Muti, Hannah Sophie, Heij, Lara Rosaline, Keller, Gisela, Kohlruss, Meike, Langer, Rupert ORCID: 0000-0001-9491-3609, Dislich, Bastian, Cheong, Jae-Ho, Kim, Young-Woo, Kim, Hyunki, Kook, Myeong-Cherl, Cunningham, David ORCID: 0000-0001-5158-1069, Allum, William H., Langley, Ruth E., Nankivell, Matthew G., Quirke, Philip ORCID: 0000-0002-3597-5444, Hayden, Jeremy, West, Nicholas P., Irvine, Andrew J., Yoshikawa, Takaki, Oshima, Takashi, Huss, Ralf, Grosser, Bianca, Roviello, Franco, d'Ignazio, Alessia, Quaas, Alexander, Alakus, Hakan, Tan, Xiuxiang, Pearson, Alexander, Luedde, Tom, Ebert, Matthias P., Jaeger, Dirk, Trautwein, Christian, Gaisa, Nadine Therese, Grabsch, Heike, I and Kather, Jakob Nikolas (2021). Correspondence to: Dr Jakob Nikolas Kather, @jnkath For the Genomic Data Commons data portal see https://portal.gdc.cancer.gov See Online for appendix. Lancet Digit. Health, 3 (10). S. E654 - 11. AMSTERDAM: ELSEVIER. ISSN 2589-7500

Full text not available from this repository.

Abstract

Background Response to immunotherapy in gastric cancer is associated with microsatellite instability (or mismatch repair deficiency) and Epstein-Barr virus (EBV) positivity. We therefore aimed to develop and validate deep learning based classifiers to detect microsatellite instability and EBV status from routine histology slides. Methods In this retrospective, multicentre study, we collected tissue samples from ten cohorts of patients with gastric cancer from seven countries (South Korea, Switzerland, Japan, Italy, Germany, the UK and the USA). We trained a deep learning-based classifier to detect microsatellite instability and EBV positivity from digitised, haematoxylin and eosin stained resection slides without annotating tumour containing regions. The performance of the classifier was assessed by within-cohort cross-validation in all ten cohorts and by external validation, for which we split the cohorts into a five-cohort training dataset and a five-cohort test dataset. We measured the area under the receiver operating curve (AUROC) for detection of microsatellite instability and EBV status. Microsatellite instability and EBV status were determined to be detectable if the lower bound of the 95% CI for the AUROC was above 0middot5. Findings Across the ten cohorts, our analysis included 2823 patients with known microsatellite instability status and 2685 patients with known EBV status. In the within-cohort cross-validation, the deep learning-based classifier could detect microsatellite instability status in nine of ten cohorts, with AUROCs ranging from 0middot597 (95% CI 0middot522-0middot737) to 0middot836 (0middot795-0middot880) and EBV status in five of eight cohorts, with AUROCs ranging from 0middot819 (0middot752-0middot841) to 0middot897 (0middot513-0middot966). Training a classifier on the pooled training dataset and testing it on the five remaining cohorts resulted in high classification performance with AUROCs ranging from 0middot723 (95% CI 0middot676-0middot794) to 0middot863 (0middot747-0middot969) for detection of microsatellite instability and from 0middot672 (0middot403-0middot989) to 0middot859 (0middot823-0middot919) for detection of EBV status. Interpretation Classifiers became increasingly robust when trained on pooled cohorts. After prospective validation, this deep learning-based tissue classification system could be used as an inexpensive predictive biomarker for immunotherapy in gastric cancer. Funding German Cancer Aid and German Federal Ministry of Health. Copyright (c) 2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.

Item Type: Journal Article
Creators:
CreatorsEmailORCIDORCID Put Code
Muti, Hannah SophieUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Heij, Lara RosalineUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Keller, GiselaUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Kohlruss, MeikeUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Langer, RupertUNSPECIFIEDorcid.org/0000-0001-9491-3609UNSPECIFIED
Dislich, BastianUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Cheong, Jae-HoUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Kim, Young-WooUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Kim, HyunkiUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Kook, Myeong-CherlUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Cunningham, DavidUNSPECIFIEDorcid.org/0000-0001-5158-1069UNSPECIFIED
Allum, William H.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Langley, Ruth E.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Nankivell, Matthew G.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Quirke, PhilipUNSPECIFIEDorcid.org/0000-0002-3597-5444UNSPECIFIED
Hayden, JeremyUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
West, Nicholas P.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Irvine, Andrew J.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Yoshikawa, TakakiUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Oshima, TakashiUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Huss, RalfUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Grosser, BiancaUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Roviello, FrancoUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
d'Ignazio, AlessiaUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Quaas, AlexanderUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Alakus, HakanUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Tan, XiuxiangUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Pearson, AlexanderUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Luedde, TomUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Ebert, Matthias P.UNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Jaeger, DirkUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Trautwein, ChristianUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Gaisa, Nadine ThereseUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Grabsch, Heike, IUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
Kather, Jakob NikolasUNSPECIFIEDUNSPECIFIEDUNSPECIFIED
URN: urn:nbn:de:hbz:38-605957
DOI: 10.1016/S2589-7500(21)00133-3
Journal or Publication Title: Lancet Digit. Health
Volume: 3
Number: 10
Page Range: S. E654 - 11
Date: 2021
Publisher: ELSEVIER
Place of Publication: AMSTERDAM
ISSN: 2589-7500
Language: English
Faculty: Unspecified
Divisions: Unspecified
Subjects: no entry
Uncontrolled Keywords:
KeywordsLanguage
MICROSATELLITE INSTABILITY; GASTRIC-CANCERMultiple languages
Medical Informatics; Medicine, General & InternalMultiple languages
URI: http://kups.ub.uni-koeln.de/id/eprint/60595

Downloads

Downloads per month over past year

Altmetric

Export

Actions (login required)

View Item View Item