Accuracy of Deep Learning-Aided Detection of Microsatellite Instability in Colorectal Cancer: A Systematic Review and Meta-Analysis
Clinical Colorectal Cancer · 2026
DOI: 10.1016/j.clcc.2025.09.004 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift Clinical Colorectal Cancer, 2026. Verfasst von Zihong Wang und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medical Engineering, The First Affiliated Hospital (Southwest Hospital) of Army Medical University, Chongqing 400038, China. Hauptthemen laut MEDLINE: Microsatellite Instability, Colorectal Neoplasms, Deep Learning. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 56 Literaturangaben.
Abstract (Originalsprache)
PURPOSE: We synthesized the evidence on the accuracy of Deep learning in detecting colorectal cancer microsatellite instability to contribute to the development and updating of intelligent detection tools. METHODS: A systematic search was conducted in PubMed, Web of Science, Embase, and Cochrane Library from their inception until December 1, 2024. The included studies were assessed for the risk of bias using the Quality Assessment of Diagnostic Accuracy Studies-2. Only the diagnostic fourfold tables in the validation set underwent meta-analyses. RESULTS: Thirty original studies were finally included. The pooled specificity, sensitivity and area under the summary receiver operating characteristic curve of pathology slice-based deep learning were 0.86 (95% CI, 0.80-0.90), 0.90 (95% CI, 0.85-0.93) and 0.94 (95% CI 0.30-1.00), respectively. For the external validation of pathology slice-based deep learning, the pooled specificity, sensitivity and summary receiver operating characteristic curve were 0.84 (95% CI, 0.76-0.89), 0.88 (95% CI, 0.83-0.92) and 0.93 (95% CI, 1.00-0.00), respectively. The number of included studies on MRI-, colonoscopy-, and Raman spectrum-based deep learning was extremely small. CONCLUSIONS: Pathology slice-based deep learning is accurate for colorectal cancer microsatellite detection and has potential for future software development, but imaging image-based deep learning needs more validation. TRIAL REGISTRATION: This study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020). The protocol was prospectively registered with PROSPERO (CRD42025629990).
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Medical Engineering, The First Affiliated Hospital (Southwest Hospital) of Army Medical University, Chongqing 400038, China.
- Department of Medical Engineering, The First Affiliated Hospital (Southwest Hospital) of Army Medical University, Chongqing 400038, China. Electronic address: hanjc168@tmmu.edu.cn.
Zitieren
APA
Wang, Z., Chang, F., Wu, H., Yuan, X., Chen, Y., & Han, J. (2026). Accuracy of Deep Learning-Aided Detection of Microsatellite Instability in Colorectal Cancer: A Systematic Review and Meta-Analysis. Clinical Colorectal Cancer. https://doi.org/10.1016/j.clcc.2025.09.004
Vancouver
Wang Z, Chang F, Wu H, Yuan X, Chen Y, Han J. Accuracy of Deep Learning-Aided Detection of Microsatellite Instability in Colorectal Cancer: A Systematic Review and Meta-Analysis. Clinical Colorectal Cancer. 2026. doi:10.1016/j.clcc.2025.09.004
BibTeX
@article{Wang2026,
author = {Wang, Zihong and Chang, Feiba and Wu, Hongbo and Yuan, Xianju and Chen, Yong and Han, Jinchuan},
title = {Accuracy of Deep Learning-Aided Detection of Microsatellite Instability in Colorectal Cancer: A Systematic Review and Meta-Analysis},
journal = {Clinical Colorectal Cancer},
year = {2026},
doi = {10.1016/j.clcc.2025.09.004}
}
EUVIMED