# Clinically informed intermediate reasoning enables generalizable prostate cancer prognostication through machine learning in limited settings

## TL;DR

A machine learning approach improves prostate cancer prognosis by using biopsy images and clinical reasoning, working well even with limited data.

## Contribution

A data-efficient pipeline with clinically informed reasoning for prostate cancer prognosis across diverse clinical settings.

## Key findings

- The pipeline resolved dual-domain shifts across specimen types and institutions.
- The approach achieved consistent external validation and outperformed Gleason grading.
- It provides an interpretable framework for clinical decision-making in limited settings.

## Abstract

Machine learning has shown promise in medical image classification. However, its generalizability remains challenging. Here, we show that data-efficient pre-surgical prognostication of prostate cancer from biopsy specimens is enabled by versatile feature extraction from whole-mount histopathology and a clinically informed intermediate reasoning step. With data from multiple institutions, our pipeline resolved dual-domain shifts across specimen types and institutions and achieved consistent external validation, reinforced by comprehensive analyses of generalizability. This highlights the robustness of our prognostic approach when compared to the Gleason grading system. We establish an equitable, interpretable, and clinically applicable framework, supporting actionable decisions for prognosis and treatment planning, even in limited real-world clinical environments.