Abstract:Simple summary Detecting deficient mismatch repair (dMMR) in patients with colorectal cancer is essential for clinical decision-making, including evaluation of prognosis, guidance of adjuvant chemotherapy and immunotherapy, and primary screening for Lynch syndrome. However, outside of tertiary care centers, existing detection methods are not widely disseminated and highly depend on the experienced pathologist. Therefore, it is of great clinical significance to develop a broadly accessible and low-cost tool for dMMR prediction, particularly prior to surgery. In this study, we developed a convenient and reliable model for predicting dMMR status in CRC patients on routine preoperative characterization utilizing multiple machine learning algorithms. This model will work as an automated screening tool for identifying patients suitable for mismatch repair testing and consequently for improving the detection rate of dMMR, while reducing unnecessary labor and cost in patients with proficient mismatch repair. Background Deficient mismatch repair (dMMR) indicates a sustained anti-tumor immune response and has a favorable prognosis in patients with colorectal cancer (CRC). Although all CRC patients are recommended to undergo dMMR testing after surgery, current diagnostic approaches are not available for all country hospitals and patients. Therefore, efficient and low-cost predictive models for dMMR, especially for preoperative evaluations, are warranted. Methods A large scale of 5596 CRC patients who underwent surgical resection and mismatch repair testing were enrolled and randomly divided into training and validation cohorts. The clinical features exploited for predicting dMMR comprised the demographic characteristics, preoperative laboratory data, and tumor burden information. Machine learning (ML) methods involving eight basic algorithms, ensemble learning methods, and fusion algorithms were adopted with 10-fold cross-validation, and their performance was evaluated based on the area under the receiver operating characteristic curve (AUC) and calibration curves. The clinical net benefits were assessed using a decision curve analysis (DCA), and a nomogram was developed to facilitate model clinical practicality. Results All models achieved an AUC of nearly 0.80 in the validation cohort, with the stacking model exhibiting the best performance (AUC = 0.832). Logistical DCA revealed that the stacking model yielded more clinical net benefits than the conventional regression models. In the subgroup analysis, the stacking model also predicted dMMR regardless of the clinical stage. The nomogram showed a favorable consistence with the actual outcome in the calibration curve. Conclusion With the aid of ML algorithms, we developed a novel and robust model for predicting dMMR in CRC patients with satisfactory discriminative performance and designed a user-friendly and convenient nomogram.

Performance evaluation of predictive models for detecting MMR gene mutations associated with Lynch syndrome in cancer patients in a Chinese cohort in Taiwan

The Prevalence of Germline Mutations in Chinese Colorectal Cancer Patients with Mismatch Repair Deficiency.

Germline Mutations in Chinese Colorectal Cancer Patients with Mismatch Repair Deficiency.

Performance of PREMM(1,2,6), MMRpredict, and MMRpro in detecting Lynch syndrome among endometrial cancer cases

Mismatch Repair (MMR) Gene Mutation Carriers Have Favorable Outcome in Colorectal and Endometrial Cancer: A Prospective Cohort Study

B-210 Spectrum of Germline Mutation with Characteristic MLH1 Hotspots and High Concordance with MMR Deficiency Phenotype in Lynch Syndrome Patients in Taiwan

Mismatch repair gene mutation analysis and colonoscopy surveillance in Chinese lynch syndrome families

Evaluating the performance of models for predicting the BRCA germline mutations in Han Chinese familial breast cancer patients

Prevalence and Spectrum of DNA Mismatch Repair Gene Variation in the General Chinese Population

Clinical features and mismatch repair genes analyses of Chinese suspected hereditary non-polyposis colorectal cancer: a cost-effective screening strategy proposal.

Development and Interpretation of a Clinicopathological-Based Model for the Identification of Microsatellite Instability in Colorectal Cancer

Genetic spectrum of familial colorectal cancer based on Chinese clinical criteria of Lynch syndrome: study based on a multicenter population

Mutation Spectrum of Germline Cancer Susceptibility Genes among Unselected Chinese Colorectal Cancer Patients

Cancer Risk Prediction for Mismatch Repair Germline Mutation in Chinese Hereditary Non-Polyposis Colorectal Cancer Pedigree

Development and External Validation of a Novel Nomogram for Screening Chinese Lynch Syndrome: Based on a Multicenter, Population Study

Testing Region Selection and Prognostic Analysis of MLH1 Promoter Methylation in Colorectal Cancer in China

Potentially Functional Polymorphisms in Methionyl-tRNA Synthetase Related Gene Are Associated with Gastric Cancer in a Chinese Population

Lynch syndrome pre-screening and comprehensive characterization in a multi-center large cohort of Chinese patients with colorectal cancer

Clinical Relevance of Pathogenic Germline Variants in Mismatch Repair Genes in Chinese Breast Cancer Patients

Application of machine learning in the prediction of deficient mismatch repair in patients with colorectal cancer based on routine preoperative characterization

Variants of DNA Mismatch Repair Genes Derived from 33,998 Chinese Individuals with and Without Cancer Reveal Their Highly Ethnic-Specific Nature.