Comprehensive Summary
In this study, Chaa et al takes a novel approach to predict prostate cancer (PCa) metastasis through histopathology-based artificial intelligence (AI) algorithms. Risk scores were generated from histopathologic data utilizing whole slide images (WSIs) or tissue microarrays (TMAs) from nationwide cohorts with PCa. The AI algorithm displayed norinferior performance when compared to currently utilized genomic classifiers when predicting risk of metastasis. While this tool is not intended to replace traditional staging and grading as performed by a pathologist, it serves as proof that histopathology-based algorithms my be utilized to predict metastasis risk.
Outcomes and Implications
As the utilization of AI algorithms continues to become more widespread, this novel investigation offers a path for expanding the reach of pathology studies to those without access to on-site pathologists. This tool may provide better risk stratification and support clinical decisions for the benefit of the patient. Such a tool may furthermore be used to identify early aspects of the disease course that may pose risk of metastasis, leading to early intervention and better patient outcomes.