PuSH - Publikationsserver des Helmholtz Zentrums München

Wagner, S. ; Reisenbüchler, D. ; West, N.P.* ; Niehues, J.M.* ; Zhu, J.* ; Foersch, S.* ; Veldhuizen, G.P.* ; Quirke, P.* ; Grabsch, H.I.* ; van den Brandt, P.A.* ; Hutchins, G.G.A.* ; Richman, S.D.* ; Yuan, T.* ; Langer, R.* ; Jenniskens, J.C.A.* ; Offermans, K.* ; Mueller, W.* ; Gray, R.* ; Gruber, S.B.* ; Greenson, J.K.* ; Rennert, G.* ; Bonner, J.D.* ; Schmolze, D.* ; Jonnagaddala, J.* ; Hawkins, N.J.* ; Ward, R.L.* ; Morton, D.* ; Seymour, M.* ; Magill, L.* ; Nowak, M.* ; Hay, J.* ; Koelzer, V.H.* ; Church, D.N.* ; Church, D.* ; Domingo, E.* ; Edwards, J.* ; Glimelius, B.* ; Gogenur, I.* ; Harkin, A.* ; Iveson, T.* ; Jaeger, E.* ; Kelly, C.* ; Kerr, R.* ; Maka, N.* ; Morgan, H.* ; Oien, K.* ; Orange, C.* ; Palles, C.* ; Roxburgh, C.* ; Sansom, O.* ; Saunders, M.* ; Tomlinson, I.* ; Matek, C. ; Geppert, C.* ; Peng, C.* ; Zhi, C.* ; Ouyang, X.* ; James, J.A.* ; Loughrey, M.B.* ; Salto-Tellez, M.* ; Brenner, H.* ; Hoffmeister, M.* ; Truhn, D.* ; Schnabel, J.A. ; Boxberg, M.* ; Peng, T. ; Kather, J.N.*

Transformer-based biomarker prediction from colorectal cancer histology: A large-scale multicentric study.

Cancer Cell 41, 1650-1661.e4 (2023)
Verlagsversion DOI PMC
Open Access Gold (Paid Option)
Creative Commons Lizenzvertrag
Deep learning (DL) can accelerate the prediction of prognostic biomarkers from routine pathology slides in colorectal cancer (CRC). However, current approaches rely on convolutional neural networks (CNNs) and have mostly been validated on small patient cohorts. Here, we develop a new transformer-based pipeline for end-to-end biomarker prediction from pathology slides by combining a pre-trained transformer encoder with a transformer network for patch aggregation. Our transformer-based approach substantially improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. After training and evaluating on a large multicenter cohort of over 13,000 patients from 16 colorectal cancer cohorts, we achieve a sensitivity of 0.99 with a negative predictive value of over 0.99 for prediction of microsatellite instability (MSI) on surgical resection specimens. We demonstrate that resection specimen-only training reaches clinical-grade performance on endoscopic biopsy tissue, solving a long-standing diagnostic problem.
Altmetric
Weitere Metriken?
[➜Einloggen]
Zusatzinfos bearbeiten [➜Einloggen]
Publikationstyp Artikel: Journalartikel
Dokumenttyp Wissenschaftlicher Artikel
Korrespondenzautor
Schlagwörter Artificial Intelligence ; Biomarker ; Colorectal Cancer ; Deep Learning ; Microsatellite Instability ; Multiple Instance Learning ; Transformer; Colon-cancer; Microsatellite Instability; Survival; Decision; Model
ISSN (print) / ISBN 1535-6108
e-ISSN 1878-3686
Zeitschrift Cancer Cell
Quellenangaben Band: 41, Heft: 9, Seiten: 1650-1661.e4 Artikelnummer: , Supplement: ,
Verlag Cell Press
Verlagsort Cambridge, Mass.
Nichtpatentliteratur Publikationen
Begutachtungsstatus Peer reviewed
Institut(e) Helmholtz Artifical Intelligence Cooperation Unit (HAICU)
Institute of Computational Biology (ICB)
Institute of AI for Health (AIH)
Institute for Machine Learning in Biomed Imaging (IML)
Förderungen NCI NIH HHS
Medical Research Council
Cancer Research UK
Department of Health