Diagnostic accuracy of deep learning for microsatellite instability prediction in gastric cancer: A systematic review and meta-analysis
- 1Qinghai University
Abstract
Deep learning (DL) applied to haematoxylin and eosin (H&E)-stained pathology images can predict microsatellite instability (MSI) in gastric cancer (GC). We evaluated its diagnostic accuracy and discriminative performance. PubMed, Embase, Cochrane Library and Web of Science were searched for studies applying DL to H&E images for MSI prediction in GC. The protocol was prospectively registered in PROSPERO (CRD420261503140). Risk of bias was assessed with QUADAS-2. Pooled sensitivity and specificity were estimated with a bivariate random-effects model, with subgroup analyses by validation strategy and data source; slide-level areas under the curve (AUCs) were pooled after logit transformation. Prespecified scenario analyses assessed robustness. Ten studies were included; seven independent datasets from six studies entered the primary synthesis. Pooled sensitivity, specificity and summary receiver operating characteristic AUC were 0.78 (95% CI 0.63-0.88), 0.85 (95% CI 0.77-0.90) and 0.89 (95% CI 0.86-0.91); nine studies contributed to the secondary AUC synthesis (pooled AUC 0.85, 95% CI 0.80-0.89). Hospital-derived cohorts showed numerically higher sensitivity (0.86) and diagnostic odds ratio (39.8) than TCGA-derived cohorts (0.71 and 11.5), but subgroup differences were limited by the small number of studies and substantial heterogeneity. DL models show good diagnostic accuracy for MSI prediction in GC, but clinical utility still requires confirmation in standardised multicentre, independent external and prospective studies.
Related
- MLH1 AND MSH2 PROTEIN EXPRESSION LOSS AND MICROSATELLITE INSTABILITY IN COLORECTAL ADENOMAS OF PATIENTS WITH LYNCH SYNDROME: A SYSTEMATIC REVIEW AND META-ANALYSIS Perda de expressão das proteínas MLH1 e MSH2 e instabilidade de microssatélites em adenomas colorretais de pacientes com Síndrome de Lynch: uma Revisão Sistemática e Meta-Análise
- Diagnostic accuracy of the appendicitis inflammatory response (AIR) score for acute appendicitis during pregnancy: A systematic review and diagnostic meta-analysis
- Diagnostic Accuracy and Renal Safety of Carbon Dioxide Angiography in Lower Limb Angiography: A Systematic Review and Meta-Analysis
- Potential utility of more subtle microsatellite instability for predicting response to nivolumab in gastric cancer patients
- Transthoracic echocardiography vs. cardiac magnetic resonance imaging for the diagnosis of left ventricular thrombus: an updated systematic review and diagnostic accuracy meta-analysis
BibTeX
@misc{du2026diagnostic,
title = {{Diagnostic accuracy of deep learning for microsatellite instability prediction in gastric cancer: A systematic review and meta-analysis}},
author = {Y. Du and H. Liao and J. Li and Y. He and J. Di},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.12.26365351},
url = {https://www.medrxiv.org/content/10.64898/2026.10.12.26365351v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Diagnostic accuracy of deep learning for microsatellite instability prediction in gastric cancer: A systematic review and meta-analysis
AU - Y. Du
AU - H. Liao
AU - J. Li
AU - Y. He
AU - J. Di
PY - 2026
DA - 2026/10/12
PB - medRxiv
DO - 10.64898/2026.10.12.26365351
UR - https://www.medrxiv.org/content/10.64898/2026.10.12.26365351v1
AB - Deep learning (DL) applied to haematoxylin and eosin (H&E)-stained pathology images can predict microsatellite instability (MSI) in gastric cancer (GC). We evaluated its diagnostic accuracy and discriminative performance. PubMed, Embase, Cochrane Library and Web of Science were searched for studies applying DL to H&E images for MSI prediction in GC. The protocol was prospectively registered in PROSPERO (CRD420261503140). Risk of bias was assessed with QUADAS-2. Pooled sensitivity and specificity were estimated with a bivariate random-effects model, with subgroup analyses by validation strategy and data source; slide-level areas under the curve (AUCs) were pooled after logit transformation. Prespecified scenario analyses assessed robustness. Ten studies were included; seven independent datasets from six studies entered the primary synthesis. Pooled sensitivity, specificity and summary receiver operating characteristic AUC were 0.78 (95% CI 0.63-0.88), 0.85 (95% CI 0.77-0.90) and 0.89 (95% CI 0.86-0.91); nine studies contributed to the secondary AUC synthesis (pooled AUC 0.85, 95% CI 0.80-0.89). Hospital-derived cohorts showed numerically higher sensitivity (0.86) and diagnostic odds ratio (39.8) than TCGA-derived cohorts (0.71 and 11.5), but subgroup differences were limited by the small number of studies and substantial heterogeneity. DL models show good diagnostic accuracy for MSI prediction in GC, but clinical utility still requires confirmation in standardised multicentre, independent external and prospective studies.
ER -