Background: Ulcerative colitis has a serious impact on the quality of life of patients and is more likely to progress to colon cancer. Therefore, early diagnosis and timely intervention are of considerable importance.
Methods: Gene expression data of active ulcerative colitis were downloaded from the Gene Expression Omnibus (GEO) database, and genes with significant differential expression were identified. Biochemical markers with diagnostic significance were selected through machine learning methods. The expression differences of the selected markers between colon adenocarcinoma (COAD) and healthy control groups in The Cancer Genome Atlas (TCGA) database were analyzed to evaluate their diagnostic value. In addition, the correlation between the selected markers and clinical indicators, as well as their predictive efficacy for the survival of COAD patients, was explored.
Results: Through machine learning and LASSO regression analysis, UGT2A3 was finally determined as a diagnostic marker for ulcerative colitis. It demonstrated high diagnostic accuracy in both the training set and the external validation set. Furthermore, UGT2A3 was significantly downregulated in COAD tissues compared to normal control tissues. The ROC curve suggested that UGT2A3 could serve as a diagnostic marker for COAD with excellent performance, achieving an AUC of 0.969. Immune infiltration analysis indicated a significant negative correlation between the expression of UGT2A3 and neutrophils. Correlation analysis suggested a link between UGT2A3 and the pathological classification of colon cancer. Survival analysis showed that UGT2A3 is negatively correlated with OS, PPS, and RFS in colon cancer.
Conclusion: The author identified UGT2A3 as a diagnostic marker for ulcerative colitis through bioinformatics methods, and verified its significant downregulation in colon cancer, as well as its predictive role in the survival of COAD patients. These findings suggest that UGT2A3 may serve not only as a diagnostic marker for ulcerative colitis and colon cancer but also as a potential prognostic indicator for colon cancer.
Keywords: UGT2A3; colon cancer; diagnostic; machining learning; ulcerative colitis.
Copyright © 2024 Chen.