A new framework combining machine learning and radiomics will help distinguish between low- and high-risk prostate cancer, according to new research published in Scientific Reports.
"By rigorously and systematically combining machine learning with radiomics, our goal is to provide radiologists and clinical personnel with a sound prediction tool that can eventually translate to more effective and personalized patient care,” said lead author Gaurav Pandey, PhD, of the Icahn School of Medicine at Mount Sinai, in a prepared statement.
AI continues to be an instrumental tool in the diagnosis of many cancers, including cervical, uterine and lung.
Pandey and colleagues developed the method to allow for radiologists to accurately identify treatment options for prostate cancer patients. This may decrease the chance for unnecessary clinical intervention.
The current method used to assess prostate cancer risk—Prostate Imaging Reporting and Data System, version 2 (PI-RADS v2)—is subjective, as it uses a five-point scoring system that grades lesions found on MRI. The PI-RADS v2 typically leads to different scores and interpretations among radiologists, potentially generating unnecessary interventions.
Pandey et al. combined radiomics, which uses algorithms to extract large amounts of quantitative characteristics from medical images, with machine learning for their framework. Radiomics has already shown effectiveness in improving and offering complimentary information to radiologists regarding benign and malignant breast tumors.
Their method used 110 radiomic features that were interpreted using a Quadratic kernel-based Support Vector Machine (QSVM) in a final cohort of 54 individuals.
Importantly, compared to the PI-RADS v2 method, the machine learning method performed with “reasonably high precision” or predictive value and high recall or sensitivity (0.86 and 0.72 for the high and lower-risk classes, respectively) for grading prostate cancer patients in an independent validation set. Though, the PI-RADS v2 classification had a higher overall area under the curve (AUC) than the machine-learning classifier (0.73 versus 0.71).
"The pathway to predicting prostate cancer progression with high accuracy is ever improving, and we believe our objective framework is a much-needed advancement,” Pandey said in the same statement.
As a senior news writer for TriMed, Subrata covers cardiology, clinical innovation and healthcare business. She has a master’s degree in communication management and 12 years of experience in journalism and public relations.