Breast Cancer

At the 2016 San Antonio Breast Cancer Symposium, investigators presented findings on PFS benefits of adding an mTOR inhibitor to anti-hormonal therapy in advanced disease in postmenopausal women; the lack of clarity on the optimal duration of extended AI therapy, also in postmenopausal women; and using CTCs in helping predict breast cancer outcomes in the neoadjuvant setting.

Fulvestrant plus everolimus improves PFS in HR+, HER2- advanced breast cancer

Key clinical point This study provides further evidence of the benefits of adding an mTOR inhibitor to anti-hormonal therapy in postmenopausal women with advanced breast cancer resistant to aromatase inhibitors. Major finding Fulvestrant plus everolimus was associated with a PFS of 10.4 months, vs 5.1 months for fulvestrant alone. Data source Randomized phase 2 trial of 131 women with HR-positive, HER2-negative locally advanced of metastatic breast cancer resistant to aromatase inhibitors. Disclosures Sponsored by PrECOG with financial support from Novartis. Dr Kornblum reported having no conflicts of interest.

Adding everolimus to fulvestrant doubled median progression-free survival (PFS) among postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) metastatic breast cancer resistant to therapy with an aromatase inhibitor (AI), Noah S Kornblum, MD, reported during his presentation at the meeting of the PrECOG 0102 trial findings.

In the randomized phase 2 trial, the combination of the mammalian target of rapamycin (mTOR) inhibitor everolimus with the selective estrogen receptor down-regulator (SERD) fulvestrant was associated with a median PFS of 10.4 months, compared with 5.1 months for fulvestrant plus placebo, said Dr Kornblum, of Montefiore Einstein Center for Cancer Care, New York. The findings provide additional evidence that adding everolimus to anti-estrogen therapy in AI-resistant disease improves clinical outcomes,” he added.

Most women with HR-positive breast cancer who are treated with an AI will eventually develop resistance to those agents. Strategies for overcoming resistance include the addition of everolimus to a steroid AI, exemestane, as has been demonstrated in the BOLERO-2 trial. “Another strategy for overcoming AI resistance is by more completely blocking estrogen-receptor signaling through the use of a selective estrogen receptor down-regulator, which may result in more complete blockade of the ER signaling pathway than a steroidal AI such as exemestane,” Dr Kornblum said.

To test this hypothesis, the investigators enrolled 131 postmenopausal women with inoperable locally advanced or metastatic HR-positive, HER2-negative breast cancer resistant to AIs. AI resistance was defined as relapse while receiving adjuvant AI therapy, and/or progression after one or more AIs for metastatic disease. The patients could have had no more than one prior chemotherapy regimen for metastatic disease.

The patients were stratified by Eastern Cooperative Oncology Group performance status, presence of measurable disease, and previous chemotherapy status, and were then randomized to receive either high-dose fulvestrant (500 mg on day 1 and 15 of cycle 1, and then on day 1 of cycles 2-12) plus oral everolimus 10 mg/day, or fulvestrant and placebo. The trial had an induction phase, in which patients were treated until evidence of progressive disease or unacceptable toxicity for a maximum of 12 28-day cycles, and a continuation phase in which patients who had neither disease progression nor experienced unacceptable toxicities could have their data unblinded and could continue on the fulvestrant-everolimus combination.

The trial did not include the use of corticosteroid-containing mouthwash for prevention of treatment-associated stomatitis, because it was designed before the evidence of the benefit of such prophylaxis became public, Dr Kornblum said.

The primary endpoint of PFS by investigator assessment was significantly better with the fulvestrant-everolimus group at 10.4 months, compared with 5.1 months for the fulvestrant-placebo group. The hazard ratio was 0.60 (P = .02). There was no difference in overall survival (OS), however. Median OS in the combination group was 24.8 months, compared with not yet reached in the placebo arm (not statistically significant).

The combination therapy was associated with more grade 3 adverse events than the fulvestrant-placebo combination (48% vs 14%, respectively). The most common grade 3 adverse events occurring in more than 5% of patients were stomatitis, pneumonitis, fatigue, and hyperglycemia. Overall, the safety profile of the combination was consistent with that seen in BOLERO-2, Dr Kornblum said. In all, 10% of patients assigned to the combination and 12% assigned to placebo withdrew from the study because of adverse events; these patients were included in the analysis, which was by intention to treat.

— Neil Osterweil

Still no clarity on duration of extended AI therapy

Key clinical point The optimal duration of aromatase inhibitor (AI) therapy following 5 years of endocrine therapy in postmenopausal women is still unknown. Major finding There were no significant differences in disease-free or overall survival in three studies investigating extended AI therapy. Data source Randomized phase 2 NSABP B-42 with 3996 patients; randomized phase 3 DATA study with 1912 patients; randomized phase 3 IDEAL trial with 1824 patients. Disclosures NSABP B-42 was sponsored by PrECOG with financial support from Novartis. Dr Mamounas reported having no conflicts of interest. The DATA trial was sponsored by the Dutch Breast Cancer Research Group and Novartis. Dr Tjan-Heijnen reported nothing to disclose. IDEAL was supported by the Dutch Breast Cancer Research Group and Novartis. Dr Blok reported nothing to disclose.

When does adjuvant therapy with an aromatase inhibitor become too much of a good thing? Or, put another way, what’s the optimal duration of extended aromatase inhibitor therapy? That’s the question that three clinical trials have tried – but largely failed – to answer.

For example, the randomized, double-blinded NSABP B-42 trial, comparing extended therapy with letrozole in postmenopausal women with hormone receptor-positive (HR-positive) breast cancer who have completed previous adjuvant therapy with an aromatase inhibitor (AI) showed no difference in disease-free survival (DFS) after 7 years of follow-up between women treated with extended letrozole or placebo.

“Our findings suggest that careful assessment of potential risks and benefits is required before recommending extended letrozole therapy to patients with early-stage breast cancer, including patient and tumor characteristics such as age and nodal status, existing comorbidities, information on bone mineral density, and tolerance of the aromatase inhibitor in the initial years,” Eleftherios P Mamounas, MD, of NRG Oncology/NSABP, said at the symposium.

DATA data
In the DATA study, also presented at the meeting, investigators from the Netherlands compared 6 years of anastrozole with 3 years of anastrozole after 2 or 3 years of adjuvant tamoxifen for postmenopausal women with estrogen receptor-positive (ER-positive), and/or progesterone receptor-positive (PR-positive) breast cancer. They found that “adapted” DFS (DFS starting 3 years after randomization) and adapted overall survival (OS) were similar between the two groups. “The findings of the DATA study do not support extended adjuvant AI use after 5 years of sequential endocrine therapy for all postmenopausal hormone receptor-positive breast cancer patients,” said Vivianne Tjan-Heijnen, MD, of Maastricht University Medical Center in the Netherlands.

Less than IDEAL
In the optimistically named IDEAL trial, a separate team of investigators, also from the Netherlands, looked at the relative merits of continuing adjuvant therapy with letrozole for 2.5 or 5 years after 5 years of adjuvant therapy with tamoxifen, an AI, or a combination in postmenopausal women with HR-positive breast cancer. They found no differences in either DFS or OS between patients treated for 5 years or those treated for only half that long. “We conclude that there is no benefit of extending AI-based therapy longer than two-and-a-half years,” said Erik Blok, MD, of Leiden University Medical Center in the Netherlands.

Give what, to whom, for how long?
Results of the trials raise more questions than they answer, said Michael Gnant, MD, of the Medical University of Vienna, the invited discussant. “Essentially, these three trials did not reach the necessary statistical significance levels to demonstrate a clear benefit for the respective AI extension,” he said, adding that he does not think that other agents used in luminal breast cancer would help. “Based on their tolerability profile, and in part also on financial toxicity, I don’t think that the promising agents we explore in many situations for the treatment of hormone receptor-positive breast cancer will realistically be used in the extended adjuvant setting,” he said.

New strategies are needed for targeting the chronic part of luminal breast cancer recurrence risk, Dr Gnant noted. Using endocrine therapies in that setting would likely be ineffective. Instead, agents that could directly target dormant cancer stem cells would “eliminate the source of late metastases for good.”

The best evidence to date clearly points to individualized treatment plans for patients, Dr Gnant said. For example, for a patient who has had 2-5 years of tamoxifen, an AI for 2.5-5 additional years can help prevent recurrences, provided the patient has risk factors for recurrence and excellent bone health. “Based on the trial results, it is more complex for a patient who comes off initial or sequential AI. There are factors favoring the extension of AI treatment, and other factors to speak against such extension. I suggest to start with patient features at this time,” he said.

Currently, the main factor driving the choice of extended AI therapy will be how well the patient has tolerated AIs in the first years of therapy and whether she is at increased risk for fractures, suggesting younger age as a factor favoring extended AI use. Patients with higher clinicopathologic risk factors such as node positivity or more luminal type tumors, as well as higher risk according to genomic studies, might also benefit from extended AI therapy, he said.

Biomarkers needed
“What the data from these and other trials tell us is that endocrine therapy is not for everyone. We need biomarkers that can tell us who should be getting extended endocrine therapy, be it 10 years or even a longer duration of time, versus a subgroup that might do very well with 5 five years of AI,” Aditya Bardia, MBBS, MPH, of the breast cancer division at Massachusetts General Hospital Cancer Center in Boston, said in an interview.

There are several such biomarkers under investigation, but they need validation and testing in large scale clinical trials before they find their way into day-to-practice, said Dr Bardia, who was not involved in the studies.

— Neil Osterweil

CTCs help predict breast cancer outcomes in neoadjuvant setting

Key clinical point CTCs are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy. Major finding OS was associated with the presence of at least 2 CTCs at baseline (HR, 2.6 for 2 CTCs; 3.84 for 3-4 CTCs; and 6.25 for 5 or more CTCs). Data source Meta-analysis of data for 2156 patients. Disclosures Supported by a research grant from Janssen Diagnostics. Dr Bidard reported having no disclosures.

Circulating tumor cells (CTCs) are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy, according to findings from an international meta-analysis of individual patient data. The cells, which can be measured using a Food and Drug Administration-approved assay, are known to seed distant metastases and to be prognostic before and during therapy for patients with metastatic breast cancer, and prognostic before adjuvant therapy for patients with nonmetastatic breast cancer.

However, findings in the neoadjuvant setting have been varied, Francois-Clement Bidard, MD, of Institut Curie, Paris, reported at the symposium.

In the IMENEO study, CTCs were useful, independent of pathologic complete response, for predicting overall survival and distant disease-free survival in the neoadjuvant setting. In addition, the findings showed for the first time that CTCs also predict locoregional relapse-free survival. Based on an analysis of data from 2156 patients from 21 studies at 16 centers in 10 countries, the CTC positivity rates using thresholds of one or more, two or more, and five or more, respectively, were 25%, 13%, and 6% in 1574 patients tested at baseline; 17%, 6%, and 3% in 290 tested after neoadjuvant chemotherapy; 15%, 5%, and 1% in 1,200 tested before surgery; and 11%, 4%, and 1% in 285 tested after surgery, Dr Bidard said.

Prior to neoadjuvant chemotherapy, at least one CTC was found in 19%, 22%, 24%, 29%, and 41% of cT1, T2, T3, T4a-c, and T4d breast cancers, respectively, and this was marginally associated with hormone receptor negativity, he said, noting that later CTC detection rates were not associated with any patient baseline characteristics. Nearly one in four patients (24%) achieved pathologic complete response, but this was not associated at any time point with CTC count.

For the primary study endpoint of overall survival, a significant association was found with the presence of at least two CTCs at baseline (hazard ratio, 2.6 for two CTCs; 3.84 for three to four CTCs; and 6.25 for five or more CTCs). Similar associations were found for distant disease-free survival (hazard ratios, 2.4, 3.4, and 5.0, respectively) and for locoregional relapse-free interval with two CTCs and five or more CTCs (hazard ratios, 2.4 and 4.2, respectively).

Similar results were found using later time points, such as after the start of neoadjuvant chemotherapy or before surgery, he said.

On multivariate analysis, baseline CTC detection using any of the thresholds remained an independent predictor of overall and distant disease-free survival and locoregional relapse-free interval when considered together with pathologic complete response, cT, cN, and tumor subtype, suggesting that CTC measurement adds value to comprehensive prognostic models. That is, they complement rather than duplicate usual prognostic factors and pathologic complete response rates to better predict outcomes in patients with early breast cancer in the neoadjuvant setting, Dr Bidard said.

— Sharon Worcester

PIK3 inhibitor gives slight PFS edge at high cost for advanced breast cancer

Key clinical point The PI3K inhibitor buparlisib plus fulvestrant slightly prolonged progression-free survival of HR+/HER2- breast cancer pretreated with an aromatase inhibitor and mTOR inhibitor. Major finding The combination met its primary endpoint of better PFS than fulvestrant/placebo, but with high liver toxicity and mood disorders. Data source Randomized phase 3 trial of 432 women with HR-positive, HER2-negative, AI-pretreated breast cancer that progressed on or after mTOR inhibitor therapy. Disclosures Novartis sponsored the study. Dr Di Leo disclosed consulting and lecture fees from the company, and Dr O’Regan disclosed contracted research support. Dr Arteaga reported no disclosures relevant to the study.

A combination of a PI3K inhibitor and selective estrogen receptor down-regulator (SERD) met its primary endpoint of 2.1 months better progression-free survival (PFS) in postmenopausal women with locally advanced or metastatic breast cancer who were quickly running out of other treatment options, but the small gain in PFS came at a very high price in terms of toxicities, including mood disorders that may have led to patient suicide attempts, according to investigators.

The BELLE-3 trial looked at the combination of the SERD fulvestrant and an experimental inhibitor of the PI3 kinase, buparlisib, in postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) breast cancer treated with an aromatase inhibitor (AI) who experienced disease progression either on or after receiving therapy with an inhibitor of the mammalian target of rapamycin complex 1 (mTORC1).

The combination of fulvestrant and buparlisib was associated with a median PFS of 3.9 months, compared with 1.8 months for fulvestrant and placebo (P less than .001), Angelo Di Leo, MD, of Ospedale Misericordia e Dolce in Prato, Italy, reported at the symposium. Objective response rates (ORR) were low, at 7.8% in the combination arm, and 2.1% in the fulvestrant-plus-placebo arm. Although the PFS difference was statistically significant, “the higher rate of toxicity in patients receiving buparlisib and fulvestrant, including transaminase elevations and mood disorders, may represent a clinically relevant challenge for future development of this compound in this particular group of patients,” Dr Di Leo said.

Blocks AKT pathway
The preclinical rationale for the use of a P13K inhibitor after disease progression on mTORC1 inhibitor is that current mTOR inhibitors such as everolimus have a feedback mechanism that activates the AKT pathway, and that the use of P13K inhibitors can “abrogate or attenuate this activation, potentially blocking that pathway,” explained coinvestigator Ruth O’Regan, MD, head of the division of hematology and oncology at the University of Wisconsin-Madison School of Medicine and Public Health. Dr O’Regan discussed the BELLE-3 findings in a briefing before Dr Di Leo’s presentation of the data in general session.

In BELLE-3, 432 postmenopausal women with HR-positive, HER2-negative, AI-pretreated, locally advanced or metastatic breast cancer that had progressed on or after treatment with an mTOR inhibitor as the last line of therapy were enrolled. The patients were stratified by the presence or absence of visceral disease and then randomized on a 2:1 basis to fulvestrant 500 mg daily plus either buparlisib 100 mg/day (289 patients) or placebo (143). The primary endpoint of investigator-assessed PFS favored the addition of buparlisib, with a hazard ratio for progression of 0.67 (P less than .001). PFS results by independent central review were similar (HR 0.57, P less than .001).

The ORR for the buparlisib-fulvestrant combination, 7.6%, consisted of 0.3% complete responses and 7.3% partial responses. The ORR for the placebo-fulvestrant combination, 2.1%, was composed entirely of partial responses. The respective clinical benefit rates, defined as a combination of complete and partial responses and stable disease, were 24.6% and 15.4, respectively.

The benefit of buparlisib was evidently entirely among patients with visceral disease, with a PFS of 3.1 months, compared with 1.5 months. In contrast, PFS among patients with no visceral disease was 4.2 months, compared with 4.1 months, respectively, and was not significant. In addition, the P13K inhibitor seemed to benefit patients with PIK3CA mutations detected in either the primary tumor or in circulating DNA samples, but not patients with wild-type PIK3CA.

Depression, anxiety with combination

Patients assigned to buparlisib-fulvestrant group had substantially higher proportions of alanine aminotransferase and aspartate aminotransferase elevations compared with patients in the placebo-fulvestrant group, as well as more reported depression and anxiety. Three patients in the buparlisib arm attempted suicide. There were no reported suicide attempts in the placebo arm.

Dr O’Regan said at the briefing that mood disorders are known adverse events associated with buparlisib, and that patients with psychiatric disorders were excluded from the trial.

— Neil Osterweil

At the 2016 San Antonio Breast Cancer Symposium, investigators presented findings on PFS benefits of adding an mTOR inhibitor to anti-hormonal therapy in advanced disease in postmenopausal women; the lack of clarity on the optimal duration of extended AI therapy, also in postmenopausal women; and using CTCs in helping predict breast cancer outcomes in the neoadjuvant setting.

Fulvestrant plus everolimus improves PFS in HR+, HER2- advanced breast cancer

Key clinical point This study provides further evidence of the benefits of adding an mTOR inhibitor to anti-hormonal therapy in postmenopausal women with advanced breast cancer resistant to aromatase inhibitors. Major finding Fulvestrant plus everolimus was associated with a PFS of 10.4 months, vs 5.1 months for fulvestrant alone. Data source Randomized phase 2 trial of 131 women with HR-positive, HER2-negative locally advanced of metastatic breast cancer resistant to aromatase inhibitors. Disclosures Sponsored by PrECOG with financial support from Novartis. Dr Kornblum reported having no conflicts of interest.

Adding everolimus to fulvestrant doubled median progression-free survival (PFS) among postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) metastatic breast cancer resistant to therapy with an aromatase inhibitor (AI), Noah S Kornblum, MD, reported during his presentation at the meeting of the PrECOG 0102 trial findings.

In the randomized phase 2 trial, the combination of the mammalian target of rapamycin (mTOR) inhibitor everolimus with the selective estrogen receptor down-regulator (SERD) fulvestrant was associated with a median PFS of 10.4 months, compared with 5.1 months for fulvestrant plus placebo, said Dr Kornblum, of Montefiore Einstein Center for Cancer Care, New York. The findings provide additional evidence that adding everolimus to anti-estrogen therapy in AI-resistant disease improves clinical outcomes,” he added.

Most women with HR-positive breast cancer who are treated with an AI will eventually develop resistance to those agents. Strategies for overcoming resistance include the addition of everolimus to a steroid AI, exemestane, as has been demonstrated in the BOLERO-2 trial. “Another strategy for overcoming AI resistance is by more completely blocking estrogen-receptor signaling through the use of a selective estrogen receptor down-regulator, which may result in more complete blockade of the ER signaling pathway than a steroidal AI such as exemestane,” Dr Kornblum said.

To test this hypothesis, the investigators enrolled 131 postmenopausal women with inoperable locally advanced or metastatic HR-positive, HER2-negative breast cancer resistant to AIs. AI resistance was defined as relapse while receiving adjuvant AI therapy, and/or progression after one or more AIs for metastatic disease. The patients could have had no more than one prior chemotherapy regimen for metastatic disease.

The patients were stratified by Eastern Cooperative Oncology Group performance status, presence of measurable disease, and previous chemotherapy status, and were then randomized to receive either high-dose fulvestrant (500 mg on day 1 and 15 of cycle 1, and then on day 1 of cycles 2-12) plus oral everolimus 10 mg/day, or fulvestrant and placebo. The trial had an induction phase, in which patients were treated until evidence of progressive disease or unacceptable toxicity for a maximum of 12 28-day cycles, and a continuation phase in which patients who had neither disease progression nor experienced unacceptable toxicities could have their data unblinded and could continue on the fulvestrant-everolimus combination.

The trial did not include the use of corticosteroid-containing mouthwash for prevention of treatment-associated stomatitis, because it was designed before the evidence of the benefit of such prophylaxis became public, Dr Kornblum said.

The primary endpoint of PFS by investigator assessment was significantly better with the fulvestrant-everolimus group at 10.4 months, compared with 5.1 months for the fulvestrant-placebo group. The hazard ratio was 0.60 (P = .02). There was no difference in overall survival (OS), however. Median OS in the combination group was 24.8 months, compared with not yet reached in the placebo arm (not statistically significant).

The combination therapy was associated with more grade 3 adverse events than the fulvestrant-placebo combination (48% vs 14%, respectively). The most common grade 3 adverse events occurring in more than 5% of patients were stomatitis, pneumonitis, fatigue, and hyperglycemia. Overall, the safety profile of the combination was consistent with that seen in BOLERO-2, Dr Kornblum said. In all, 10% of patients assigned to the combination and 12% assigned to placebo withdrew from the study because of adverse events; these patients were included in the analysis, which was by intention to treat.

— Neil Osterweil

Still no clarity on duration of extended AI therapy

Key clinical point The optimal duration of aromatase inhibitor (AI) therapy following 5 years of endocrine therapy in postmenopausal women is still unknown. Major finding There were no significant differences in disease-free or overall survival in three studies investigating extended AI therapy. Data source Randomized phase 2 NSABP B-42 with 3996 patients; randomized phase 3 DATA study with 1912 patients; randomized phase 3 IDEAL trial with 1824 patients. Disclosures NSABP B-42 was sponsored by PrECOG with financial support from Novartis. Dr Mamounas reported having no conflicts of interest. The DATA trial was sponsored by the Dutch Breast Cancer Research Group and Novartis. Dr Tjan-Heijnen reported nothing to disclose. IDEAL was supported by the Dutch Breast Cancer Research Group and Novartis. Dr Blok reported nothing to disclose.

When does adjuvant therapy with an aromatase inhibitor become too much of a good thing? Or, put another way, what’s the optimal duration of extended aromatase inhibitor therapy? That’s the question that three clinical trials have tried – but largely failed – to answer.

For example, the randomized, double-blinded NSABP B-42 trial, comparing extended therapy with letrozole in postmenopausal women with hormone receptor-positive (HR-positive) breast cancer who have completed previous adjuvant therapy with an aromatase inhibitor (AI) showed no difference in disease-free survival (DFS) after 7 years of follow-up between women treated with extended letrozole or placebo.

“Our findings suggest that careful assessment of potential risks and benefits is required before recommending extended letrozole therapy to patients with early-stage breast cancer, including patient and tumor characteristics such as age and nodal status, existing comorbidities, information on bone mineral density, and tolerance of the aromatase inhibitor in the initial years,” Eleftherios P Mamounas, MD, of NRG Oncology/NSABP, said at the symposium.

DATA data
In the DATA study, also presented at the meeting, investigators from the Netherlands compared 6 years of anastrozole with 3 years of anastrozole after 2 or 3 years of adjuvant tamoxifen for postmenopausal women with estrogen receptor-positive (ER-positive), and/or progesterone receptor-positive (PR-positive) breast cancer. They found that “adapted” DFS (DFS starting 3 years after randomization) and adapted overall survival (OS) were similar between the two groups. “The findings of the DATA study do not support extended adjuvant AI use after 5 years of sequential endocrine therapy for all postmenopausal hormone receptor-positive breast cancer patients,” said Vivianne Tjan-Heijnen, MD, of Maastricht University Medical Center in the Netherlands.

Less than IDEAL
In the optimistically named IDEAL trial, a separate team of investigators, also from the Netherlands, looked at the relative merits of continuing adjuvant therapy with letrozole for 2.5 or 5 years after 5 years of adjuvant therapy with tamoxifen, an AI, or a combination in postmenopausal women with HR-positive breast cancer. They found no differences in either DFS or OS between patients treated for 5 years or those treated for only half that long. “We conclude that there is no benefit of extending AI-based therapy longer than two-and-a-half years,” said Erik Blok, MD, of Leiden University Medical Center in the Netherlands.

Give what, to whom, for how long?
Results of the trials raise more questions than they answer, said Michael Gnant, MD, of the Medical University of Vienna, the invited discussant. “Essentially, these three trials did not reach the necessary statistical significance levels to demonstrate a clear benefit for the respective AI extension,” he said, adding that he does not think that other agents used in luminal breast cancer would help. “Based on their tolerability profile, and in part also on financial toxicity, I don’t think that the promising agents we explore in many situations for the treatment of hormone receptor-positive breast cancer will realistically be used in the extended adjuvant setting,” he said.

New strategies are needed for targeting the chronic part of luminal breast cancer recurrence risk, Dr Gnant noted. Using endocrine therapies in that setting would likely be ineffective. Instead, agents that could directly target dormant cancer stem cells would “eliminate the source of late metastases for good.”

The best evidence to date clearly points to individualized treatment plans for patients, Dr Gnant said. For example, for a patient who has had 2-5 years of tamoxifen, an AI for 2.5-5 additional years can help prevent recurrences, provided the patient has risk factors for recurrence and excellent bone health. “Based on the trial results, it is more complex for a patient who comes off initial or sequential AI. There are factors favoring the extension of AI treatment, and other factors to speak against such extension. I suggest to start with patient features at this time,” he said.

Currently, the main factor driving the choice of extended AI therapy will be how well the patient has tolerated AIs in the first years of therapy and whether she is at increased risk for fractures, suggesting younger age as a factor favoring extended AI use. Patients with higher clinicopathologic risk factors such as node positivity or more luminal type tumors, as well as higher risk according to genomic studies, might also benefit from extended AI therapy, he said.

Biomarkers needed
“What the data from these and other trials tell us is that endocrine therapy is not for everyone. We need biomarkers that can tell us who should be getting extended endocrine therapy, be it 10 years or even a longer duration of time, versus a subgroup that might do very well with 5 five years of AI,” Aditya Bardia, MBBS, MPH, of the breast cancer division at Massachusetts General Hospital Cancer Center in Boston, said in an interview.

There are several such biomarkers under investigation, but they need validation and testing in large scale clinical trials before they find their way into day-to-practice, said Dr Bardia, who was not involved in the studies.

— Neil Osterweil

CTCs help predict breast cancer outcomes in neoadjuvant setting

Key clinical point CTCs are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy. Major finding OS was associated with the presence of at least 2 CTCs at baseline (HR, 2.6 for 2 CTCs; 3.84 for 3-4 CTCs; and 6.25 for 5 or more CTCs). Data source Meta-analysis of data for 2156 patients. Disclosures Supported by a research grant from Janssen Diagnostics. Dr Bidard reported having no disclosures.

Circulating tumor cells (CTCs) are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy, according to findings from an international meta-analysis of individual patient data. The cells, which can be measured using a Food and Drug Administration-approved assay, are known to seed distant metastases and to be prognostic before and during therapy for patients with metastatic breast cancer, and prognostic before adjuvant therapy for patients with nonmetastatic breast cancer.

However, findings in the neoadjuvant setting have been varied, Francois-Clement Bidard, MD, of Institut Curie, Paris, reported at the symposium.

In the IMENEO study, CTCs were useful, independent of pathologic complete response, for predicting overall survival and distant disease-free survival in the neoadjuvant setting. In addition, the findings showed for the first time that CTCs also predict locoregional relapse-free survival. Based on an analysis of data from 2156 patients from 21 studies at 16 centers in 10 countries, the CTC positivity rates using thresholds of one or more, two or more, and five or more, respectively, were 25%, 13%, and 6% in 1574 patients tested at baseline; 17%, 6%, and 3% in 290 tested after neoadjuvant chemotherapy; 15%, 5%, and 1% in 1,200 tested before surgery; and 11%, 4%, and 1% in 285 tested after surgery, Dr Bidard said.

Prior to neoadjuvant chemotherapy, at least one CTC was found in 19%, 22%, 24%, 29%, and 41% of cT1, T2, T3, T4a-c, and T4d breast cancers, respectively, and this was marginally associated with hormone receptor negativity, he said, noting that later CTC detection rates were not associated with any patient baseline characteristics. Nearly one in four patients (24%) achieved pathologic complete response, but this was not associated at any time point with CTC count.

For the primary study endpoint of overall survival, a significant association was found with the presence of at least two CTCs at baseline (hazard ratio, 2.6 for two CTCs; 3.84 for three to four CTCs; and 6.25 for five or more CTCs). Similar associations were found for distant disease-free survival (hazard ratios, 2.4, 3.4, and 5.0, respectively) and for locoregional relapse-free interval with two CTCs and five or more CTCs (hazard ratios, 2.4 and 4.2, respectively).

Similar results were found using later time points, such as after the start of neoadjuvant chemotherapy or before surgery, he said.

On multivariate analysis, baseline CTC detection using any of the thresholds remained an independent predictor of overall and distant disease-free survival and locoregional relapse-free interval when considered together with pathologic complete response, cT, cN, and tumor subtype, suggesting that CTC measurement adds value to comprehensive prognostic models. That is, they complement rather than duplicate usual prognostic factors and pathologic complete response rates to better predict outcomes in patients with early breast cancer in the neoadjuvant setting, Dr Bidard said.

— Sharon Worcester

PIK3 inhibitor gives slight PFS edge at high cost for advanced breast cancer

Key clinical point The PI3K inhibitor buparlisib plus fulvestrant slightly prolonged progression-free survival of HR+/HER2- breast cancer pretreated with an aromatase inhibitor and mTOR inhibitor. Major finding The combination met its primary endpoint of better PFS than fulvestrant/placebo, but with high liver toxicity and mood disorders. Data source Randomized phase 3 trial of 432 women with HR-positive, HER2-negative, AI-pretreated breast cancer that progressed on or after mTOR inhibitor therapy. Disclosures Novartis sponsored the study. Dr Di Leo disclosed consulting and lecture fees from the company, and Dr O’Regan disclosed contracted research support. Dr Arteaga reported no disclosures relevant to the study.

A combination of a PI3K inhibitor and selective estrogen receptor down-regulator (SERD) met its primary endpoint of 2.1 months better progression-free survival (PFS) in postmenopausal women with locally advanced or metastatic breast cancer who were quickly running out of other treatment options, but the small gain in PFS came at a very high price in terms of toxicities, including mood disorders that may have led to patient suicide attempts, according to investigators.

The BELLE-3 trial looked at the combination of the SERD fulvestrant and an experimental inhibitor of the PI3 kinase, buparlisib, in postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) breast cancer treated with an aromatase inhibitor (AI) who experienced disease progression either on or after receiving therapy with an inhibitor of the mammalian target of rapamycin complex 1 (mTORC1).

The combination of fulvestrant and buparlisib was associated with a median PFS of 3.9 months, compared with 1.8 months for fulvestrant and placebo (P less than .001), Angelo Di Leo, MD, of Ospedale Misericordia e Dolce in Prato, Italy, reported at the symposium. Objective response rates (ORR) were low, at 7.8% in the combination arm, and 2.1% in the fulvestrant-plus-placebo arm. Although the PFS difference was statistically significant, “the higher rate of toxicity in patients receiving buparlisib and fulvestrant, including transaminase elevations and mood disorders, may represent a clinically relevant challenge for future development of this compound in this particular group of patients,” Dr Di Leo said.

Blocks AKT pathway
The preclinical rationale for the use of a P13K inhibitor after disease progression on mTORC1 inhibitor is that current mTOR inhibitors such as everolimus have a feedback mechanism that activates the AKT pathway, and that the use of P13K inhibitors can “abrogate or attenuate this activation, potentially blocking that pathway,” explained coinvestigator Ruth O’Regan, MD, head of the division of hematology and oncology at the University of Wisconsin-Madison School of Medicine and Public Health. Dr O’Regan discussed the BELLE-3 findings in a briefing before Dr Di Leo’s presentation of the data in general session.

In BELLE-3, 432 postmenopausal women with HR-positive, HER2-negative, AI-pretreated, locally advanced or metastatic breast cancer that had progressed on or after treatment with an mTOR inhibitor as the last line of therapy were enrolled. The patients were stratified by the presence or absence of visceral disease and then randomized on a 2:1 basis to fulvestrant 500 mg daily plus either buparlisib 100 mg/day (289 patients) or placebo (143). The primary endpoint of investigator-assessed PFS favored the addition of buparlisib, with a hazard ratio for progression of 0.67 (P less than .001). PFS results by independent central review were similar (HR 0.57, P less than .001).

The ORR for the buparlisib-fulvestrant combination, 7.6%, consisted of 0.3% complete responses and 7.3% partial responses. The ORR for the placebo-fulvestrant combination, 2.1%, was composed entirely of partial responses. The respective clinical benefit rates, defined as a combination of complete and partial responses and stable disease, were 24.6% and 15.4, respectively.

The benefit of buparlisib was evidently entirely among patients with visceral disease, with a PFS of 3.1 months, compared with 1.5 months. In contrast, PFS among patients with no visceral disease was 4.2 months, compared with 4.1 months, respectively, and was not significant. In addition, the P13K inhibitor seemed to benefit patients with PIK3CA mutations detected in either the primary tumor or in circulating DNA samples, but not patients with wild-type PIK3CA.

Depression, anxiety with combination

Patients assigned to buparlisib-fulvestrant group had substantially higher proportions of alanine aminotransferase and aspartate aminotransferase elevations compared with patients in the placebo-fulvestrant group, as well as more reported depression and anxiety. Three patients in the buparlisib arm attempted suicide. There were no reported suicide attempts in the placebo arm.

Dr O’Regan said at the briefing that mood disorders are known adverse events associated with buparlisib, and that patients with psychiatric disorders were excluded from the trial.

— Neil Osterweil

At the 2016 San Antonio Breast Cancer Symposium, investigators presented findings on PFS benefits of adding an mTOR inhibitor to anti-hormonal therapy in advanced disease in postmenopausal women; the lack of clarity on the optimal duration of extended AI therapy, also in postmenopausal women; and using CTCs in helping predict breast cancer outcomes in the neoadjuvant setting.

Fulvestrant plus everolimus improves PFS in HR+, HER2- advanced breast cancer

Key clinical point This study provides further evidence of the benefits of adding an mTOR inhibitor to anti-hormonal therapy in postmenopausal women with advanced breast cancer resistant to aromatase inhibitors. Major finding Fulvestrant plus everolimus was associated with a PFS of 10.4 months, vs 5.1 months for fulvestrant alone. Data source Randomized phase 2 trial of 131 women with HR-positive, HER2-negative locally advanced of metastatic breast cancer resistant to aromatase inhibitors. Disclosures Sponsored by PrECOG with financial support from Novartis. Dr Kornblum reported having no conflicts of interest.

Adding everolimus to fulvestrant doubled median progression-free survival (PFS) among postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) metastatic breast cancer resistant to therapy with an aromatase inhibitor (AI), Noah S Kornblum, MD, reported during his presentation at the meeting of the PrECOG 0102 trial findings.

In the randomized phase 2 trial, the combination of the mammalian target of rapamycin (mTOR) inhibitor everolimus with the selective estrogen receptor down-regulator (SERD) fulvestrant was associated with a median PFS of 10.4 months, compared with 5.1 months for fulvestrant plus placebo, said Dr Kornblum, of Montefiore Einstein Center for Cancer Care, New York. The findings provide additional evidence that adding everolimus to anti-estrogen therapy in AI-resistant disease improves clinical outcomes,” he added.

Most women with HR-positive breast cancer who are treated with an AI will eventually develop resistance to those agents. Strategies for overcoming resistance include the addition of everolimus to a steroid AI, exemestane, as has been demonstrated in the BOLERO-2 trial. “Another strategy for overcoming AI resistance is by more completely blocking estrogen-receptor signaling through the use of a selective estrogen receptor down-regulator, which may result in more complete blockade of the ER signaling pathway than a steroidal AI such as exemestane,” Dr Kornblum said.

To test this hypothesis, the investigators enrolled 131 postmenopausal women with inoperable locally advanced or metastatic HR-positive, HER2-negative breast cancer resistant to AIs. AI resistance was defined as relapse while receiving adjuvant AI therapy, and/or progression after one or more AIs for metastatic disease. The patients could have had no more than one prior chemotherapy regimen for metastatic disease.

The patients were stratified by Eastern Cooperative Oncology Group performance status, presence of measurable disease, and previous chemotherapy status, and were then randomized to receive either high-dose fulvestrant (500 mg on day 1 and 15 of cycle 1, and then on day 1 of cycles 2-12) plus oral everolimus 10 mg/day, or fulvestrant and placebo. The trial had an induction phase, in which patients were treated until evidence of progressive disease or unacceptable toxicity for a maximum of 12 28-day cycles, and a continuation phase in which patients who had neither disease progression nor experienced unacceptable toxicities could have their data unblinded and could continue on the fulvestrant-everolimus combination.

The trial did not include the use of corticosteroid-containing mouthwash for prevention of treatment-associated stomatitis, because it was designed before the evidence of the benefit of such prophylaxis became public, Dr Kornblum said.

The primary endpoint of PFS by investigator assessment was significantly better with the fulvestrant-everolimus group at 10.4 months, compared with 5.1 months for the fulvestrant-placebo group. The hazard ratio was 0.60 (P = .02). There was no difference in overall survival (OS), however. Median OS in the combination group was 24.8 months, compared with not yet reached in the placebo arm (not statistically significant).

The combination therapy was associated with more grade 3 adverse events than the fulvestrant-placebo combination (48% vs 14%, respectively). The most common grade 3 adverse events occurring in more than 5% of patients were stomatitis, pneumonitis, fatigue, and hyperglycemia. Overall, the safety profile of the combination was consistent with that seen in BOLERO-2, Dr Kornblum said. In all, 10% of patients assigned to the combination and 12% assigned to placebo withdrew from the study because of adverse events; these patients were included in the analysis, which was by intention to treat.

— Neil Osterweil

Still no clarity on duration of extended AI therapy

Key clinical point The optimal duration of aromatase inhibitor (AI) therapy following 5 years of endocrine therapy in postmenopausal women is still unknown. Major finding There were no significant differences in disease-free or overall survival in three studies investigating extended AI therapy. Data source Randomized phase 2 NSABP B-42 with 3996 patients; randomized phase 3 DATA study with 1912 patients; randomized phase 3 IDEAL trial with 1824 patients. Disclosures NSABP B-42 was sponsored by PrECOG with financial support from Novartis. Dr Mamounas reported having no conflicts of interest. The DATA trial was sponsored by the Dutch Breast Cancer Research Group and Novartis. Dr Tjan-Heijnen reported nothing to disclose. IDEAL was supported by the Dutch Breast Cancer Research Group and Novartis. Dr Blok reported nothing to disclose.

When does adjuvant therapy with an aromatase inhibitor become too much of a good thing? Or, put another way, what’s the optimal duration of extended aromatase inhibitor therapy? That’s the question that three clinical trials have tried – but largely failed – to answer.

For example, the randomized, double-blinded NSABP B-42 trial, comparing extended therapy with letrozole in postmenopausal women with hormone receptor-positive (HR-positive) breast cancer who have completed previous adjuvant therapy with an aromatase inhibitor (AI) showed no difference in disease-free survival (DFS) after 7 years of follow-up between women treated with extended letrozole or placebo.

“Our findings suggest that careful assessment of potential risks and benefits is required before recommending extended letrozole therapy to patients with early-stage breast cancer, including patient and tumor characteristics such as age and nodal status, existing comorbidities, information on bone mineral density, and tolerance of the aromatase inhibitor in the initial years,” Eleftherios P Mamounas, MD, of NRG Oncology/NSABP, said at the symposium.

DATA data
In the DATA study, also presented at the meeting, investigators from the Netherlands compared 6 years of anastrozole with 3 years of anastrozole after 2 or 3 years of adjuvant tamoxifen for postmenopausal women with estrogen receptor-positive (ER-positive), and/or progesterone receptor-positive (PR-positive) breast cancer. They found that “adapted” DFS (DFS starting 3 years after randomization) and adapted overall survival (OS) were similar between the two groups. “The findings of the DATA study do not support extended adjuvant AI use after 5 years of sequential endocrine therapy for all postmenopausal hormone receptor-positive breast cancer patients,” said Vivianne Tjan-Heijnen, MD, of Maastricht University Medical Center in the Netherlands.

Less than IDEAL
In the optimistically named IDEAL trial, a separate team of investigators, also from the Netherlands, looked at the relative merits of continuing adjuvant therapy with letrozole for 2.5 or 5 years after 5 years of adjuvant therapy with tamoxifen, an AI, or a combination in postmenopausal women with HR-positive breast cancer. They found no differences in either DFS or OS between patients treated for 5 years or those treated for only half that long. “We conclude that there is no benefit of extending AI-based therapy longer than two-and-a-half years,” said Erik Blok, MD, of Leiden University Medical Center in the Netherlands.

Give what, to whom, for how long?
Results of the trials raise more questions than they answer, said Michael Gnant, MD, of the Medical University of Vienna, the invited discussant. “Essentially, these three trials did not reach the necessary statistical significance levels to demonstrate a clear benefit for the respective AI extension,” he said, adding that he does not think that other agents used in luminal breast cancer would help. “Based on their tolerability profile, and in part also on financial toxicity, I don’t think that the promising agents we explore in many situations for the treatment of hormone receptor-positive breast cancer will realistically be used in the extended adjuvant setting,” he said.

New strategies are needed for targeting the chronic part of luminal breast cancer recurrence risk, Dr Gnant noted. Using endocrine therapies in that setting would likely be ineffective. Instead, agents that could directly target dormant cancer stem cells would “eliminate the source of late metastases for good.”

The best evidence to date clearly points to individualized treatment plans for patients, Dr Gnant said. For example, for a patient who has had 2-5 years of tamoxifen, an AI for 2.5-5 additional years can help prevent recurrences, provided the patient has risk factors for recurrence and excellent bone health. “Based on the trial results, it is more complex for a patient who comes off initial or sequential AI. There are factors favoring the extension of AI treatment, and other factors to speak against such extension. I suggest to start with patient features at this time,” he said.

Currently, the main factor driving the choice of extended AI therapy will be how well the patient has tolerated AIs in the first years of therapy and whether she is at increased risk for fractures, suggesting younger age as a factor favoring extended AI use. Patients with higher clinicopathologic risk factors such as node positivity or more luminal type tumors, as well as higher risk according to genomic studies, might also benefit from extended AI therapy, he said.

Biomarkers needed
“What the data from these and other trials tell us is that endocrine therapy is not for everyone. We need biomarkers that can tell us who should be getting extended endocrine therapy, be it 10 years or even a longer duration of time, versus a subgroup that might do very well with 5 five years of AI,” Aditya Bardia, MBBS, MPH, of the breast cancer division at Massachusetts General Hospital Cancer Center in Boston, said in an interview.

There are several such biomarkers under investigation, but they need validation and testing in large scale clinical trials before they find their way into day-to-practice, said Dr Bardia, who was not involved in the studies.

— Neil Osterweil

CTCs help predict breast cancer outcomes in neoadjuvant setting

Key clinical point CTCs are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy. Major finding OS was associated with the presence of at least 2 CTCs at baseline (HR, 2.6 for 2 CTCs; 3.84 for 3-4 CTCs; and 6.25 for 5 or more CTCs). Data source Meta-analysis of data for 2156 patients. Disclosures Supported by a research grant from Janssen Diagnostics. Dr Bidard reported having no disclosures.

Circulating tumor cells (CTCs) are a useful prognostic biomarker in early breast cancer patients treated with neoadjuvant chemotherapy, according to findings from an international meta-analysis of individual patient data. The cells, which can be measured using a Food and Drug Administration-approved assay, are known to seed distant metastases and to be prognostic before and during therapy for patients with metastatic breast cancer, and prognostic before adjuvant therapy for patients with nonmetastatic breast cancer.

However, findings in the neoadjuvant setting have been varied, Francois-Clement Bidard, MD, of Institut Curie, Paris, reported at the symposium.

In the IMENEO study, CTCs were useful, independent of pathologic complete response, for predicting overall survival and distant disease-free survival in the neoadjuvant setting. In addition, the findings showed for the first time that CTCs also predict locoregional relapse-free survival. Based on an analysis of data from 2156 patients from 21 studies at 16 centers in 10 countries, the CTC positivity rates using thresholds of one or more, two or more, and five or more, respectively, were 25%, 13%, and 6% in 1574 patients tested at baseline; 17%, 6%, and 3% in 290 tested after neoadjuvant chemotherapy; 15%, 5%, and 1% in 1,200 tested before surgery; and 11%, 4%, and 1% in 285 tested after surgery, Dr Bidard said.

Prior to neoadjuvant chemotherapy, at least one CTC was found in 19%, 22%, 24%, 29%, and 41% of cT1, T2, T3, T4a-c, and T4d breast cancers, respectively, and this was marginally associated with hormone receptor negativity, he said, noting that later CTC detection rates were not associated with any patient baseline characteristics. Nearly one in four patients (24%) achieved pathologic complete response, but this was not associated at any time point with CTC count.

For the primary study endpoint of overall survival, a significant association was found with the presence of at least two CTCs at baseline (hazard ratio, 2.6 for two CTCs; 3.84 for three to four CTCs; and 6.25 for five or more CTCs). Similar associations were found for distant disease-free survival (hazard ratios, 2.4, 3.4, and 5.0, respectively) and for locoregional relapse-free interval with two CTCs and five or more CTCs (hazard ratios, 2.4 and 4.2, respectively).

Similar results were found using later time points, such as after the start of neoadjuvant chemotherapy or before surgery, he said.

On multivariate analysis, baseline CTC detection using any of the thresholds remained an independent predictor of overall and distant disease-free survival and locoregional relapse-free interval when considered together with pathologic complete response, cT, cN, and tumor subtype, suggesting that CTC measurement adds value to comprehensive prognostic models. That is, they complement rather than duplicate usual prognostic factors and pathologic complete response rates to better predict outcomes in patients with early breast cancer in the neoadjuvant setting, Dr Bidard said.

— Sharon Worcester

PIK3 inhibitor gives slight PFS edge at high cost for advanced breast cancer

Key clinical point The PI3K inhibitor buparlisib plus fulvestrant slightly prolonged progression-free survival of HR+/HER2- breast cancer pretreated with an aromatase inhibitor and mTOR inhibitor. Major finding The combination met its primary endpoint of better PFS than fulvestrant/placebo, but with high liver toxicity and mood disorders. Data source Randomized phase 3 trial of 432 women with HR-positive, HER2-negative, AI-pretreated breast cancer that progressed on or after mTOR inhibitor therapy. Disclosures Novartis sponsored the study. Dr Di Leo disclosed consulting and lecture fees from the company, and Dr O’Regan disclosed contracted research support. Dr Arteaga reported no disclosures relevant to the study.

A combination of a PI3K inhibitor and selective estrogen receptor down-regulator (SERD) met its primary endpoint of 2.1 months better progression-free survival (PFS) in postmenopausal women with locally advanced or metastatic breast cancer who were quickly running out of other treatment options, but the small gain in PFS came at a very high price in terms of toxicities, including mood disorders that may have led to patient suicide attempts, according to investigators.

The BELLE-3 trial looked at the combination of the SERD fulvestrant and an experimental inhibitor of the PI3 kinase, buparlisib, in postmenopausal women with hormone receptor-positive (HR-positive), human epidermal growth factor receptor 2-negative (HER2-negative) breast cancer treated with an aromatase inhibitor (AI) who experienced disease progression either on or after receiving therapy with an inhibitor of the mammalian target of rapamycin complex 1 (mTORC1).

The combination of fulvestrant and buparlisib was associated with a median PFS of 3.9 months, compared with 1.8 months for fulvestrant and placebo (P less than .001), Angelo Di Leo, MD, of Ospedale Misericordia e Dolce in Prato, Italy, reported at the symposium. Objective response rates (ORR) were low, at 7.8% in the combination arm, and 2.1% in the fulvestrant-plus-placebo arm. Although the PFS difference was statistically significant, “the higher rate of toxicity in patients receiving buparlisib and fulvestrant, including transaminase elevations and mood disorders, may represent a clinically relevant challenge for future development of this compound in this particular group of patients,” Dr Di Leo said.

Blocks AKT pathway
The preclinical rationale for the use of a P13K inhibitor after disease progression on mTORC1 inhibitor is that current mTOR inhibitors such as everolimus have a feedback mechanism that activates the AKT pathway, and that the use of P13K inhibitors can “abrogate or attenuate this activation, potentially blocking that pathway,” explained coinvestigator Ruth O’Regan, MD, head of the division of hematology and oncology at the University of Wisconsin-Madison School of Medicine and Public Health. Dr O’Regan discussed the BELLE-3 findings in a briefing before Dr Di Leo’s presentation of the data in general session.

In BELLE-3, 432 postmenopausal women with HR-positive, HER2-negative, AI-pretreated, locally advanced or metastatic breast cancer that had progressed on or after treatment with an mTOR inhibitor as the last line of therapy were enrolled. The patients were stratified by the presence or absence of visceral disease and then randomized on a 2:1 basis to fulvestrant 500 mg daily plus either buparlisib 100 mg/day (289 patients) or placebo (143). The primary endpoint of investigator-assessed PFS favored the addition of buparlisib, with a hazard ratio for progression of 0.67 (P less than .001). PFS results by independent central review were similar (HR 0.57, P less than .001).

The ORR for the buparlisib-fulvestrant combination, 7.6%, consisted of 0.3% complete responses and 7.3% partial responses. The ORR for the placebo-fulvestrant combination, 2.1%, was composed entirely of partial responses. The respective clinical benefit rates, defined as a combination of complete and partial responses and stable disease, were 24.6% and 15.4, respectively.

The benefit of buparlisib was evidently entirely among patients with visceral disease, with a PFS of 3.1 months, compared with 1.5 months. In contrast, PFS among patients with no visceral disease was 4.2 months, compared with 4.1 months, respectively, and was not significant. In addition, the P13K inhibitor seemed to benefit patients with PIK3CA mutations detected in either the primary tumor or in circulating DNA samples, but not patients with wild-type PIK3CA.

Depression, anxiety with combination

Patients assigned to buparlisib-fulvestrant group had substantially higher proportions of alanine aminotransferase and aspartate aminotransferase elevations compared with patients in the placebo-fulvestrant group, as well as more reported depression and anxiety. Three patients in the buparlisib arm attempted suicide. There were no reported suicide attempts in the placebo arm.

Dr O’Regan said at the briefing that mood disorders are known adverse events associated with buparlisib, and that patients with psychiatric disorders were excluded from the trial.

— Neil Osterweil

Cyclophosphamide, an agent used to treat various malignant and autoimmune disorders, can cause severe hyponatremia with seizures in rare cases. The exact mechanism of cyclophosphamide-induced hyponatremia is poorly understood, but is thought to occur from a drug- associated antidiuretic hormone (ADH) release leading to free water retention.¹ This unusual phenomenon of cyclophosphamide-associated syndrome of inappropriate antidiuretic hormone secretion (SIADH) has been described only in case reports, most of which reported the development of severe hyponatremia within a week after administration of cyclophosphamide.^2-5 We report a unique case of a patient who developed severe, symptomatic hyponatremia with seizures, with her serum sodium decreasing from 137 mEq to 112 mEq within 30 hours after her first dose of low-dose cyclophosphamide (600 mg/m²).

Case presentation and summary

A 68-year-old white woman with a history of bilateral invasive ductal carcinoma of the breast (status-post bilateral mastectomy) presented to the emergency department (ED) at our facility with new onset seizure. The patient had been diagnosed 8 months earlier with stage I (T1c, N0, M0) poorly differentiated infiltrating ductal carcinoma (triple negative) of the left breast for which she underwent left segmental mastectomy about 1 month after diagnosis. She was subsequently found to have progressive disease with stage IIIC (T2, N3, and M0) infiltrating ductal carcinoma with lobular features (ER/PR+, Her2) of the right breast. She underwent a right modified radical mastectomy 5 months after her stage IIIC breast cancer diagnosis. She received her first cycle of adjuvant chemotherapy with intravenous doxorubicin (60 mg/m²) and cyclophosphamide (600 mg/m²), which included pre-hydration, a day before presenting to our facility.

According to the patient’s family who provided the initial history, the patient reported tightness in her left arm while sitting at the dinner table. She was confused and subsequently had jerking movement of her right upper extremity with left facial twitching which lasted about 40 seconds. There was no loss of consciousness, or bowel or bladder control. She became unresponsive after the episode. Review of systems was negative except for a report of nausea a few hours before the onset of seizures, which resolved with ondansetron. Her past medical history was significant for breast cancer as already mentioned, seasonal allergic rhinitis, and hypertension. Home medications included hydrochlorothiazide 12.5 mg oral daily, aspirin 81 mg oral daily, and fexofenadine and loratadine oral daily as needed for allergies. There were no other significant surgical history other than already stated. The patient lived at home with her family and was independent with her instrumental activities of daily living. She is a former smoker of tobacco and quit smoking 30 years ago.

On arrival at our facility, the patient had normal vital signs. Significant findings on physical examination were an elderly female who seemed somnolent; not able to follow commands with a documented Glasgow Coma Scale of 10 with eyes opening spontaneously, incomprehensible sounds, and flexion withdrawal from pain as her best responses. She had an increased tone in her left upper extremity and had a brisk, deep tendon reflexes without clonus or 3+ (range, 0-5+, with 2+ being normal). The remainder of her physical exam was unremarkable. Laboratory testing revealed a glucose level of 120 mg/dL (normal, 65-110 mg/dL), sodium of 112 mEq/L (normal, 135-145 mmol/L), and chloride of 78 mEq/L (normal, 95-105 mmol/L). Serum osmolality and urine osmolality were 242 mOsm/kg (normal, 282-295 mOsm/kg) and 449 mOsm/kg (normal, 500-800 mOsm/kg) respectively, indicative of suboptimally dilute urine despite relatively low serum osmolality or SIADH. Urine electrolytes were not obtained.

Imaging studies including computed-tomography scans of the head and chest x-ray performed in the ED were unremarkable. After a phenytoin load, an electroencephalogram was obtained which showed diffuse encephalopathy without active seizure foci. A non-contrast magnetic-resonance imaging (MRI) of the brain was performed but it failed to show acute infarct, mass, mass effect, or brain herniation. There was nonspecific white matter abnormality with compromise of the bilateral cerebral hemispheres, calloseptal junction, left posterior pillar, and bilateral anterior pillars of the fornix, possibly representative of chronic white matter microvascular ischemic changes or less likely vasculitis or demyelination. Correction of her hyponatremia with normal saline was started in the ED with a change in serum sodium from 112 mEq/L to 115 mEq/L within 2 hours. She was admitted to the intensive care unit (ICU) where her sodium correction with normal saline and free water restriction was continued with a goal correction rate of 8-12 mEq/L in 24 hours. The patient’s serum sodium as well as level of consciousness improved gradually over the course of her ICU stay. After 64 hours in the hospital, her sodium had corrected to 137 mEq/L (normal, 135-145 mmol/L; Figure). She was then alert and oriented to person, place, and time. All motor findings noted on presentation had resolved. Her saline infusion was discontinued and serum sodium remained within normal range. She was discharged to a rehabilitation facility. Her hydrochlorothiazide was also discontinued.

Discussion

Hyponatremia is a common finding in cancer patients caused usually by paraneoplastic syndrome, chemotherapy, immunotherapy, or other associated treatment.⁶ SIADH is a frequent cause of hyponatremia in cancer patients and should be suspected in patients with hyponatremia, hypo-osmolality, and a urine osmolality above 100 mOsmol/kg.⁷

Our patient’s presentation and laboratory findings suggested SIADH as the likely cause of hyponatremia with a low sodium, a serum osmolality 242 mOsm/kg and urine osmolality of 449 mOsm/kg.^8-10 She had no known underlying contributory comorbid condition relating to her serum lipids, thyroid, adrenal, kidney, or heart to date. Her use of a thiazide diuretic was the only confounding factor. The most plausible cause of hyponatremia/SIADH in our patient was likely cyclophosphamide based on her history, timeline of symptoms, and the absence of other possible causes. Though the mechanism for many of the previously mentioned etiologies are known, the mechanism of cyclophosphamide-induced SIADH is difficult to elucidate since the imminent complication of hemorrhagic cystitis means patients receiving this drug are often aggressively hydrated to prevent this complication.^11,12 The result is that there is marked retention of water leading to potentially fatal hyponatremia in selected cases.¹¹ This phenomenon has been fairly well described in patients receiving doses of 6 g/m² as given in the STAMP protocol for stem cell mobilization or at doses of 30-50 mg/kg used to treat malignancy.¹² Our patient clearly falls in this category given that she received a dose of 600 mg/m². We found no evidence in her history to suggest post-operative, genetic or other cause for her hyponatremia. Our case mirrors a report by Koo and colleagues who described severe hyponatremia occurring within 24 hours following a single dose of intravenous cyclophosphamide 700 mg followed by saline infusion.¹³ In the case reported by Jayachandra and colleagues in which suspected cyclophosphamide-induced hyponatremia led to seizures, the patient received 500 mg IV of cyclophosphamide and had serum sodium as low as 106 mEq/L within a 24-hour period,² similar to our patient.

There is a paucity of data on cyclophosphamide-induced SIADH. The mechanism by which cyclophosphamide causes SIADH is currently unknown. In addition, there are currently no set criteria that help identify at-risk patients who may develop such an event, including the dosage of cyclophosphamide that may trigger the SIADH, because lower doses of the drug have been associated with this complication.¹⁴

In a retrospective analysis by Lee and colleagues, cyclophosphamide-induced hyponatremia was found to be associated with male sex on a univariate analysis, but no risk factors were found in a multivariate analysis.¹⁵ It is likely that the concomitant use of diuretics, hydration, and high-dose cyclophosphamide contributed to hyponatremia/SIADH in our patient, though it is not clear through what mechanism. Harlow and colleagues proposed a mechanism for this phenomenon in 1979 based on the autopsy of a patient who had received high-dose cyclophosphamide involving degranulation of hypothalamic neurosecretory organelles and loss of Herring’s bodies. They inferred that metabolites of cyclophosphamide indirectly triggered inappropriate secretion of antidiuretic hormone as seen with a use of the structurally related analogue ifosfamide,¹⁶ but to our knowledge, this has yet to be replicated. Cyclophosphamide metabolite may have a direct tubular effect on the collecting duct epithelium leading to water retention¹⁵ as established by Campbell and colleagues. In one case, an established diabetes insipidus patient developed cyclophosphamide-induced antidiuresis without vasopressin secretion.¹⁷ It is imperative that the scientific community conduct research into the risk factors, underlying mechanisms, and methods of prevention to reduce and/or eliminate SIADH associated with use of cyclophosphamide.

Cyclophosphamide, an agent used to treat various malignant and autoimmune disorders, can cause severe hyponatremia with seizures in rare cases. The exact mechanism of cyclophosphamide-induced hyponatremia is poorly understood, but is thought to occur from a drug- associated antidiuretic hormone (ADH) release leading to free water retention.¹ This unusual phenomenon of cyclophosphamide-associated syndrome of inappropriate antidiuretic hormone secretion (SIADH) has been described only in case reports, most of which reported the development of severe hyponatremia within a week after administration of cyclophosphamide.^2-5 We report a unique case of a patient who developed severe, symptomatic hyponatremia with seizures, with her serum sodium decreasing from 137 mEq to 112 mEq within 30 hours after her first dose of low-dose cyclophosphamide (600 mg/m²).

Case presentation and summary

A 68-year-old white woman with a history of bilateral invasive ductal carcinoma of the breast (status-post bilateral mastectomy) presented to the emergency department (ED) at our facility with new onset seizure. The patient had been diagnosed 8 months earlier with stage I (T1c, N0, M0) poorly differentiated infiltrating ductal carcinoma (triple negative) of the left breast for which she underwent left segmental mastectomy about 1 month after diagnosis. She was subsequently found to have progressive disease with stage IIIC (T2, N3, and M0) infiltrating ductal carcinoma with lobular features (ER/PR+, Her2) of the right breast. She underwent a right modified radical mastectomy 5 months after her stage IIIC breast cancer diagnosis. She received her first cycle of adjuvant chemotherapy with intravenous doxorubicin (60 mg/m²) and cyclophosphamide (600 mg/m²), which included pre-hydration, a day before presenting to our facility.

According to the patient’s family who provided the initial history, the patient reported tightness in her left arm while sitting at the dinner table. She was confused and subsequently had jerking movement of her right upper extremity with left facial twitching which lasted about 40 seconds. There was no loss of consciousness, or bowel or bladder control. She became unresponsive after the episode. Review of systems was negative except for a report of nausea a few hours before the onset of seizures, which resolved with ondansetron. Her past medical history was significant for breast cancer as already mentioned, seasonal allergic rhinitis, and hypertension. Home medications included hydrochlorothiazide 12.5 mg oral daily, aspirin 81 mg oral daily, and fexofenadine and loratadine oral daily as needed for allergies. There were no other significant surgical history other than already stated. The patient lived at home with her family and was independent with her instrumental activities of daily living. She is a former smoker of tobacco and quit smoking 30 years ago.

On arrival at our facility, the patient had normal vital signs. Significant findings on physical examination were an elderly female who seemed somnolent; not able to follow commands with a documented Glasgow Coma Scale of 10 with eyes opening spontaneously, incomprehensible sounds, and flexion withdrawal from pain as her best responses. She had an increased tone in her left upper extremity and had a brisk, deep tendon reflexes without clonus or 3+ (range, 0-5+, with 2+ being normal). The remainder of her physical exam was unremarkable. Laboratory testing revealed a glucose level of 120 mg/dL (normal, 65-110 mg/dL), sodium of 112 mEq/L (normal, 135-145 mmol/L), and chloride of 78 mEq/L (normal, 95-105 mmol/L). Serum osmolality and urine osmolality were 242 mOsm/kg (normal, 282-295 mOsm/kg) and 449 mOsm/kg (normal, 500-800 mOsm/kg) respectively, indicative of suboptimally dilute urine despite relatively low serum osmolality or SIADH. Urine electrolytes were not obtained.

Imaging studies including computed-tomography scans of the head and chest x-ray performed in the ED were unremarkable. After a phenytoin load, an electroencephalogram was obtained which showed diffuse encephalopathy without active seizure foci. A non-contrast magnetic-resonance imaging (MRI) of the brain was performed but it failed to show acute infarct, mass, mass effect, or brain herniation. There was nonspecific white matter abnormality with compromise of the bilateral cerebral hemispheres, calloseptal junction, left posterior pillar, and bilateral anterior pillars of the fornix, possibly representative of chronic white matter microvascular ischemic changes or less likely vasculitis or demyelination. Correction of her hyponatremia with normal saline was started in the ED with a change in serum sodium from 112 mEq/L to 115 mEq/L within 2 hours. She was admitted to the intensive care unit (ICU) where her sodium correction with normal saline and free water restriction was continued with a goal correction rate of 8-12 mEq/L in 24 hours. The patient’s serum sodium as well as level of consciousness improved gradually over the course of her ICU stay. After 64 hours in the hospital, her sodium had corrected to 137 mEq/L (normal, 135-145 mmol/L; Figure). She was then alert and oriented to person, place, and time. All motor findings noted on presentation had resolved. Her saline infusion was discontinued and serum sodium remained within normal range. She was discharged to a rehabilitation facility. Her hydrochlorothiazide was also discontinued.

Discussion

Hyponatremia is a common finding in cancer patients caused usually by paraneoplastic syndrome, chemotherapy, immunotherapy, or other associated treatment.⁶ SIADH is a frequent cause of hyponatremia in cancer patients and should be suspected in patients with hyponatremia, hypo-osmolality, and a urine osmolality above 100 mOsmol/kg.⁷

Our patient’s presentation and laboratory findings suggested SIADH as the likely cause of hyponatremia with a low sodium, a serum osmolality 242 mOsm/kg and urine osmolality of 449 mOsm/kg.^8-10 She had no known underlying contributory comorbid condition relating to her serum lipids, thyroid, adrenal, kidney, or heart to date. Her use of a thiazide diuretic was the only confounding factor. The most plausible cause of hyponatremia/SIADH in our patient was likely cyclophosphamide based on her history, timeline of symptoms, and the absence of other possible causes. Though the mechanism for many of the previously mentioned etiologies are known, the mechanism of cyclophosphamide-induced SIADH is difficult to elucidate since the imminent complication of hemorrhagic cystitis means patients receiving this drug are often aggressively hydrated to prevent this complication.^11,12 The result is that there is marked retention of water leading to potentially fatal hyponatremia in selected cases.¹¹ This phenomenon has been fairly well described in patients receiving doses of 6 g/m² as given in the STAMP protocol for stem cell mobilization or at doses of 30-50 mg/kg used to treat malignancy.¹² Our patient clearly falls in this category given that she received a dose of 600 mg/m². We found no evidence in her history to suggest post-operative, genetic or other cause for her hyponatremia. Our case mirrors a report by Koo and colleagues who described severe hyponatremia occurring within 24 hours following a single dose of intravenous cyclophosphamide 700 mg followed by saline infusion.¹³ In the case reported by Jayachandra and colleagues in which suspected cyclophosphamide-induced hyponatremia led to seizures, the patient received 500 mg IV of cyclophosphamide and had serum sodium as low as 106 mEq/L within a 24-hour period,² similar to our patient.

There is a paucity of data on cyclophosphamide-induced SIADH. The mechanism by which cyclophosphamide causes SIADH is currently unknown. In addition, there are currently no set criteria that help identify at-risk patients who may develop such an event, including the dosage of cyclophosphamide that may trigger the SIADH, because lower doses of the drug have been associated with this complication.¹⁴

In a retrospective analysis by Lee and colleagues, cyclophosphamide-induced hyponatremia was found to be associated with male sex on a univariate analysis, but no risk factors were found in a multivariate analysis.¹⁵ It is likely that the concomitant use of diuretics, hydration, and high-dose cyclophosphamide contributed to hyponatremia/SIADH in our patient, though it is not clear through what mechanism. Harlow and colleagues proposed a mechanism for this phenomenon in 1979 based on the autopsy of a patient who had received high-dose cyclophosphamide involving degranulation of hypothalamic neurosecretory organelles and loss of Herring’s bodies. They inferred that metabolites of cyclophosphamide indirectly triggered inappropriate secretion of antidiuretic hormone as seen with a use of the structurally related analogue ifosfamide,¹⁶ but to our knowledge, this has yet to be replicated. Cyclophosphamide metabolite may have a direct tubular effect on the collecting duct epithelium leading to water retention¹⁵ as established by Campbell and colleagues. In one case, an established diabetes insipidus patient developed cyclophosphamide-induced antidiuresis without vasopressin secretion.¹⁷ It is imperative that the scientific community conduct research into the risk factors, underlying mechanisms, and methods of prevention to reduce and/or eliminate SIADH associated with use of cyclophosphamide.

Cyclophosphamide, an agent used to treat various malignant and autoimmune disorders, can cause severe hyponatremia with seizures in rare cases. The exact mechanism of cyclophosphamide-induced hyponatremia is poorly understood, but is thought to occur from a drug- associated antidiuretic hormone (ADH) release leading to free water retention.¹ This unusual phenomenon of cyclophosphamide-associated syndrome of inappropriate antidiuretic hormone secretion (SIADH) has been described only in case reports, most of which reported the development of severe hyponatremia within a week after administration of cyclophosphamide.^2-5 We report a unique case of a patient who developed severe, symptomatic hyponatremia with seizures, with her serum sodium decreasing from 137 mEq to 112 mEq within 30 hours after her first dose of low-dose cyclophosphamide (600 mg/m²).

Case presentation and summary

A 68-year-old white woman with a history of bilateral invasive ductal carcinoma of the breast (status-post bilateral mastectomy) presented to the emergency department (ED) at our facility with new onset seizure. The patient had been diagnosed 8 months earlier with stage I (T1c, N0, M0) poorly differentiated infiltrating ductal carcinoma (triple negative) of the left breast for which she underwent left segmental mastectomy about 1 month after diagnosis. She was subsequently found to have progressive disease with stage IIIC (T2, N3, and M0) infiltrating ductal carcinoma with lobular features (ER/PR+, Her2) of the right breast. She underwent a right modified radical mastectomy 5 months after her stage IIIC breast cancer diagnosis. She received her first cycle of adjuvant chemotherapy with intravenous doxorubicin (60 mg/m²) and cyclophosphamide (600 mg/m²), which included pre-hydration, a day before presenting to our facility.

According to the patient’s family who provided the initial history, the patient reported tightness in her left arm while sitting at the dinner table. She was confused and subsequently had jerking movement of her right upper extremity with left facial twitching which lasted about 40 seconds. There was no loss of consciousness, or bowel or bladder control. She became unresponsive after the episode. Review of systems was negative except for a report of nausea a few hours before the onset of seizures, which resolved with ondansetron. Her past medical history was significant for breast cancer as already mentioned, seasonal allergic rhinitis, and hypertension. Home medications included hydrochlorothiazide 12.5 mg oral daily, aspirin 81 mg oral daily, and fexofenadine and loratadine oral daily as needed for allergies. There were no other significant surgical history other than already stated. The patient lived at home with her family and was independent with her instrumental activities of daily living. She is a former smoker of tobacco and quit smoking 30 years ago.

On arrival at our facility, the patient had normal vital signs. Significant findings on physical examination were an elderly female who seemed somnolent; not able to follow commands with a documented Glasgow Coma Scale of 10 with eyes opening spontaneously, incomprehensible sounds, and flexion withdrawal from pain as her best responses. She had an increased tone in her left upper extremity and had a brisk, deep tendon reflexes without clonus or 3+ (range, 0-5+, with 2+ being normal). The remainder of her physical exam was unremarkable. Laboratory testing revealed a glucose level of 120 mg/dL (normal, 65-110 mg/dL), sodium of 112 mEq/L (normal, 135-145 mmol/L), and chloride of 78 mEq/L (normal, 95-105 mmol/L). Serum osmolality and urine osmolality were 242 mOsm/kg (normal, 282-295 mOsm/kg) and 449 mOsm/kg (normal, 500-800 mOsm/kg) respectively, indicative of suboptimally dilute urine despite relatively low serum osmolality or SIADH. Urine electrolytes were not obtained.

Imaging studies including computed-tomography scans of the head and chest x-ray performed in the ED were unremarkable. After a phenytoin load, an electroencephalogram was obtained which showed diffuse encephalopathy without active seizure foci. A non-contrast magnetic-resonance imaging (MRI) of the brain was performed but it failed to show acute infarct, mass, mass effect, or brain herniation. There was nonspecific white matter abnormality with compromise of the bilateral cerebral hemispheres, calloseptal junction, left posterior pillar, and bilateral anterior pillars of the fornix, possibly representative of chronic white matter microvascular ischemic changes or less likely vasculitis or demyelination. Correction of her hyponatremia with normal saline was started in the ED with a change in serum sodium from 112 mEq/L to 115 mEq/L within 2 hours. She was admitted to the intensive care unit (ICU) where her sodium correction with normal saline and free water restriction was continued with a goal correction rate of 8-12 mEq/L in 24 hours. The patient’s serum sodium as well as level of consciousness improved gradually over the course of her ICU stay. After 64 hours in the hospital, her sodium had corrected to 137 mEq/L (normal, 135-145 mmol/L; Figure). She was then alert and oriented to person, place, and time. All motor findings noted on presentation had resolved. Her saline infusion was discontinued and serum sodium remained within normal range. She was discharged to a rehabilitation facility. Her hydrochlorothiazide was also discontinued.

Discussion

Hyponatremia is a common finding in cancer patients caused usually by paraneoplastic syndrome, chemotherapy, immunotherapy, or other associated treatment.⁶ SIADH is a frequent cause of hyponatremia in cancer patients and should be suspected in patients with hyponatremia, hypo-osmolality, and a urine osmolality above 100 mOsmol/kg.⁷

Our patient’s presentation and laboratory findings suggested SIADH as the likely cause of hyponatremia with a low sodium, a serum osmolality 242 mOsm/kg and urine osmolality of 449 mOsm/kg.^8-10 She had no known underlying contributory comorbid condition relating to her serum lipids, thyroid, adrenal, kidney, or heart to date. Her use of a thiazide diuretic was the only confounding factor. The most plausible cause of hyponatremia/SIADH in our patient was likely cyclophosphamide based on her history, timeline of symptoms, and the absence of other possible causes. Though the mechanism for many of the previously mentioned etiologies are known, the mechanism of cyclophosphamide-induced SIADH is difficult to elucidate since the imminent complication of hemorrhagic cystitis means patients receiving this drug are often aggressively hydrated to prevent this complication.^11,12 The result is that there is marked retention of water leading to potentially fatal hyponatremia in selected cases.¹¹ This phenomenon has been fairly well described in patients receiving doses of 6 g/m² as given in the STAMP protocol for stem cell mobilization or at doses of 30-50 mg/kg used to treat malignancy.¹² Our patient clearly falls in this category given that she received a dose of 600 mg/m². We found no evidence in her history to suggest post-operative, genetic or other cause for her hyponatremia. Our case mirrors a report by Koo and colleagues who described severe hyponatremia occurring within 24 hours following a single dose of intravenous cyclophosphamide 700 mg followed by saline infusion.¹³ In the case reported by Jayachandra and colleagues in which suspected cyclophosphamide-induced hyponatremia led to seizures, the patient received 500 mg IV of cyclophosphamide and had serum sodium as low as 106 mEq/L within a 24-hour period,² similar to our patient.

There is a paucity of data on cyclophosphamide-induced SIADH. The mechanism by which cyclophosphamide causes SIADH is currently unknown. In addition, there are currently no set criteria that help identify at-risk patients who may develop such an event, including the dosage of cyclophosphamide that may trigger the SIADH, because lower doses of the drug have been associated with this complication.¹⁴

In a retrospective analysis by Lee and colleagues, cyclophosphamide-induced hyponatremia was found to be associated with male sex on a univariate analysis, but no risk factors were found in a multivariate analysis.¹⁵ It is likely that the concomitant use of diuretics, hydration, and high-dose cyclophosphamide contributed to hyponatremia/SIADH in our patient, though it is not clear through what mechanism. Harlow and colleagues proposed a mechanism for this phenomenon in 1979 based on the autopsy of a patient who had received high-dose cyclophosphamide involving degranulation of hypothalamic neurosecretory organelles and loss of Herring’s bodies. They inferred that metabolites of cyclophosphamide indirectly triggered inappropriate secretion of antidiuretic hormone as seen with a use of the structurally related analogue ifosfamide,¹⁶ but to our knowledge, this has yet to be replicated. Cyclophosphamide metabolite may have a direct tubular effect on the collecting duct epithelium leading to water retention¹⁵ as established by Campbell and colleagues. In one case, an established diabetes insipidus patient developed cyclophosphamide-induced antidiuresis without vasopressin secretion.¹⁷ It is imperative that the scientific community conduct research into the risk factors, underlying mechanisms, and methods of prevention to reduce and/or eliminate SIADH associated with use of cyclophosphamide.

Pathologic and, increasingly, molecular analysis of tumor tissue biopsies is the gold standard in initial diagnosis of cancer, but liquid biopsies, which analyze tumor-derived material circulating in the bloodstream are gaining traction. Here, we discuss the current state of development of this complementary and potentially alternative approach to tumor analysis.

Liquid biopsy gaining traction

Biopsies enable oncologists to gather information about a potential or established tumor, including confirmation of the presence of cancerous tissue and determination of its histological characteristics, such as tumor grade and stage, as well as its molecular features, such as the presence of certain gene mutations. Ultimately, this information can be put to use in determining the most appropriate course of treatment.

The current gold standard is a tissue biopsy that typically involves an invasive procedure to permit the collection of a piece of tumor tissue. Yet, tissue biopsies are not always feasible because of the location of the tumor or the poor performance status of many patients with advanced disease. They also provide only a snapshot of the disease at the time at which they were taken and don’t necessarily reflect the genetic heterogeneity or evolution of a tumor over time.

The detection of components that are derived from the tumor circulating in the blood of cancer patients had fueled the idea of blood-based diagnostics in oncology – so-called liquid biopsies. These have rapidly gained traction in the past several decades as a less expensive (the cost of performing genomic analyses on blood samples is at least an order of magnitude less than on tissue samples), less invasive (requiring only a simple blood draw) alternative source of information about tumors.¹

As researchers have refined the ability to exploit liquid biopsies, commercial interest has been piqued. More than 35 companies within the United States alone are developing liquid biopsies, and it’s easy to see why with a market projected to be in the many billions of dollars.²

Seeking out tumor clues in the blood

Liquid biopsies consist of a 10-15 mL blood sample drawn into a tube that contains an anticoagulant and it can contain several different types of tumor-associated material. Thus far, two components – circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA) – have formed the cornerstone of liquid biopsies. At present, it is not clear whether these components are released randomly, as a by-product of tumor cell death or if they are released as part of a specific biologic process, such as for the colonization of metastatic sites. It reality, it may be a little of both, and active dissemination may be particularly relevant for CTCs, among which are postulated to be a population of cancer stem cells that can initiate distant metastases.^3,4

The discovery of CTCs dates back to the 1860s, when cells that were morphologically identical to the tumor were identified in the blood of a patient with metastatic cancer. Their potential significance was not fully realized until a few decades ago, when they were found to exist from early on in the course of disease.^3,4

CTCs, which can be either single cells or clusters of cells known as microemboli, have a short half-life in the bloodstream – less than 2 ½ hours – and are also extremely rare (1 mL of blood contains 1-10 CTCs) against a background of many millions of normal cells. Thus the detection and isolation of CTCs presents a significant challenge. More than 40 different platforms are being developed for the isolation and enrichment of CTCs. For the most part, these use a method called positive selection to pick out CTCs.^1,3,4

Positive selection exploits the biological or physical properties that are specific to CTCs and absent in normal cells, for example, the presence of a specific tumor-associated antigen on their surface or differences in size, density or electric charge. The limitations of this method are that, not only do you need to know something about CTCs to begin to understand what makes them truly unique and ensure only isolation of CTCs, but their phenotype is also thought to be continually changing.^1,3,4

In recent years, the focus has shifted toward technologies that use negative depletion, meaning that they target the other types of cells in the blood sample and filter those away until only the CTCs are left behind. The most advanced are devices that use microfluidic technology to sort the cells, such as the CTC-iChip system being developed by researchers at Massachusetts General Hospital in Boston.⁵

ctDNA consists of small fragments of nucleic acids that are not contained within a cell or associated with cell fragments and is thought to be present in 50%-90% of patients, depending on the type of cancer they have. ctDNA has a similarly short half-life in the circulation to CTCs and, like CTCs, ctDNA is present at very low levels in the bloodstream. Although levels of ctDNA have been shown to increase with increasing tumor burden, it is still often obscured by the presence of other cell-free DNA derived from non-tumor cells.

ctDNA can be distinguished from other cell-free DNA by the presence of somatic mutations and a number of highly sensitive methods have been developed to detect them, including the amplification-refractory mutation system (ARMS); digital polymerase chain reaction; and the beads, emulsification, amplification, and magnetics (BEAMing) system. Next-generation sequencing technologies, including tagged-amplicon deep sequencing (TAm-Seq), the Safe-Sequencing System (Safe-SeqS), and cancer personalized profiling by deep sequencing (CAPP-seq), can also be used and the race for ever more sensitive analytical tools is ongoing.^1,3,4,6

Applying liquid biopsies now and in the future

There are a plethora of potential applications for liquid biopsies^3,7 (Figure 1), and probably the most exciting among them is the potential for screening for and early detection of cancer. The fact that ctDNA and CTCs have both been shown to be present from the earliest stages of disease has sparked interest in the possibility of developing simple blood tests to identify tumors before they become detectable by other methods and at a point at which they may be curable.

Given that both are present at such low levels within the circulation and are particularly sparse at earlier stages of disease, current technologies may lack the specificity and sensitivity for this application at present. However, numerous clinical trials are ongoing.

For CTCs, simple enumeration has been the most extensively investigated application to date. Numerous studies have shown that the number of CTCs in the bloodstream has prognostic significance in various different tumor types. Three such studies led to the first regulatory approval for a CTC detection system (Table 1 and Table 2).^8-10

One area in which liquid biopsies could really come into their own is in providing more real-time analysis of tumors. This is something that has proven particularly challenging with tissue biopsies because repeating these invasive procedures is problematic. But the ease of repeat blood draws means that serial liquid biopsies could be performed and might offer the possibility of monitoring disease progression and evolution over the course of disease and particularly in response to treatment.

Indeed, studies have shown that in addition to baseline CTC counts, changes in CTC number during treatment are also prognostic. There was improved survival among patients whose CTC counts decreased below a threshold value during treatment and vice versa. This is also an approved use for CellSearch though at present it is not widely clinically implemented.¹²

Clinical utility remains elusive

The ultimate goal would be for liquid biopsies to have an impact on treatment decisions, allowing oncologists to change management strategy based on predicted sensitivity or resistance to therapy, so-called clinical utility. Thus far, clinical utility has proved elusive, though liquid biopsies using ctDNA to evaluate tumor genotype have come closest.

The Cobas EGFR Mutation Test v2 recently became the first ctDNA-based liquid biopsy to receive regulatory approval. It was approved as a companion diagnostic to identify patients with advanced non–small-cell lung cancer (NSCLC) who have specific mutations in the epidermal growth factor receptor (EGFR) gene and are therefore eligible for treatment with the EGFR inhibitor erlotinib.¹³

Approval was based on comparison of EGFR mutation identification rates using plasma ctDNA samples and tumor tissue samples from patients enrolled in the phase 3 ENSURE trial, which compared the efficacy of erlotinib with chemotherapy as first-line therapy in patients with advanced NSCLC. Of the 217 patients enrolled in the trial, 98.6% of patients had both tumor biopsy and plasma ctDNA samples available for testing. Concordance between the two types of biopsy in identifying patients with EGFR mutations was high and patients with EGFR positivity according to liquid biopsy results demonstrated improved progression-free survival when treated with erlotinib.¹⁴

The results of a large-scale genomic analysis of various different types of tumors using ctDNA were also recently presented at the 2016 American Society of Clinical Oncology meeting. Blood samples from more than 15,000 patients with 50 different tumor types, including advanced lung cancer (37%), breast cancer (14%), and CRC (10%), were collected and compared with either available tumor biopsy samples from the same cases (n = 398) or, in the majority of cases, with The Cancer Genome Atlas database, which uses tumor biopsies to perform genome-wide sequencing studies. Both types of biopsy revealed very similar mutation patterns when the Guardant360 next-generation sequencing test, which targets 70 genes, was applied. In particular, when EGFR, BRAF, KRAS, ALK, RET, and ROS1 mutations were identified by tumor tissue biopsy, the same mutations were reported in 94%-100% of plasma samples.¹⁵

Studies of the clinical utility of ctDNA and CTCs are among ongoing clinical trials of liquid biopsies (Tables 2 and 3). The potential for using CTCs to guide treatment decisions has become particularly relevant in breast cancer in light of results showing that patients with primary tumors that are negative for human epidermal growth factor receptor 2 (HER2) amplification, an important biomarker in breast cancer, may have CTCs that are HER2-positive, in up to 30% of cases. These patients may therefore still benefit from HER2-targeted therapy.¹⁶

The DETECT studies are the first phase 3 trials in which treatment decisions are being based on the phenotypic characteristics of CTCs. DETECT III (NCT01619111) is comparing lapatinib in combination with standard therapy with standard therapy alone in patients with HER2-negative metastatic breast cancer who have HER2-positive CTCs, whereas DETECT IV (NCT02035813) is enrolling patients with HER2-negative, hormone receptor-positive metastatic breast cancer and persistent HER2-negative CTCs to receive standard endocrine therapy and the mammalian target of rapamycin inhibitor everolimus.

Other targets and sources for liquid biopsy

Another approach to liquid biopsies that is also beginning to take off is to collect tumor-derived exosomes from the bloodstream. Exosomes are tiny, fluid-filled, membrane-bound sacks that bud off from the surface of a cell to expel waste or to transport cargo from one cell to another. DNA, RNA, and protein can be extracted from tumor-derived exosomes and could also serve as molecular biomarkers relating to the cancer cells from which they came.^6,7

Exosome Diagnostics is bringing the first exosome-based diagnostic tests to the market and recently teamed up with Amgen for the development of these liquid biopsies.¹⁷ In January 2016, they launched ExoDx Lung (ALK), for detection of EML4-ALK gene fusions in patients with NSCLC, using a proprietary platform for the isolation of RNA from exosomes. Data that was presented at several different conferences in 2015 demonstrated a sensitivity of 88% and specificity of 100% for this diagnostic when compared with tissue ALK status in NSCLC patients receiving a second-generation ALK inhibitor following progression on prior ALK inhibitor therapy.¹⁸

In September, they subsequently announced the launch of a test that analyses genetic information from exosomes collected from a urine sample taken from prostate cancer patients. Using a 3-gene signature, in combination with a proprietary algorithm, this diagnostic generates a score assessing a prostate cancer patient’s risk for higher grade, more aggressive disease. It is designed to complement the prostate-specific antigen score and has demonstrated accuracy in ruling out the presence of high-grade cancer before an initial biopsy in more than 1,500 patients.¹⁹

Pathologic and, increasingly, molecular analysis of tumor tissue biopsies is the gold standard in initial diagnosis of cancer, but liquid biopsies, which analyze tumor-derived material circulating in the bloodstream are gaining traction. Here, we discuss the current state of development of this complementary and potentially alternative approach to tumor analysis.

Liquid biopsy gaining traction

Biopsies enable oncologists to gather information about a potential or established tumor, including confirmation of the presence of cancerous tissue and determination of its histological characteristics, such as tumor grade and stage, as well as its molecular features, such as the presence of certain gene mutations. Ultimately, this information can be put to use in determining the most appropriate course of treatment.

The current gold standard is a tissue biopsy that typically involves an invasive procedure to permit the collection of a piece of tumor tissue. Yet, tissue biopsies are not always feasible because of the location of the tumor or the poor performance status of many patients with advanced disease. They also provide only a snapshot of the disease at the time at which they were taken and don’t necessarily reflect the genetic heterogeneity or evolution of a tumor over time.

The detection of components that are derived from the tumor circulating in the blood of cancer patients had fueled the idea of blood-based diagnostics in oncology – so-called liquid biopsies. These have rapidly gained traction in the past several decades as a less expensive (the cost of performing genomic analyses on blood samples is at least an order of magnitude less than on tissue samples), less invasive (requiring only a simple blood draw) alternative source of information about tumors.¹

As researchers have refined the ability to exploit liquid biopsies, commercial interest has been piqued. More than 35 companies within the United States alone are developing liquid biopsies, and it’s easy to see why with a market projected to be in the many billions of dollars.²

Seeking out tumor clues in the blood

Liquid biopsies consist of a 10-15 mL blood sample drawn into a tube that contains an anticoagulant and it can contain several different types of tumor-associated material. Thus far, two components – circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA) – have formed the cornerstone of liquid biopsies. At present, it is not clear whether these components are released randomly, as a by-product of tumor cell death or if they are released as part of a specific biologic process, such as for the colonization of metastatic sites. It reality, it may be a little of both, and active dissemination may be particularly relevant for CTCs, among which are postulated to be a population of cancer stem cells that can initiate distant metastases.^3,4

The discovery of CTCs dates back to the 1860s, when cells that were morphologically identical to the tumor were identified in the blood of a patient with metastatic cancer. Their potential significance was not fully realized until a few decades ago, when they were found to exist from early on in the course of disease.^3,4

CTCs, which can be either single cells or clusters of cells known as microemboli, have a short half-life in the bloodstream – less than 2 ½ hours – and are also extremely rare (1 mL of blood contains 1-10 CTCs) against a background of many millions of normal cells. Thus the detection and isolation of CTCs presents a significant challenge. More than 40 different platforms are being developed for the isolation and enrichment of CTCs. For the most part, these use a method called positive selection to pick out CTCs.^1,3,4

Positive selection exploits the biological or physical properties that are specific to CTCs and absent in normal cells, for example, the presence of a specific tumor-associated antigen on their surface or differences in size, density or electric charge. The limitations of this method are that, not only do you need to know something about CTCs to begin to understand what makes them truly unique and ensure only isolation of CTCs, but their phenotype is also thought to be continually changing.^1,3,4

In recent years, the focus has shifted toward technologies that use negative depletion, meaning that they target the other types of cells in the blood sample and filter those away until only the CTCs are left behind. The most advanced are devices that use microfluidic technology to sort the cells, such as the CTC-iChip system being developed by researchers at Massachusetts General Hospital in Boston.⁵

ctDNA consists of small fragments of nucleic acids that are not contained within a cell or associated with cell fragments and is thought to be present in 50%-90% of patients, depending on the type of cancer they have. ctDNA has a similarly short half-life in the circulation to CTCs and, like CTCs, ctDNA is present at very low levels in the bloodstream. Although levels of ctDNA have been shown to increase with increasing tumor burden, it is still often obscured by the presence of other cell-free DNA derived from non-tumor cells.

ctDNA can be distinguished from other cell-free DNA by the presence of somatic mutations and a number of highly sensitive methods have been developed to detect them, including the amplification-refractory mutation system (ARMS); digital polymerase chain reaction; and the beads, emulsification, amplification, and magnetics (BEAMing) system. Next-generation sequencing technologies, including tagged-amplicon deep sequencing (TAm-Seq), the Safe-Sequencing System (Safe-SeqS), and cancer personalized profiling by deep sequencing (CAPP-seq), can also be used and the race for ever more sensitive analytical tools is ongoing.^1,3,4,6

Applying liquid biopsies now and in the future

There are a plethora of potential applications for liquid biopsies^3,7 (Figure 1), and probably the most exciting among them is the potential for screening for and early detection of cancer. The fact that ctDNA and CTCs have both been shown to be present from the earliest stages of disease has sparked interest in the possibility of developing simple blood tests to identify tumors before they become detectable by other methods and at a point at which they may be curable.

Given that both are present at such low levels within the circulation and are particularly sparse at earlier stages of disease, current technologies may lack the specificity and sensitivity for this application at present. However, numerous clinical trials are ongoing.

For CTCs, simple enumeration has been the most extensively investigated application to date. Numerous studies have shown that the number of CTCs in the bloodstream has prognostic significance in various different tumor types. Three such studies led to the first regulatory approval for a CTC detection system (Table 1 and Table 2).^8-10

One area in which liquid biopsies could really come into their own is in providing more real-time analysis of tumors. This is something that has proven particularly challenging with tissue biopsies because repeating these invasive procedures is problematic. But the ease of repeat blood draws means that serial liquid biopsies could be performed and might offer the possibility of monitoring disease progression and evolution over the course of disease and particularly in response to treatment.

Indeed, studies have shown that in addition to baseline CTC counts, changes in CTC number during treatment are also prognostic. There was improved survival among patients whose CTC counts decreased below a threshold value during treatment and vice versa. This is also an approved use for CellSearch though at present it is not widely clinically implemented.¹²

Clinical utility remains elusive

The ultimate goal would be for liquid biopsies to have an impact on treatment decisions, allowing oncologists to change management strategy based on predicted sensitivity or resistance to therapy, so-called clinical utility. Thus far, clinical utility has proved elusive, though liquid biopsies using ctDNA to evaluate tumor genotype have come closest.

The Cobas EGFR Mutation Test v2 recently became the first ctDNA-based liquid biopsy to receive regulatory approval. It was approved as a companion diagnostic to identify patients with advanced non–small-cell lung cancer (NSCLC) who have specific mutations in the epidermal growth factor receptor (EGFR) gene and are therefore eligible for treatment with the EGFR inhibitor erlotinib.¹³

Approval was based on comparison of EGFR mutation identification rates using plasma ctDNA samples and tumor tissue samples from patients enrolled in the phase 3 ENSURE trial, which compared the efficacy of erlotinib with chemotherapy as first-line therapy in patients with advanced NSCLC. Of the 217 patients enrolled in the trial, 98.6% of patients had both tumor biopsy and plasma ctDNA samples available for testing. Concordance between the two types of biopsy in identifying patients with EGFR mutations was high and patients with EGFR positivity according to liquid biopsy results demonstrated improved progression-free survival when treated with erlotinib.¹⁴

The results of a large-scale genomic analysis of various different types of tumors using ctDNA were also recently presented at the 2016 American Society of Clinical Oncology meeting. Blood samples from more than 15,000 patients with 50 different tumor types, including advanced lung cancer (37%), breast cancer (14%), and CRC (10%), were collected and compared with either available tumor biopsy samples from the same cases (n = 398) or, in the majority of cases, with The Cancer Genome Atlas database, which uses tumor biopsies to perform genome-wide sequencing studies. Both types of biopsy revealed very similar mutation patterns when the Guardant360 next-generation sequencing test, which targets 70 genes, was applied. In particular, when EGFR, BRAF, KRAS, ALK, RET, and ROS1 mutations were identified by tumor tissue biopsy, the same mutations were reported in 94%-100% of plasma samples.¹⁵

Studies of the clinical utility of ctDNA and CTCs are among ongoing clinical trials of liquid biopsies (Tables 2 and 3). The potential for using CTCs to guide treatment decisions has become particularly relevant in breast cancer in light of results showing that patients with primary tumors that are negative for human epidermal growth factor receptor 2 (HER2) amplification, an important biomarker in breast cancer, may have CTCs that are HER2-positive, in up to 30% of cases. These patients may therefore still benefit from HER2-targeted therapy.¹⁶

The DETECT studies are the first phase 3 trials in which treatment decisions are being based on the phenotypic characteristics of CTCs. DETECT III (NCT01619111) is comparing lapatinib in combination with standard therapy with standard therapy alone in patients with HER2-negative metastatic breast cancer who have HER2-positive CTCs, whereas DETECT IV (NCT02035813) is enrolling patients with HER2-negative, hormone receptor-positive metastatic breast cancer and persistent HER2-negative CTCs to receive standard endocrine therapy and the mammalian target of rapamycin inhibitor everolimus.

Other targets and sources for liquid biopsy

Another approach to liquid biopsies that is also beginning to take off is to collect tumor-derived exosomes from the bloodstream. Exosomes are tiny, fluid-filled, membrane-bound sacks that bud off from the surface of a cell to expel waste or to transport cargo from one cell to another. DNA, RNA, and protein can be extracted from tumor-derived exosomes and could also serve as molecular biomarkers relating to the cancer cells from which they came.^6,7

Exosome Diagnostics is bringing the first exosome-based diagnostic tests to the market and recently teamed up with Amgen for the development of these liquid biopsies.¹⁷ In January 2016, they launched ExoDx Lung (ALK), for detection of EML4-ALK gene fusions in patients with NSCLC, using a proprietary platform for the isolation of RNA from exosomes. Data that was presented at several different conferences in 2015 demonstrated a sensitivity of 88% and specificity of 100% for this diagnostic when compared with tissue ALK status in NSCLC patients receiving a second-generation ALK inhibitor following progression on prior ALK inhibitor therapy.¹⁸

In September, they subsequently announced the launch of a test that analyses genetic information from exosomes collected from a urine sample taken from prostate cancer patients. Using a 3-gene signature, in combination with a proprietary algorithm, this diagnostic generates a score assessing a prostate cancer patient’s risk for higher grade, more aggressive disease. It is designed to complement the prostate-specific antigen score and has demonstrated accuracy in ruling out the presence of high-grade cancer before an initial biopsy in more than 1,500 patients.¹⁹

Pathologic and, increasingly, molecular analysis of tumor tissue biopsies is the gold standard in initial diagnosis of cancer, but liquid biopsies, which analyze tumor-derived material circulating in the bloodstream are gaining traction. Here, we discuss the current state of development of this complementary and potentially alternative approach to tumor analysis.

Liquid biopsy gaining traction

Biopsies enable oncologists to gather information about a potential or established tumor, including confirmation of the presence of cancerous tissue and determination of its histological characteristics, such as tumor grade and stage, as well as its molecular features, such as the presence of certain gene mutations. Ultimately, this information can be put to use in determining the most appropriate course of treatment.

The current gold standard is a tissue biopsy that typically involves an invasive procedure to permit the collection of a piece of tumor tissue. Yet, tissue biopsies are not always feasible because of the location of the tumor or the poor performance status of many patients with advanced disease. They also provide only a snapshot of the disease at the time at which they were taken and don’t necessarily reflect the genetic heterogeneity or evolution of a tumor over time.

The detection of components that are derived from the tumor circulating in the blood of cancer patients had fueled the idea of blood-based diagnostics in oncology – so-called liquid biopsies. These have rapidly gained traction in the past several decades as a less expensive (the cost of performing genomic analyses on blood samples is at least an order of magnitude less than on tissue samples), less invasive (requiring only a simple blood draw) alternative source of information about tumors.¹

As researchers have refined the ability to exploit liquid biopsies, commercial interest has been piqued. More than 35 companies within the United States alone are developing liquid biopsies, and it’s easy to see why with a market projected to be in the many billions of dollars.²

Seeking out tumor clues in the blood

Liquid biopsies consist of a 10-15 mL blood sample drawn into a tube that contains an anticoagulant and it can contain several different types of tumor-associated material. Thus far, two components – circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA) – have formed the cornerstone of liquid biopsies. At present, it is not clear whether these components are released randomly, as a by-product of tumor cell death or if they are released as part of a specific biologic process, such as for the colonization of metastatic sites. It reality, it may be a little of both, and active dissemination may be particularly relevant for CTCs, among which are postulated to be a population of cancer stem cells that can initiate distant metastases.^3,4

The discovery of CTCs dates back to the 1860s, when cells that were morphologically identical to the tumor were identified in the blood of a patient with metastatic cancer. Their potential significance was not fully realized until a few decades ago, when they were found to exist from early on in the course of disease.^3,4

CTCs, which can be either single cells or clusters of cells known as microemboli, have a short half-life in the bloodstream – less than 2 ½ hours – and are also extremely rare (1 mL of blood contains 1-10 CTCs) against a background of many millions of normal cells. Thus the detection and isolation of CTCs presents a significant challenge. More than 40 different platforms are being developed for the isolation and enrichment of CTCs. For the most part, these use a method called positive selection to pick out CTCs.^1,3,4

Positive selection exploits the biological or physical properties that are specific to CTCs and absent in normal cells, for example, the presence of a specific tumor-associated antigen on their surface or differences in size, density or electric charge. The limitations of this method are that, not only do you need to know something about CTCs to begin to understand what makes them truly unique and ensure only isolation of CTCs, but their phenotype is also thought to be continually changing.^1,3,4

In recent years, the focus has shifted toward technologies that use negative depletion, meaning that they target the other types of cells in the blood sample and filter those away until only the CTCs are left behind. The most advanced are devices that use microfluidic technology to sort the cells, such as the CTC-iChip system being developed by researchers at Massachusetts General Hospital in Boston.⁵

ctDNA consists of small fragments of nucleic acids that are not contained within a cell or associated with cell fragments and is thought to be present in 50%-90% of patients, depending on the type of cancer they have. ctDNA has a similarly short half-life in the circulation to CTCs and, like CTCs, ctDNA is present at very low levels in the bloodstream. Although levels of ctDNA have been shown to increase with increasing tumor burden, it is still often obscured by the presence of other cell-free DNA derived from non-tumor cells.

ctDNA can be distinguished from other cell-free DNA by the presence of somatic mutations and a number of highly sensitive methods have been developed to detect them, including the amplification-refractory mutation system (ARMS); digital polymerase chain reaction; and the beads, emulsification, amplification, and magnetics (BEAMing) system. Next-generation sequencing technologies, including tagged-amplicon deep sequencing (TAm-Seq), the Safe-Sequencing System (Safe-SeqS), and cancer personalized profiling by deep sequencing (CAPP-seq), can also be used and the race for ever more sensitive analytical tools is ongoing.^1,3,4,6

Applying liquid biopsies now and in the future

There are a plethora of potential applications for liquid biopsies^3,7 (Figure 1), and probably the most exciting among them is the potential for screening for and early detection of cancer. The fact that ctDNA and CTCs have both been shown to be present from the earliest stages of disease has sparked interest in the possibility of developing simple blood tests to identify tumors before they become detectable by other methods and at a point at which they may be curable.

Given that both are present at such low levels within the circulation and are particularly sparse at earlier stages of disease, current technologies may lack the specificity and sensitivity for this application at present. However, numerous clinical trials are ongoing.

For CTCs, simple enumeration has been the most extensively investigated application to date. Numerous studies have shown that the number of CTCs in the bloodstream has prognostic significance in various different tumor types. Three such studies led to the first regulatory approval for a CTC detection system (Table 1 and Table 2).^8-10

One area in which liquid biopsies could really come into their own is in providing more real-time analysis of tumors. This is something that has proven particularly challenging with tissue biopsies because repeating these invasive procedures is problematic. But the ease of repeat blood draws means that serial liquid biopsies could be performed and might offer the possibility of monitoring disease progression and evolution over the course of disease and particularly in response to treatment.

Indeed, studies have shown that in addition to baseline CTC counts, changes in CTC number during treatment are also prognostic. There was improved survival among patients whose CTC counts decreased below a threshold value during treatment and vice versa. This is also an approved use for CellSearch though at present it is not widely clinically implemented.¹²

Clinical utility remains elusive

The ultimate goal would be for liquid biopsies to have an impact on treatment decisions, allowing oncologists to change management strategy based on predicted sensitivity or resistance to therapy, so-called clinical utility. Thus far, clinical utility has proved elusive, though liquid biopsies using ctDNA to evaluate tumor genotype have come closest.

The Cobas EGFR Mutation Test v2 recently became the first ctDNA-based liquid biopsy to receive regulatory approval. It was approved as a companion diagnostic to identify patients with advanced non–small-cell lung cancer (NSCLC) who have specific mutations in the epidermal growth factor receptor (EGFR) gene and are therefore eligible for treatment with the EGFR inhibitor erlotinib.¹³

Approval was based on comparison of EGFR mutation identification rates using plasma ctDNA samples and tumor tissue samples from patients enrolled in the phase 3 ENSURE trial, which compared the efficacy of erlotinib with chemotherapy as first-line therapy in patients with advanced NSCLC. Of the 217 patients enrolled in the trial, 98.6% of patients had both tumor biopsy and plasma ctDNA samples available for testing. Concordance between the two types of biopsy in identifying patients with EGFR mutations was high and patients with EGFR positivity according to liquid biopsy results demonstrated improved progression-free survival when treated with erlotinib.¹⁴

The results of a large-scale genomic analysis of various different types of tumors using ctDNA were also recently presented at the 2016 American Society of Clinical Oncology meeting. Blood samples from more than 15,000 patients with 50 different tumor types, including advanced lung cancer (37%), breast cancer (14%), and CRC (10%), were collected and compared with either available tumor biopsy samples from the same cases (n = 398) or, in the majority of cases, with The Cancer Genome Atlas database, which uses tumor biopsies to perform genome-wide sequencing studies. Both types of biopsy revealed very similar mutation patterns when the Guardant360 next-generation sequencing test, which targets 70 genes, was applied. In particular, when EGFR, BRAF, KRAS, ALK, RET, and ROS1 mutations were identified by tumor tissue biopsy, the same mutations were reported in 94%-100% of plasma samples.¹⁵

Studies of the clinical utility of ctDNA and CTCs are among ongoing clinical trials of liquid biopsies (Tables 2 and 3). The potential for using CTCs to guide treatment decisions has become particularly relevant in breast cancer in light of results showing that patients with primary tumors that are negative for human epidermal growth factor receptor 2 (HER2) amplification, an important biomarker in breast cancer, may have CTCs that are HER2-positive, in up to 30% of cases. These patients may therefore still benefit from HER2-targeted therapy.¹⁶

The DETECT studies are the first phase 3 trials in which treatment decisions are being based on the phenotypic characteristics of CTCs. DETECT III (NCT01619111) is comparing lapatinib in combination with standard therapy with standard therapy alone in patients with HER2-negative metastatic breast cancer who have HER2-positive CTCs, whereas DETECT IV (NCT02035813) is enrolling patients with HER2-negative, hormone receptor-positive metastatic breast cancer and persistent HER2-negative CTCs to receive standard endocrine therapy and the mammalian target of rapamycin inhibitor everolimus.

Other targets and sources for liquid biopsy

Another approach to liquid biopsies that is also beginning to take off is to collect tumor-derived exosomes from the bloodstream. Exosomes are tiny, fluid-filled, membrane-bound sacks that bud off from the surface of a cell to expel waste or to transport cargo from one cell to another. DNA, RNA, and protein can be extracted from tumor-derived exosomes and could also serve as molecular biomarkers relating to the cancer cells from which they came.^6,7

Exosome Diagnostics is bringing the first exosome-based diagnostic tests to the market and recently teamed up with Amgen for the development of these liquid biopsies.¹⁷ In January 2016, they launched ExoDx Lung (ALK), for detection of EML4-ALK gene fusions in patients with NSCLC, using a proprietary platform for the isolation of RNA from exosomes. Data that was presented at several different conferences in 2015 demonstrated a sensitivity of 88% and specificity of 100% for this diagnostic when compared with tissue ALK status in NSCLC patients receiving a second-generation ALK inhibitor following progression on prior ALK inhibitor therapy.¹⁸

In September, they subsequently announced the launch of a test that analyses genetic information from exosomes collected from a urine sample taken from prostate cancer patients. Using a 3-gene signature, in combination with a proprietary algorithm, this diagnostic generates a score assessing a prostate cancer patient’s risk for higher grade, more aggressive disease. It is designed to complement the prostate-specific antigen score and has demonstrated accuracy in ruling out the presence of high-grade cancer before an initial biopsy in more than 1,500 patients.¹⁹