Artificial intelligence at the intersection of pathology and radiology in prostate cancer

Abstract

Pathologic grading plays a key role in prostate cancer risk stratification and treatment selection, traditionally assessed from systemic core needle biopsies sampled throughout the prostate gland. Multiparametric magnetic resonance imaging (mpMRI) has become a well-established clinical tool for detecting and localizing prostate cancer. However, both pathologic and radiologic assessment suffer from poor reproducibility among readers. Artificial intelligence (AI) methods show promise in aiding the detection and assessment of imaging-based tasks, dependent on the curation of high-quality training sets. This review provides an overview of recent advances in AI applied to mpMRI and digital pathology in prostate cancer which enable advanced characterization of disease through combined radiology-pathology assessment.

Figure

Example mpMRI (T2-weighted and ADC images) with corresponding whole mount hematoxylin-eosin (H&E) section demonstrating two pathologically-defined tumors. Overall assessment in this patient resulted in Gleason grade 3+4 score assignment, resulting in weak labels derived from total tumor extent. However, detailed pathologic assessment demonstrates majority of Gleason pattern 4 is located in the anterior portion of the tumor, with predominately Gleason pattern 3 throughout remainder of the tumor field. Opportunities for improved spatial learning include density mapping of dominant pathologic grading and exclusion of non-cancerous structures within tumor field.