Current risk stratification methods for localized prostate cancer (PCa) are reliant on clinical and pathological variables that do not easily account for location of cancer spread. Prostate MRI is a helpful tool to identify anatomic extra-prostatic cancer spread (EPE) and seminal vesicle invasion (SVI) but is subject to radiologist expertise and inter-observer variation. We report deep learning models which provide objective end-to-end evaluation of EPE and SVI on prostate MRI. Both EPE and SVI models demonstrate high discriminatory ability on three held-out test sets spanning different clinical settings, equipment manufacturers, and MRI magnet strengths. Interpretability studies suggest both EPE and SVI models identify clinically-relevant anatomic regions. Lastly, we show that classification of EPE and SVI by our models is independently associated with increased risk of biochemical recurrence (BCR) following localized treatment. Furthermore, we demonstrate that our models can be easily integrated to well-established risk stratification methods (NCCN and UCSF-CAPRA) for improved ability to identify high risk PCa phenotypes.

The authors have declared no competing interest.

This study did not receive any funding

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

This study was approved by the institutional review board (IRB 2000027592) at Yale University.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors