TY - JOUR
T1 - Clinical-Grade Detection of Microsatellite Instability in Colorectal Tumors by Deep Learning
AU - Echle, Amelie
AU - Grabsch, Heike Irmgard
AU - Quirke, Philip
AU - van den Brandt, Piet A.
AU - West, Nicholas P.
AU - Hutchins, Gordon G. A.
AU - Heij, Lara R.
AU - Tan, Xiuxiang
AU - Richman, Susan D.
AU - Krause, Jeremias
AU - Alwers, Elizabeth
AU - Jenniskens, Josien
AU - Offermans, Kelly
AU - Gray, Richard
AU - Brenner, Hermann
AU - Chang-Claude, Jenny
AU - Trautwein, Christian
AU - Pearson, Alexander T.
AU - Boor, Peter
AU - Luedde, Tom
AU - Gaisa, Nadine Therese
AU - Hoffmeister, Michael
AU - Kather, Jakob Nikolas
N1 - Funding Information:
The authors are grateful to all investigators and contributing pathologists from the TCGA study (more information on http://portal.gdc.cancer.gov ), the Rainbow-TMA consortium in the Netherlands (listed in Supplementary Table 6 ), the DACHS consortium in Germany, the QUASAR consortium, and the YCR-BCIP consortium in the United Kingdom. Collection and testing of the YCR-BCIP cases was funded by Yorkshire Cancer Research L386 and L394 as part of previous studies. Philip Quirke is an NIHR Senior Investigator. We thank the Yorkshire and Humber Histopathologists who facilitated the collection of the YCR BCIP cases.
Funding Information:
Funding This study was primarily funded by the authors’ academic institutions. These authors are supported by additional grants: Peter Boor: German Research Foundation (SFB/TRR57, SFB/TRR219, BO3755/3-1, and BO3755/6-1), the German Federal Ministry of Education and Research (BMBF: STOP-FSGS-01GM1901A), and the German Federal Ministry of Economic Affairs and Energy (BMWi: EMPAIA project). Alexander T. Pearson: National Institutes of Health / National Institute of Dental and Craniofacial Research (K08-DE026500), Institutional Research Grant (IRG-16-222-56) from the American Cancer Society , Cancer Research Foundation Research Grant, and the University of Chicago Medicine Comprehensive Cancer Center Support Grant (P30-CA14599). Tom Luedde: Horizon 2020 through the European Research Council Consolidator Grant PhaseControl (771083), a Mildred Scheel–Endowed Professorship from the German Cancer Aid (Deutsche Krebshilfe), the German Research Foundation (SFB CRC1382/P01, SFB-TRR57/P06, LU 1360/3-1), the Ernst-Jung-Foundation Hamburg , and the Interdisciplinary Center of Clinical Research) at RWTH Aachen. Jakob Nikolas Kather: RWTH University Aachen (START 2018-691906), Max-Eder-Programme of the German Cancer Aid (Deutsche Krebshilfe, 70113864).
Funding Information:
Funding This study was primarily funded by the authors? academic institutions. These authors are supported by additional grants: Peter Boor: German Research Foundation (SFB/TRR57, SFB/TRR219, BO3755/3-1, and BO3755/6-1), the German Federal Ministry of Education and Research (BMBF: STOP-FSGS-01GM1901A), and the German Federal Ministry of Economic Affairs and Energy (BMWi: EMPAIA project). Alexander T. Pearson: National Institutes of Health/National Institute of Dental and Craniofacial Research (K08-DE026500), Institutional Research Grant (IRG-16-222-56) from the American Cancer Society, Cancer Research Foundation Research Grant, and the University of Chicago Medicine Comprehensive Cancer Center Support Grant (P30-CA14599). Tom Luedde: Horizon 2020 through the European Research Council Consolidator Grant PhaseControl (771083), a Mildred Scheel?Endowed Professorship from the German Cancer Aid (Deutsche Krebshilfe), the German Research Foundation (SFB CRC1382/P01, SFB-TRR57/P06, LU 1360/3-1), the Ernst-Jung-Foundation Hamburg, and the Interdisciplinary Center of Clinical Research) at RWTH Aachen. Jakob Nikolas Kather: RWTH University Aachen (START 2018-691906), Max-Eder-Programme of the German Cancer Aid (Deutsche Krebshilfe, 70113864).
Publisher Copyright:
© 2020 The Authors
PY - 2020/10
Y1 - 2020/10
N2 - BACKGROUND & AIMS: Microsatellite instability (MSI) and mismatch-repair deficiency (dMMR) in colorectal tumors are used to select treatment for patients. Deep learning can detect MSI and dMMR in tumor samples on routine histology slides faster and less expensively than molecular assays. However, clinical application of this technology requires high performance and multisite validation, which have not yet been performed. METHODS: We collected H&E-stained slides and findings from molecular analyses for MSI and dMMR from 8836 colorectal tumors (of all stages) included in the MSI-DETECT consortium study, from Germany, the Netherlands, the United Kingdom, and the United States. Specimens with dMMR were identified by immunohistochemistry analyses of tissue microarrays for loss of MLH1, MSH2, MSH6, and/or PMS2. Specimens with MSI were identified by genetic analyses. We trained a deep-learning detector to identify samples with MSI from these slides; performance was assessed by cross-validation (N = 6406 specimens) and validated in an external cohort (n = 771 specimens). Prespecified endpoints were area under the receiver operating characteristic (AUROC) curve and area under the precision-recall curve (AUPRC). RESULTS: The deep-learning detector identified specimens with dMMR or MSI with a mean AUROC curve of 0.92 (lower bound, 0.91; upper bound, 0.93) and an AUPRC of 0.63 (range, 0.59-0.65), or 67% specificity and 95% sensitivity, in the cross-validation development cohort. In the validation cohort, the classifier identified samples with dMMR with an AUROC of 0.95 (range, 0.92-0.96) without image preprocessing and an AUROC of 0.96 (range, 0.93-0.98) after color normalization. CONCLUSIONS: We developed a deep-learning system that detects colorectal cancer specimens with dMMR or MSI using H&E-stained slides; it detected tissues with dMMR with an AUROC of 0.96 in a large, international validation cohort. This system might be used for high-throughput, low-cost evaluation of colorectal tissue specimens.
AB - BACKGROUND & AIMS: Microsatellite instability (MSI) and mismatch-repair deficiency (dMMR) in colorectal tumors are used to select treatment for patients. Deep learning can detect MSI and dMMR in tumor samples on routine histology slides faster and less expensively than molecular assays. However, clinical application of this technology requires high performance and multisite validation, which have not yet been performed. METHODS: We collected H&E-stained slides and findings from molecular analyses for MSI and dMMR from 8836 colorectal tumors (of all stages) included in the MSI-DETECT consortium study, from Germany, the Netherlands, the United Kingdom, and the United States. Specimens with dMMR were identified by immunohistochemistry analyses of tissue microarrays for loss of MLH1, MSH2, MSH6, and/or PMS2. Specimens with MSI were identified by genetic analyses. We trained a deep-learning detector to identify samples with MSI from these slides; performance was assessed by cross-validation (N = 6406 specimens) and validated in an external cohort (n = 771 specimens). Prespecified endpoints were area under the receiver operating characteristic (AUROC) curve and area under the precision-recall curve (AUPRC). RESULTS: The deep-learning detector identified specimens with dMMR or MSI with a mean AUROC curve of 0.92 (lower bound, 0.91; upper bound, 0.93) and an AUPRC of 0.63 (range, 0.59-0.65), or 67% specificity and 95% sensitivity, in the cross-validation development cohort. In the validation cohort, the classifier identified samples with dMMR with an AUROC of 0.95 (range, 0.92-0.96) without image preprocessing and an AUROC of 0.96 (range, 0.93-0.98) after color normalization. CONCLUSIONS: We developed a deep-learning system that detects colorectal cancer specimens with dMMR or MSI using H&E-stained slides; it detected tissues with dMMR with an AUROC of 0.96 in a large, international validation cohort. This system might be used for high-throughput, low-cost evaluation of colorectal tissue specimens.
KW - biomarker
KW - cancer immunotherapy
KW - Lynch syndrome
KW - mutation
KW - MISMATCH-REPAIR GENES
KW - MOLECULAR CHARACTERIZATION
KW - CANCER
KW - COLON
U2 - 10.1053/j.gastro.2020.06.021
DO - 10.1053/j.gastro.2020.06.021
M3 - Article
C2 - 32562722
SN - 0016-5085
VL - 159
SP - 1406-1416.e11
JO - Gastroenterology
JF - Gastroenterology
IS - 4
ER -