Researchers Develop Machine Learning Tool to Predict Mesothelioma Risk

“How long do I have?” is among the most important questions asked by those diagnosed with malignant pleural mesothelioma. Though the asbestos-related disease is always considered fatal, there are numerous variables that affect each individual’s prognosis. In an attempt to improve clinicians’ ability to answer their patients’ questions and offer appropriate treatment, a group of researchers from Memorial Sloan Kettering developed a next-generation sequencing-derived machine-learning risk prediction model that has proven to be remarkably accurate.

Current Risk Stratification Tools for Mesothelioma are Inadequate

Memorial Sloan Kettering is one of the world’s premier cancer hospitals, and researchers there treat many patients diagnosed with malignant mesothelioma. Out of concern for the inadequacies of current risk stratification tools, scientists there set out to develop a machine-learning tool that they called OncoCast-MPM to create a more accurate model for patient prognosis.

To accomplish their goals, they conducted a retrospective study that included clinical, pathological, and routine next-generation sequencing data from 194 patients treated at their facility over a five-year period.  Rather than relying on subjective data such as patient-reported pain or lab tests that can have rapidly changing results, they relied entirely on clinical and pathological features as well as molecular profiling, with 274 variables input in the OncoCast-MPM model. They used those features to provide individualized risk scores for each patient within the cohort and then compared their scores to actual survival or validation of their results.

Mesothelioma Patients Stratified into Risk Groups with Significant Survival Differences

The researchers found that the OncoCast-MPM model was able to accurately separate patients for survival as early as one year into diagnosis. Its use requires no special testing to generate risk scores, making it available for use for any patient in the world, for free.  In assessing the impact of their model, the researchers wrote, “In the absence of highly effective therapies with valid biomarkers in a very heterogeneous disease, efforts to provide accurate and personalized prognoses become essential to provide individualized patient care, prioritize high-risk patients for clinical trials, evaluate real-world data, and generate accurately matched historical control groups. OncoCast-MPM might fill this unmet need.”

If you or someone you love has been diagnosed with malignant mesothelioma, finding the answers you need can be hard. For help and access to essential resources, contact the Patient Advocates at Mesothelioma.net today at 1-800-692-8608.

Written by Terri Oppenheimer

Terri Heimann Oppenheimer is the head writer of our Mesothelioma.net news blog. She graduated from the College of William and Mary with a degree in English. Terri believes that knowledge is power and she is committed to sharing news about the impact of mesothelioma, the latest research and medical breakthroughs, and victims’ stories.

Learn more about and contact Terri