Related Experiment Video
Updated: Jul 8, 2025
Author Spotlight: Unveiling Transmembrane Protein Family-Related Markers in Gastric Cancer and Implications for Targeted Therapies
Published on: September 15, 2023
Identification of Comprehensive Biomarkers in Patients With Mismatch Repair-Deficient Colon Adenocarcinoma Based on
Zhengjun Li1, Linxin Teng2, Zhiwei Pan3
1College of Health Economics Management, Nanjing University of Chinese Medicine, Nanjing, Jiangsu, China; Jiangsu Provincial Engineering Research Center of TCM External Medication Development and Application, Nanjing, China.
Abstract:
Immunocheckpoint inhibitors have shown impressive efficacy in patients with colon cancer and other types of solid tumor that are mismatch repair-deficient (dMMR). Currently, PCR-capillary electrophoresis is one of the mainstream detection methods for dMMR, but its accuracy is still limited by germline mismatch repair (MMR) mutations, the functional redundancy of the MMR system, and abnormal methylation of MutL Homolog 1 promoter. Therefore, this study aimed to develop new biomarkers for dMMR based on artificial intelligence (AI) and pathologic images, which may help to improve the detection accuracy. To screen for the differential expression genes (DEGs) in dMMR patients and validate their diagnostic and prognostic efficiency, we used the expression profile data from the Cancer Genome Atlas (TCGA). The results showed that the expression of Immunoglobulin Lambda Joining 3 in dMMR patients was significantly downregulated and negatively correlated with the prognosis. Meanwhile, our diagnostic models based on pathologic image features showed good performance with area under the curves (AUCs) of 0.73, 0.86, and 0.81 in the training, test, and external validation sets (Jiangsu Traditional Chinese Medicine Hospital cohort). Based on gene expression and pathologic characteristics, we developed an effective prognosis model for dMMR patients through multiple Cox regression analysis (with AUC values of 0.88, 0.89, and 0.88 at 1-, 3-, and 5-year intervals, respectively). In conclusion, our results showed that Immunoglobulin Lambda Joining 3 and nucleus shape-related parameters (such as nuclear texture, nuclear eccentricity, nuclear size, and nuclear pixel intensity) were independent diagnostic and prognostic factors, suggesting that they could be used as new biomarkers for dMMR patients.
Insights
New biomarkers, including Immunoglobulin Lambda Joining 3 and nuclear features identified by AI, can improve the detection and prognosis of mismatch repair-deficient (dMMR) solid tumors. These findings offer a more accurate approach beyond current methods for dMMR patient management.
Area of Science:
- Oncology
- Genomics
- Artificial Intelligence in Medicine
Background:
- Immunocheckpoint inhibitors demonstrate significant efficacy in mismatch repair-deficient (dMMR) solid tumors.
- Current dMMR detection methods like PCR-capillary electrophoresis have limitations, including accuracy issues related to genetic factors and MMR system complexity.
- There is a need for improved diagnostic and prognostic biomarkers for dMMR cancers.
Purpose of the Study:
- To develop novel artificial intelligence (AI)-driven biomarkers using pathologic images for enhanced dMMR detection.
- To identify and validate differential gene expression (DEGs) associated with dMMR status.
- To create accurate diagnostic and prognostic models for dMMR patients.
Main Methods:
- Utilized Cancer Genome Atlas (TCGA) expression profile data to screen for DEGs in dMMR patients.
- Developed AI-based diagnostic models analyzing pathologic image features.
- Constructed a prognostic model using multiple Cox regression analysis incorporating gene expression and pathologic characteristics.
Main Results:
- Identified significant downregulation of Immunoglobulin Lambda Joining 3 in dMMR patients, negatively correlating with prognosis.
- AI diagnostic models achieved high performance (AUCs 0.73-0.86) across training, test, and external validation sets.
- Developed a robust prognosis model with excellent predictive accuracy (AUCs 0.88-0.89) at 1-, 3-, and 5-year intervals.
Conclusions:
- Immunoglobulin Lambda Joining 3 and AI-identified nuclear shape parameters (texture, eccentricity, size, intensity) are independent diagnostic and prognostic factors for dMMR.
- These factors represent promising novel biomarkers for improving dMMR patient identification and management.
- The study highlights the potential of integrating AI with pathology and genomics for precision oncology.
