Researchers have discovered a novel machine learning framework that differentiates between low and high risk for prostate cancer more precisely than ever before.
The framework is specifically designed to help physicians and radiologists more accurately identify treatment options for patients with prostate cancer, thereby reducing the likelihood of unnecessary clinical interventions.
The research team of the Icahn School of Medicine at Mount Sinai and the University of Southern California's Keck School of Medicine (USC), who made the discovery in their report that prostate cancer was a major cause of cancer deaths, was second only to lung cancer ,
While recent advances in prostate cancer research have saved many lives, objective predictive tools have remained unfulfilled.
Currently, the standard methods for assessing prostate cancer risk are multiparametric magnetic resonance imaging (mpMRI), which detects prostate lesions and the Prostate Imaging Reporting and Data System, Version 2 (PI-RADS v2), a five-point scoring system classifies the lesions found in the mpMRI.
Together, these tools are designed to predict the likelihood of clinically significant prostate cancer. However, the rating of PI-RADS v2 is subjective and does not clearly distinguish between mean and malignant levels of cancer (scores 3, 4, and 5), often leading to different interpretations in clinicians.
Assistant Professor of Genetics and Genomics at The School, Gaurav Pandey, said by consistently and systematically combining machine learning with Radiomics, their goal is to provide radiologists and clinical staff with a solid predictive tool that ultimately becomes more effective The publication Bino Varghese also said that the way to predict prostate cancer progression with high accuracy is getting better, and they believe their objective framework is a much needed step forward.