This study investigates the use of deep learning to predict biochemical recurrence of prostate cancer after radical prostatectomy. A deep learning system was trained on H&E-stained microarray cores to predict recurrence, utilizing a nested case-control study of 685 patients and validated on an independent cohort of 204 patients. The biomarker shows strong correlation with biochemical recurrence in both cohorts, and concept-based explanations reveal tissue patterns interpretable by pathologists, suggesting predictive power beyond the ISUP grading system.
Publisher
Communications Medicine
Published On
Jun 08, 2022
Authors
Hans Pinckaers, Jolique van Ipenburg, Jonathan Melamed, Angelo De Marzo, Elizabeth A. Platz, Bram van Ginneken, Jeroen van der Laak, Geert Litjens
Tags
deep learning
prostate cancer
biochemical recurrence
predictive modeling
pathology
Related Publications
Explore these studies to deepen your understanding of the subject.
