pipette
ESEspañol

Comparative analysis of end-to-end learning and foundation models for clinical-grade prostate pathology

Nita Mulliqi, Anders Blilie, Xiaoyi Ji, Kelvin Szolnoky, Henrik Olsson, Sol Erika Boman, Matteo Titus, Geraldine Martinez Gonzalez, Julia Anna Mielcarz, Masi Valkonen, Einar G. Gudlaugsson, Svein Reidar Kjosavik, J. Asenjo, Marcello Gambacorta, Paolo Libretti, Marcin Braun, Radzisław Kordek, Roman Łowicki, Kristina Hotakainen, Päivi O. Väre, Bodil Ginnerup Pedersen, Karina Dalsgaard Sørensen, Benedicte Parm Ulhøi, Pekka Ruusuvuori, Brett Delahunt, Hemamali Samaratunga, Toyonori Tsuzuki, Emiel A. M. Janssen, Lars Egevad, Martin Eklund, Kimmo Kartasalo

Peer-reviewed journalReal-world use

In the authors' words

Abstract Foundation models (FMs) using self-supervised pre-training are widely advocated as a universal solution for pathology, yet their advantages over task-specific (TS) models remain unclear. Here we show, using over 100,000 prostate biopsies from 6268 patients across 15 sites, that while FMs demonstrate utility in data-scarce scenarios, their performance converges with—and is in some cases surpassed by—TS models with sufficient labeled training data. Concordance with pathologists’ grading reaches quadratic Cohen’s kappas of 0.91 (95% CI 0.90-0.92; all models) on internal and 0.87 (0.86-0.88; TS), 0.86 (0.84-0.87; UNI FM) and 0.85 (0.84-0.86; Virchow2 FM) on external cohorts at full training. Task-specific training markedly reduces misgrading, misdiagnosis of challenging morphologies, and technical variability. Notably, FMs consume up to 35 times more energy. We advocate for considering both FMs and end-to-end learning for robust, resource-efficient, and clinically applicable AI pathology.

Main resultThe abstract does not state a limitation.

Appeared: Wednesday, October 28. Nature Communications. Peer-reviewed journal.

DOI: 10.1038/s41467-026-78996-7