User login
GENEVA – Radiogenomic MRI signatures may be able to identify anaplastic lymphoma kinase (ALK)–positive brain metastases in non–small cell lung cancer (NSCLC), offering a minimally invasive option that could allow for initiation of treatment while waiting for molecular results, according to investigators.
In the future, artificial intelligence may be able to detect these imaging patterns, allowing for rapid and accurate mutation subtyping, reported lead author Shweta Wadhwa, MD, of Tata Memorial Centre in Mumbai, India, who presented the findings at the European Lung Cancer Conference.
“Radiogenomics is a concept used to associate genetic information with medical images,” Dr. Wadhwa explained at the meeting presented by the European Society for Medical Oncology. “It creates imaging biomarkers noninvasively without using biopsy. … The aim of my study was to analyze certain MRI data genomic parameters and correlate with the ALK mutation status.”
Dr. Wadhwa and her colleagues retrospectively analyzed data from 75 patients with ALK-positive NSCLC who underwent multiparametric MRI at the time of diagnosis. Univariate logistic regression analysis was conducted to look for associations between ALK mutation status and various clinical factors, including sex, age, smoking, histology, TNM stage, and imaging characteristics.
Out of 75 patients, 46 were ALK positive and 29 were ALK negative. Analysis showed that ALK positivity was associated with a variety of lesion morphology characteristics. ALK-positive lesions more often exhibited a fuzzy and infiltrative T2w border with hypointense peripheral solid rim, compared with ALK-negative lesions, which frequently had a well-defined T2w border with no solid rim (P less than .001). On T1w, most ALK-positive lesions were heterogeneous, whereas ALK-negative lesions were predominantly hypointense (P less than .001). Diffusion-weighted images showed that ALK-positive lesions often had peripheral restriction of the solid rim, compared with ALK-negative lesions, which were associated with central restriction (P = .001). MRI also revealed that about half of ALK-positive patients (54.3%) had meningeal involvement, compared with just 17.2% of ALK-negative patients (P = .02). ALK positivity was also associated with younger age and lack of smoking history. Considering these findings, Dr. Wadhwa concluded that “radiogenomics has a potential role in personalized management of ALK-positive NSCLC brain metastases.”
In an interview, Dr. Wadhwa provided more insight regarding the clinical need for this technology. “We have to wait for 10 days [for molecular diagnostic results], and ALK is usually aggressive disease, so if we wait for 10 days, patients can undergo rapid progression.”
Dr. Wadhwa noted that these results are similar to that of her colleague, Abhishek Mahajan, MD, who recently published results showing potential for radiogenomic detection of epidermal growth factor receptor (EGFR) status. According to Dr. Wadhwa, the two investigators plan to build on their collective findings in an effort to automate radiogenomic detection of NSCLC mutation subtypes.
“My upcoming project with my coinvestigator is to take a bigger sample,” Dr. Wadhwa said. “We will be further generalizing [this process] to all patients in a prospective study. We will also be sending this to the University of Pennsylvania for automatic brain segmentation.” Dr. Wadhwa estimated that adding automation will provide an accuracy rate of around 90%.
“We will train the computer accordingly,” Dr. Wadhwa said, “and then the computer will tell us, yes, this is ALK positive, this is EGFR positive.”
The investigators reported no external study funding and reported no conflicts of interest.
SOURCE: Wadhwa S et al. ELCC 2019, Abstract 55O.
GENEVA – Radiogenomic MRI signatures may be able to identify anaplastic lymphoma kinase (ALK)–positive brain metastases in non–small cell lung cancer (NSCLC), offering a minimally invasive option that could allow for initiation of treatment while waiting for molecular results, according to investigators.
In the future, artificial intelligence may be able to detect these imaging patterns, allowing for rapid and accurate mutation subtyping, reported lead author Shweta Wadhwa, MD, of Tata Memorial Centre in Mumbai, India, who presented the findings at the European Lung Cancer Conference.
“Radiogenomics is a concept used to associate genetic information with medical images,” Dr. Wadhwa explained at the meeting presented by the European Society for Medical Oncology. “It creates imaging biomarkers noninvasively without using biopsy. … The aim of my study was to analyze certain MRI data genomic parameters and correlate with the ALK mutation status.”
Dr. Wadhwa and her colleagues retrospectively analyzed data from 75 patients with ALK-positive NSCLC who underwent multiparametric MRI at the time of diagnosis. Univariate logistic regression analysis was conducted to look for associations between ALK mutation status and various clinical factors, including sex, age, smoking, histology, TNM stage, and imaging characteristics.
Out of 75 patients, 46 were ALK positive and 29 were ALK negative. Analysis showed that ALK positivity was associated with a variety of lesion morphology characteristics. ALK-positive lesions more often exhibited a fuzzy and infiltrative T2w border with hypointense peripheral solid rim, compared with ALK-negative lesions, which frequently had a well-defined T2w border with no solid rim (P less than .001). On T1w, most ALK-positive lesions were heterogeneous, whereas ALK-negative lesions were predominantly hypointense (P less than .001). Diffusion-weighted images showed that ALK-positive lesions often had peripheral restriction of the solid rim, compared with ALK-negative lesions, which were associated with central restriction (P = .001). MRI also revealed that about half of ALK-positive patients (54.3%) had meningeal involvement, compared with just 17.2% of ALK-negative patients (P = .02). ALK positivity was also associated with younger age and lack of smoking history. Considering these findings, Dr. Wadhwa concluded that “radiogenomics has a potential role in personalized management of ALK-positive NSCLC brain metastases.”
In an interview, Dr. Wadhwa provided more insight regarding the clinical need for this technology. “We have to wait for 10 days [for molecular diagnostic results], and ALK is usually aggressive disease, so if we wait for 10 days, patients can undergo rapid progression.”
Dr. Wadhwa noted that these results are similar to that of her colleague, Abhishek Mahajan, MD, who recently published results showing potential for radiogenomic detection of epidermal growth factor receptor (EGFR) status. According to Dr. Wadhwa, the two investigators plan to build on their collective findings in an effort to automate radiogenomic detection of NSCLC mutation subtypes.
“My upcoming project with my coinvestigator is to take a bigger sample,” Dr. Wadhwa said. “We will be further generalizing [this process] to all patients in a prospective study. We will also be sending this to the University of Pennsylvania for automatic brain segmentation.” Dr. Wadhwa estimated that adding automation will provide an accuracy rate of around 90%.
“We will train the computer accordingly,” Dr. Wadhwa said, “and then the computer will tell us, yes, this is ALK positive, this is EGFR positive.”
The investigators reported no external study funding and reported no conflicts of interest.
SOURCE: Wadhwa S et al. ELCC 2019, Abstract 55O.
GENEVA – Radiogenomic MRI signatures may be able to identify anaplastic lymphoma kinase (ALK)–positive brain metastases in non–small cell lung cancer (NSCLC), offering a minimally invasive option that could allow for initiation of treatment while waiting for molecular results, according to investigators.
In the future, artificial intelligence may be able to detect these imaging patterns, allowing for rapid and accurate mutation subtyping, reported lead author Shweta Wadhwa, MD, of Tata Memorial Centre in Mumbai, India, who presented the findings at the European Lung Cancer Conference.
“Radiogenomics is a concept used to associate genetic information with medical images,” Dr. Wadhwa explained at the meeting presented by the European Society for Medical Oncology. “It creates imaging biomarkers noninvasively without using biopsy. … The aim of my study was to analyze certain MRI data genomic parameters and correlate with the ALK mutation status.”
Dr. Wadhwa and her colleagues retrospectively analyzed data from 75 patients with ALK-positive NSCLC who underwent multiparametric MRI at the time of diagnosis. Univariate logistic regression analysis was conducted to look for associations between ALK mutation status and various clinical factors, including sex, age, smoking, histology, TNM stage, and imaging characteristics.
Out of 75 patients, 46 were ALK positive and 29 were ALK negative. Analysis showed that ALK positivity was associated with a variety of lesion morphology characteristics. ALK-positive lesions more often exhibited a fuzzy and infiltrative T2w border with hypointense peripheral solid rim, compared with ALK-negative lesions, which frequently had a well-defined T2w border with no solid rim (P less than .001). On T1w, most ALK-positive lesions were heterogeneous, whereas ALK-negative lesions were predominantly hypointense (P less than .001). Diffusion-weighted images showed that ALK-positive lesions often had peripheral restriction of the solid rim, compared with ALK-negative lesions, which were associated with central restriction (P = .001). MRI also revealed that about half of ALK-positive patients (54.3%) had meningeal involvement, compared with just 17.2% of ALK-negative patients (P = .02). ALK positivity was also associated with younger age and lack of smoking history. Considering these findings, Dr. Wadhwa concluded that “radiogenomics has a potential role in personalized management of ALK-positive NSCLC brain metastases.”
In an interview, Dr. Wadhwa provided more insight regarding the clinical need for this technology. “We have to wait for 10 days [for molecular diagnostic results], and ALK is usually aggressive disease, so if we wait for 10 days, patients can undergo rapid progression.”
Dr. Wadhwa noted that these results are similar to that of her colleague, Abhishek Mahajan, MD, who recently published results showing potential for radiogenomic detection of epidermal growth factor receptor (EGFR) status. According to Dr. Wadhwa, the two investigators plan to build on their collective findings in an effort to automate radiogenomic detection of NSCLC mutation subtypes.
“My upcoming project with my coinvestigator is to take a bigger sample,” Dr. Wadhwa said. “We will be further generalizing [this process] to all patients in a prospective study. We will also be sending this to the University of Pennsylvania for automatic brain segmentation.” Dr. Wadhwa estimated that adding automation will provide an accuracy rate of around 90%.
“We will train the computer accordingly,” Dr. Wadhwa said, “and then the computer will tell us, yes, this is ALK positive, this is EGFR positive.”
The investigators reported no external study funding and reported no conflicts of interest.
SOURCE: Wadhwa S et al. ELCC 2019, Abstract 55O.
REPORTING FROM ELCC 2019