placebo plus chemotherapy. There were similar numbers of grade 3/4 adverse events and comparable adverse events leading to discontinuation of treatment in both arms.
The atezolizumab monotherapy arm had fewer adverse events leading to withdrawal from any treatment, when compared with the placebo-chemotherapy arm – 7% and 34%, respectively. Two patients in the atezolizumab arm and one in the placebo-chemotherapy died of treatment-related causes.
Atezolizumab-chemo vs. placebo-chemo
Matthew D. Galsky, MD, of Mount Sinai Health System and Icahn School of Medicine at Mount Sinai in New York, presented the second interim OS comparison of atezolizumab plus chemotherapy with placebo plus chemotherapy (Abstract CT042).
The primary analysis had shown a statistically significant improvement in PFS for patients on atezolizumab-chemotherapy, in comparison with placebo-chemotherapy, with encouraging OS improvement, but the boundary for declaring significance for the OS endpoint was not crossed (Lancet. 2020 May 16;395[10236]:1547-1557).
Because IMvigor130 included both patients who received cisplatin and patients who investigators deemed cisplatin ineligible, the second interim analysis included an exploratory analysis of whether there was a difference in outcome between patients who received or did not receive cisplatin.
At a median follow-up of 13.3 months, the median OS was not significantly different in the atezolizumab-chemotherapy arm (n = 451) and the placebo-chemotherapy arm (n = 400) – 16.1 months and 13.4 months, respectively (HR, 0.84; 95% CI, 0.71-1.00; P = .026).
There were no clinically or pathologically defined subgroups that experienced an OS benefit from atezolizumab-chemotherapy over placebo-chemotherapy.
As for subsequent nonprotocol therapy, 24% of the placebo-chemotherapy arm received an immune checkpoint inhibitor at progression, as did 7% of the atezolizumab-chemotherapy arm. There was no difference in receipt of an immune checkpoint inhibitor post progression among patients treated with cisplatin versus carboplatin.
The benefit of combining atezolizumab with chemotherapy appeared more substantial with cisplatin-based chemotherapy than with carboplatin-based treatment. With cisplatin, the median OS was 21.6 months for the atezolizumab-chemotherapy arm and 14.6 months for the placebo-chemotherapy arm. With carboplatin, the median OS was 14.3 months and 13.0 months, respectively.
PD-L1 status was prognostic for patients who received cisplatin, with lower OS being observed for patients with PD-L1 IC0/1 status and higher OS observed for patients with PD-L1 IC2/3 status. Atezolizumab plus cisplatin-based chemotherapy appeared superior to cisplatin-based chemotherapy alone in both PD-L1–low and –high groups.
Atezolizumab did not seem to benefit patients who were treated with carboplatin, and PD-L1 status did not seem to influence OS among the carboplatin-treated patients.
Although similar ORR results were seen with cisplatin and carboplatin, there appeared to be a longer median DOR among cisplatin-treated patients who received atezolizumab than among those who did not – 13.2 months and 8.3 months, respectively.
No such benefit from atezolizumab was seen in carboplatin-treated patients. The median DOR was 8.1 months among patients who received atezolizumab and 7.1 months among those who did not.
The overall safety profile for atezolizumab plus chemotherapy was consistent with prior reports of the combination. Treatment-related grade 3-5 adverse events were similar on the atezolizumab-chemotherapy arm and the placebo-chemotherapy arm.
The present and future
The investigators who presented the second interim analysis for OS of the IMvigor130 trial were appropriately modest in their conclusions. After all, the prespecified boundary for significant improvement in OS for the addition of atezolizumab to chemotherapy was not crossed. No change in guideline-based clinical practice would be appropriate at the present time.
The various exploratory analyses are hypothesis generating and invite potential mechanistic explanations. However, given the nonrandom allocation of patients to cisplatin- or carboplatin-based chemotherapy, unrecognized variables may have influenced any appearance of a difference in OS between the regimens.
In IMvigor130, treatment was given until unacceptable toxicity or disease progression. It is uncertain whether the current National Comprehensive Cancer Network category 1 recommendation of chemotherapy induction followed by immune checkpoint inhibitor maintenance therapy will prove superior to the IMvigor130 strategy.
Clearly – and concordant with current treatment guidelines – atezolizumab monotherapy can benefit some patients, though the response rate for atezolizumab monotherapy was lower than for chemotherapy (23.4% vs. 44.1%).
As noted by the session chair, Marina Chiara Garassino, MD, of the University of Chicago, the OS curves were initially superior for chemotherapy over atezolizumab. However, the apparent early OS benefit for chemotherapy dissipated over time and, among responders to atezolizumab, response duration was considerably longer than for chemotherapy.
IMvigor130 will ultimately have a final OS analysis to clarify the relative benefits of the various treatment strategies. Fortunately, this large phase 3 study will yield a treasure trove of data to inform future research and build on the advances of recent years for patients with advanced urothelial cancer.
IMvigor130 is sponsored by Hoffmann-La Roche. Dr. Davis, Dr. Galsky, and Dr. Garassino disclosed relationships with Hoffmann-La Roche and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
A second interim overall survival (OS) analysis suggested that atezolizumab monotherapy provides a clinical benefit as first-line treatment for mUC patients with PD-L1–expressing immune cells representing at least 5% of the tumor area (IC2/3), including patients who are cisplatin ineligible.
The analysis also suggested that atezolizumab plus chemotherapy produces similar OS results as chemotherapy plus placebo, but patients receiving atezolizumab may do better with cisplatin-based chemotherapy than with carboplatin-based chemotherapy.
These results were reported in two presentations at the American Association for Cancer Research Annual Meeting 2021: Week 1.
Current guidelines from the National Comprehensive Cancer Network and the European Society for Medical Oncology recommend atezolizumab monotherapy for cisplatin-ineligible patients with mUC and PD-L1 IC2/3.
The ongoing phase 3 IMvigor130 trial was designed to compare atezolizumab plus gemcitabine/platinum chemotherapy, atezolizumab monotherapy, and placebo plus chemotherapy. Platinum-based chemotherapy included either cisplatin or carboplatin, per investigator choice.
Coprimary endpoints for IMvigor130 were progression-free survival (PFS) and OS for atezolizumab plus chemotherapy versus placebo plus chemotherapy. The hierarchical study design dictated that OS would only be assessed for the comparison of atezolizumab monotherapy versus placebo-chemotherapy in the overall and PD-L1 IC2/3 populations if there was statistical improvement in OS for the atezolizumab-chemotherapy arm over the placebo-chemotherapy arm.
Secondary endpoints were overall response rate (ORR; per RECIST 1.1), duration of response (DOR) for all patients, and PFS for the comparison between atezolizumab monotherapy and placebo-chemotherapy. Exploratory analyses were performed on cisplatin-ineligible patients by PD-L1 status.
At the time of the primary analysis, an OS benefit for atezolizumab-chemotherapy over placebo-chemotherapy was not observed. Therefore, the OS benefit of atezolizumab monotherapy versus placebo-chemotherapy was not assessed. However, a trend toward improved OS was noted with atezolizumab for PD-L1 IC2/3 patients, including cisplatin-ineligible patients.
Atezolizumab vs. placebo-chemo
Ian D. Davis, MBBS, PhD, of Monash University in Melbourne, presented the second interim analysis of OS with atezolizumab monotherapy versus placebo plus chemotherapy (Abstract CT040).
The median follow-up was 14.9 months for atezolizumab monotherapy (n = 360) and 11.8 months for placebo-chemotherapy (n = 359). The median OS was 15.2 months and 13.1 months, respectively (hazard ratio, 0.99; 95% confidence interval, 0.83-1.19). There was no apparent OS benefit of atezolizumab for any clinically selected subgroup.
The ORR was 23.4% for atezolizumab monotherapy and 44.1% for placebo-chemotherapy. The median DOR was more than 3.5 times longer for atezolizumab monotherapy than for placebo-chemotherapy – 29.6 months and 8.1 months, respectively.
Although there was no formal statistical comparison, exploratory subgroup analyses demonstrated that the median OS for the PD-L1 IC2/3 patients appeared higher in the atezolizumab monotherapy arm than in the placebo-chemotherapy arm – 27.5 months and 16.7 months, respectively.
Similarly, the median OS for cisplatin-ineligible PD-L1 IC2/3 patients appeared higher for atezolizumab monotherapy than for placebo-chemotherapy – 18.6 months and 10.0 months, respectively.
In terms of safety, atezolizumab monotherapy compared favorably with placebo plus chemotherapy. There were similar numbers of grade 3/4 adverse events and comparable adverse events leading to discontinuation of treatment in both arms.
The atezolizumab monotherapy arm had fewer adverse events leading to withdrawal from any treatment, when compared with the placebo-chemotherapy arm – 7% and 34%, respectively. Two patients in the atezolizumab arm and one in the placebo-chemotherapy died of treatment-related causes.
Atezolizumab-chemo vs. placebo-chemo
Matthew D. Galsky, MD, of Mount Sinai Health System and Icahn School of Medicine at Mount Sinai in New York, presented the second interim OS comparison of atezolizumab plus chemotherapy with placebo plus chemotherapy (Abstract CT042).
The primary analysis had shown a statistically significant improvement in PFS for patients on atezolizumab-chemotherapy, in comparison with placebo-chemotherapy, with encouraging OS improvement, but the boundary for declaring significance for the OS endpoint was not crossed (Lancet. 2020 May 16;395[10236]:1547-1557).
Because IMvigor130 included both patients who received cisplatin and patients who investigators deemed cisplatin ineligible, the second interim analysis included an exploratory analysis of whether there was a difference in outcome between patients who received or did not receive cisplatin.
At a median follow-up of 13.3 months, the median OS was not significantly different in the atezolizumab-chemotherapy arm (n = 451) and the placebo-chemotherapy arm (n = 400) – 16.1 months and 13.4 months, respectively (HR, 0.84; 95% CI, 0.71-1.00; P = .026).
There were no clinically or pathologically defined subgroups that experienced an OS benefit from atezolizumab-chemotherapy over placebo-chemotherapy.
As for subsequent nonprotocol therapy, 24% of the placebo-chemotherapy arm received an immune checkpoint inhibitor at progression, as did 7% of the atezolizumab-chemotherapy arm. There was no difference in receipt of an immune checkpoint inhibitor post progression among patients treated with cisplatin versus carboplatin.
The benefit of combining atezolizumab with chemotherapy appeared more substantial with cisplatin-based chemotherapy than with carboplatin-based treatment. With cisplatin, the median OS was 21.6 months for the atezolizumab-chemotherapy arm and 14.6 months for the placebo-chemotherapy arm. With carboplatin, the median OS was 14.3 months and 13.0 months, respectively.
PD-L1 status was prognostic for patients who received cisplatin, with lower OS being observed for patients with PD-L1 IC0/1 status and higher OS observed for patients with PD-L1 IC2/3 status. Atezolizumab plus cisplatin-based chemotherapy appeared superior to cisplatin-based chemotherapy alone in both PD-L1–low and –high groups.
Atezolizumab did not seem to benefit patients who were treated with carboplatin, and PD-L1 status did not seem to influence OS among the carboplatin-treated patients.
Although similar ORR results were seen with cisplatin and carboplatin, there appeared to be a longer median DOR among cisplatin-treated patients who received atezolizumab than among those who did not – 13.2 months and 8.3 months, respectively.
No such benefit from atezolizumab was seen in carboplatin-treated patients. The median DOR was 8.1 months among patients who received atezolizumab and 7.1 months among those who did not.
The overall safety profile for atezolizumab plus chemotherapy was consistent with prior reports of the combination. Treatment-related grade 3-5 adverse events were similar on the atezolizumab-chemotherapy arm and the placebo-chemotherapy arm.
The present and future
The investigators who presented the second interim analysis for OS of the IMvigor130 trial were appropriately modest in their conclusions. After all, the prespecified boundary for significant improvement in OS for the addition of atezolizumab to chemotherapy was not crossed. No change in guideline-based clinical practice would be appropriate at the present time.
The various exploratory analyses are hypothesis generating and invite potential mechanistic explanations. However, given the nonrandom allocation of patients to cisplatin- or carboplatin-based chemotherapy, unrecognized variables may have influenced any appearance of a difference in OS between the regimens.
In IMvigor130, treatment was given until unacceptable toxicity or disease progression. It is uncertain whether the current National Comprehensive Cancer Network category 1 recommendation of chemotherapy induction followed by immune checkpoint inhibitor maintenance therapy will prove superior to the IMvigor130 strategy.
Clearly – and concordant with current treatment guidelines – atezolizumab monotherapy can benefit some patients, though the response rate for atezolizumab monotherapy was lower than for chemotherapy (23.4% vs. 44.1%).
As noted by the session chair, Marina Chiara Garassino, MD, of the University of Chicago, the OS curves were initially superior for chemotherapy over atezolizumab. However, the apparent early OS benefit for chemotherapy dissipated over time and, among responders to atezolizumab, response duration was considerably longer than for chemotherapy.
IMvigor130 will ultimately have a final OS analysis to clarify the relative benefits of the various treatment strategies. Fortunately, this large phase 3 study will yield a treasure trove of data to inform future research and build on the advances of recent years for patients with advanced urothelial cancer.
IMvigor130 is sponsored by Hoffmann-La Roche. Dr. Davis, Dr. Galsky, and Dr. Garassino disclosed relationships with Hoffmann-La Roche and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
AACR 2021
Leveraging the microbiome to enhance cancer treatment
Andrea Facciabene, PhD, of the University of Pennsylvania, Philadelphia, and colleagues conducted a preclinical study in which vancomycin enhanced the efficacy of radiotherapy against melanoma and lung cancer. Now, researchers are conducting a clinical trial to determine if vancomycin can have the same effect in patients with non–small cell lung cancer.
Dr. Facciabene reviewed this research at the AACR Virtual Special Conference: Radiation Science and Medicine.
According to Dr. Facciabene, “gut microbiota” includes the more than 1,000 different strains of bacteria living in human intestines. He indicated that the average human has 10 times more bacteria than cells in the body and 150 times more genes in the gut microbiome than in the human genome.
In healthy individuals, the gut microbiota play a key role in intestinal function and digestive processes, modulation of hormones and vitamin secretion, energy extraction from food, and development and maintenance of a balanced immune system.
“Dysbiosis” is the term applied to a change in the composition, diversity, or metabolites of the microbiome from a healthy pattern to one associated with disease. Antibiotic therapy is a classic cause of dysbiosis, and dysbiosis has been implicated in a variety of inflammatory diseases.
The mechanisms by which the gut microbiome could influence systemic immunity is not known but is relevant to cancer therapy response. Augmenting the frequency and durability of response to immune-targeted treatments – potentially by manipulating the influence of gut microbiota on the immune system – could be highly impactful.
Gut microbiota and radiation-induced cell death
Immunogenic cell death – a process by which tumors die and release their intracellular molecular contents – is one of the mechanisms by which radiotherapy kills cancer cells.
Tumor cells succumbing to immunogenic cell death stimulate antigen presenting cells, such as dendritic cells, that engulf tumor antigens and cross-present them to CD8+ cytotoxic T lymphocytes. This process culminates in the generation of a specific immune response capable of killing the malignant cells in the irradiated area, but it also impacts distant nonirradiated tumors – an abscopal effect.
Dr. Facciabene and colleagues hypothesized that alterations of the gut microbiota could have an impact on the effect of radiotherapy. To investigate this, they studied mouse models of melanoma.
The team allowed B16-OVA tumors to grow for 9-12 days, then delivered a single dose of radiotherapy (21 Gy) to one – but not all – tumors. Simultaneously with the delivery of radiotherapy, the investigators started some animals on oral vancomycin. The team chose vancomycin because its effects are localized and impact the gut microbiota directly, without any known systemic effects.
Results showed that vancomycin significantly augmented the impact of radiotherapy in the irradiated area and was associated with regression of remote tumors.
The effects of the combination treatment on tumor volume were significantly greater than the effects of either treatment alone. Since manipulation of the gut microbiome potentiated radiotherapy effects both locally and distantly, the investigators concluded that immunogenic cell death may be involved in both the local and abscopal effects of radiotherapy.
When the experiment was repeated with a lung tumor model, similar findings were observed.
Involvement of cytotoxic T cells and interferon-gamma
Dr. Facciabene and colleagues found that the irradiated and unirradiated B16 OVA melanoma tumors treated with the radiotherapy-vancomycin combination were infiltrated by CD3+ and CD8+ T cells.
The investigators selectively depleted CD8+ T cells by pretreating the mice with an anti-CD8 monoclonal antibody. Depletion of CD8+ cells prior to administering radiotherapy plus vancomycin abrogated the antitumor effects of the combination treatment, demonstrating that the CD8+ T cells were required.
To characterize the antigen specificity of the tumor-infiltrating CD8+ T cells, Dr. Facciabene and colleagues used OVA MHC class 1 tetramer. Tumors from mice treated with vancomycin alone, radiotherapy alone, or the combination were dissected. Individual dendritic cells were assayed for OVA tetramer by flow cytometry.
The investigators found that tumors from mice treated with radiotherapy plus vancomycin had a significantly higher number of OVA-specific CD8+ T cells, in comparison with untreated tumors or tumors treated with either vancomycin alone or radiotherapy alone. Since antibody that impaired recognition of MHC class I peptides by T cells ablated the effect, it was clear that antigen recognition was vital.
Interferon-gamma (IFN-gamma) is known to play a critical role in both differentiation and effector functions of CD8+ cytolytic T cells in the antitumor immune response. To determine whether IFN-gamma is involved in the antitumor effects of the radiotherapy-vancomycin combination, the investigators measured intratumoral expression of IFN-gamma in the tumors 5 days after radiotherapy.
IFN-gamma messenger RNA expression levels were significantly elevated in the combination treatment group when compared with either treatment alone. In B16-OVA melanoma–challenged knockout mice, the enhancement of the radiotherapy effects by vancomycin was ablated.
The investigators concluded that vancomycin remodels the tumor microenvironment and increases the functionality of tumor-infiltrating, tumor-specific, CD8+ T cells. Furthermore, IFN-gamma is required to augment the radiotherapy-induced immune effect against the tumor.
Potential biochemical mediators of immune effects
The gut microbiota aid host digestion and generate a large repertoire of metabolites after defermentation of fiber. Short-chain fatty acids (SCFAs) constitute the major products of bacterial fermentation.
Acetic acid, propionic acid, and butyric acid represent 95% of total SCFAs present in the intestine. SCFAs are known to directly modulate cytokine production and dendritic cell function.
In their study, Dr. Facciabene and colleagues focused on butyric acid. Using mass spectroscopy, they demonstrated that vancomycin treatment reduces butyrate concentrations in tumor and tumor-draining lymph nodes by eradicating the major families of SCFA-producing Clostridia species.
To test whether supplementing butyrate could influence the synergy of the radiotherapy-vancomycin combination in vivo, the investigators added sodium butyrate to the mice’s drinking water when starting vancomycin treatment. The team then challenged the mice with B16-OVA tumors and treated them with radiotherapy.
In agreement with the group’s prior findings, vancomycin enhanced the tumor-inhibitory effects of radiotherapy, but dietary butyrate inhibited the benefit. The investigators found a significant decrease in the population of B16-OVA–presenting dendritic cells in the lymph nodes of mice receiving the supplemental butyrate.
Dr. Facciabene said these findings were supported by a recent publication. The authors observed that butyrate inhibited type I IFN expression in dendritic cells and radiotherapy-induced, tumor-specific cytotoxic T-cell immune responses without directly protecting tumor cells from the cytotoxic effects of radiotherapy.
Wide-ranging implications
Overall, Dr. Facciabene’s research has shown that:
- Vancomycin significantly enhances the tumor inhibitory effect of targeted radiation, including abscopal effects.
- The synergistic effects are dependent upon IFN-gamma and CD8+ cells.
- Depletion of some gut microbiome species increases antigen presentation by dendritic cells. This is mediated by SCFAs produced by certain bacterial families.
- There are promising new strategies to improve responses to radiotherapy, including targeting gut microbiota.
A clinical trial (NCT03546829) of vancomycin plus stereotactic body radiation in patients with locally advanced non–small cell lung cancer has been launched to investigate these findings further. Early data analysis has shown a significant impact of vancomycin on several species of gut microbiota, according to Dr. Facciabene.
Revolutionary results from immune-targeted therapy in the recent past have highlighted the important role the immune system can play in fighting cancer. Still, up to one-third of cancer patients fail to respond to overtly immune-targeted therapy.
The ability to inhibit cancer cells from evading immune surveillance by using new adjuvants – including those acting on non-traditional targets like gut microbiota – could herald the next major advances in cancer therapy. During his presentation, Dr. Facciabene gave participants an enticing hint of what could be coming for cancer patients in the years ahead.
Dr. Facciabene reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Andrea Facciabene, PhD, of the University of Pennsylvania, Philadelphia, and colleagues conducted a preclinical study in which vancomycin enhanced the efficacy of radiotherapy against melanoma and lung cancer. Now, researchers are conducting a clinical trial to determine if vancomycin can have the same effect in patients with non–small cell lung cancer.
Dr. Facciabene reviewed this research at the AACR Virtual Special Conference: Radiation Science and Medicine.
According to Dr. Facciabene, “gut microbiota” includes the more than 1,000 different strains of bacteria living in human intestines. He indicated that the average human has 10 times more bacteria than cells in the body and 150 times more genes in the gut microbiome than in the human genome.
In healthy individuals, the gut microbiota play a key role in intestinal function and digestive processes, modulation of hormones and vitamin secretion, energy extraction from food, and development and maintenance of a balanced immune system.
“Dysbiosis” is the term applied to a change in the composition, diversity, or metabolites of the microbiome from a healthy pattern to one associated with disease. Antibiotic therapy is a classic cause of dysbiosis, and dysbiosis has been implicated in a variety of inflammatory diseases.
The mechanisms by which the gut microbiome could influence systemic immunity is not known but is relevant to cancer therapy response. Augmenting the frequency and durability of response to immune-targeted treatments – potentially by manipulating the influence of gut microbiota on the immune system – could be highly impactful.
Gut microbiota and radiation-induced cell death
Immunogenic cell death – a process by which tumors die and release their intracellular molecular contents – is one of the mechanisms by which radiotherapy kills cancer cells.
Tumor cells succumbing to immunogenic cell death stimulate antigen presenting cells, such as dendritic cells, that engulf tumor antigens and cross-present them to CD8+ cytotoxic T lymphocytes. This process culminates in the generation of a specific immune response capable of killing the malignant cells in the irradiated area, but it also impacts distant nonirradiated tumors – an abscopal effect.
Dr. Facciabene and colleagues hypothesized that alterations of the gut microbiota could have an impact on the effect of radiotherapy. To investigate this, they studied mouse models of melanoma.
The team allowed B16-OVA tumors to grow for 9-12 days, then delivered a single dose of radiotherapy (21 Gy) to one – but not all – tumors. Simultaneously with the delivery of radiotherapy, the investigators started some animals on oral vancomycin. The team chose vancomycin because its effects are localized and impact the gut microbiota directly, without any known systemic effects.
Results showed that vancomycin significantly augmented the impact of radiotherapy in the irradiated area and was associated with regression of remote tumors.
The effects of the combination treatment on tumor volume were significantly greater than the effects of either treatment alone. Since manipulation of the gut microbiome potentiated radiotherapy effects both locally and distantly, the investigators concluded that immunogenic cell death may be involved in both the local and abscopal effects of radiotherapy.
When the experiment was repeated with a lung tumor model, similar findings were observed.
Involvement of cytotoxic T cells and interferon-gamma
Dr. Facciabene and colleagues found that the irradiated and unirradiated B16 OVA melanoma tumors treated with the radiotherapy-vancomycin combination were infiltrated by CD3+ and CD8+ T cells.
The investigators selectively depleted CD8+ T cells by pretreating the mice with an anti-CD8 monoclonal antibody. Depletion of CD8+ cells prior to administering radiotherapy plus vancomycin abrogated the antitumor effects of the combination treatment, demonstrating that the CD8+ T cells were required.
To characterize the antigen specificity of the tumor-infiltrating CD8+ T cells, Dr. Facciabene and colleagues used OVA MHC class 1 tetramer. Tumors from mice treated with vancomycin alone, radiotherapy alone, or the combination were dissected. Individual dendritic cells were assayed for OVA tetramer by flow cytometry.
The investigators found that tumors from mice treated with radiotherapy plus vancomycin had a significantly higher number of OVA-specific CD8+ T cells, in comparison with untreated tumors or tumors treated with either vancomycin alone or radiotherapy alone. Since antibody that impaired recognition of MHC class I peptides by T cells ablated the effect, it was clear that antigen recognition was vital.
Interferon-gamma (IFN-gamma) is known to play a critical role in both differentiation and effector functions of CD8+ cytolytic T cells in the antitumor immune response. To determine whether IFN-gamma is involved in the antitumor effects of the radiotherapy-vancomycin combination, the investigators measured intratumoral expression of IFN-gamma in the tumors 5 days after radiotherapy.
IFN-gamma messenger RNA expression levels were significantly elevated in the combination treatment group when compared with either treatment alone. In B16-OVA melanoma–challenged knockout mice, the enhancement of the radiotherapy effects by vancomycin was ablated.
The investigators concluded that vancomycin remodels the tumor microenvironment and increases the functionality of tumor-infiltrating, tumor-specific, CD8+ T cells. Furthermore, IFN-gamma is required to augment the radiotherapy-induced immune effect against the tumor.
Potential biochemical mediators of immune effects
The gut microbiota aid host digestion and generate a large repertoire of metabolites after defermentation of fiber. Short-chain fatty acids (SCFAs) constitute the major products of bacterial fermentation.
Acetic acid, propionic acid, and butyric acid represent 95% of total SCFAs present in the intestine. SCFAs are known to directly modulate cytokine production and dendritic cell function.
In their study, Dr. Facciabene and colleagues focused on butyric acid. Using mass spectroscopy, they demonstrated that vancomycin treatment reduces butyrate concentrations in tumor and tumor-draining lymph nodes by eradicating the major families of SCFA-producing Clostridia species.
To test whether supplementing butyrate could influence the synergy of the radiotherapy-vancomycin combination in vivo, the investigators added sodium butyrate to the mice’s drinking water when starting vancomycin treatment. The team then challenged the mice with B16-OVA tumors and treated them with radiotherapy.
In agreement with the group’s prior findings, vancomycin enhanced the tumor-inhibitory effects of radiotherapy, but dietary butyrate inhibited the benefit. The investigators found a significant decrease in the population of B16-OVA–presenting dendritic cells in the lymph nodes of mice receiving the supplemental butyrate.
Dr. Facciabene said these findings were supported by a recent publication. The authors observed that butyrate inhibited type I IFN expression in dendritic cells and radiotherapy-induced, tumor-specific cytotoxic T-cell immune responses without directly protecting tumor cells from the cytotoxic effects of radiotherapy.
Wide-ranging implications
Overall, Dr. Facciabene’s research has shown that:
- Vancomycin significantly enhances the tumor inhibitory effect of targeted radiation, including abscopal effects.
- The synergistic effects are dependent upon IFN-gamma and CD8+ cells.
- Depletion of some gut microbiome species increases antigen presentation by dendritic cells. This is mediated by SCFAs produced by certain bacterial families.
- There are promising new strategies to improve responses to radiotherapy, including targeting gut microbiota.
A clinical trial (NCT03546829) of vancomycin plus stereotactic body radiation in patients with locally advanced non–small cell lung cancer has been launched to investigate these findings further. Early data analysis has shown a significant impact of vancomycin on several species of gut microbiota, according to Dr. Facciabene.
Revolutionary results from immune-targeted therapy in the recent past have highlighted the important role the immune system can play in fighting cancer. Still, up to one-third of cancer patients fail to respond to overtly immune-targeted therapy.
The ability to inhibit cancer cells from evading immune surveillance by using new adjuvants – including those acting on non-traditional targets like gut microbiota – could herald the next major advances in cancer therapy. During his presentation, Dr. Facciabene gave participants an enticing hint of what could be coming for cancer patients in the years ahead.
Dr. Facciabene reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Andrea Facciabene, PhD, of the University of Pennsylvania, Philadelphia, and colleagues conducted a preclinical study in which vancomycin enhanced the efficacy of radiotherapy against melanoma and lung cancer. Now, researchers are conducting a clinical trial to determine if vancomycin can have the same effect in patients with non–small cell lung cancer.
Dr. Facciabene reviewed this research at the AACR Virtual Special Conference: Radiation Science and Medicine.
According to Dr. Facciabene, “gut microbiota” includes the more than 1,000 different strains of bacteria living in human intestines. He indicated that the average human has 10 times more bacteria than cells in the body and 150 times more genes in the gut microbiome than in the human genome.
In healthy individuals, the gut microbiota play a key role in intestinal function and digestive processes, modulation of hormones and vitamin secretion, energy extraction from food, and development and maintenance of a balanced immune system.
“Dysbiosis” is the term applied to a change in the composition, diversity, or metabolites of the microbiome from a healthy pattern to one associated with disease. Antibiotic therapy is a classic cause of dysbiosis, and dysbiosis has been implicated in a variety of inflammatory diseases.
The mechanisms by which the gut microbiome could influence systemic immunity is not known but is relevant to cancer therapy response. Augmenting the frequency and durability of response to immune-targeted treatments – potentially by manipulating the influence of gut microbiota on the immune system – could be highly impactful.
Gut microbiota and radiation-induced cell death
Immunogenic cell death – a process by which tumors die and release their intracellular molecular contents – is one of the mechanisms by which radiotherapy kills cancer cells.
Tumor cells succumbing to immunogenic cell death stimulate antigen presenting cells, such as dendritic cells, that engulf tumor antigens and cross-present them to CD8+ cytotoxic T lymphocytes. This process culminates in the generation of a specific immune response capable of killing the malignant cells in the irradiated area, but it also impacts distant nonirradiated tumors – an abscopal effect.
Dr. Facciabene and colleagues hypothesized that alterations of the gut microbiota could have an impact on the effect of radiotherapy. To investigate this, they studied mouse models of melanoma.
The team allowed B16-OVA tumors to grow for 9-12 days, then delivered a single dose of radiotherapy (21 Gy) to one – but not all – tumors. Simultaneously with the delivery of radiotherapy, the investigators started some animals on oral vancomycin. The team chose vancomycin because its effects are localized and impact the gut microbiota directly, without any known systemic effects.
Results showed that vancomycin significantly augmented the impact of radiotherapy in the irradiated area and was associated with regression of remote tumors.
The effects of the combination treatment on tumor volume were significantly greater than the effects of either treatment alone. Since manipulation of the gut microbiome potentiated radiotherapy effects both locally and distantly, the investigators concluded that immunogenic cell death may be involved in both the local and abscopal effects of radiotherapy.
When the experiment was repeated with a lung tumor model, similar findings were observed.
Involvement of cytotoxic T cells and interferon-gamma
Dr. Facciabene and colleagues found that the irradiated and unirradiated B16 OVA melanoma tumors treated with the radiotherapy-vancomycin combination were infiltrated by CD3+ and CD8+ T cells.
The investigators selectively depleted CD8+ T cells by pretreating the mice with an anti-CD8 monoclonal antibody. Depletion of CD8+ cells prior to administering radiotherapy plus vancomycin abrogated the antitumor effects of the combination treatment, demonstrating that the CD8+ T cells were required.
To characterize the antigen specificity of the tumor-infiltrating CD8+ T cells, Dr. Facciabene and colleagues used OVA MHC class 1 tetramer. Tumors from mice treated with vancomycin alone, radiotherapy alone, or the combination were dissected. Individual dendritic cells were assayed for OVA tetramer by flow cytometry.
The investigators found that tumors from mice treated with radiotherapy plus vancomycin had a significantly higher number of OVA-specific CD8+ T cells, in comparison with untreated tumors or tumors treated with either vancomycin alone or radiotherapy alone. Since antibody that impaired recognition of MHC class I peptides by T cells ablated the effect, it was clear that antigen recognition was vital.
Interferon-gamma (IFN-gamma) is known to play a critical role in both differentiation and effector functions of CD8+ cytolytic T cells in the antitumor immune response. To determine whether IFN-gamma is involved in the antitumor effects of the radiotherapy-vancomycin combination, the investigators measured intratumoral expression of IFN-gamma in the tumors 5 days after radiotherapy.
IFN-gamma messenger RNA expression levels were significantly elevated in the combination treatment group when compared with either treatment alone. In B16-OVA melanoma–challenged knockout mice, the enhancement of the radiotherapy effects by vancomycin was ablated.
The investigators concluded that vancomycin remodels the tumor microenvironment and increases the functionality of tumor-infiltrating, tumor-specific, CD8+ T cells. Furthermore, IFN-gamma is required to augment the radiotherapy-induced immune effect against the tumor.
Potential biochemical mediators of immune effects
The gut microbiota aid host digestion and generate a large repertoire of metabolites after defermentation of fiber. Short-chain fatty acids (SCFAs) constitute the major products of bacterial fermentation.
Acetic acid, propionic acid, and butyric acid represent 95% of total SCFAs present in the intestine. SCFAs are known to directly modulate cytokine production and dendritic cell function.
In their study, Dr. Facciabene and colleagues focused on butyric acid. Using mass spectroscopy, they demonstrated that vancomycin treatment reduces butyrate concentrations in tumor and tumor-draining lymph nodes by eradicating the major families of SCFA-producing Clostridia species.
To test whether supplementing butyrate could influence the synergy of the radiotherapy-vancomycin combination in vivo, the investigators added sodium butyrate to the mice’s drinking water when starting vancomycin treatment. The team then challenged the mice with B16-OVA tumors and treated them with radiotherapy.
In agreement with the group’s prior findings, vancomycin enhanced the tumor-inhibitory effects of radiotherapy, but dietary butyrate inhibited the benefit. The investigators found a significant decrease in the population of B16-OVA–presenting dendritic cells in the lymph nodes of mice receiving the supplemental butyrate.
Dr. Facciabene said these findings were supported by a recent publication. The authors observed that butyrate inhibited type I IFN expression in dendritic cells and radiotherapy-induced, tumor-specific cytotoxic T-cell immune responses without directly protecting tumor cells from the cytotoxic effects of radiotherapy.
Wide-ranging implications
Overall, Dr. Facciabene’s research has shown that:
- Vancomycin significantly enhances the tumor inhibitory effect of targeted radiation, including abscopal effects.
- The synergistic effects are dependent upon IFN-gamma and CD8+ cells.
- Depletion of some gut microbiome species increases antigen presentation by dendritic cells. This is mediated by SCFAs produced by certain bacterial families.
- There are promising new strategies to improve responses to radiotherapy, including targeting gut microbiota.
A clinical trial (NCT03546829) of vancomycin plus stereotactic body radiation in patients with locally advanced non–small cell lung cancer has been launched to investigate these findings further. Early data analysis has shown a significant impact of vancomycin on several species of gut microbiota, according to Dr. Facciabene.
Revolutionary results from immune-targeted therapy in the recent past have highlighted the important role the immune system can play in fighting cancer. Still, up to one-third of cancer patients fail to respond to overtly immune-targeted therapy.
The ability to inhibit cancer cells from evading immune surveillance by using new adjuvants – including those acting on non-traditional targets like gut microbiota – could herald the next major advances in cancer therapy. During his presentation, Dr. Facciabene gave participants an enticing hint of what could be coming for cancer patients in the years ahead.
Dr. Facciabene reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM AACR: RADIATION SCIENCE AND MEDICINE
Steroid-refractory pneumonitis from ICIs: Experience at major centers
Pneumonitis is an uncommon and potentially life-threatening complication of immune checkpoint inhibitor (ICI) therapy. A fraction of patients with ICI-related pneumonitis fail to respond to initial therapy with high-dose systemic steroids.
The recently published experiences at two major cancer centers shed light on the outcomes from treatment and can provide some advice to clinicians for dealing with affected patients.
The Johns Hopkins experience
Because ICI-related pneumonitis typically improves within 48-72 hours of steroid therapy, at Johns Hopkins University, Baltimore, steroid-refractory pneumonitis is defined as pneumonitis that demonstrates no clinical improvement after high-dose corticosteroids for 2-14 days. If the immune toxicity–specialized, multidisciplinary management team implements additional immunosuppressive therapy, that is regarded as confirmatory evidence.
Aanika Balaji, a medical student at Johns Hopkins University, and colleagues retrospectively summarized the clinical course of 12 patients with ICI-related pneumonitis between 2011 and 2020. Clinical improvement with subsequent treatment was evidenced by reduction in either level of care or oxygen requirements.
Three-quarters of the patients were current or former smokers, and the same proportion had lung cancer. Most patients (91.6%) had received chemotherapy, 58.3% had prior chest radiotherapy, and 58.3% had achieved partial response or stable disease with an ICI.
Steroid-refractory ICI-related pneumonitis developed between 40 and 127 days (median, 85 days) after the first dose of ICI therapy. Subsequent immunosuppressive management included IVIg, infliximab, or the combination, in addition to ICU-level supportive care.
Among the seven patients who received IVIg alone, two patients (29%) achieved clinical improvement and hospital discharge. The remainder died.
The two patients treated with infliximab and the three patients treated with sequential IVIg and infliximab died. All deaths were attributed to ICI-related pneumonitis or infectious complications.
Overall, clinically relevant findings were:
- Steroid-refractory ICI-related pneumonitis was seen in 18.5% of patients referred for multidisciplinary care.
- Steroid-refractory ICI-related pneumonitis occurred at a median of 85 days into a patient’s ICI treatment.
- Some patients improved clinically after IVIg therapy, but mortality was high overall.
- Infliximab therapy, alone or in combination with IVIg, was ineffective.
The Memorial Sloan Kettering experience
Jason Beattie, MD, of Memorial Sloan Kettering Cancer Center, New York, and colleagues performed a retrospective study of patients who had pneumonitis after ICI therapy and/or received immune modulator therapy after corticosteroids in the setting of ICI cancer treatment.
Manual record review was performed to exclude cases of pneumonitis from other causes. The period reviewed was roughly contemporaneous with the Johns Hopkins series.
Patients with ICI-related pneumonitis were divided into “steroid refractory” (i.e., no response to high-dose corticosteroids) or “steroid resistant” (i.e., initial response, followed by worsening) categories.
The researchers identified 26 patients with ICI-related pneumonitis, all of whom had advanced malignancy (8 lung cancer, 4 malignant melanoma, 4 renal cell cancer, and 10 “other” cancers).
A majority of patients (85%) were current or former smokers, 73% had received ICI monotherapy, 35% had received prior chest radiation at a median interval of 4.9 months prior to pneumonitis onset, and 27% had preexisting pulmonary disease.
Twelve patients (46%) had steroid-refractory ICI-related pneumonitis, and 14 (54%) had steroid-resistant ICI-related pneumonitis.
The two groups differed in time to pneumonitis onset (a median of 68 days in the refractory group and 182 days in the resistant group) and time to immune modulator therapy after beginning steroids (median 7 days and 2.9 months, respectively). In the steroid-refractory cases, pneumonitis was more severe.
In addition to corticosteroids, most patients received infliximab monotherapy or infliximab with mycophenolate mofetil. In contrast to the Johns Hopkins series, IVIg was not used in the Memorial Sloan Kettering cases.
Outcomes from immune modulators were graded based on clinical evidence (progress notes, oxygen requirements, level of care, radiologic information, etc.) of resolution of pneumonitis on imaging at least 8 weeks after cessation of steroids and immune modulator therapy, durable improvement for at least 8 weeks after immune modulator therapy, transient improvement followed by pneumonitis relapse or inadequate follow-up because of death or hospice referral, or no improvement.
Ten patients (38%) had durable improvement of ICI-related pneumonitis, of whom three (12%) had complete resolution. Two of the patients with complete resolution had steroid-refractory pneumonitis, both of whom had received infliximab followed by mycophenolate mofetil.
Among the seven patients with durable improvement, four remained alive on immune modulators. Time to resolution of pneumonitis was protracted, ranging from 2.3 months to 8.4 months in the steroid-refractory patients.
Durable response was less common with steroid-refractory (25%) than steroid-resistant (50%) disease, with a significant difference in 90-day survival of 25% and 71%, respectively.
Among the 13 (50%) patients with transient improvement in ICI-related pneumonitis, 8 ultimately died, either because of recurrent ICI-related pneumonitis or infection. All three patients with no improvement from immune modulators died.
The 90-day all-cause mortality was 50%, with durable pneumonitis improvement and freedom from severe infectious complications occurring in only about a third of patients.
Lessons for clinicians
The National Comprehensive Cancer Network, the Society for Immunotherapy of Cancer, and the European Society of Medical Oncology have all published guidelines and recommendations for immunosuppression for steroid-refractory adverse events from ICIs.
Unfortunately, there is little experience with steroid-unresponsive ICI-related pneumonitis. The ideal sequence, dose, and duration of additional immune modulator therapy for ICI-related pneumonitis are unclear and may differ from the approaches to other immune-related toxicities.
This is important because, as suggested in an editorial by Margaret Gatti-Mays, MD, and James L. Gulley, MD, PhD, it is likely that ICI-related pneumonitis will be seen more in routine practice than in clinical trial populations. In addition, across all tumor types, ICI-related pneumonitis is the most common cause of ICI-associated death from toxicity.
The retrospective studies from Johns Hopkins and Memorial Sloan Kettering constitute the largest published experience with ICI-related pneumonitis and yield important clinical insights.
Uniform definitions of potentially important patient subgroups (e.g., steroid refractory vs. steroid resistant) are needed. The steroid-refractory and steroid-resistant subgroups have distinctly different clinical features and outcomes. Uniformity in the subgroup definitions would be a useful starting point from both clinical and research perspectives.
Preferred treatment choices need to be tested systematically in multi-institutional studies. Any potential impact of treatment for ICI-related pneumonitis on antitumor immune control should be identified.
Endpoints of interest need to be defined and measured prospectively. All-cause mortality after 90 days is important, but, as the authors of both reviews noted, there are vitally important narratives and differences in functionality that are completely concealed by restricting the focus to mortality.
Potential causal relationships with antecedent exposure to tobacco, radiation, intrathoracic tumor burden, or other factors need to be defined.
Clinicians need predictive biomarkers for ICI-related pneumonitis (e.g., in peripheral blood, pulmonary function testing, or bronchoscopy specimens). At-risk patients may benefit from early intervention.
The limitations of single-institution record reviews in guiding real-world patient management notwithstanding, these reviews illustrate the value of registries and prospective studies to guide the path forward. Taking these next steps will ensure for our patients that the success of immune-targeted therapy against their cancer never becomes a Pyrrhic victory.
The Johns Hopkins investigators and the editorialists reported having no disclosures. The Memorial Sloan Kettering investigators disclosed relationships with Targeted Oncology, Merck, Array BioPharma, Novartis, and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Pneumonitis is an uncommon and potentially life-threatening complication of immune checkpoint inhibitor (ICI) therapy. A fraction of patients with ICI-related pneumonitis fail to respond to initial therapy with high-dose systemic steroids.
The recently published experiences at two major cancer centers shed light on the outcomes from treatment and can provide some advice to clinicians for dealing with affected patients.
The Johns Hopkins experience
Because ICI-related pneumonitis typically improves within 48-72 hours of steroid therapy, at Johns Hopkins University, Baltimore, steroid-refractory pneumonitis is defined as pneumonitis that demonstrates no clinical improvement after high-dose corticosteroids for 2-14 days. If the immune toxicity–specialized, multidisciplinary management team implements additional immunosuppressive therapy, that is regarded as confirmatory evidence.
Aanika Balaji, a medical student at Johns Hopkins University, and colleagues retrospectively summarized the clinical course of 12 patients with ICI-related pneumonitis between 2011 and 2020. Clinical improvement with subsequent treatment was evidenced by reduction in either level of care or oxygen requirements.
Three-quarters of the patients were current or former smokers, and the same proportion had lung cancer. Most patients (91.6%) had received chemotherapy, 58.3% had prior chest radiotherapy, and 58.3% had achieved partial response or stable disease with an ICI.
Steroid-refractory ICI-related pneumonitis developed between 40 and 127 days (median, 85 days) after the first dose of ICI therapy. Subsequent immunosuppressive management included IVIg, infliximab, or the combination, in addition to ICU-level supportive care.
Among the seven patients who received IVIg alone, two patients (29%) achieved clinical improvement and hospital discharge. The remainder died.
The two patients treated with infliximab and the three patients treated with sequential IVIg and infliximab died. All deaths were attributed to ICI-related pneumonitis or infectious complications.
Overall, clinically relevant findings were:
- Steroid-refractory ICI-related pneumonitis was seen in 18.5% of patients referred for multidisciplinary care.
- Steroid-refractory ICI-related pneumonitis occurred at a median of 85 days into a patient’s ICI treatment.
- Some patients improved clinically after IVIg therapy, but mortality was high overall.
- Infliximab therapy, alone or in combination with IVIg, was ineffective.
The Memorial Sloan Kettering experience
Jason Beattie, MD, of Memorial Sloan Kettering Cancer Center, New York, and colleagues performed a retrospective study of patients who had pneumonitis after ICI therapy and/or received immune modulator therapy after corticosteroids in the setting of ICI cancer treatment.
Manual record review was performed to exclude cases of pneumonitis from other causes. The period reviewed was roughly contemporaneous with the Johns Hopkins series.
Patients with ICI-related pneumonitis were divided into “steroid refractory” (i.e., no response to high-dose corticosteroids) or “steroid resistant” (i.e., initial response, followed by worsening) categories.
The researchers identified 26 patients with ICI-related pneumonitis, all of whom had advanced malignancy (8 lung cancer, 4 malignant melanoma, 4 renal cell cancer, and 10 “other” cancers).
A majority of patients (85%) were current or former smokers, 73% had received ICI monotherapy, 35% had received prior chest radiation at a median interval of 4.9 months prior to pneumonitis onset, and 27% had preexisting pulmonary disease.
Twelve patients (46%) had steroid-refractory ICI-related pneumonitis, and 14 (54%) had steroid-resistant ICI-related pneumonitis.
The two groups differed in time to pneumonitis onset (a median of 68 days in the refractory group and 182 days in the resistant group) and time to immune modulator therapy after beginning steroids (median 7 days and 2.9 months, respectively). In the steroid-refractory cases, pneumonitis was more severe.
In addition to corticosteroids, most patients received infliximab monotherapy or infliximab with mycophenolate mofetil. In contrast to the Johns Hopkins series, IVIg was not used in the Memorial Sloan Kettering cases.
Outcomes from immune modulators were graded based on clinical evidence (progress notes, oxygen requirements, level of care, radiologic information, etc.) of resolution of pneumonitis on imaging at least 8 weeks after cessation of steroids and immune modulator therapy, durable improvement for at least 8 weeks after immune modulator therapy, transient improvement followed by pneumonitis relapse or inadequate follow-up because of death or hospice referral, or no improvement.
Ten patients (38%) had durable improvement of ICI-related pneumonitis, of whom three (12%) had complete resolution. Two of the patients with complete resolution had steroid-refractory pneumonitis, both of whom had received infliximab followed by mycophenolate mofetil.
Among the seven patients with durable improvement, four remained alive on immune modulators. Time to resolution of pneumonitis was protracted, ranging from 2.3 months to 8.4 months in the steroid-refractory patients.
Durable response was less common with steroid-refractory (25%) than steroid-resistant (50%) disease, with a significant difference in 90-day survival of 25% and 71%, respectively.
Among the 13 (50%) patients with transient improvement in ICI-related pneumonitis, 8 ultimately died, either because of recurrent ICI-related pneumonitis or infection. All three patients with no improvement from immune modulators died.
The 90-day all-cause mortality was 50%, with durable pneumonitis improvement and freedom from severe infectious complications occurring in only about a third of patients.
Lessons for clinicians
The National Comprehensive Cancer Network, the Society for Immunotherapy of Cancer, and the European Society of Medical Oncology have all published guidelines and recommendations for immunosuppression for steroid-refractory adverse events from ICIs.
Unfortunately, there is little experience with steroid-unresponsive ICI-related pneumonitis. The ideal sequence, dose, and duration of additional immune modulator therapy for ICI-related pneumonitis are unclear and may differ from the approaches to other immune-related toxicities.
This is important because, as suggested in an editorial by Margaret Gatti-Mays, MD, and James L. Gulley, MD, PhD, it is likely that ICI-related pneumonitis will be seen more in routine practice than in clinical trial populations. In addition, across all tumor types, ICI-related pneumonitis is the most common cause of ICI-associated death from toxicity.
The retrospective studies from Johns Hopkins and Memorial Sloan Kettering constitute the largest published experience with ICI-related pneumonitis and yield important clinical insights.
Uniform definitions of potentially important patient subgroups (e.g., steroid refractory vs. steroid resistant) are needed. The steroid-refractory and steroid-resistant subgroups have distinctly different clinical features and outcomes. Uniformity in the subgroup definitions would be a useful starting point from both clinical and research perspectives.
Preferred treatment choices need to be tested systematically in multi-institutional studies. Any potential impact of treatment for ICI-related pneumonitis on antitumor immune control should be identified.
Endpoints of interest need to be defined and measured prospectively. All-cause mortality after 90 days is important, but, as the authors of both reviews noted, there are vitally important narratives and differences in functionality that are completely concealed by restricting the focus to mortality.
Potential causal relationships with antecedent exposure to tobacco, radiation, intrathoracic tumor burden, or other factors need to be defined.
Clinicians need predictive biomarkers for ICI-related pneumonitis (e.g., in peripheral blood, pulmonary function testing, or bronchoscopy specimens). At-risk patients may benefit from early intervention.
The limitations of single-institution record reviews in guiding real-world patient management notwithstanding, these reviews illustrate the value of registries and prospective studies to guide the path forward. Taking these next steps will ensure for our patients that the success of immune-targeted therapy against their cancer never becomes a Pyrrhic victory.
The Johns Hopkins investigators and the editorialists reported having no disclosures. The Memorial Sloan Kettering investigators disclosed relationships with Targeted Oncology, Merck, Array BioPharma, Novartis, and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Pneumonitis is an uncommon and potentially life-threatening complication of immune checkpoint inhibitor (ICI) therapy. A fraction of patients with ICI-related pneumonitis fail to respond to initial therapy with high-dose systemic steroids.
The recently published experiences at two major cancer centers shed light on the outcomes from treatment and can provide some advice to clinicians for dealing with affected patients.
The Johns Hopkins experience
Because ICI-related pneumonitis typically improves within 48-72 hours of steroid therapy, at Johns Hopkins University, Baltimore, steroid-refractory pneumonitis is defined as pneumonitis that demonstrates no clinical improvement after high-dose corticosteroids for 2-14 days. If the immune toxicity–specialized, multidisciplinary management team implements additional immunosuppressive therapy, that is regarded as confirmatory evidence.
Aanika Balaji, a medical student at Johns Hopkins University, and colleagues retrospectively summarized the clinical course of 12 patients with ICI-related pneumonitis between 2011 and 2020. Clinical improvement with subsequent treatment was evidenced by reduction in either level of care or oxygen requirements.
Three-quarters of the patients were current or former smokers, and the same proportion had lung cancer. Most patients (91.6%) had received chemotherapy, 58.3% had prior chest radiotherapy, and 58.3% had achieved partial response or stable disease with an ICI.
Steroid-refractory ICI-related pneumonitis developed between 40 and 127 days (median, 85 days) after the first dose of ICI therapy. Subsequent immunosuppressive management included IVIg, infliximab, or the combination, in addition to ICU-level supportive care.
Among the seven patients who received IVIg alone, two patients (29%) achieved clinical improvement and hospital discharge. The remainder died.
The two patients treated with infliximab and the three patients treated with sequential IVIg and infliximab died. All deaths were attributed to ICI-related pneumonitis or infectious complications.
Overall, clinically relevant findings were:
- Steroid-refractory ICI-related pneumonitis was seen in 18.5% of patients referred for multidisciplinary care.
- Steroid-refractory ICI-related pneumonitis occurred at a median of 85 days into a patient’s ICI treatment.
- Some patients improved clinically after IVIg therapy, but mortality was high overall.
- Infliximab therapy, alone or in combination with IVIg, was ineffective.
The Memorial Sloan Kettering experience
Jason Beattie, MD, of Memorial Sloan Kettering Cancer Center, New York, and colleagues performed a retrospective study of patients who had pneumonitis after ICI therapy and/or received immune modulator therapy after corticosteroids in the setting of ICI cancer treatment.
Manual record review was performed to exclude cases of pneumonitis from other causes. The period reviewed was roughly contemporaneous with the Johns Hopkins series.
Patients with ICI-related pneumonitis were divided into “steroid refractory” (i.e., no response to high-dose corticosteroids) or “steroid resistant” (i.e., initial response, followed by worsening) categories.
The researchers identified 26 patients with ICI-related pneumonitis, all of whom had advanced malignancy (8 lung cancer, 4 malignant melanoma, 4 renal cell cancer, and 10 “other” cancers).
A majority of patients (85%) were current or former smokers, 73% had received ICI monotherapy, 35% had received prior chest radiation at a median interval of 4.9 months prior to pneumonitis onset, and 27% had preexisting pulmonary disease.
Twelve patients (46%) had steroid-refractory ICI-related pneumonitis, and 14 (54%) had steroid-resistant ICI-related pneumonitis.
The two groups differed in time to pneumonitis onset (a median of 68 days in the refractory group and 182 days in the resistant group) and time to immune modulator therapy after beginning steroids (median 7 days and 2.9 months, respectively). In the steroid-refractory cases, pneumonitis was more severe.
In addition to corticosteroids, most patients received infliximab monotherapy or infliximab with mycophenolate mofetil. In contrast to the Johns Hopkins series, IVIg was not used in the Memorial Sloan Kettering cases.
Outcomes from immune modulators were graded based on clinical evidence (progress notes, oxygen requirements, level of care, radiologic information, etc.) of resolution of pneumonitis on imaging at least 8 weeks after cessation of steroids and immune modulator therapy, durable improvement for at least 8 weeks after immune modulator therapy, transient improvement followed by pneumonitis relapse or inadequate follow-up because of death or hospice referral, or no improvement.
Ten patients (38%) had durable improvement of ICI-related pneumonitis, of whom three (12%) had complete resolution. Two of the patients with complete resolution had steroid-refractory pneumonitis, both of whom had received infliximab followed by mycophenolate mofetil.
Among the seven patients with durable improvement, four remained alive on immune modulators. Time to resolution of pneumonitis was protracted, ranging from 2.3 months to 8.4 months in the steroid-refractory patients.
Durable response was less common with steroid-refractory (25%) than steroid-resistant (50%) disease, with a significant difference in 90-day survival of 25% and 71%, respectively.
Among the 13 (50%) patients with transient improvement in ICI-related pneumonitis, 8 ultimately died, either because of recurrent ICI-related pneumonitis or infection. All three patients with no improvement from immune modulators died.
The 90-day all-cause mortality was 50%, with durable pneumonitis improvement and freedom from severe infectious complications occurring in only about a third of patients.
Lessons for clinicians
The National Comprehensive Cancer Network, the Society for Immunotherapy of Cancer, and the European Society of Medical Oncology have all published guidelines and recommendations for immunosuppression for steroid-refractory adverse events from ICIs.
Unfortunately, there is little experience with steroid-unresponsive ICI-related pneumonitis. The ideal sequence, dose, and duration of additional immune modulator therapy for ICI-related pneumonitis are unclear and may differ from the approaches to other immune-related toxicities.
This is important because, as suggested in an editorial by Margaret Gatti-Mays, MD, and James L. Gulley, MD, PhD, it is likely that ICI-related pneumonitis will be seen more in routine practice than in clinical trial populations. In addition, across all tumor types, ICI-related pneumonitis is the most common cause of ICI-associated death from toxicity.
The retrospective studies from Johns Hopkins and Memorial Sloan Kettering constitute the largest published experience with ICI-related pneumonitis and yield important clinical insights.
Uniform definitions of potentially important patient subgroups (e.g., steroid refractory vs. steroid resistant) are needed. The steroid-refractory and steroid-resistant subgroups have distinctly different clinical features and outcomes. Uniformity in the subgroup definitions would be a useful starting point from both clinical and research perspectives.
Preferred treatment choices need to be tested systematically in multi-institutional studies. Any potential impact of treatment for ICI-related pneumonitis on antitumor immune control should be identified.
Endpoints of interest need to be defined and measured prospectively. All-cause mortality after 90 days is important, but, as the authors of both reviews noted, there are vitally important narratives and differences in functionality that are completely concealed by restricting the focus to mortality.
Potential causal relationships with antecedent exposure to tobacco, radiation, intrathoracic tumor burden, or other factors need to be defined.
Clinicians need predictive biomarkers for ICI-related pneumonitis (e.g., in peripheral blood, pulmonary function testing, or bronchoscopy specimens). At-risk patients may benefit from early intervention.
The limitations of single-institution record reviews in guiding real-world patient management notwithstanding, these reviews illustrate the value of registries and prospective studies to guide the path forward. Taking these next steps will ensure for our patients that the success of immune-targeted therapy against their cancer never becomes a Pyrrhic victory.
The Johns Hopkins investigators and the editorialists reported having no disclosures. The Memorial Sloan Kettering investigators disclosed relationships with Targeted Oncology, Merck, Array BioPharma, Novartis, and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
mCODE: Improving data sharing to enhance cancer care
An initiative designed to improve sharing of patient data may provide “tremendous benefits” in cancer care and research, according to authors of a review article.
The goals of the initiative, called Minimal Common Oncology Data Elements (mCODE), were to identify the data elements in electronic health records that are “essential” for making treatment decisions and create “a standardized computable data format” that would improve the exchange of data across EHRs, according to the mCODE website.
Travis J. Osterman, DO, of Vanderbilt University Medical Center in Nashville, Tenn., and colleagues described the mCODE initiative in a review published in JCO Clinical Cancer Informatics.
At present, commercially available EHRs are poorly designed to support modern oncology workflow, requiring laborious data entry and lacking a common library of oncology-specific discrete data elements. As an example, most EHRs poorly support the needs of precision oncology and clinical genetics, since next-generation sequencing and genetic test results are almost universally reported in PDF files.
In addition, basic, operational oncology data (e.g., cancer staging, adverse event documentation, response to treatment, etc.) are captured in EHRs primarily as an unstructured narrative.
Computable, analytical data are found for only the small percentage of patients in clinical trials. Even then, some degree of manual data abstraction is regularly required.
Interoperability of EHRs between practices and health care institutions is often so poor that the transfer of basic cancer-related information as analyzable data is difficult or even impossible.
Making progress: The 21st Century Cures Act
The American Society of Clinical Oncology has a more than 15-year history of developing oncology data standards. Unfortunately, progress in implementing these standards has been glacially slow. Impediments have included:
- A lack of conformance with clinical workflows.
- Failure to test standards on specific-use cases during pilot testing.
- A focus on data exchange, rather than the practical impediments to data entry.
- Poor engagement with EHR vendors in distributing clinical information modules with an oncology-specific focus
- Instability of data interoperability technologies.
The 21st Century Cures Act, which became law in December 2016, mandated improvement in the interoperability of health information through the development of data standards and application programming interfaces.
In early 2020, final rules for implementation required technology vendors to employ application programming interfaces using a single interoperability resource. In addition, payers were required to use the United States Core Data for Interoperability Standard for data exchange. These requirements were intended to provide patients with access to their own health care data “without special effort.”
As a fortunate byproduct, since EHR vendors are required to implement application program interfaces using the Health Level Seven International (HL7) Fast Healthcare Interoperability Resource (FHIR) Specification, the final rules could enable systems like mCODE to be more easily integrated with existing EHRs.
Lessons from CancerLinQ
ASCO created the health technology platform CancerLinQ in 2014, envisioning that it could become an oncology-focused learning health system – a system in which internal data and experience are systematically integrated with external evidence, allowing knowledge to be put into practice.
CancerLinQ extracts data from EHRs and other sources via direct software connections. CancerLinQ then aggregates, harmonizes, and normalizes the data in a cloud-based environment.
The data are available to participating practices for quality improvement in patient care and secondary research. In 2020, records of cancer patients in the CancerLinQ database surpassed 2 million.
CancerLinQ has been successful. However, because of the nature of the EHR ecosystem and the scope and variability of data capture by clinicians, supporting a true learning health system has proven to be a formidable task. Postprocessing manual review using trained human curators is laborious and unsustainable.
The CancerLinQ experience illustrated that basic cancer-pertinent data should be standardized in the EHR and collected prospectively.
The mCODE model
The mCODE initiative seeks to facilitate progress in care quality, clinical research, and health care policy by developing and maintaining a standard, computable, interoperable data format.
Guiding principles that were adopted early in mCODE’s development included:
- A collaborative, noncommercial, use case–driven developmental model.
- Iterative processes.
- User-driven development, refinement, and maintenance.
- Low ongoing maintenance requirements.
A foundational moment in mCODE’s development involved achieving consensus among stakeholders that the project would fail if EHR vendors required additional data entry by users.
After pilot work, a real-world endpoints project, working-group deliberation, public comment, and refinement, the final data standard included six primary domains: patient, disease, laboratory data/vital signs, genomics, treatment, and outcome.
Each domain is further divided into several concepts with specific associated data elements. The data elements are modeled into value sets that specify the possible values for the data element.
To test mCODE, eight organizations representing oncology EHR vendors, standards developers, and research organizations participated in a cancer interoperability track. The comments helped refine mCODE version 1.0, which was released in March 2020 and is accessible via the mCODE website.
Additions will likely be reviewed by a technical review group after external piloting of new use cases.
Innovation, not regulation
Every interaction between a patient and care provider yields information that could lead to improved safety and better outcomes. To be successful, the information must be collected in a computable format so it can be aggregated with data from other patients, analyzed without manual curation, and shared through interoperable systems. Those data should also be secure enough to protect the privacy of individual patients.
mCODE is a consensus data standard for oncology that provides an infrastructure to share patient data between oncology practices and health care systems while promising little to no additional data entry on the part of clinicians. Adoption by sites will be critical, however.
Publishing the standard through the HL7 FHIR technology demonstrated to EHR vendors and regulatory agencies the stability of HL7, an essential requirement for its incorporation into software.
EHR vendors and others are engaged in the CodeX HL7 FHIR Accelerator to design projects to expand and/or modify mCODE. Their creativity and innovativeness via the external advisory mCODE council and/or CodeX will be encouraged to help mCODE reach its full potential.
As part of CodeX, the Community of Practice, an open forum for end users, was established to provide regular updates about mCODE-related initiatives and use cases to solicit in-progress input, according to Robert S. Miller, MD, medical director of CancerLinQ and an author of the mCODE review.
For mCODE to be embraced by all stakeholders, there should be no additional regulations. By engaging stakeholders in an enterprise that supports innovation and collaboration – without additional regulation – mCODE could maximize the potential of EHRs that, until now, have assisted us only marginally in accomplishing those goals.
mCODE is a joint venture of ASCO/CancerLinQ, the Alliance for Clinical Trials in Oncology Foundation, the MITRE Corporation, the American Society for Radiation Oncology, and the Society of Surgical Oncology.
Dr. Osterman disclosed a grant from the National Cancer Institute and relationships with Infostratix, eHealth, AstraZeneca, Outcomes Insights, Biodesix, MD Outlook, GenomOncology, Cota Healthcare, GE Healthcare, and Microsoft. Dr. Miller and the third review author disclosed no conflicts of interest.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
An initiative designed to improve sharing of patient data may provide “tremendous benefits” in cancer care and research, according to authors of a review article.
The goals of the initiative, called Minimal Common Oncology Data Elements (mCODE), were to identify the data elements in electronic health records that are “essential” for making treatment decisions and create “a standardized computable data format” that would improve the exchange of data across EHRs, according to the mCODE website.
Travis J. Osterman, DO, of Vanderbilt University Medical Center in Nashville, Tenn., and colleagues described the mCODE initiative in a review published in JCO Clinical Cancer Informatics.
At present, commercially available EHRs are poorly designed to support modern oncology workflow, requiring laborious data entry and lacking a common library of oncology-specific discrete data elements. As an example, most EHRs poorly support the needs of precision oncology and clinical genetics, since next-generation sequencing and genetic test results are almost universally reported in PDF files.
In addition, basic, operational oncology data (e.g., cancer staging, adverse event documentation, response to treatment, etc.) are captured in EHRs primarily as an unstructured narrative.
Computable, analytical data are found for only the small percentage of patients in clinical trials. Even then, some degree of manual data abstraction is regularly required.
Interoperability of EHRs between practices and health care institutions is often so poor that the transfer of basic cancer-related information as analyzable data is difficult or even impossible.
Making progress: The 21st Century Cures Act
The American Society of Clinical Oncology has a more than 15-year history of developing oncology data standards. Unfortunately, progress in implementing these standards has been glacially slow. Impediments have included:
- A lack of conformance with clinical workflows.
- Failure to test standards on specific-use cases during pilot testing.
- A focus on data exchange, rather than the practical impediments to data entry.
- Poor engagement with EHR vendors in distributing clinical information modules with an oncology-specific focus
- Instability of data interoperability technologies.
The 21st Century Cures Act, which became law in December 2016, mandated improvement in the interoperability of health information through the development of data standards and application programming interfaces.
In early 2020, final rules for implementation required technology vendors to employ application programming interfaces using a single interoperability resource. In addition, payers were required to use the United States Core Data for Interoperability Standard for data exchange. These requirements were intended to provide patients with access to their own health care data “without special effort.”
As a fortunate byproduct, since EHR vendors are required to implement application program interfaces using the Health Level Seven International (HL7) Fast Healthcare Interoperability Resource (FHIR) Specification, the final rules could enable systems like mCODE to be more easily integrated with existing EHRs.
Lessons from CancerLinQ
ASCO created the health technology platform CancerLinQ in 2014, envisioning that it could become an oncology-focused learning health system – a system in which internal data and experience are systematically integrated with external evidence, allowing knowledge to be put into practice.
CancerLinQ extracts data from EHRs and other sources via direct software connections. CancerLinQ then aggregates, harmonizes, and normalizes the data in a cloud-based environment.
The data are available to participating practices for quality improvement in patient care and secondary research. In 2020, records of cancer patients in the CancerLinQ database surpassed 2 million.
CancerLinQ has been successful. However, because of the nature of the EHR ecosystem and the scope and variability of data capture by clinicians, supporting a true learning health system has proven to be a formidable task. Postprocessing manual review using trained human curators is laborious and unsustainable.
The CancerLinQ experience illustrated that basic cancer-pertinent data should be standardized in the EHR and collected prospectively.
The mCODE model
The mCODE initiative seeks to facilitate progress in care quality, clinical research, and health care policy by developing and maintaining a standard, computable, interoperable data format.
Guiding principles that were adopted early in mCODE’s development included:
- A collaborative, noncommercial, use case–driven developmental model.
- Iterative processes.
- User-driven development, refinement, and maintenance.
- Low ongoing maintenance requirements.
A foundational moment in mCODE’s development involved achieving consensus among stakeholders that the project would fail if EHR vendors required additional data entry by users.
After pilot work, a real-world endpoints project, working-group deliberation, public comment, and refinement, the final data standard included six primary domains: patient, disease, laboratory data/vital signs, genomics, treatment, and outcome.
Each domain is further divided into several concepts with specific associated data elements. The data elements are modeled into value sets that specify the possible values for the data element.
To test mCODE, eight organizations representing oncology EHR vendors, standards developers, and research organizations participated in a cancer interoperability track. The comments helped refine mCODE version 1.0, which was released in March 2020 and is accessible via the mCODE website.
Additions will likely be reviewed by a technical review group after external piloting of new use cases.
Innovation, not regulation
Every interaction between a patient and care provider yields information that could lead to improved safety and better outcomes. To be successful, the information must be collected in a computable format so it can be aggregated with data from other patients, analyzed without manual curation, and shared through interoperable systems. Those data should also be secure enough to protect the privacy of individual patients.
mCODE is a consensus data standard for oncology that provides an infrastructure to share patient data between oncology practices and health care systems while promising little to no additional data entry on the part of clinicians. Adoption by sites will be critical, however.
Publishing the standard through the HL7 FHIR technology demonstrated to EHR vendors and regulatory agencies the stability of HL7, an essential requirement for its incorporation into software.
EHR vendors and others are engaged in the CodeX HL7 FHIR Accelerator to design projects to expand and/or modify mCODE. Their creativity and innovativeness via the external advisory mCODE council and/or CodeX will be encouraged to help mCODE reach its full potential.
As part of CodeX, the Community of Practice, an open forum for end users, was established to provide regular updates about mCODE-related initiatives and use cases to solicit in-progress input, according to Robert S. Miller, MD, medical director of CancerLinQ and an author of the mCODE review.
For mCODE to be embraced by all stakeholders, there should be no additional regulations. By engaging stakeholders in an enterprise that supports innovation and collaboration – without additional regulation – mCODE could maximize the potential of EHRs that, until now, have assisted us only marginally in accomplishing those goals.
mCODE is a joint venture of ASCO/CancerLinQ, the Alliance for Clinical Trials in Oncology Foundation, the MITRE Corporation, the American Society for Radiation Oncology, and the Society of Surgical Oncology.
Dr. Osterman disclosed a grant from the National Cancer Institute and relationships with Infostratix, eHealth, AstraZeneca, Outcomes Insights, Biodesix, MD Outlook, GenomOncology, Cota Healthcare, GE Healthcare, and Microsoft. Dr. Miller and the third review author disclosed no conflicts of interest.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
An initiative designed to improve sharing of patient data may provide “tremendous benefits” in cancer care and research, according to authors of a review article.
The goals of the initiative, called Minimal Common Oncology Data Elements (mCODE), were to identify the data elements in electronic health records that are “essential” for making treatment decisions and create “a standardized computable data format” that would improve the exchange of data across EHRs, according to the mCODE website.
Travis J. Osterman, DO, of Vanderbilt University Medical Center in Nashville, Tenn., and colleagues described the mCODE initiative in a review published in JCO Clinical Cancer Informatics.
At present, commercially available EHRs are poorly designed to support modern oncology workflow, requiring laborious data entry and lacking a common library of oncology-specific discrete data elements. As an example, most EHRs poorly support the needs of precision oncology and clinical genetics, since next-generation sequencing and genetic test results are almost universally reported in PDF files.
In addition, basic, operational oncology data (e.g., cancer staging, adverse event documentation, response to treatment, etc.) are captured in EHRs primarily as an unstructured narrative.
Computable, analytical data are found for only the small percentage of patients in clinical trials. Even then, some degree of manual data abstraction is regularly required.
Interoperability of EHRs between practices and health care institutions is often so poor that the transfer of basic cancer-related information as analyzable data is difficult or even impossible.
Making progress: The 21st Century Cures Act
The American Society of Clinical Oncology has a more than 15-year history of developing oncology data standards. Unfortunately, progress in implementing these standards has been glacially slow. Impediments have included:
- A lack of conformance with clinical workflows.
- Failure to test standards on specific-use cases during pilot testing.
- A focus on data exchange, rather than the practical impediments to data entry.
- Poor engagement with EHR vendors in distributing clinical information modules with an oncology-specific focus
- Instability of data interoperability technologies.
The 21st Century Cures Act, which became law in December 2016, mandated improvement in the interoperability of health information through the development of data standards and application programming interfaces.
In early 2020, final rules for implementation required technology vendors to employ application programming interfaces using a single interoperability resource. In addition, payers were required to use the United States Core Data for Interoperability Standard for data exchange. These requirements were intended to provide patients with access to their own health care data “without special effort.”
As a fortunate byproduct, since EHR vendors are required to implement application program interfaces using the Health Level Seven International (HL7) Fast Healthcare Interoperability Resource (FHIR) Specification, the final rules could enable systems like mCODE to be more easily integrated with existing EHRs.
Lessons from CancerLinQ
ASCO created the health technology platform CancerLinQ in 2014, envisioning that it could become an oncology-focused learning health system – a system in which internal data and experience are systematically integrated with external evidence, allowing knowledge to be put into practice.
CancerLinQ extracts data from EHRs and other sources via direct software connections. CancerLinQ then aggregates, harmonizes, and normalizes the data in a cloud-based environment.
The data are available to participating practices for quality improvement in patient care and secondary research. In 2020, records of cancer patients in the CancerLinQ database surpassed 2 million.
CancerLinQ has been successful. However, because of the nature of the EHR ecosystem and the scope and variability of data capture by clinicians, supporting a true learning health system has proven to be a formidable task. Postprocessing manual review using trained human curators is laborious and unsustainable.
The CancerLinQ experience illustrated that basic cancer-pertinent data should be standardized in the EHR and collected prospectively.
The mCODE model
The mCODE initiative seeks to facilitate progress in care quality, clinical research, and health care policy by developing and maintaining a standard, computable, interoperable data format.
Guiding principles that were adopted early in mCODE’s development included:
- A collaborative, noncommercial, use case–driven developmental model.
- Iterative processes.
- User-driven development, refinement, and maintenance.
- Low ongoing maintenance requirements.
A foundational moment in mCODE’s development involved achieving consensus among stakeholders that the project would fail if EHR vendors required additional data entry by users.
After pilot work, a real-world endpoints project, working-group deliberation, public comment, and refinement, the final data standard included six primary domains: patient, disease, laboratory data/vital signs, genomics, treatment, and outcome.
Each domain is further divided into several concepts with specific associated data elements. The data elements are modeled into value sets that specify the possible values for the data element.
To test mCODE, eight organizations representing oncology EHR vendors, standards developers, and research organizations participated in a cancer interoperability track. The comments helped refine mCODE version 1.0, which was released in March 2020 and is accessible via the mCODE website.
Additions will likely be reviewed by a technical review group after external piloting of new use cases.
Innovation, not regulation
Every interaction between a patient and care provider yields information that could lead to improved safety and better outcomes. To be successful, the information must be collected in a computable format so it can be aggregated with data from other patients, analyzed without manual curation, and shared through interoperable systems. Those data should also be secure enough to protect the privacy of individual patients.
mCODE is a consensus data standard for oncology that provides an infrastructure to share patient data between oncology practices and health care systems while promising little to no additional data entry on the part of clinicians. Adoption by sites will be critical, however.
Publishing the standard through the HL7 FHIR technology demonstrated to EHR vendors and regulatory agencies the stability of HL7, an essential requirement for its incorporation into software.
EHR vendors and others are engaged in the CodeX HL7 FHIR Accelerator to design projects to expand and/or modify mCODE. Their creativity and innovativeness via the external advisory mCODE council and/or CodeX will be encouraged to help mCODE reach its full potential.
As part of CodeX, the Community of Practice, an open forum for end users, was established to provide regular updates about mCODE-related initiatives and use cases to solicit in-progress input, according to Robert S. Miller, MD, medical director of CancerLinQ and an author of the mCODE review.
For mCODE to be embraced by all stakeholders, there should be no additional regulations. By engaging stakeholders in an enterprise that supports innovation and collaboration – without additional regulation – mCODE could maximize the potential of EHRs that, until now, have assisted us only marginally in accomplishing those goals.
mCODE is a joint venture of ASCO/CancerLinQ, the Alliance for Clinical Trials in Oncology Foundation, the MITRE Corporation, the American Society for Radiation Oncology, and the Society of Surgical Oncology.
Dr. Osterman disclosed a grant from the National Cancer Institute and relationships with Infostratix, eHealth, AstraZeneca, Outcomes Insights, Biodesix, MD Outlook, GenomOncology, Cota Healthcare, GE Healthcare, and Microsoft. Dr. Miller and the third review author disclosed no conflicts of interest.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM JCO CLINICAL CANCER INFORMATICS
Genetic testing for breast and ovarian cancer: What has changed and what still needs to change?
Investigators found racial and ethnic disparities in genetic testing as well as “persistent underuse” of testing in patients with ovarian cancer.
The team also discovered that most pathogenic variant (PV) results were in 20 genes associated with breast and/or ovarian cancer, and testing other genes largely revealed variants of uncertain significance (VUS).
Allison W. Kurian, MD, of Stanford (Calif.) University, and colleagues recounted these findings in the Journal of Clinical Oncology.
Because of improvements in sequencing technology, competition among commercial purveyors, and declining cost, genetic testing has been increasingly available to clinicians for patient management and cancer prevention (JAMA. 2015 Sep 8;314[10]:997-8). Although germline testing can guide therapy for several solid tumors, there is little research about how often and how well it is used in practice.
For their study, Dr. Kurian and colleagues used a SEER Genetic Testing Linkage Demonstration Project in a population-based assessment of testing for cancer risk. The investigators analyzed 7-year trends in testing among all women diagnosed with breast or ovarian cancer in Georgia or California from 2013 to 2017, reviewing testing patterns and result interpretation from 2012 to 2019.
Before analyzing the data, the investigators made the following hypotheses:
- Multigene panels (MGP) would entirely replace testing for BRCA1/2 only.
- Testing underutilization in patients with ovarian cancer would improve over time.
- More patients would be tested at lower levels of pretest risk for PVs.
- Sociodemographic differences in testing trends would not be observed.
- Detection of PVs and VUS would increase.
- Racial and ethnic disparities in rates of VUS would diminish.
Study conduct
The investigators examined genetic tests performed from 2012 through the beginning of 2019 at major commercial laboratories and linked that information with data in the SEER registries in Georgia and California on all breast and ovarian cancer patients diagnosed between 2013 and 2017. There were few criteria for exclusion.
Genetic testing results were categorized as identifying a PV or likely PV, VUS, or benign or likely benign mutation by American College of Medical Genetics criteria. When a patient had genetic testing on more than one occasion, the most recent test was used.
If a PV was identified, the types of PVs were grouped according to the level of evidence that supported pathogenicity into the following categories:
- BRCA1 or BRCA2 mutations.
- PVs in other genes designated by the National Comprehensive Cancer Network as associated with breast or ovarian cancer (e.g., ATM, BARD1, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PALB2, MS2, PTEN, RAD51C, RAD51D, STK11, and TP53).
- PVs in other actionable genes (e.g., APC, BMPR1A, MEN1, MUTYH, NF2, RB1, RET, SDHAF2, SDHB, SDHC, SDHD, SMAD4, TSC1, TSC2, and VHL).
- Any other tested genes.
The investigators also tabulated instances in which genetic testing identified a VUS in any gene but no PV. If a VUS was identified originally and was reclassified more recently into the “PV/likely PV” or “benign/likely benign” categories, only the resolved categorization was recorded.
The authors evaluated clinical and sociodemographic correlates of testing trends for breast and ovarian cancer, assessing the relationship between race, age, and geographic site in receipt of any test or type of test.
Among laboratories, the investigators examined trends in the number of genes tested, associations with sociodemographic factors, categories of test results, and whether trends differed by race or ethnicity.
Findings, by hypothesis
Hypothesis #1: MGP will entirely replace testing for BRCA1/2 only.
About 25% of tested patients with breast cancer diagnosed in early 2013 received MGP, compared with more than 80% of those diagnosed in late 2017.
The trend for ovarian cancer was similar. About 40% of patients diagnosed in early 2013 received MGP, compared with more than 90% diagnosed in late 2017. These trends were similar in California and Georgia.
From 2012 to 2019, there was a consistent upward trend in gene number for patients with breast cancer (mean, 19) or ovarian cancer (mean, 21), from approximately 10 genes to 35 genes.
Hypothesis #2: Underutilization of testing in patients with ovarian cancer will improve.
Among the 187,535 patients with breast cancer and the 14,689 patients with ovarian cancer diagnosed in Georgia or California from 2013 through 2017, on average, testing rates increased 2% per year.
In all, 25.2% of breast cancer patients and 34.3% of ovarian cancer patients had genetic testing on one (87.3%) or more (12.7%) occasions.
Prior research suggested that, in 2013 and 2014, 31% of women with ovarian cancer had genetic testing (JAMA Oncol. 2018 Aug 1;4[8]:1066-72/ J Clin Oncol. 2019 May 20;37[15]:1305-15).
The investigators therefore concluded that underutilization of genetic testing in ovarian cancer did not improve substantially during the 7-year interval analyzed.
Hypothesis #3: More patients will be tested at lower levels of pretest risk.
These data were more difficult to abstract from the SEER database, but older patients were more likely to be tested in later years.
In patients older than 60 years of age (who accounted for more than 50% of both cancer cohorts), testing rates increased from 11.1% to 14.9% for breast cancer and 25.3% to 31.4% for ovarian cancer. By contrast, patients younger than 45 years of age were less than 15% of the sample and had lower testing rates over time.
There were no substantial changes in testing rates by other clinical variables. Therefore, in concert with the age-related testing trends, it is likely that women were tested for genetic mutations at increasingly lower levels of pretest risk.
Hypothesis #4: Sociodemographic differences in testing trends will not be observed.
Among patients with breast cancer, approximately 31% of those who had genetic testing were uninsured, 31% had Medicaid, and 26% had private insurance, Medicare, or other insurance.
For patients with ovarian cancer, approximately 28% were uninsured, 27% had Medicaid, and 39% had private insurance, Medicare, or other insurance.
The authors had previously found that less testing was associated with Black race, greater poverty, and less insurance coverage (J Clin Oncol. 2019 May 20;37[15]:1305-15). However, they noted no changes in testing rates by sociodemographic variables over time.
Hypothesis #5: Detection of both PVs and VUS will increase.
The proportion of tested breast cancer patients with PVs in BRCA1/2 decreased from 7.5% to 5.0% (P < .001), whereas PV yield for the two other clinically salient categories (breast or ovarian and other actionable genes) increased.
The proportion of PVs in any breast or ovarian gene increased from 1.3% to 4.6%, and the proportion in any other actionable gene increased from 0.3% to 1.3%.
For breast cancer patients, VUS-only rates increased from 8.5% in early 2013 to 22.4% in late 2017.
For ovarian cancer patients, the yield of PVs in BRCA1/2 decreased from 15.7% to 12.4% (P < .001), whereas the PV yield for breast or ovarian genes increased from 3.9% to 4.3%, and the yield for other actionable genes increased from 0.3% to 2.0%.
In ovarian cancer patients, the PV or VUS-only result rate increased from 30.8% in early 2013 to 43.0% in late 2017, entirely due to the increase in VUS-only rates. VUS were identified in 8.1% of patients diagnosed in early 2013 and increased to 28.3% in patients diagnosed in late 2017.
Hypothesis #6: Racial or ethnic disparities in rates of VUS will diminish.
Among patients with breast cancer, racial or ethnic differences in PV rates were small and did not change over time. For patients with ovarian cancer, PV rates across racial or ethnic groups diminished over time.
However, for both breast and ovarian cancer patients, there were large differences in VUS-only rates by race and ethnicity that persisted during the interval studied.
In 2017, for patients with breast cancer, VUS-only rates were substantially higher in Asian (42.4%), Black (36.6%), and Hispanic (27.7%) patients than in non-Hispanic White patients (24.5%, P < .001).
Similar trends were noted for patients with ovarian cancer. VUS-only rates were substantially higher in Asian (47.8%), Black (46.0%), and Hispanic (36.8%) patients than in non-Hispanic White patients (24.6%, P < .001).
Multivariable logistic regressions were performed separately for tested patients with breast cancer and ovarian cancer, and the results showed no significant interaction between race or ethnicity and date. Therefore, there was no significant change in racial or ethnic differences in VUS-only results across the study period.
Where these findings leave clinicians in 2021
Among the patients studied, there was:
- Marked expansion in the number of genes sequenced.
- A likely modest trend toward testing patients with lower pretest risk of a PV.
- No sociodemographic differences in testing trends.
- A small increase in PV rates and a substantial increase in VUS-only rates.
- Near-complete replacement of selective testing by MGP.
For patients with breast cancer, the proportion of all PVs that were in BRCA1/2 fell substantially. Adoption of MGP testing doubled the probability of detecting a PV in other tested genes. Most of the increase was in genes with an established breast or ovarian cancer association, with fewer PVs found in other actionable genes and very few PVs in other tested genes.
Contrary to their hypothesis, the authors observed a sustained undertesting of patients with ovarian cancer. Only 34.3% performed versus nearly 100% recommended, with little change since 2014.
This finding is surprising – and tremendously disappointing – since the prevalence of BRCA1/2 PVs is higher in ovarian cancer than in other cancers (Gynecol Oncol. 2017 Nov;147[2]:375-380), and germline-targeted therapy with PARP inhibitors has been approved for use since 2014.
Furthermore, insurance carriers provide coverage for genetic testing in most patients with carcinoma of the ovary, fallopian tube, and/or peritoneum.
Action plans: Less could be more
During the period analyzed, the increase in VUS-only results dramatically outpaced the increase in PVs.
Since there is a substantially larger volume of clinical genetic testing in non-Hispanic White patients with breast or ovarian cancer, the spectrum of normal variation is less well-defined in other racial or ethnic groups.
The study showed a widening of the “racial-ethnic VUS gap,” with Black and Asian patients having nearly twofold more VUS, although they were not tested for more genes than non-Hispanic White patients.
This is problematic on several levels. Identification of a VUS is challenging for communicating results to and recommending cascade testing for family members.
There is worrisome information regarding overtreatment or counseling of VUS patients about their results. For example, the PROMPT registry showed that 10%-15% of women with PV/VUS in genes not associated with a high risk of ovarian cancer underwent oophorectomy without a clear indication for the procedure.
Although population-based testing might augment the available data on the spectrum of normal variation in racial and ethnic minorities, it would likely exacerbate the proliferation of VUS over PVs.
It is essential to accelerate ongoing approaches to VUS reclassification.
In addition, the authors suggest that it may be time to reverse the trend in increasing the number of genes tested in MGPs. Their rationale is that, in Georgia and California, most PVs among patients with breast and ovarian cancer were identified in 20 genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PMS2, PALB2, PTEN, RAD51C, RAD51D, STK11, and TP53).
If the Georgia and California data are representative of a more generalized pattern, a panel of 20 breast cancer– and/or ovarian cancer–associated genes may be ideal for maximizing the yield of clinically relevant PVs and minimizing VUS results for all patients.
Finally, defining the patient, clinician, and health care system factors that impede widespread genetic testing for ovarian cancer patients must be prioritized. As the authors suggest, quality improvement efforts should focus on getting a lot closer to testing rates of 100% for patients with ovarian cancer and building the database that will help sort VUS in minority patients into their proper context of pathogenicity, rather than adding more genes per test.
This research was supported by the National Cancer Institute, the Centers for Disease Control and Prevention, and the California Department of Public Health. The authors disclosed relationships with Myriad Genetics, Ambry Genetics, Color Genomics, GeneDx/BioReference, InVitae, Genentech, Genomic Health, Roche/Genentech, Oncoquest, Tesaro, and Karyopharm Therapeutics.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Investigators found racial and ethnic disparities in genetic testing as well as “persistent underuse” of testing in patients with ovarian cancer.
The team also discovered that most pathogenic variant (PV) results were in 20 genes associated with breast and/or ovarian cancer, and testing other genes largely revealed variants of uncertain significance (VUS).
Allison W. Kurian, MD, of Stanford (Calif.) University, and colleagues recounted these findings in the Journal of Clinical Oncology.
Because of improvements in sequencing technology, competition among commercial purveyors, and declining cost, genetic testing has been increasingly available to clinicians for patient management and cancer prevention (JAMA. 2015 Sep 8;314[10]:997-8). Although germline testing can guide therapy for several solid tumors, there is little research about how often and how well it is used in practice.
For their study, Dr. Kurian and colleagues used a SEER Genetic Testing Linkage Demonstration Project in a population-based assessment of testing for cancer risk. The investigators analyzed 7-year trends in testing among all women diagnosed with breast or ovarian cancer in Georgia or California from 2013 to 2017, reviewing testing patterns and result interpretation from 2012 to 2019.
Before analyzing the data, the investigators made the following hypotheses:
- Multigene panels (MGP) would entirely replace testing for BRCA1/2 only.
- Testing underutilization in patients with ovarian cancer would improve over time.
- More patients would be tested at lower levels of pretest risk for PVs.
- Sociodemographic differences in testing trends would not be observed.
- Detection of PVs and VUS would increase.
- Racial and ethnic disparities in rates of VUS would diminish.
Study conduct
The investigators examined genetic tests performed from 2012 through the beginning of 2019 at major commercial laboratories and linked that information with data in the SEER registries in Georgia and California on all breast and ovarian cancer patients diagnosed between 2013 and 2017. There were few criteria for exclusion.
Genetic testing results were categorized as identifying a PV or likely PV, VUS, or benign or likely benign mutation by American College of Medical Genetics criteria. When a patient had genetic testing on more than one occasion, the most recent test was used.
If a PV was identified, the types of PVs were grouped according to the level of evidence that supported pathogenicity into the following categories:
- BRCA1 or BRCA2 mutations.
- PVs in other genes designated by the National Comprehensive Cancer Network as associated with breast or ovarian cancer (e.g., ATM, BARD1, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PALB2, MS2, PTEN, RAD51C, RAD51D, STK11, and TP53).
- PVs in other actionable genes (e.g., APC, BMPR1A, MEN1, MUTYH, NF2, RB1, RET, SDHAF2, SDHB, SDHC, SDHD, SMAD4, TSC1, TSC2, and VHL).
- Any other tested genes.
The investigators also tabulated instances in which genetic testing identified a VUS in any gene but no PV. If a VUS was identified originally and was reclassified more recently into the “PV/likely PV” or “benign/likely benign” categories, only the resolved categorization was recorded.
The authors evaluated clinical and sociodemographic correlates of testing trends for breast and ovarian cancer, assessing the relationship between race, age, and geographic site in receipt of any test or type of test.
Among laboratories, the investigators examined trends in the number of genes tested, associations with sociodemographic factors, categories of test results, and whether trends differed by race or ethnicity.
Findings, by hypothesis
Hypothesis #1: MGP will entirely replace testing for BRCA1/2 only.
About 25% of tested patients with breast cancer diagnosed in early 2013 received MGP, compared with more than 80% of those diagnosed in late 2017.
The trend for ovarian cancer was similar. About 40% of patients diagnosed in early 2013 received MGP, compared with more than 90% diagnosed in late 2017. These trends were similar in California and Georgia.
From 2012 to 2019, there was a consistent upward trend in gene number for patients with breast cancer (mean, 19) or ovarian cancer (mean, 21), from approximately 10 genes to 35 genes.
Hypothesis #2: Underutilization of testing in patients with ovarian cancer will improve.
Among the 187,535 patients with breast cancer and the 14,689 patients with ovarian cancer diagnosed in Georgia or California from 2013 through 2017, on average, testing rates increased 2% per year.
In all, 25.2% of breast cancer patients and 34.3% of ovarian cancer patients had genetic testing on one (87.3%) or more (12.7%) occasions.
Prior research suggested that, in 2013 and 2014, 31% of women with ovarian cancer had genetic testing (JAMA Oncol. 2018 Aug 1;4[8]:1066-72/ J Clin Oncol. 2019 May 20;37[15]:1305-15).
The investigators therefore concluded that underutilization of genetic testing in ovarian cancer did not improve substantially during the 7-year interval analyzed.
Hypothesis #3: More patients will be tested at lower levels of pretest risk.
These data were more difficult to abstract from the SEER database, but older patients were more likely to be tested in later years.
In patients older than 60 years of age (who accounted for more than 50% of both cancer cohorts), testing rates increased from 11.1% to 14.9% for breast cancer and 25.3% to 31.4% for ovarian cancer. By contrast, patients younger than 45 years of age were less than 15% of the sample and had lower testing rates over time.
There were no substantial changes in testing rates by other clinical variables. Therefore, in concert with the age-related testing trends, it is likely that women were tested for genetic mutations at increasingly lower levels of pretest risk.
Hypothesis #4: Sociodemographic differences in testing trends will not be observed.
Among patients with breast cancer, approximately 31% of those who had genetic testing were uninsured, 31% had Medicaid, and 26% had private insurance, Medicare, or other insurance.
For patients with ovarian cancer, approximately 28% were uninsured, 27% had Medicaid, and 39% had private insurance, Medicare, or other insurance.
The authors had previously found that less testing was associated with Black race, greater poverty, and less insurance coverage (J Clin Oncol. 2019 May 20;37[15]:1305-15). However, they noted no changes in testing rates by sociodemographic variables over time.
Hypothesis #5: Detection of both PVs and VUS will increase.
The proportion of tested breast cancer patients with PVs in BRCA1/2 decreased from 7.5% to 5.0% (P < .001), whereas PV yield for the two other clinically salient categories (breast or ovarian and other actionable genes) increased.
The proportion of PVs in any breast or ovarian gene increased from 1.3% to 4.6%, and the proportion in any other actionable gene increased from 0.3% to 1.3%.
For breast cancer patients, VUS-only rates increased from 8.5% in early 2013 to 22.4% in late 2017.
For ovarian cancer patients, the yield of PVs in BRCA1/2 decreased from 15.7% to 12.4% (P < .001), whereas the PV yield for breast or ovarian genes increased from 3.9% to 4.3%, and the yield for other actionable genes increased from 0.3% to 2.0%.
In ovarian cancer patients, the PV or VUS-only result rate increased from 30.8% in early 2013 to 43.0% in late 2017, entirely due to the increase in VUS-only rates. VUS were identified in 8.1% of patients diagnosed in early 2013 and increased to 28.3% in patients diagnosed in late 2017.
Hypothesis #6: Racial or ethnic disparities in rates of VUS will diminish.
Among patients with breast cancer, racial or ethnic differences in PV rates were small and did not change over time. For patients with ovarian cancer, PV rates across racial or ethnic groups diminished over time.
However, for both breast and ovarian cancer patients, there were large differences in VUS-only rates by race and ethnicity that persisted during the interval studied.
In 2017, for patients with breast cancer, VUS-only rates were substantially higher in Asian (42.4%), Black (36.6%), and Hispanic (27.7%) patients than in non-Hispanic White patients (24.5%, P < .001).
Similar trends were noted for patients with ovarian cancer. VUS-only rates were substantially higher in Asian (47.8%), Black (46.0%), and Hispanic (36.8%) patients than in non-Hispanic White patients (24.6%, P < .001).
Multivariable logistic regressions were performed separately for tested patients with breast cancer and ovarian cancer, and the results showed no significant interaction between race or ethnicity and date. Therefore, there was no significant change in racial or ethnic differences in VUS-only results across the study period.
Where these findings leave clinicians in 2021
Among the patients studied, there was:
- Marked expansion in the number of genes sequenced.
- A likely modest trend toward testing patients with lower pretest risk of a PV.
- No sociodemographic differences in testing trends.
- A small increase in PV rates and a substantial increase in VUS-only rates.
- Near-complete replacement of selective testing by MGP.
For patients with breast cancer, the proportion of all PVs that were in BRCA1/2 fell substantially. Adoption of MGP testing doubled the probability of detecting a PV in other tested genes. Most of the increase was in genes with an established breast or ovarian cancer association, with fewer PVs found in other actionable genes and very few PVs in other tested genes.
Contrary to their hypothesis, the authors observed a sustained undertesting of patients with ovarian cancer. Only 34.3% performed versus nearly 100% recommended, with little change since 2014.
This finding is surprising – and tremendously disappointing – since the prevalence of BRCA1/2 PVs is higher in ovarian cancer than in other cancers (Gynecol Oncol. 2017 Nov;147[2]:375-380), and germline-targeted therapy with PARP inhibitors has been approved for use since 2014.
Furthermore, insurance carriers provide coverage for genetic testing in most patients with carcinoma of the ovary, fallopian tube, and/or peritoneum.
Action plans: Less could be more
During the period analyzed, the increase in VUS-only results dramatically outpaced the increase in PVs.
Since there is a substantially larger volume of clinical genetic testing in non-Hispanic White patients with breast or ovarian cancer, the spectrum of normal variation is less well-defined in other racial or ethnic groups.
The study showed a widening of the “racial-ethnic VUS gap,” with Black and Asian patients having nearly twofold more VUS, although they were not tested for more genes than non-Hispanic White patients.
This is problematic on several levels. Identification of a VUS is challenging for communicating results to and recommending cascade testing for family members.
There is worrisome information regarding overtreatment or counseling of VUS patients about their results. For example, the PROMPT registry showed that 10%-15% of women with PV/VUS in genes not associated with a high risk of ovarian cancer underwent oophorectomy without a clear indication for the procedure.
Although population-based testing might augment the available data on the spectrum of normal variation in racial and ethnic minorities, it would likely exacerbate the proliferation of VUS over PVs.
It is essential to accelerate ongoing approaches to VUS reclassification.
In addition, the authors suggest that it may be time to reverse the trend in increasing the number of genes tested in MGPs. Their rationale is that, in Georgia and California, most PVs among patients with breast and ovarian cancer were identified in 20 genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PMS2, PALB2, PTEN, RAD51C, RAD51D, STK11, and TP53).
If the Georgia and California data are representative of a more generalized pattern, a panel of 20 breast cancer– and/or ovarian cancer–associated genes may be ideal for maximizing the yield of clinically relevant PVs and minimizing VUS results for all patients.
Finally, defining the patient, clinician, and health care system factors that impede widespread genetic testing for ovarian cancer patients must be prioritized. As the authors suggest, quality improvement efforts should focus on getting a lot closer to testing rates of 100% for patients with ovarian cancer and building the database that will help sort VUS in minority patients into their proper context of pathogenicity, rather than adding more genes per test.
This research was supported by the National Cancer Institute, the Centers for Disease Control and Prevention, and the California Department of Public Health. The authors disclosed relationships with Myriad Genetics, Ambry Genetics, Color Genomics, GeneDx/BioReference, InVitae, Genentech, Genomic Health, Roche/Genentech, Oncoquest, Tesaro, and Karyopharm Therapeutics.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Investigators found racial and ethnic disparities in genetic testing as well as “persistent underuse” of testing in patients with ovarian cancer.
The team also discovered that most pathogenic variant (PV) results were in 20 genes associated with breast and/or ovarian cancer, and testing other genes largely revealed variants of uncertain significance (VUS).
Allison W. Kurian, MD, of Stanford (Calif.) University, and colleagues recounted these findings in the Journal of Clinical Oncology.
Because of improvements in sequencing technology, competition among commercial purveyors, and declining cost, genetic testing has been increasingly available to clinicians for patient management and cancer prevention (JAMA. 2015 Sep 8;314[10]:997-8). Although germline testing can guide therapy for several solid tumors, there is little research about how often and how well it is used in practice.
For their study, Dr. Kurian and colleagues used a SEER Genetic Testing Linkage Demonstration Project in a population-based assessment of testing for cancer risk. The investigators analyzed 7-year trends in testing among all women diagnosed with breast or ovarian cancer in Georgia or California from 2013 to 2017, reviewing testing patterns and result interpretation from 2012 to 2019.
Before analyzing the data, the investigators made the following hypotheses:
- Multigene panels (MGP) would entirely replace testing for BRCA1/2 only.
- Testing underutilization in patients with ovarian cancer would improve over time.
- More patients would be tested at lower levels of pretest risk for PVs.
- Sociodemographic differences in testing trends would not be observed.
- Detection of PVs and VUS would increase.
- Racial and ethnic disparities in rates of VUS would diminish.
Study conduct
The investigators examined genetic tests performed from 2012 through the beginning of 2019 at major commercial laboratories and linked that information with data in the SEER registries in Georgia and California on all breast and ovarian cancer patients diagnosed between 2013 and 2017. There were few criteria for exclusion.
Genetic testing results were categorized as identifying a PV or likely PV, VUS, or benign or likely benign mutation by American College of Medical Genetics criteria. When a patient had genetic testing on more than one occasion, the most recent test was used.
If a PV was identified, the types of PVs were grouped according to the level of evidence that supported pathogenicity into the following categories:
- BRCA1 or BRCA2 mutations.
- PVs in other genes designated by the National Comprehensive Cancer Network as associated with breast or ovarian cancer (e.g., ATM, BARD1, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PALB2, MS2, PTEN, RAD51C, RAD51D, STK11, and TP53).
- PVs in other actionable genes (e.g., APC, BMPR1A, MEN1, MUTYH, NF2, RB1, RET, SDHAF2, SDHB, SDHC, SDHD, SMAD4, TSC1, TSC2, and VHL).
- Any other tested genes.
The investigators also tabulated instances in which genetic testing identified a VUS in any gene but no PV. If a VUS was identified originally and was reclassified more recently into the “PV/likely PV” or “benign/likely benign” categories, only the resolved categorization was recorded.
The authors evaluated clinical and sociodemographic correlates of testing trends for breast and ovarian cancer, assessing the relationship between race, age, and geographic site in receipt of any test or type of test.
Among laboratories, the investigators examined trends in the number of genes tested, associations with sociodemographic factors, categories of test results, and whether trends differed by race or ethnicity.
Findings, by hypothesis
Hypothesis #1: MGP will entirely replace testing for BRCA1/2 only.
About 25% of tested patients with breast cancer diagnosed in early 2013 received MGP, compared with more than 80% of those diagnosed in late 2017.
The trend for ovarian cancer was similar. About 40% of patients diagnosed in early 2013 received MGP, compared with more than 90% diagnosed in late 2017. These trends were similar in California and Georgia.
From 2012 to 2019, there was a consistent upward trend in gene number for patients with breast cancer (mean, 19) or ovarian cancer (mean, 21), from approximately 10 genes to 35 genes.
Hypothesis #2: Underutilization of testing in patients with ovarian cancer will improve.
Among the 187,535 patients with breast cancer and the 14,689 patients with ovarian cancer diagnosed in Georgia or California from 2013 through 2017, on average, testing rates increased 2% per year.
In all, 25.2% of breast cancer patients and 34.3% of ovarian cancer patients had genetic testing on one (87.3%) or more (12.7%) occasions.
Prior research suggested that, in 2013 and 2014, 31% of women with ovarian cancer had genetic testing (JAMA Oncol. 2018 Aug 1;4[8]:1066-72/ J Clin Oncol. 2019 May 20;37[15]:1305-15).
The investigators therefore concluded that underutilization of genetic testing in ovarian cancer did not improve substantially during the 7-year interval analyzed.
Hypothesis #3: More patients will be tested at lower levels of pretest risk.
These data were more difficult to abstract from the SEER database, but older patients were more likely to be tested in later years.
In patients older than 60 years of age (who accounted for more than 50% of both cancer cohorts), testing rates increased from 11.1% to 14.9% for breast cancer and 25.3% to 31.4% for ovarian cancer. By contrast, patients younger than 45 years of age were less than 15% of the sample and had lower testing rates over time.
There were no substantial changes in testing rates by other clinical variables. Therefore, in concert with the age-related testing trends, it is likely that women were tested for genetic mutations at increasingly lower levels of pretest risk.
Hypothesis #4: Sociodemographic differences in testing trends will not be observed.
Among patients with breast cancer, approximately 31% of those who had genetic testing were uninsured, 31% had Medicaid, and 26% had private insurance, Medicare, or other insurance.
For patients with ovarian cancer, approximately 28% were uninsured, 27% had Medicaid, and 39% had private insurance, Medicare, or other insurance.
The authors had previously found that less testing was associated with Black race, greater poverty, and less insurance coverage (J Clin Oncol. 2019 May 20;37[15]:1305-15). However, they noted no changes in testing rates by sociodemographic variables over time.
Hypothesis #5: Detection of both PVs and VUS will increase.
The proportion of tested breast cancer patients with PVs in BRCA1/2 decreased from 7.5% to 5.0% (P < .001), whereas PV yield for the two other clinically salient categories (breast or ovarian and other actionable genes) increased.
The proportion of PVs in any breast or ovarian gene increased from 1.3% to 4.6%, and the proportion in any other actionable gene increased from 0.3% to 1.3%.
For breast cancer patients, VUS-only rates increased from 8.5% in early 2013 to 22.4% in late 2017.
For ovarian cancer patients, the yield of PVs in BRCA1/2 decreased from 15.7% to 12.4% (P < .001), whereas the PV yield for breast or ovarian genes increased from 3.9% to 4.3%, and the yield for other actionable genes increased from 0.3% to 2.0%.
In ovarian cancer patients, the PV or VUS-only result rate increased from 30.8% in early 2013 to 43.0% in late 2017, entirely due to the increase in VUS-only rates. VUS were identified in 8.1% of patients diagnosed in early 2013 and increased to 28.3% in patients diagnosed in late 2017.
Hypothesis #6: Racial or ethnic disparities in rates of VUS will diminish.
Among patients with breast cancer, racial or ethnic differences in PV rates were small and did not change over time. For patients with ovarian cancer, PV rates across racial or ethnic groups diminished over time.
However, for both breast and ovarian cancer patients, there were large differences in VUS-only rates by race and ethnicity that persisted during the interval studied.
In 2017, for patients with breast cancer, VUS-only rates were substantially higher in Asian (42.4%), Black (36.6%), and Hispanic (27.7%) patients than in non-Hispanic White patients (24.5%, P < .001).
Similar trends were noted for patients with ovarian cancer. VUS-only rates were substantially higher in Asian (47.8%), Black (46.0%), and Hispanic (36.8%) patients than in non-Hispanic White patients (24.6%, P < .001).
Multivariable logistic regressions were performed separately for tested patients with breast cancer and ovarian cancer, and the results showed no significant interaction between race or ethnicity and date. Therefore, there was no significant change in racial or ethnic differences in VUS-only results across the study period.
Where these findings leave clinicians in 2021
Among the patients studied, there was:
- Marked expansion in the number of genes sequenced.
- A likely modest trend toward testing patients with lower pretest risk of a PV.
- No sociodemographic differences in testing trends.
- A small increase in PV rates and a substantial increase in VUS-only rates.
- Near-complete replacement of selective testing by MGP.
For patients with breast cancer, the proportion of all PVs that were in BRCA1/2 fell substantially. Adoption of MGP testing doubled the probability of detecting a PV in other tested genes. Most of the increase was in genes with an established breast or ovarian cancer association, with fewer PVs found in other actionable genes and very few PVs in other tested genes.
Contrary to their hypothesis, the authors observed a sustained undertesting of patients with ovarian cancer. Only 34.3% performed versus nearly 100% recommended, with little change since 2014.
This finding is surprising – and tremendously disappointing – since the prevalence of BRCA1/2 PVs is higher in ovarian cancer than in other cancers (Gynecol Oncol. 2017 Nov;147[2]:375-380), and germline-targeted therapy with PARP inhibitors has been approved for use since 2014.
Furthermore, insurance carriers provide coverage for genetic testing in most patients with carcinoma of the ovary, fallopian tube, and/or peritoneum.
Action plans: Less could be more
During the period analyzed, the increase in VUS-only results dramatically outpaced the increase in PVs.
Since there is a substantially larger volume of clinical genetic testing in non-Hispanic White patients with breast or ovarian cancer, the spectrum of normal variation is less well-defined in other racial or ethnic groups.
The study showed a widening of the “racial-ethnic VUS gap,” with Black and Asian patients having nearly twofold more VUS, although they were not tested for more genes than non-Hispanic White patients.
This is problematic on several levels. Identification of a VUS is challenging for communicating results to and recommending cascade testing for family members.
There is worrisome information regarding overtreatment or counseling of VUS patients about their results. For example, the PROMPT registry showed that 10%-15% of women with PV/VUS in genes not associated with a high risk of ovarian cancer underwent oophorectomy without a clear indication for the procedure.
Although population-based testing might augment the available data on the spectrum of normal variation in racial and ethnic minorities, it would likely exacerbate the proliferation of VUS over PVs.
It is essential to accelerate ongoing approaches to VUS reclassification.
In addition, the authors suggest that it may be time to reverse the trend in increasing the number of genes tested in MGPs. Their rationale is that, in Georgia and California, most PVs among patients with breast and ovarian cancer were identified in 20 genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CDH1, CHEK2, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PMS2, PALB2, PTEN, RAD51C, RAD51D, STK11, and TP53).
If the Georgia and California data are representative of a more generalized pattern, a panel of 20 breast cancer– and/or ovarian cancer–associated genes may be ideal for maximizing the yield of clinically relevant PVs and minimizing VUS results for all patients.
Finally, defining the patient, clinician, and health care system factors that impede widespread genetic testing for ovarian cancer patients must be prioritized. As the authors suggest, quality improvement efforts should focus on getting a lot closer to testing rates of 100% for patients with ovarian cancer and building the database that will help sort VUS in minority patients into their proper context of pathogenicity, rather than adding more genes per test.
This research was supported by the National Cancer Institute, the Centers for Disease Control and Prevention, and the California Department of Public Health. The authors disclosed relationships with Myriad Genetics, Ambry Genetics, Color Genomics, GeneDx/BioReference, InVitae, Genentech, Genomic Health, Roche/Genentech, Oncoquest, Tesaro, and Karyopharm Therapeutics.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM THE JOURNAL OF CLINICAL ONCOLOGY
Enfortumab vedotin offers hope to poor-prognosis patients with advanced urothelial cancer
Approximately half of all patients with locally advanced or metastatic urothelial cancer (la/mUC) are ineligible to receive cisplatin-based chemotherapy. They face poor outlooks and extremely limited treatment options.
A new study indicates that enfortumab vedotin (EV) can cause major, prolonged responses in most patients in that unfortunate setting.
EV is an antibody-drug conjugate directed against nectin-4, an immunoglobulin-like cell adhesion molecule that is highly expressed in UC, obviating the need for testing prior to treatment. It is internalized in malignant cells, with release of the active moiety (monomethyl auristatin E; MMAE). MMAE causes microtubule disruption, with resultant cell-cycle arrest and apoptosis.
EV received accelerated approval from the Food and Drug Administration in December 2019 after publication of the results from cohort 1 of the open-label, single-arm, phase 2 EV-201 study.
Arjun V. Balar, MD, of the Perlmutter Cancer Center at New York University Langone Health, presented results from cohort 2 of EV-201 – the cisplatin-ineligible cohort – at the 2021 Genitourinary Cancer Symposium (Abstract 394).
EV in patients ineligible for platinum-based therapy
Patients in cohort 2 of EV-201 had received immune checkpoint inhibitor therapy for la/mUC. They received EV in the FDA-approved dose for cohort 1: 1.25 mg/kg EV on days 1, 8, and 15 of a 28-day cycle.
Patients experienced disease progression during or following their most recent treatment. Patients with more than two neuropathies, active central nervous system metastases, and uncontrolled diabetes mellitus were excluded.
“Platinum ineligible” was defined as a creatinine clearance between 30-59 cm3/min, Eastern Cooperative Oncology Group performance status (ECOG PS) 2, or hearing loss of grade 2 or greater.
The primary endpoint for cohort 2 was confirmed overall response rate (ORR) per RECIST 1.1 by blinded independent central review. Secondary endpoints were duration of response, progression-free survival, overall survival, and safety.
There were 91 patients enrolled. Two patients never received EV treatment because of deterioration after registration. The median treatment duration among the remaining 89 patients was 6.0 months (range, 0.3-24.6).
Impressive results in poor-risk patients
The patients in EV-201 cohort 2 were elderly (median age, 75 years; range, 49-90) with comorbidities. The primary reasons for platinum-ineligibility were creatinine clearance less than 60 mL/min (66%), grade 2 or greater hearing loss (15%), and ECOG PS 2 (7%); 12% of patients met more than one criterion for platinum ineligibility.
The primary tumor site was in the upper urinary tract in 43% of patients, and 79% had visceral metastases, including 24% with liver involvement.
The confirmed ORR was 52% (95% confidence interval, 40.8-62.4), with 20% complete responses. There were responses in all subgroups, including patients with primary tumor sites in the upper tract (ORR, 61%), those with liver metastasis (ORR, 48%), and patients who had not responded to immune checkpoint inhibitors (ORR, 48%).
A total of 88% of patients had some decrease in measurable tumor diameters, generally within a few weeks of treatment initiation (median time to response, 1.8 months). The rapid response to treatment was especially important to patients having cancer-associated pain.
The median progression-free and overall survival durations were 5.8 months (95% CI, 5.0-8.3) and 14.7 months (95% CI, 10.5-18.2), respectively. The median response duration was 10.9 months (95% CI, 5.78-NR). More than 25% of responses extended beyond 12 months.
About 82% of patients in cohort 2 discontinued treatment, most commonly because of disease progression (51%). The second most common reason was the development of treatment-related adverse events (TRAE; 24%).
Drilling down on treatment-related adverse events
As might be expected for cisplatin-ineligible patients, adverse events were higher for patients in cohort 2 than for cohort 1 and led to treatment discontinuation in 16% of patients overall.
TRAEs over grade 3 occurred in 55% of patients. TRAEs of special interest included rash (61% overall; 17% ≥ grade 3), peripheral neuropathy (54% overall; 8% ≥ grade 3), and hyperglycemia (10% overall; 6% ≥ grade 3). Dose reductions, interruptions, and physical therapy were helpful.
Twenty percent of patients with TRAE hyperglycemia had hyperglycemia at baseline, and 30% of TRAEs were in patients with high body mass index (BMI).
There were four treatment-related deaths, all in patients 75 years or older with multiple comorbidities. Three of the four deaths occurred within 30 days of first EV dose in patients with BMI of 30 or greater (acute kidney injury, metabolic acidosis, and multiple organ dysfunction syndrome). The remaining death occurred more than 30 days after the last dose (pneumonitis).
Context and caution
The authors concluded that EV produced durable responses in platinum-ineligible patients with la/mUC, including 20% complete responses. Safety was felt to be as expected, given the known toxicities of the agent and the compromised medical condition of the patients studied.
The study discussant, Arlene O. Siefker-Radtke, MD, of the University of Texas MD Anderson Cancer Center, Houston, agreed that EV fills an unmet need, showing impressive responses in patients with visceral, liver, and bone metastases. She agreed that EV should be investigated across the spectrum of urothelial cancer.
Dr. Siefker-Radtke reminded attendees that the FDA package insert for EV described a 48% increase in the area under the concentration-time curve concentration of the MMAE active moiety in patients with mild hepatic impairment and that EV use should be avoided entirely in patients with moderate to severe liver disease.
She speculated whether augmented toxicity in patients with a high BMI could be attributable to clinically occult impaired hepatic function from fatty liver infiltration.
She indicated that clinicians should monitor closely patients with higher BMI and grade 3-4 hyperglycemia or elevated hemoglobin A1c levels and advised holding EV in patients who develop:
- Glucose levels above 250 mg/dL
- Peeling skin or bullous skin lesions. These lesions can be heralded by a diffuse erythematous or papillary rash in the preceding weeks.
- Grade 3 diarrhea or mucosal membrane toxicity of other types.
Notwithstanding concerns about toxicity and the need for monitoring, EV merits continued study in combination with other agents and in additional settings in the clinical spectrum of urothelial cancer. It is an important new option for oncologists caring for patients with urothelial cancer.
The EV-201 study was funded by Seagen. Dr. Balar and Dr. Siefker-Radtke disclosed relationships with Seagen and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Approximately half of all patients with locally advanced or metastatic urothelial cancer (la/mUC) are ineligible to receive cisplatin-based chemotherapy. They face poor outlooks and extremely limited treatment options.
A new study indicates that enfortumab vedotin (EV) can cause major, prolonged responses in most patients in that unfortunate setting.
EV is an antibody-drug conjugate directed against nectin-4, an immunoglobulin-like cell adhesion molecule that is highly expressed in UC, obviating the need for testing prior to treatment. It is internalized in malignant cells, with release of the active moiety (monomethyl auristatin E; MMAE). MMAE causes microtubule disruption, with resultant cell-cycle arrest and apoptosis.
EV received accelerated approval from the Food and Drug Administration in December 2019 after publication of the results from cohort 1 of the open-label, single-arm, phase 2 EV-201 study.
Arjun V. Balar, MD, of the Perlmutter Cancer Center at New York University Langone Health, presented results from cohort 2 of EV-201 – the cisplatin-ineligible cohort – at the 2021 Genitourinary Cancer Symposium (Abstract 394).
EV in patients ineligible for platinum-based therapy
Patients in cohort 2 of EV-201 had received immune checkpoint inhibitor therapy for la/mUC. They received EV in the FDA-approved dose for cohort 1: 1.25 mg/kg EV on days 1, 8, and 15 of a 28-day cycle.
Patients experienced disease progression during or following their most recent treatment. Patients with more than two neuropathies, active central nervous system metastases, and uncontrolled diabetes mellitus were excluded.
“Platinum ineligible” was defined as a creatinine clearance between 30-59 cm3/min, Eastern Cooperative Oncology Group performance status (ECOG PS) 2, or hearing loss of grade 2 or greater.
The primary endpoint for cohort 2 was confirmed overall response rate (ORR) per RECIST 1.1 by blinded independent central review. Secondary endpoints were duration of response, progression-free survival, overall survival, and safety.
There were 91 patients enrolled. Two patients never received EV treatment because of deterioration after registration. The median treatment duration among the remaining 89 patients was 6.0 months (range, 0.3-24.6).
Impressive results in poor-risk patients
The patients in EV-201 cohort 2 were elderly (median age, 75 years; range, 49-90) with comorbidities. The primary reasons for platinum-ineligibility were creatinine clearance less than 60 mL/min (66%), grade 2 or greater hearing loss (15%), and ECOG PS 2 (7%); 12% of patients met more than one criterion for platinum ineligibility.
The primary tumor site was in the upper urinary tract in 43% of patients, and 79% had visceral metastases, including 24% with liver involvement.
The confirmed ORR was 52% (95% confidence interval, 40.8-62.4), with 20% complete responses. There were responses in all subgroups, including patients with primary tumor sites in the upper tract (ORR, 61%), those with liver metastasis (ORR, 48%), and patients who had not responded to immune checkpoint inhibitors (ORR, 48%).
A total of 88% of patients had some decrease in measurable tumor diameters, generally within a few weeks of treatment initiation (median time to response, 1.8 months). The rapid response to treatment was especially important to patients having cancer-associated pain.
The median progression-free and overall survival durations were 5.8 months (95% CI, 5.0-8.3) and 14.7 months (95% CI, 10.5-18.2), respectively. The median response duration was 10.9 months (95% CI, 5.78-NR). More than 25% of responses extended beyond 12 months.
About 82% of patients in cohort 2 discontinued treatment, most commonly because of disease progression (51%). The second most common reason was the development of treatment-related adverse events (TRAE; 24%).
Drilling down on treatment-related adverse events
As might be expected for cisplatin-ineligible patients, adverse events were higher for patients in cohort 2 than for cohort 1 and led to treatment discontinuation in 16% of patients overall.
TRAEs over grade 3 occurred in 55% of patients. TRAEs of special interest included rash (61% overall; 17% ≥ grade 3), peripheral neuropathy (54% overall; 8% ≥ grade 3), and hyperglycemia (10% overall; 6% ≥ grade 3). Dose reductions, interruptions, and physical therapy were helpful.
Twenty percent of patients with TRAE hyperglycemia had hyperglycemia at baseline, and 30% of TRAEs were in patients with high body mass index (BMI).
There were four treatment-related deaths, all in patients 75 years or older with multiple comorbidities. Three of the four deaths occurred within 30 days of first EV dose in patients with BMI of 30 or greater (acute kidney injury, metabolic acidosis, and multiple organ dysfunction syndrome). The remaining death occurred more than 30 days after the last dose (pneumonitis).
Context and caution
The authors concluded that EV produced durable responses in platinum-ineligible patients with la/mUC, including 20% complete responses. Safety was felt to be as expected, given the known toxicities of the agent and the compromised medical condition of the patients studied.
The study discussant, Arlene O. Siefker-Radtke, MD, of the University of Texas MD Anderson Cancer Center, Houston, agreed that EV fills an unmet need, showing impressive responses in patients with visceral, liver, and bone metastases. She agreed that EV should be investigated across the spectrum of urothelial cancer.
Dr. Siefker-Radtke reminded attendees that the FDA package insert for EV described a 48% increase in the area under the concentration-time curve concentration of the MMAE active moiety in patients with mild hepatic impairment and that EV use should be avoided entirely in patients with moderate to severe liver disease.
She speculated whether augmented toxicity in patients with a high BMI could be attributable to clinically occult impaired hepatic function from fatty liver infiltration.
She indicated that clinicians should monitor closely patients with higher BMI and grade 3-4 hyperglycemia or elevated hemoglobin A1c levels and advised holding EV in patients who develop:
- Glucose levels above 250 mg/dL
- Peeling skin or bullous skin lesions. These lesions can be heralded by a diffuse erythematous or papillary rash in the preceding weeks.
- Grade 3 diarrhea or mucosal membrane toxicity of other types.
Notwithstanding concerns about toxicity and the need for monitoring, EV merits continued study in combination with other agents and in additional settings in the clinical spectrum of urothelial cancer. It is an important new option for oncologists caring for patients with urothelial cancer.
The EV-201 study was funded by Seagen. Dr. Balar and Dr. Siefker-Radtke disclosed relationships with Seagen and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Approximately half of all patients with locally advanced or metastatic urothelial cancer (la/mUC) are ineligible to receive cisplatin-based chemotherapy. They face poor outlooks and extremely limited treatment options.
A new study indicates that enfortumab vedotin (EV) can cause major, prolonged responses in most patients in that unfortunate setting.
EV is an antibody-drug conjugate directed against nectin-4, an immunoglobulin-like cell adhesion molecule that is highly expressed in UC, obviating the need for testing prior to treatment. It is internalized in malignant cells, with release of the active moiety (monomethyl auristatin E; MMAE). MMAE causes microtubule disruption, with resultant cell-cycle arrest and apoptosis.
EV received accelerated approval from the Food and Drug Administration in December 2019 after publication of the results from cohort 1 of the open-label, single-arm, phase 2 EV-201 study.
Arjun V. Balar, MD, of the Perlmutter Cancer Center at New York University Langone Health, presented results from cohort 2 of EV-201 – the cisplatin-ineligible cohort – at the 2021 Genitourinary Cancer Symposium (Abstract 394).
EV in patients ineligible for platinum-based therapy
Patients in cohort 2 of EV-201 had received immune checkpoint inhibitor therapy for la/mUC. They received EV in the FDA-approved dose for cohort 1: 1.25 mg/kg EV on days 1, 8, and 15 of a 28-day cycle.
Patients experienced disease progression during or following their most recent treatment. Patients with more than two neuropathies, active central nervous system metastases, and uncontrolled diabetes mellitus were excluded.
“Platinum ineligible” was defined as a creatinine clearance between 30-59 cm3/min, Eastern Cooperative Oncology Group performance status (ECOG PS) 2, or hearing loss of grade 2 or greater.
The primary endpoint for cohort 2 was confirmed overall response rate (ORR) per RECIST 1.1 by blinded independent central review. Secondary endpoints were duration of response, progression-free survival, overall survival, and safety.
There were 91 patients enrolled. Two patients never received EV treatment because of deterioration after registration. The median treatment duration among the remaining 89 patients was 6.0 months (range, 0.3-24.6).
Impressive results in poor-risk patients
The patients in EV-201 cohort 2 were elderly (median age, 75 years; range, 49-90) with comorbidities. The primary reasons for platinum-ineligibility were creatinine clearance less than 60 mL/min (66%), grade 2 or greater hearing loss (15%), and ECOG PS 2 (7%); 12% of patients met more than one criterion for platinum ineligibility.
The primary tumor site was in the upper urinary tract in 43% of patients, and 79% had visceral metastases, including 24% with liver involvement.
The confirmed ORR was 52% (95% confidence interval, 40.8-62.4), with 20% complete responses. There were responses in all subgroups, including patients with primary tumor sites in the upper tract (ORR, 61%), those with liver metastasis (ORR, 48%), and patients who had not responded to immune checkpoint inhibitors (ORR, 48%).
A total of 88% of patients had some decrease in measurable tumor diameters, generally within a few weeks of treatment initiation (median time to response, 1.8 months). The rapid response to treatment was especially important to patients having cancer-associated pain.
The median progression-free and overall survival durations were 5.8 months (95% CI, 5.0-8.3) and 14.7 months (95% CI, 10.5-18.2), respectively. The median response duration was 10.9 months (95% CI, 5.78-NR). More than 25% of responses extended beyond 12 months.
About 82% of patients in cohort 2 discontinued treatment, most commonly because of disease progression (51%). The second most common reason was the development of treatment-related adverse events (TRAE; 24%).
Drilling down on treatment-related adverse events
As might be expected for cisplatin-ineligible patients, adverse events were higher for patients in cohort 2 than for cohort 1 and led to treatment discontinuation in 16% of patients overall.
TRAEs over grade 3 occurred in 55% of patients. TRAEs of special interest included rash (61% overall; 17% ≥ grade 3), peripheral neuropathy (54% overall; 8% ≥ grade 3), and hyperglycemia (10% overall; 6% ≥ grade 3). Dose reductions, interruptions, and physical therapy were helpful.
Twenty percent of patients with TRAE hyperglycemia had hyperglycemia at baseline, and 30% of TRAEs were in patients with high body mass index (BMI).
There were four treatment-related deaths, all in patients 75 years or older with multiple comorbidities. Three of the four deaths occurred within 30 days of first EV dose in patients with BMI of 30 or greater (acute kidney injury, metabolic acidosis, and multiple organ dysfunction syndrome). The remaining death occurred more than 30 days after the last dose (pneumonitis).
Context and caution
The authors concluded that EV produced durable responses in platinum-ineligible patients with la/mUC, including 20% complete responses. Safety was felt to be as expected, given the known toxicities of the agent and the compromised medical condition of the patients studied.
The study discussant, Arlene O. Siefker-Radtke, MD, of the University of Texas MD Anderson Cancer Center, Houston, agreed that EV fills an unmet need, showing impressive responses in patients with visceral, liver, and bone metastases. She agreed that EV should be investigated across the spectrum of urothelial cancer.
Dr. Siefker-Radtke reminded attendees that the FDA package insert for EV described a 48% increase in the area under the concentration-time curve concentration of the MMAE active moiety in patients with mild hepatic impairment and that EV use should be avoided entirely in patients with moderate to severe liver disease.
She speculated whether augmented toxicity in patients with a high BMI could be attributable to clinically occult impaired hepatic function from fatty liver infiltration.
She indicated that clinicians should monitor closely patients with higher BMI and grade 3-4 hyperglycemia or elevated hemoglobin A1c levels and advised holding EV in patients who develop:
- Glucose levels above 250 mg/dL
- Peeling skin or bullous skin lesions. These lesions can be heralded by a diffuse erythematous or papillary rash in the preceding weeks.
- Grade 3 diarrhea or mucosal membrane toxicity of other types.
Notwithstanding concerns about toxicity and the need for monitoring, EV merits continued study in combination with other agents and in additional settings in the clinical spectrum of urothelial cancer. It is an important new option for oncologists caring for patients with urothelial cancer.
The EV-201 study was funded by Seagen. Dr. Balar and Dr. Siefker-Radtke disclosed relationships with Seagen and many other companies.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM GUCS 2021
How has the pandemic affected rural and urban cancer patients?
Research has shown that, compared with their urban counterparts, rural cancer patients have higher cancer-related mortality and other negative treatment outcomes.
Among other explanations, the disparity has been attributed to lower education and income levels, medical and behavioral risk factors, differences in health literacy, and lower confidence in the medical system among rural residents (JCO Oncol Pract. 2020 Jul;16(7):422-30).
A new survey has provided some insight into how the COVID-19 pandemic has impacted rural and urban cancer patients differently.
The survey showed that urban patients were more likely to report changes to their daily lives, thought themselves more likely to become infected with SARS-CoV-2, and were more likely to take measures to mitigate the risk of infection. However, there were no major differences between urban and rural patients with regard to changes in social interaction.
Bailee Daniels of the University of Utah in Salt Lake City, presented these results at the AACR Virtual Meeting: COVID-19 and Cancer (Abstract S04-03).
The COVID-19 and Oncology Patient Experience Consortium
Ms. Daniels explained that the COVID-19 and Oncology Patient Experience (COPES) Consortium was created to investigate various aspects of the patient experience during the pandemic. Three cancer centers – Moffitt Cancer Center, Huntsman Cancer Institute, and the Sylvester Comprehensive Cancer Center – participate in COPES.
At Huntsman, investigators studied social and health behaviors of cancer patients to assess whether there was a difference between those from rural and urban areas. The researchers looked at the impact of the pandemic on psychosocial outcomes, preventive measures patients implemented, and their perceptions of the risk of SARS-CoV-2 infection.
The team’s hypothesis was that rural patients might be more vulnerable than urban patients to the effects of social isolation, emotional distress, and health-adverse behaviors, but the investigators noted that there has been no prior research on the topic.
Assessing behaviors, attitudes, and outcomes
Between August and September 2020, the researchers surveyed 1,328 adult cancer patients who had visited Huntsman in the previous 4 years and who were enrolled in Huntsman’s Total Cancer Care or Precision Exercise Prescription studies.
Patients completed questionnaires that encompassed demographic and clinical factors, employment status, health behaviors, and infection preventive measures. Questionnaires were provided in electronic, paper, or phone-based formats. Information regarding age, race, ethnicity, and tumor stage was abstracted from Huntsman’s electronic health record.
Modifications in daily life and social interaction were assessed on a 5-point scale. Changes in exercise habits and alcohol consumption were assessed on a 3-point scale. Infection mitigation measures (the use of face masks and hand sanitizer) and perceptions about the likelihood of SARS-CoV-2 infection were measured.
The rural-urban community area codes system, which classifies U.S. census tracts by measures of population density, urbanization, and daily commuting, was utilized to categorize patients into rural and urban residences.
Characteristics of urban and rural cancer patients
There were 997 urban and 331 rural participants. The mean age was 60.1 years in the urban population and 62.6 years in the rural population (P = .01). There were no urban-rural differences in sex, ethnicity, cancer stage, or body mass index.
More urban than rural participants were employed full- or part-time (45% vs. 37%; P = .045). The rural counties had more patients who were not currently employed, primarily due to retirement (77% vs. 69% urban; P < .001).
“No health insurance coverage” was reported by 2% of urban and 4% of rural participants (P = .009), and 85% of all patients reported “good” to “excellent” overall health. Cancer patients in rural counties were significantly more likely to have ever smoked (37% vs. 25% urban; P = .001). In addition, alcohol consumption in the previous year was higher in rural patients. “Every day to less than once monthly” alcohol usage was reported by 44% of urban and 60% of rural patients (P < .001).
Changes in daily life and health-related behavior during the pandemic
Urban patients were more likely to report changes in their daily lives due to the pandemic. Specifically, 35% of urban patients and 26% of rural patients said the pandemic had changed their daily life “a lot” (P = .001).
However, there were no major differences between urban and rural patients when it came to changes in social interaction in the past month or feeling lonely in the past month (P = .45 and P = .88, respectively). Similarly, there were no significant differences for changes in alcohol consumption between the groups (P = .90).
Changes in exercise habits due to the pandemic were more common among patients in urban counties (51% vs. 39% rural; P < .001), though similar percentages of patients reported exercising less (44% urban vs. 45% rural) or more frequently (24% urban vs. 20% rural).
In terms of infection mitigation measures, urban patients were more likely to use face masks “very often” (83% vs. 66% rural; P < .001), while hand sanitizer was used “very often” among 66% of urban and 57% of rural participants (P = .05).
Urban participants were more likely than were their rural counterparts to think themselves “somewhat” or “very” likely to develop COVID-19 (22% vs. 14%; P = .04).
It might be short-sighted for oncology and public health specialists to be dismissive of differences in infection mitigation behaviors and perceptions of vulnerability to SARS-CoV-2 infection. Those behaviors and perceptions of risk could lead to lower vaccination rates in rural areas. If that occurs, there would be major negative consequences for the long-term health of rural communities and their medically vulnerable residents.
Future directions
Although the first 6 months of the COVID-19 pandemic had disparate effects on cancer patients living in rural and urban counties, the reasons for the disparities are complex and not easily explained by this study.
It is possible that sequential administration of the survey during the pandemic would have uncovered greater variances in attitude and health-related behaviors.
As Ms. Daniels noted, when the survey was performed, Utah had not experienced a high frequency of COVID-19 cases. Furthermore, different levels of restrictions were implemented on a county-by-county basis, potentially influencing patients’ behaviors, psychosocial adjustment, and perceptions of risk.
In addition, there may have been differences in unmeasured endpoints (infection rates, medical care utilization via telemedicine, hospitalization rates, late effects, and mortality) between the urban and rural populations.
As the investigators concluded, further research is needed to better characterize the pandemic’s short- and long-term effects on cancer patients in rural and urban settings and appropriate interventions. Such studies may yield insights into the various facets of the well-documented “rural health gap” in cancer outcomes and interventions that could narrow the gap in spheres beyond the COVID-19 pandemic.
Ms. Daniels reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Research has shown that, compared with their urban counterparts, rural cancer patients have higher cancer-related mortality and other negative treatment outcomes.
Among other explanations, the disparity has been attributed to lower education and income levels, medical and behavioral risk factors, differences in health literacy, and lower confidence in the medical system among rural residents (JCO Oncol Pract. 2020 Jul;16(7):422-30).
A new survey has provided some insight into how the COVID-19 pandemic has impacted rural and urban cancer patients differently.
The survey showed that urban patients were more likely to report changes to their daily lives, thought themselves more likely to become infected with SARS-CoV-2, and were more likely to take measures to mitigate the risk of infection. However, there were no major differences between urban and rural patients with regard to changes in social interaction.
Bailee Daniels of the University of Utah in Salt Lake City, presented these results at the AACR Virtual Meeting: COVID-19 and Cancer (Abstract S04-03).
The COVID-19 and Oncology Patient Experience Consortium
Ms. Daniels explained that the COVID-19 and Oncology Patient Experience (COPES) Consortium was created to investigate various aspects of the patient experience during the pandemic. Three cancer centers – Moffitt Cancer Center, Huntsman Cancer Institute, and the Sylvester Comprehensive Cancer Center – participate in COPES.
At Huntsman, investigators studied social and health behaviors of cancer patients to assess whether there was a difference between those from rural and urban areas. The researchers looked at the impact of the pandemic on psychosocial outcomes, preventive measures patients implemented, and their perceptions of the risk of SARS-CoV-2 infection.
The team’s hypothesis was that rural patients might be more vulnerable than urban patients to the effects of social isolation, emotional distress, and health-adverse behaviors, but the investigators noted that there has been no prior research on the topic.
Assessing behaviors, attitudes, and outcomes
Between August and September 2020, the researchers surveyed 1,328 adult cancer patients who had visited Huntsman in the previous 4 years and who were enrolled in Huntsman’s Total Cancer Care or Precision Exercise Prescription studies.
Patients completed questionnaires that encompassed demographic and clinical factors, employment status, health behaviors, and infection preventive measures. Questionnaires were provided in electronic, paper, or phone-based formats. Information regarding age, race, ethnicity, and tumor stage was abstracted from Huntsman’s electronic health record.
Modifications in daily life and social interaction were assessed on a 5-point scale. Changes in exercise habits and alcohol consumption were assessed on a 3-point scale. Infection mitigation measures (the use of face masks and hand sanitizer) and perceptions about the likelihood of SARS-CoV-2 infection were measured.
The rural-urban community area codes system, which classifies U.S. census tracts by measures of population density, urbanization, and daily commuting, was utilized to categorize patients into rural and urban residences.
Characteristics of urban and rural cancer patients
There were 997 urban and 331 rural participants. The mean age was 60.1 years in the urban population and 62.6 years in the rural population (P = .01). There were no urban-rural differences in sex, ethnicity, cancer stage, or body mass index.
More urban than rural participants were employed full- or part-time (45% vs. 37%; P = .045). The rural counties had more patients who were not currently employed, primarily due to retirement (77% vs. 69% urban; P < .001).
“No health insurance coverage” was reported by 2% of urban and 4% of rural participants (P = .009), and 85% of all patients reported “good” to “excellent” overall health. Cancer patients in rural counties were significantly more likely to have ever smoked (37% vs. 25% urban; P = .001). In addition, alcohol consumption in the previous year was higher in rural patients. “Every day to less than once monthly” alcohol usage was reported by 44% of urban and 60% of rural patients (P < .001).
Changes in daily life and health-related behavior during the pandemic
Urban patients were more likely to report changes in their daily lives due to the pandemic. Specifically, 35% of urban patients and 26% of rural patients said the pandemic had changed their daily life “a lot” (P = .001).
However, there were no major differences between urban and rural patients when it came to changes in social interaction in the past month or feeling lonely in the past month (P = .45 and P = .88, respectively). Similarly, there were no significant differences for changes in alcohol consumption between the groups (P = .90).
Changes in exercise habits due to the pandemic were more common among patients in urban counties (51% vs. 39% rural; P < .001), though similar percentages of patients reported exercising less (44% urban vs. 45% rural) or more frequently (24% urban vs. 20% rural).
In terms of infection mitigation measures, urban patients were more likely to use face masks “very often” (83% vs. 66% rural; P < .001), while hand sanitizer was used “very often” among 66% of urban and 57% of rural participants (P = .05).
Urban participants were more likely than were their rural counterparts to think themselves “somewhat” or “very” likely to develop COVID-19 (22% vs. 14%; P = .04).
It might be short-sighted for oncology and public health specialists to be dismissive of differences in infection mitigation behaviors and perceptions of vulnerability to SARS-CoV-2 infection. Those behaviors and perceptions of risk could lead to lower vaccination rates in rural areas. If that occurs, there would be major negative consequences for the long-term health of rural communities and their medically vulnerable residents.
Future directions
Although the first 6 months of the COVID-19 pandemic had disparate effects on cancer patients living in rural and urban counties, the reasons for the disparities are complex and not easily explained by this study.
It is possible that sequential administration of the survey during the pandemic would have uncovered greater variances in attitude and health-related behaviors.
As Ms. Daniels noted, when the survey was performed, Utah had not experienced a high frequency of COVID-19 cases. Furthermore, different levels of restrictions were implemented on a county-by-county basis, potentially influencing patients’ behaviors, psychosocial adjustment, and perceptions of risk.
In addition, there may have been differences in unmeasured endpoints (infection rates, medical care utilization via telemedicine, hospitalization rates, late effects, and mortality) between the urban and rural populations.
As the investigators concluded, further research is needed to better characterize the pandemic’s short- and long-term effects on cancer patients in rural and urban settings and appropriate interventions. Such studies may yield insights into the various facets of the well-documented “rural health gap” in cancer outcomes and interventions that could narrow the gap in spheres beyond the COVID-19 pandemic.
Ms. Daniels reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Research has shown that, compared with their urban counterparts, rural cancer patients have higher cancer-related mortality and other negative treatment outcomes.
Among other explanations, the disparity has been attributed to lower education and income levels, medical and behavioral risk factors, differences in health literacy, and lower confidence in the medical system among rural residents (JCO Oncol Pract. 2020 Jul;16(7):422-30).
A new survey has provided some insight into how the COVID-19 pandemic has impacted rural and urban cancer patients differently.
The survey showed that urban patients were more likely to report changes to their daily lives, thought themselves more likely to become infected with SARS-CoV-2, and were more likely to take measures to mitigate the risk of infection. However, there were no major differences between urban and rural patients with regard to changes in social interaction.
Bailee Daniels of the University of Utah in Salt Lake City, presented these results at the AACR Virtual Meeting: COVID-19 and Cancer (Abstract S04-03).
The COVID-19 and Oncology Patient Experience Consortium
Ms. Daniels explained that the COVID-19 and Oncology Patient Experience (COPES) Consortium was created to investigate various aspects of the patient experience during the pandemic. Three cancer centers – Moffitt Cancer Center, Huntsman Cancer Institute, and the Sylvester Comprehensive Cancer Center – participate in COPES.
At Huntsman, investigators studied social and health behaviors of cancer patients to assess whether there was a difference between those from rural and urban areas. The researchers looked at the impact of the pandemic on psychosocial outcomes, preventive measures patients implemented, and their perceptions of the risk of SARS-CoV-2 infection.
The team’s hypothesis was that rural patients might be more vulnerable than urban patients to the effects of social isolation, emotional distress, and health-adverse behaviors, but the investigators noted that there has been no prior research on the topic.
Assessing behaviors, attitudes, and outcomes
Between August and September 2020, the researchers surveyed 1,328 adult cancer patients who had visited Huntsman in the previous 4 years and who were enrolled in Huntsman’s Total Cancer Care or Precision Exercise Prescription studies.
Patients completed questionnaires that encompassed demographic and clinical factors, employment status, health behaviors, and infection preventive measures. Questionnaires were provided in electronic, paper, or phone-based formats. Information regarding age, race, ethnicity, and tumor stage was abstracted from Huntsman’s electronic health record.
Modifications in daily life and social interaction were assessed on a 5-point scale. Changes in exercise habits and alcohol consumption were assessed on a 3-point scale. Infection mitigation measures (the use of face masks and hand sanitizer) and perceptions about the likelihood of SARS-CoV-2 infection were measured.
The rural-urban community area codes system, which classifies U.S. census tracts by measures of population density, urbanization, and daily commuting, was utilized to categorize patients into rural and urban residences.
Characteristics of urban and rural cancer patients
There were 997 urban and 331 rural participants. The mean age was 60.1 years in the urban population and 62.6 years in the rural population (P = .01). There were no urban-rural differences in sex, ethnicity, cancer stage, or body mass index.
More urban than rural participants were employed full- or part-time (45% vs. 37%; P = .045). The rural counties had more patients who were not currently employed, primarily due to retirement (77% vs. 69% urban; P < .001).
“No health insurance coverage” was reported by 2% of urban and 4% of rural participants (P = .009), and 85% of all patients reported “good” to “excellent” overall health. Cancer patients in rural counties were significantly more likely to have ever smoked (37% vs. 25% urban; P = .001). In addition, alcohol consumption in the previous year was higher in rural patients. “Every day to less than once monthly” alcohol usage was reported by 44% of urban and 60% of rural patients (P < .001).
Changes in daily life and health-related behavior during the pandemic
Urban patients were more likely to report changes in their daily lives due to the pandemic. Specifically, 35% of urban patients and 26% of rural patients said the pandemic had changed their daily life “a lot” (P = .001).
However, there were no major differences between urban and rural patients when it came to changes in social interaction in the past month or feeling lonely in the past month (P = .45 and P = .88, respectively). Similarly, there were no significant differences for changes in alcohol consumption between the groups (P = .90).
Changes in exercise habits due to the pandemic were more common among patients in urban counties (51% vs. 39% rural; P < .001), though similar percentages of patients reported exercising less (44% urban vs. 45% rural) or more frequently (24% urban vs. 20% rural).
In terms of infection mitigation measures, urban patients were more likely to use face masks “very often” (83% vs. 66% rural; P < .001), while hand sanitizer was used “very often” among 66% of urban and 57% of rural participants (P = .05).
Urban participants were more likely than were their rural counterparts to think themselves “somewhat” or “very” likely to develop COVID-19 (22% vs. 14%; P = .04).
It might be short-sighted for oncology and public health specialists to be dismissive of differences in infection mitigation behaviors and perceptions of vulnerability to SARS-CoV-2 infection. Those behaviors and perceptions of risk could lead to lower vaccination rates in rural areas. If that occurs, there would be major negative consequences for the long-term health of rural communities and their medically vulnerable residents.
Future directions
Although the first 6 months of the COVID-19 pandemic had disparate effects on cancer patients living in rural and urban counties, the reasons for the disparities are complex and not easily explained by this study.
It is possible that sequential administration of the survey during the pandemic would have uncovered greater variances in attitude and health-related behaviors.
As Ms. Daniels noted, when the survey was performed, Utah had not experienced a high frequency of COVID-19 cases. Furthermore, different levels of restrictions were implemented on a county-by-county basis, potentially influencing patients’ behaviors, psychosocial adjustment, and perceptions of risk.
In addition, there may have been differences in unmeasured endpoints (infection rates, medical care utilization via telemedicine, hospitalization rates, late effects, and mortality) between the urban and rural populations.
As the investigators concluded, further research is needed to better characterize the pandemic’s short- and long-term effects on cancer patients in rural and urban settings and appropriate interventions. Such studies may yield insights into the various facets of the well-documented “rural health gap” in cancer outcomes and interventions that could narrow the gap in spheres beyond the COVID-19 pandemic.
Ms. Daniels reported having no relevant disclosures.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM AACR: COVID-19 AND CANCER 2021
X-ray vision: Using AI to maximize the value of radiographic images
Artificial intelligence (AI) is expected to one day affect the entire continuum of cancer care – from screening and risk prediction to diagnosis, risk stratification, treatment selection, and follow-up, according to an expert in the field.
Hugo J.W.L. Aerts, PhD, director of the AI in Medicine Program at Brigham and Women’s Hospital in Boston, described studies using AI for some of these purposes during a presentation at the AACR Virtual Special Conference: Artificial Intelligence, Diagnosis, and Imaging (Abstract IA-06).
In one study, Dr. Aerts and colleagues set out to determine whether a convolutional neural network (CNN) could extract prognostic information from chest radiographs. The researchers tested this theory using patients from two trials – the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and the National Lung Screening Trial (NLST).
The team developed a CNN, called CXR-risk, and tested whether it could predict the longevity and prognosis of patients in the PLCO (n = 52,320) and NLST (n = 5,493) trials over a 12-year time period, based only on chest radiographs. No clinical information, demographics, radiographic interpretations, duration of follow-up, or censoring were provided to the deep-learning system.
CXR-risk output was stratified into five categories of radiographic risk scores for probability of death, from 0 (very low likelihood of mortality) to 1 (very high likelihood of mortality).
The investigators found a graded association between radiographic risk score and mortality. The very-high-risk group had mortality rates of 53.0% (PLCO) and 33.9% (NLST). In both trials, this was significantly higher than for the very-low-risk group. The unadjusted hazard ratio was 18.3 in the PCLO data set and 15.2 in the NLST data set (P < .001 for both).
This association was maintained after adjustment for radiologists’ findings (e.g., a lung nodule) and risk factors such as age, gender, and comorbid illnesses like diabetes. The adjusted HR was 4.8 in the PCLO data set and 7.0 in the NLST data set (P < .001 for both).
In both data sets, individuals in the very-high-risk group were significantly more likely to die of lung cancer. The aHR was 11.1 in the PCLO data set and 8.4 in the NSLT data set (P < .001 for both).
This might be expected for people who were interested in being screened for lung cancer. However, patients in the very-high-risk group were also more likely to die of cardiovascular illness (aHR, 3.6 for PLCO and 47.8 for NSLT; P < .001 for both) and respiratory illness (aHR, 27.5 for PLCO and 31.9 for NLST; P ≤ .001 for both).
With this information, a clinician could initiate additional testing and/or utilize more aggressive surveillance measures. If an oncologist considered therapy for a patient with newly diagnosed cancer, treatment choices and stratification for adverse events would be more intelligently planned.
Using AI to predict the risk of lung cancer
In another study, Dr. Aerts and colleagues developed and validated a CNN called CXR-LC, which was based on CXR-risk. The goal of this study was to see if CXR-LC could predict long-term incident lung cancer using data available in the EHR, including chest radiographs, age, sex, and smoking status.
The CXR-LC model was developed using data from the PLCO trial (n = 41,856) and was validated in smokers from the PLCO trial (n = 5,615; 12-year follow-up) as well as heavy smokers from the NLST trial (n = 5,493; 6-year follow-up).
Results showed that CXR-LC was able to predict which patients were at highest risk for developing lung cancer.
CXR-LC had better discrimination for incident lung cancer than did Medicare eligibility in the PLCO data set (area under the curve, 0.755 vs. 0.634; P < .001). And the performance of CXR-LC was similar to that of the PLCOM2012 risk score in both the PLCO data set (AUC, 0.755 vs. 0.751) and the NLST data set (AUC, 0.659 vs. 0.650).
When they were compared in screening populations of equal size, CXR-LC was more sensitive than Medicare eligibility criteria in the PLCO data set (74.9% vs. 63.8%; P = .012) and missed 30.7% fewer incident lung cancer diagnoses.
AI as a substitute for specialized testing and consultation
In a third study, Dr. Aerts and colleagues used a CNN to predict cardiovascular risk by assessing coronary artery calcium (CAC) from clinically obtained, readily available CT scans.
Ordinarily, identifying CAC – an accurate predictor of cardiovascular events – requires specialized expertise (manual measurement and cardiologist interpretation), time (estimated at 20 minutes/scan), and equipment (ECG-gated cardiac CT scan and special software).
In this study, the researchers used a fully end-to-end automated system with analytic time measured in less than 2 seconds.
The team trained and tuned their CNN using the Framingham Heart Study Offspring and Third Generation cohorts (n = 1,636), which included asymptomatic patients with high-quality, cardiac-gated CT scans for CAC quantification.
The researchers then tested the CNN on two asymptomatic and two symptomatic cohorts:
- Asymptomatic Framingham Heart Study participants (n = 663) in whom the outcome measures were cardiovascular disease and death.
- Asymptomatic NLST participants (n = 14,959) in whom the outcome measure was atherosclerotic cardiovascular death.
- Symptomatic PROMISE study participants with stable chest pain (n = 4,021) in whom the outcome measures were all-cause mortality, MI, and hospitalization for unstable angina.
- Symptomatic ROMICAT-II study patients with acute chest pain (n = 441) in whom the outcome measure was acute coronary syndrome at 28 days.
Among 5,521 subjects across all testing cohorts with cardiac-gated and nongated chest CT scans, the CNN and expert reader interpretations agreed on the CAC risk scores with a high level of concordance (kappa, 0.71; concordance rate, 0.79).
There was a very high Spearman’s correlation of 0.92 (P < .0001) and substantial agreement between automatically and manually calculated CAC risk groups, substantiating robust risk prediction for cardiovascular disease across multiple clinical scenarios.
Dr. Aerts commented that, among the NLST participants who had the highest risk of developing lung cancer, the risk of cardiovascular death was as high as the risk of death from lung cancer.
Using AI to assess patient outcomes
In an unpublished study, Dr. Aerts and colleagues used AI in an attempt to determine whether changes in measurements of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), and skeletal muscle mass would provide clues about treatment outcomes in lung cancer patients.
The researchers developed a deep learning model using data from 1,129 patients at Massachusetts General and Brigham and Women’s Hospitals, measuring SAT, VAT, and muscle mass. The team applied the measurement system to a population of 12,128 outpatients and calculated z scores for SAT, VAT, and muscle mass to determine “normal” values.
When they applied the norms to surgical lung cancer data sets from the Boston Lung Cancer Study (n = 437) and TRACERx study (n = 394), the researchers found that smokers had lower adiposity and lower muscle mass than never-smokers.
More importantly, over time, among lung cancer patients who lost greater than 5% of VAT, SAT, and muscle mass, those patients with the greatest SAT loss (P < .0001) or VAT loss (P = .0015) had the lowest lung cancer–specific survival in the TRACERx study. There was no significant impairment of lung cancer-specific survival for patients who experienced skeletal muscle loss (P = .23).
The same observation was made for overall survival among patients enrolled in the Boston Lung Cancer Study, using the 5% threshold. Overall survival was significantly worse with increasing VAT loss (P = .0023) and SAT loss (P = .0082) but not with increasing skeletal muscle loss (P = .3).
The investigators speculated about whether the correlation between body composition and clinical outcome could yield clues about tumor biology. To test this, the researchers used the RNA sequencing–based ORACLE risk score in lung cancer patients from TRACERx. There was a high correlation between higher ORACLE risk scores and lower VAT and SAT, suggesting that measures of adiposity on CT were reflected in tumor biology patterns on an RNA level in lung cancer patients. There was no such correlation between ORACLE risk scores and skeletal muscle mass.
Wonderment ... tempered by concern and challenges
AI has awe-inspiring potential to yield actionable and prognostically important information from data mining the EHR and extracting the vast quantities of information from images. In some cases (like CAC), it is information that is “hiding in plain sight.” However, Dr. Aerts expressed several cautions, some of which have already plagued AI.
He referenced the Gartner Hype Cycle, which provides a graphic representation of five phases in the life cycle of emerging technologies. The “innovation trigger” is followed by a “peak of inflated expectations,” a “trough of disillusionment,” a “slope of enlightenment,” and a “plateau of productivity.”
Dr. Aerts noted that, in recent years, AI has seemed to fall into the trough of disillusionment, but it may be entering the slope of enlightenment on the way to the plateau of productivity.
His research highlighted several examples of productivity in radiomics in cancer patients and those who are at high risk of developing cancer.
In Dr. Aerts’s opinion, a second concern is replication of AI research results. He noted that, among 400 published studies, only 6% of authors shared the codes that would enable their findings to be corroborated. About 30% shared test data, and 54% shared “pseudocodes,” but transparency and reproducibility are problems for the acceptance and broad implementation of AI.
Dr. Aerts endorsed the Modelhub initiative (www.modelhub.ai), a multi-institutional initiative to advance reproducibility in the AI field and advance its full potential.
However, there are additional concerns about the implementation of radiomics and, more generally, data mining from clinicians’ EHRs to personalize care.
Firstly, it may be laborious and difficult to explain complex, computer-based risk stratification models to patients. Hereditary cancer testing is an example of a risk assessment test that requires complicated explanations that many clinicians relegate to genetics counselors – when patients elect to see them. When a model is not explainable, it undermines the confidence of patients and their care providers, according to an editorial related to the CXR-LC study.
Another issue is that uptake of lung cancer screening, in practice, has been underutilized by individuals who meet current, relatively straightforward Medicare criteria. Despite the apparently better accuracy of the CXR-LC deep-learning model, its complexity and limited access could constitute an additional barrier for the at-risk individuals who should avail themselves of screening.
Furthermore, although age and gender are accurate in most circumstances, there is legitimate concern about the accuracy of, for example, smoking history data and comorbid conditions in current EHRs. Who performs the laborious curation of the input in an AI model to assure its accuracy for individual patients?
Finally, it is unclear how scalable and applicable AI will be to medically underserved populations (e.g., smaller, community-based, free-standing, socioeconomically disadvantaged or rural health care institutions). There are substantial initial and maintenance costs that may limit AI’s availability to some academic institutions and large health maintenance organizations.
As the concerns and challenges are addressed, it will be interesting to see where and when the plateau of productivity for AI in cancer care occurs. When it does, many cancer patients will benefit from enhanced care along the continuum of the complex disease they and their caregivers seek to master.
Dr. Aerts disclosed relationships with Onc.AI outside the presented work.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Artificial intelligence (AI) is expected to one day affect the entire continuum of cancer care – from screening and risk prediction to diagnosis, risk stratification, treatment selection, and follow-up, according to an expert in the field.
Hugo J.W.L. Aerts, PhD, director of the AI in Medicine Program at Brigham and Women’s Hospital in Boston, described studies using AI for some of these purposes during a presentation at the AACR Virtual Special Conference: Artificial Intelligence, Diagnosis, and Imaging (Abstract IA-06).
In one study, Dr. Aerts and colleagues set out to determine whether a convolutional neural network (CNN) could extract prognostic information from chest radiographs. The researchers tested this theory using patients from two trials – the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and the National Lung Screening Trial (NLST).
The team developed a CNN, called CXR-risk, and tested whether it could predict the longevity and prognosis of patients in the PLCO (n = 52,320) and NLST (n = 5,493) trials over a 12-year time period, based only on chest radiographs. No clinical information, demographics, radiographic interpretations, duration of follow-up, or censoring were provided to the deep-learning system.
CXR-risk output was stratified into five categories of radiographic risk scores for probability of death, from 0 (very low likelihood of mortality) to 1 (very high likelihood of mortality).
The investigators found a graded association between radiographic risk score and mortality. The very-high-risk group had mortality rates of 53.0% (PLCO) and 33.9% (NLST). In both trials, this was significantly higher than for the very-low-risk group. The unadjusted hazard ratio was 18.3 in the PCLO data set and 15.2 in the NLST data set (P < .001 for both).
This association was maintained after adjustment for radiologists’ findings (e.g., a lung nodule) and risk factors such as age, gender, and comorbid illnesses like diabetes. The adjusted HR was 4.8 in the PCLO data set and 7.0 in the NLST data set (P < .001 for both).
In both data sets, individuals in the very-high-risk group were significantly more likely to die of lung cancer. The aHR was 11.1 in the PCLO data set and 8.4 in the NSLT data set (P < .001 for both).
This might be expected for people who were interested in being screened for lung cancer. However, patients in the very-high-risk group were also more likely to die of cardiovascular illness (aHR, 3.6 for PLCO and 47.8 for NSLT; P < .001 for both) and respiratory illness (aHR, 27.5 for PLCO and 31.9 for NLST; P ≤ .001 for both).
With this information, a clinician could initiate additional testing and/or utilize more aggressive surveillance measures. If an oncologist considered therapy for a patient with newly diagnosed cancer, treatment choices and stratification for adverse events would be more intelligently planned.
Using AI to predict the risk of lung cancer
In another study, Dr. Aerts and colleagues developed and validated a CNN called CXR-LC, which was based on CXR-risk. The goal of this study was to see if CXR-LC could predict long-term incident lung cancer using data available in the EHR, including chest radiographs, age, sex, and smoking status.
The CXR-LC model was developed using data from the PLCO trial (n = 41,856) and was validated in smokers from the PLCO trial (n = 5,615; 12-year follow-up) as well as heavy smokers from the NLST trial (n = 5,493; 6-year follow-up).
Results showed that CXR-LC was able to predict which patients were at highest risk for developing lung cancer.
CXR-LC had better discrimination for incident lung cancer than did Medicare eligibility in the PLCO data set (area under the curve, 0.755 vs. 0.634; P < .001). And the performance of CXR-LC was similar to that of the PLCOM2012 risk score in both the PLCO data set (AUC, 0.755 vs. 0.751) and the NLST data set (AUC, 0.659 vs. 0.650).
When they were compared in screening populations of equal size, CXR-LC was more sensitive than Medicare eligibility criteria in the PLCO data set (74.9% vs. 63.8%; P = .012) and missed 30.7% fewer incident lung cancer diagnoses.
AI as a substitute for specialized testing and consultation
In a third study, Dr. Aerts and colleagues used a CNN to predict cardiovascular risk by assessing coronary artery calcium (CAC) from clinically obtained, readily available CT scans.
Ordinarily, identifying CAC – an accurate predictor of cardiovascular events – requires specialized expertise (manual measurement and cardiologist interpretation), time (estimated at 20 minutes/scan), and equipment (ECG-gated cardiac CT scan and special software).
In this study, the researchers used a fully end-to-end automated system with analytic time measured in less than 2 seconds.
The team trained and tuned their CNN using the Framingham Heart Study Offspring and Third Generation cohorts (n = 1,636), which included asymptomatic patients with high-quality, cardiac-gated CT scans for CAC quantification.
The researchers then tested the CNN on two asymptomatic and two symptomatic cohorts:
- Asymptomatic Framingham Heart Study participants (n = 663) in whom the outcome measures were cardiovascular disease and death.
- Asymptomatic NLST participants (n = 14,959) in whom the outcome measure was atherosclerotic cardiovascular death.
- Symptomatic PROMISE study participants with stable chest pain (n = 4,021) in whom the outcome measures were all-cause mortality, MI, and hospitalization for unstable angina.
- Symptomatic ROMICAT-II study patients with acute chest pain (n = 441) in whom the outcome measure was acute coronary syndrome at 28 days.
Among 5,521 subjects across all testing cohorts with cardiac-gated and nongated chest CT scans, the CNN and expert reader interpretations agreed on the CAC risk scores with a high level of concordance (kappa, 0.71; concordance rate, 0.79).
There was a very high Spearman’s correlation of 0.92 (P < .0001) and substantial agreement between automatically and manually calculated CAC risk groups, substantiating robust risk prediction for cardiovascular disease across multiple clinical scenarios.
Dr. Aerts commented that, among the NLST participants who had the highest risk of developing lung cancer, the risk of cardiovascular death was as high as the risk of death from lung cancer.
Using AI to assess patient outcomes
In an unpublished study, Dr. Aerts and colleagues used AI in an attempt to determine whether changes in measurements of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), and skeletal muscle mass would provide clues about treatment outcomes in lung cancer patients.
The researchers developed a deep learning model using data from 1,129 patients at Massachusetts General and Brigham and Women’s Hospitals, measuring SAT, VAT, and muscle mass. The team applied the measurement system to a population of 12,128 outpatients and calculated z scores for SAT, VAT, and muscle mass to determine “normal” values.
When they applied the norms to surgical lung cancer data sets from the Boston Lung Cancer Study (n = 437) and TRACERx study (n = 394), the researchers found that smokers had lower adiposity and lower muscle mass than never-smokers.
More importantly, over time, among lung cancer patients who lost greater than 5% of VAT, SAT, and muscle mass, those patients with the greatest SAT loss (P < .0001) or VAT loss (P = .0015) had the lowest lung cancer–specific survival in the TRACERx study. There was no significant impairment of lung cancer-specific survival for patients who experienced skeletal muscle loss (P = .23).
The same observation was made for overall survival among patients enrolled in the Boston Lung Cancer Study, using the 5% threshold. Overall survival was significantly worse with increasing VAT loss (P = .0023) and SAT loss (P = .0082) but not with increasing skeletal muscle loss (P = .3).
The investigators speculated about whether the correlation between body composition and clinical outcome could yield clues about tumor biology. To test this, the researchers used the RNA sequencing–based ORACLE risk score in lung cancer patients from TRACERx. There was a high correlation between higher ORACLE risk scores and lower VAT and SAT, suggesting that measures of adiposity on CT were reflected in tumor biology patterns on an RNA level in lung cancer patients. There was no such correlation between ORACLE risk scores and skeletal muscle mass.
Wonderment ... tempered by concern and challenges
AI has awe-inspiring potential to yield actionable and prognostically important information from data mining the EHR and extracting the vast quantities of information from images. In some cases (like CAC), it is information that is “hiding in plain sight.” However, Dr. Aerts expressed several cautions, some of which have already plagued AI.
He referenced the Gartner Hype Cycle, which provides a graphic representation of five phases in the life cycle of emerging technologies. The “innovation trigger” is followed by a “peak of inflated expectations,” a “trough of disillusionment,” a “slope of enlightenment,” and a “plateau of productivity.”
Dr. Aerts noted that, in recent years, AI has seemed to fall into the trough of disillusionment, but it may be entering the slope of enlightenment on the way to the plateau of productivity.
His research highlighted several examples of productivity in radiomics in cancer patients and those who are at high risk of developing cancer.
In Dr. Aerts’s opinion, a second concern is replication of AI research results. He noted that, among 400 published studies, only 6% of authors shared the codes that would enable their findings to be corroborated. About 30% shared test data, and 54% shared “pseudocodes,” but transparency and reproducibility are problems for the acceptance and broad implementation of AI.
Dr. Aerts endorsed the Modelhub initiative (www.modelhub.ai), a multi-institutional initiative to advance reproducibility in the AI field and advance its full potential.
However, there are additional concerns about the implementation of radiomics and, more generally, data mining from clinicians’ EHRs to personalize care.
Firstly, it may be laborious and difficult to explain complex, computer-based risk stratification models to patients. Hereditary cancer testing is an example of a risk assessment test that requires complicated explanations that many clinicians relegate to genetics counselors – when patients elect to see them. When a model is not explainable, it undermines the confidence of patients and their care providers, according to an editorial related to the CXR-LC study.
Another issue is that uptake of lung cancer screening, in practice, has been underutilized by individuals who meet current, relatively straightforward Medicare criteria. Despite the apparently better accuracy of the CXR-LC deep-learning model, its complexity and limited access could constitute an additional barrier for the at-risk individuals who should avail themselves of screening.
Furthermore, although age and gender are accurate in most circumstances, there is legitimate concern about the accuracy of, for example, smoking history data and comorbid conditions in current EHRs. Who performs the laborious curation of the input in an AI model to assure its accuracy for individual patients?
Finally, it is unclear how scalable and applicable AI will be to medically underserved populations (e.g., smaller, community-based, free-standing, socioeconomically disadvantaged or rural health care institutions). There are substantial initial and maintenance costs that may limit AI’s availability to some academic institutions and large health maintenance organizations.
As the concerns and challenges are addressed, it will be interesting to see where and when the plateau of productivity for AI in cancer care occurs. When it does, many cancer patients will benefit from enhanced care along the continuum of the complex disease they and their caregivers seek to master.
Dr. Aerts disclosed relationships with Onc.AI outside the presented work.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
Artificial intelligence (AI) is expected to one day affect the entire continuum of cancer care – from screening and risk prediction to diagnosis, risk stratification, treatment selection, and follow-up, according to an expert in the field.
Hugo J.W.L. Aerts, PhD, director of the AI in Medicine Program at Brigham and Women’s Hospital in Boston, described studies using AI for some of these purposes during a presentation at the AACR Virtual Special Conference: Artificial Intelligence, Diagnosis, and Imaging (Abstract IA-06).
In one study, Dr. Aerts and colleagues set out to determine whether a convolutional neural network (CNN) could extract prognostic information from chest radiographs. The researchers tested this theory using patients from two trials – the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and the National Lung Screening Trial (NLST).
The team developed a CNN, called CXR-risk, and tested whether it could predict the longevity and prognosis of patients in the PLCO (n = 52,320) and NLST (n = 5,493) trials over a 12-year time period, based only on chest radiographs. No clinical information, demographics, radiographic interpretations, duration of follow-up, or censoring were provided to the deep-learning system.
CXR-risk output was stratified into five categories of radiographic risk scores for probability of death, from 0 (very low likelihood of mortality) to 1 (very high likelihood of mortality).
The investigators found a graded association between radiographic risk score and mortality. The very-high-risk group had mortality rates of 53.0% (PLCO) and 33.9% (NLST). In both trials, this was significantly higher than for the very-low-risk group. The unadjusted hazard ratio was 18.3 in the PCLO data set and 15.2 in the NLST data set (P < .001 for both).
This association was maintained after adjustment for radiologists’ findings (e.g., a lung nodule) and risk factors such as age, gender, and comorbid illnesses like diabetes. The adjusted HR was 4.8 in the PCLO data set and 7.0 in the NLST data set (P < .001 for both).
In both data sets, individuals in the very-high-risk group were significantly more likely to die of lung cancer. The aHR was 11.1 in the PCLO data set and 8.4 in the NSLT data set (P < .001 for both).
This might be expected for people who were interested in being screened for lung cancer. However, patients in the very-high-risk group were also more likely to die of cardiovascular illness (aHR, 3.6 for PLCO and 47.8 for NSLT; P < .001 for both) and respiratory illness (aHR, 27.5 for PLCO and 31.9 for NLST; P ≤ .001 for both).
With this information, a clinician could initiate additional testing and/or utilize more aggressive surveillance measures. If an oncologist considered therapy for a patient with newly diagnosed cancer, treatment choices and stratification for adverse events would be more intelligently planned.
Using AI to predict the risk of lung cancer
In another study, Dr. Aerts and colleagues developed and validated a CNN called CXR-LC, which was based on CXR-risk. The goal of this study was to see if CXR-LC could predict long-term incident lung cancer using data available in the EHR, including chest radiographs, age, sex, and smoking status.
The CXR-LC model was developed using data from the PLCO trial (n = 41,856) and was validated in smokers from the PLCO trial (n = 5,615; 12-year follow-up) as well as heavy smokers from the NLST trial (n = 5,493; 6-year follow-up).
Results showed that CXR-LC was able to predict which patients were at highest risk for developing lung cancer.
CXR-LC had better discrimination for incident lung cancer than did Medicare eligibility in the PLCO data set (area under the curve, 0.755 vs. 0.634; P < .001). And the performance of CXR-LC was similar to that of the PLCOM2012 risk score in both the PLCO data set (AUC, 0.755 vs. 0.751) and the NLST data set (AUC, 0.659 vs. 0.650).
When they were compared in screening populations of equal size, CXR-LC was more sensitive than Medicare eligibility criteria in the PLCO data set (74.9% vs. 63.8%; P = .012) and missed 30.7% fewer incident lung cancer diagnoses.
AI as a substitute for specialized testing and consultation
In a third study, Dr. Aerts and colleagues used a CNN to predict cardiovascular risk by assessing coronary artery calcium (CAC) from clinically obtained, readily available CT scans.
Ordinarily, identifying CAC – an accurate predictor of cardiovascular events – requires specialized expertise (manual measurement and cardiologist interpretation), time (estimated at 20 minutes/scan), and equipment (ECG-gated cardiac CT scan and special software).
In this study, the researchers used a fully end-to-end automated system with analytic time measured in less than 2 seconds.
The team trained and tuned their CNN using the Framingham Heart Study Offspring and Third Generation cohorts (n = 1,636), which included asymptomatic patients with high-quality, cardiac-gated CT scans for CAC quantification.
The researchers then tested the CNN on two asymptomatic and two symptomatic cohorts:
- Asymptomatic Framingham Heart Study participants (n = 663) in whom the outcome measures were cardiovascular disease and death.
- Asymptomatic NLST participants (n = 14,959) in whom the outcome measure was atherosclerotic cardiovascular death.
- Symptomatic PROMISE study participants with stable chest pain (n = 4,021) in whom the outcome measures were all-cause mortality, MI, and hospitalization for unstable angina.
- Symptomatic ROMICAT-II study patients with acute chest pain (n = 441) in whom the outcome measure was acute coronary syndrome at 28 days.
Among 5,521 subjects across all testing cohorts with cardiac-gated and nongated chest CT scans, the CNN and expert reader interpretations agreed on the CAC risk scores with a high level of concordance (kappa, 0.71; concordance rate, 0.79).
There was a very high Spearman’s correlation of 0.92 (P < .0001) and substantial agreement between automatically and manually calculated CAC risk groups, substantiating robust risk prediction for cardiovascular disease across multiple clinical scenarios.
Dr. Aerts commented that, among the NLST participants who had the highest risk of developing lung cancer, the risk of cardiovascular death was as high as the risk of death from lung cancer.
Using AI to assess patient outcomes
In an unpublished study, Dr. Aerts and colleagues used AI in an attempt to determine whether changes in measurements of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), and skeletal muscle mass would provide clues about treatment outcomes in lung cancer patients.
The researchers developed a deep learning model using data from 1,129 patients at Massachusetts General and Brigham and Women’s Hospitals, measuring SAT, VAT, and muscle mass. The team applied the measurement system to a population of 12,128 outpatients and calculated z scores for SAT, VAT, and muscle mass to determine “normal” values.
When they applied the norms to surgical lung cancer data sets from the Boston Lung Cancer Study (n = 437) and TRACERx study (n = 394), the researchers found that smokers had lower adiposity and lower muscle mass than never-smokers.
More importantly, over time, among lung cancer patients who lost greater than 5% of VAT, SAT, and muscle mass, those patients with the greatest SAT loss (P < .0001) or VAT loss (P = .0015) had the lowest lung cancer–specific survival in the TRACERx study. There was no significant impairment of lung cancer-specific survival for patients who experienced skeletal muscle loss (P = .23).
The same observation was made for overall survival among patients enrolled in the Boston Lung Cancer Study, using the 5% threshold. Overall survival was significantly worse with increasing VAT loss (P = .0023) and SAT loss (P = .0082) but not with increasing skeletal muscle loss (P = .3).
The investigators speculated about whether the correlation between body composition and clinical outcome could yield clues about tumor biology. To test this, the researchers used the RNA sequencing–based ORACLE risk score in lung cancer patients from TRACERx. There was a high correlation between higher ORACLE risk scores and lower VAT and SAT, suggesting that measures of adiposity on CT were reflected in tumor biology patterns on an RNA level in lung cancer patients. There was no such correlation between ORACLE risk scores and skeletal muscle mass.
Wonderment ... tempered by concern and challenges
AI has awe-inspiring potential to yield actionable and prognostically important information from data mining the EHR and extracting the vast quantities of information from images. In some cases (like CAC), it is information that is “hiding in plain sight.” However, Dr. Aerts expressed several cautions, some of which have already plagued AI.
He referenced the Gartner Hype Cycle, which provides a graphic representation of five phases in the life cycle of emerging technologies. The “innovation trigger” is followed by a “peak of inflated expectations,” a “trough of disillusionment,” a “slope of enlightenment,” and a “plateau of productivity.”
Dr. Aerts noted that, in recent years, AI has seemed to fall into the trough of disillusionment, but it may be entering the slope of enlightenment on the way to the plateau of productivity.
His research highlighted several examples of productivity in radiomics in cancer patients and those who are at high risk of developing cancer.
In Dr. Aerts’s opinion, a second concern is replication of AI research results. He noted that, among 400 published studies, only 6% of authors shared the codes that would enable their findings to be corroborated. About 30% shared test data, and 54% shared “pseudocodes,” but transparency and reproducibility are problems for the acceptance and broad implementation of AI.
Dr. Aerts endorsed the Modelhub initiative (www.modelhub.ai), a multi-institutional initiative to advance reproducibility in the AI field and advance its full potential.
However, there are additional concerns about the implementation of radiomics and, more generally, data mining from clinicians’ EHRs to personalize care.
Firstly, it may be laborious and difficult to explain complex, computer-based risk stratification models to patients. Hereditary cancer testing is an example of a risk assessment test that requires complicated explanations that many clinicians relegate to genetics counselors – when patients elect to see them. When a model is not explainable, it undermines the confidence of patients and their care providers, according to an editorial related to the CXR-LC study.
Another issue is that uptake of lung cancer screening, in practice, has been underutilized by individuals who meet current, relatively straightforward Medicare criteria. Despite the apparently better accuracy of the CXR-LC deep-learning model, its complexity and limited access could constitute an additional barrier for the at-risk individuals who should avail themselves of screening.
Furthermore, although age and gender are accurate in most circumstances, there is legitimate concern about the accuracy of, for example, smoking history data and comorbid conditions in current EHRs. Who performs the laborious curation of the input in an AI model to assure its accuracy for individual patients?
Finally, it is unclear how scalable and applicable AI will be to medically underserved populations (e.g., smaller, community-based, free-standing, socioeconomically disadvantaged or rural health care institutions). There are substantial initial and maintenance costs that may limit AI’s availability to some academic institutions and large health maintenance organizations.
As the concerns and challenges are addressed, it will be interesting to see where and when the plateau of productivity for AI in cancer care occurs. When it does, many cancer patients will benefit from enhanced care along the continuum of the complex disease they and their caregivers seek to master.
Dr. Aerts disclosed relationships with Onc.AI outside the presented work.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers, as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
FROM AACR: AI, DIAGNOSIS, AND IMAGING 2021