Diagnostic accuracy of multidetector CT in predicting esophageal varices in cirrhotic patients: A comparison with endoscopy as the gold standard.

Abstract

Objective: To determine the diagnostic accuracy of multidetector CT in predicting the presence of esophageal varices in patients with cirrhosis taking endoscopy as gold standard. Study Design: Cross-sectional study. Setting: Department of Radiology, Pakistan Kidney and Liver Institute and Research Centre (PKLI &RC), Lahore. Period: July 2025 to December 2025. Methods: Sample size (n) of 168 patients has been calculated with 95% confidence level taking prevalence of esophageal varices 10%, multidetector CT sensitivity of 98.9% and specificity of 100% in predicting presence of esophageal varices. Sampling Technique was non-probability consecutive sampling technique. Data selection included inclusion criteria as diagnosed case of Liver Cirrhosis on gray scale ultrasound, patients of both sex, patients of age 15 to 70 years. Results: In this cross-sectional study conducted at Pakistan Kidney and Liver Institute & RC over one year, 168 cirrhotic patients (mean age 52.3 ± 10.7 years; 61.3% male) underwent both multidetector CT (MDCT) and upper endoscopy for esophageal varices (EV) evaluation. Using endoscopy as the gold standard, MDCT demonstrated exceptional diagnostic performance: sensitivity 97.1% (95% CI: 92.3–99.2%), specificity 98.4% (94.1–99.8%), positive predictive value 97.7%, negative predictive value 97.9%, and overall accuracy of 97.6% (all p < 0.001). Inter-modality agreement for variceal grading was almost perfect (Cohen’s κ = 0.91, p < 0.001). Furthermore, MDCT-detected features including paraesophageal collaterals (OR 42.6, p < 0.001), gastric varices (OR 20.1, p < 0.001), splenomegaly (OR 14.8, p < 0.001), and portal vein dilation >13 mm (OR 13.2, p < 0.001) were strongly predictive of EV presence. These findings support MDCT as a highly accurate, noninvasive, and well-tolerated alternative to endoscopy for initial EV screening and risk stratification in cirrhotic patients, particularly where endoscopic resources are limited. Conclusion: This study demonstrates that multidetector computed tomography (MDCT) is a highly accurate, noninvasive, and patient-friendly imaging modality for the detection and grading of esophageal varices in patients with liver cirrhosis. With sensitivity and specificity exceeding 97%, and near-perfect agreement with endoscopy in variceal grading (κ = 0.91), MDCT proves to be a reliable alternative to upper gastrointestinal endoscopy particularly valuable in settings where endoscopic screening is limited by availability, cost, or patient tolerance. Future multicenter studies with larger cohorts are recommended to validate these findings across diverse populations and healthcare systems.