Erdem Özkan, Abdullah Can, Nurtaç Sarıkaş, Yasin Alper Yıldız
Advanced Radiology and Imaging - 2026;3(1):15-24
Objectives: Computed tomography (CT) has become the cornerstone imaging modality for the evaluation of suspected acute appendicitis, yet real-world diagnostic performance data from single-centre experiences incorporating both CT-pathology correlation and interobserver agreement remain limited. This study aimed to evaluate the diagnostic performance of CT, using histopathology as the reference standard, to characterise individual CT imaging features, and to assess interobserver reproducibility in a consecutive adult cohort. Methods: A retrospective analysis was conducted at our institution involving 195 adult patients who underwent appendectomy after preoperative contrast-enhanced abdominal CT for clinically suspected acute appendicitis between January 2023 and February 2026. Histopathological examination served as the reference standard. Diagnostic performance metrics (sensitivity, specificity, positive and negative predictive values, overall accuracy) were calculated from the 2x2 contingency table, with 95% confidence intervals (CIs) estimated using the Wilson score method. Odds ratios (ORs) with 95% CIs were computed for each CT feature using Fisher's exact test. Receiver operating characteristic (ROC) analysis with the Youden-optimal cut-point was performed for the appendiceal diameter. Interobserver agreement for CT features was assessed in a 40-patient subsample evaluated by two radiologists with differing levels of experience, using Cohen's kappa and the intraclass correlation coefficient (ICC). Results: Acute appendicitis was histopathologically confirmed in 155 patients (79.5%). CT demonstrated a sensitivity of 98.7% (95% CI: 95.4-99.6%), a specificity of 87.5% (73.9-94.5%), a positive predictive value of 96.8% (92.8-98.6%), a negative predictive value of 94.6% (82.3-98.5%), and an overall accuracy of 96.4% (92.8-98.3%). Forty appendectomy specimens (20.5%) showed no histopathological evidence of acute appendicitis and were classified as the pathology-negative group (Path-). Periappendiceal fat stranding demonstrated the strongest independent association with confirmed appendicitis (OR: 173.83; 95% CI: 50.17-602.28; p<0.001), followed by contrast enhancement (OR: 120.00; 95% CI: 36.85-390.82; p<0.001) and wall thickening (OR: 49.23; 95% CI: 17.92-135.22; p<0.001). The presence of an appendicolith was not significantly associated with confirmed appendicitis (p=0.151). ROC analysis of the appendiceal diameter yielded an area under the ROC curve of 0.839, with an optimal cut point of 8.3 mm (sensitivity 86.5%, specificity 75.0%). All five false-positive cases were attributable to lymphoid hyperplasia. Both false-negative cases involved atypical presentations-one with early-stage disease and the other gangrenous appendicitis with an attenuated inflammatory response. Interobserver agreement for the overall CT diagnosis was perfect (Cohen's kappa coefficient=1.000), and appendiceal diameter measurement demonstrated excellent reproducibility (ICC=0.994). Conclusion: CT achieves high diagnostic accuracy for acute appendicitis in real-world clinical practice. Periappendiceal fat stranding is the single most predictive CT finding, while the presence of an appendicolith is not independently statistically significant. Gangrenous appendicitis and early-stage disease are the principal causes of CT false-negative results. Interobserver agreement for the overall diagnosis is excellent, although experience-dependent variability persists for individual secondary signs.
