article Open AccessTop 1% cited

Predicting survival from colorectal cancer histology slides using deep learning: A retrospective multicenter study

PLoS Medicine · 2019 · Vol. 16(1) · pp. e1002730–e1002730

Jakob Nikolas Kather✉(German Cancer Research Center)Johannes Krisam(Heidelberg University)Pornpimol Charoentong(German Cancer Research Center)Tom Luedde(RWTH Aachen University)Esther Herpel(Heidelberg University)Cleo‐Aron Weis(Heidelberg University)Timo Gaiser(Heidelberg University)Alexander Marx(Heidelberg University)Nektarios A. Valous(German Cancer Research Center)Dyke Ferber(German Cancer Research Center)Lina Jansen(German Cancer Research Center)Constantino Carlos Reyes‐Aldasoro(City, University of London)Inka Zörnig(German Cancer Research Center)Dirk Jäger(German Cancer Research Center)Hermann Brenner(German Cancer Research Center)Jenny Chang‐Claude(German Cancer Research Center)Michael Hoffmeister(German Cancer Research Center)Niels Halama(German Cancer Research Center)

View via DOI Full-text PDF

Abstract

In our retrospective study, we show that a CNN can assess the human tumor microenvironment and predict prognosis directly from histopathological images.

Colorectal Cancer Screening and Detection Radiomics and Machine Learning in Medical Imaging AI in cancer detectionMedicineColorectal cancerRetrospective cohort studyCancerHistologyOncologyInternal medicinePathology

MeSH terms

Deep LearningColonColoring AgentsEosine Yellowish-(YS)FemaleHematoxylinHumansImage Interpretation, Computer-AssistedMalePrognosisRectumRetrospective StudiesColorectal Neoplasms