Generalizable biomarker prediction from cancer pathology slides with self-supervised deep learning: A retrospective multi-centric study

Publikation: Beitrag in FachzeitschriftForschungsartikelBeigetragenBegutachtung

Beitragende

  • Jan Moritz Niehues - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen (Autor:in)
  • Philip Quirke - , University of Leeds (Autor:in)
  • Nicholas P. West - , University of Leeds (Autor:in)
  • Heike I. Grabsch - , University of Leeds (Autor:in)
  • Marko van Treeck - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen (Autor:in)
  • Yoni Schirris - , Klinik und Poliklinik für Kinder- und Jugendmedizin (Autor:in)
  • Gregory P. Veldhuizen - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen (Autor:in)
  • Gordon G. A. Hutchins - , University of Leeds (Autor:in)
  • Susan D. Richman - , University of Leeds (Autor:in)
  • Sebastian Foersch - , Universitätsklinikum Mainz (Autor:in)
  • Titus J. Brinker - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • Junya Fukuoka - , Nagasaki University Graduate School of Biomedical Sciences (Autor:in)
  • Andrey Bychkov - , Kameda Medical Center (Autor:in)
  • Wataru Uegami - , Kameda Medical Center (Autor:in)
  • Daniel Truhn - , Universitätsklinikum Aachen (Autor:in)
  • Hermann Brenner - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • Alexander Brobeil - , Universitätsklinikum Heidelberg (Autor:in)
  • Michael Hoffmeister - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • Jakob Nikolas Kather - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Medizinische Klinik und Poliklinik I, Universitätsklinikum Aachen, University of Leeds, Nationales Zentrum für Tumorerkrankungen (NCT) Heidelberg (Autor:in)

Abstract

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL can also predict other biomarkers with high performance and whether DL predictions generalize to external patient populations. Here, we acquire CRC tissue samples from two large multi-centric studies. We systematically compare six different state-of-the-art DL architectures to predict biomarkers from pathology slides, including MSI and mutations in BRAF, KRAS, NRAS, and PIK3CA. Using a large external validation cohort to provide a realistic evaluation setting, we show that models using self-supervised, attention-based multiple-instance learning consistently outperform previous approaches while offering explainable visualizations of the indicative regions and morphologies. While the prediction of MSI and BRAF mutations reaches a clinical-grade performance, mutation prediction of PIK3CA, KRAS, and NRAS was clinically insufficient.

Details

OriginalspracheEnglisch
Aufsatznummer100980
FachzeitschriftCell Reports : Medicine
Jahrgang4
Ausgabenummer4
PublikationsstatusVeröffentlicht - 18 Apr. 2023
Peer-Review-StatusJa

Externe IDs

PubMedCentral PMC10140458
Scopus 85152732214

Schlagworte

Ziele für nachhaltige Entwicklung

Schlagwörter

  • Humans, Retrospective Studies, Proto-Oncogene Proteins B-raf/genetics, Deep Learning, Proto-Oncogene Proteins p21(ras)/genetics, Colorectal Neoplasms/genetics, Biomarkers, Microsatellite Instability, Class I Phosphatidylinositol 3-Kinases/genetics

Bibliotheksschlagworte