DOI: 10.7197/cmj.1946373 ISSN: 1305-0028

Endoscopic Correlation of CT-Detected Gastric and Colonic Wall Thickening and Evaluation of Gastric Malignancy Risk Using a Clinical Scoring System

Mehmet Burak Kurşun, Engin Altinkaya
Show PDF Cite
Objective: To compare endoscopic confirmation rates of gastric and colonic wall thickening detected on computed tomography (CT), evaluate the diagnostic utility of CT thickening pattern (diffuse vs focal), and develop a clinical scoring model to predict the risk of gastric malignancy.Materials and Methods: This single-center retrospective cohort study included 200 patients (100 gastric, 100 colonic) with CT-reported wall thickening from March 1, 2018 to March 1, 2023, all undergoing endoscopy within 30 days. CT thickening was classified as diffuse or focal. CT–endoscopy concordance was assessed, and malignancy was confirmed by histopathology. Gastric malignancy predictors were identified via univariate and multivariate logistic regression. A clinical risk score was created from the multivariate model and internally validated with bootstrapping and ROC analysis.Results: CT–endoscopy concordance was higher in the colon group than the gastric group (36% vs. 15%; p<0.01). CT thickening pattern (diffuse vs. focal) was not independently associated with malignancy. In the gastric group, diffuse thickening was associated with lower serum albumin and total protein. Multivariate analysis identified male sex and low albumin as independent predictors of gastric malignancy. A four-parameter clinical scoring system (including sex, albumin, hemoglobin and localization) demonstrated good discrimination (AUC=0.94).Conclusion: The diagnostic yield of CT-detected gastric and colonic wall thickening varies substantially. CT-reported gastric wall thickening showed a high false-positive rate and should be interpreted in clinical context. Integrating clinical and laboratory parameters with imaging findings may improve risk stratification, and the proposed scoring system may help prioritize endoscopy.

More from our Archive