Development and validation of ‘G-CAL score’ to determine need of endoscopic ultrasound examination in patients with unexplained common bile duct dilatation

Aims To develop prediction model for ‘actionable findings’ (AF) in patients with Unexplained Common Bile Duct Dilatation (UCBDD).

Methods This is ambispective cohort study of patients who underwent evaluation for UCBDD from January 2017-January 2022. Derivation cohort consisted of 142 retrospectively recruited cases. Prospectively recruited 230 cases constituted validation cohort. Demographics, clinical presentation, liver function test (LFT), imaging findings were noted. UCBDD was defined as dilated CBD without demonstratable cause on abdominal ultrasound, computed tomography and/or magnetic resonance cholangiopancreatography (MRCP). After EUS examination, Actionable Findings (AF) requiring either pharmacotherapy or endoscopic intervention(s) were noted. Logistic regression was used to determine predictors of AF. Prediction model was developed using nomogram and validated in validation cohort.

Results AF were detected in 57.54% and 50.86% in derivation and validation cohorts respectively. On multivariate regression, abnormal Gall bladder, p- 0.026, OR-3.21, CBD diameter 9-12mm, p=0.023, OR- 3.47; CBD diameter>12mm, p- 0.004, OR-18.0; Age>55 years, p=0.001, Odd’s ratio (OR)–7.1; and abnormal LFT, p<0.0005, OR-15.82 predicted (G-CAL score) AF on EUS. Risk scores were assigned to predictors using nomogram. Nomogram performed well in terms of discrimination with area under curve 0.904, 95% CI: 0.854–0.954, p<0.001. At cut-off value of 10.90, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy of 94.18%, 58.92%, 77.88% 86.84%, 80.28% was obtained in derivation cohort. In validation cohort sensitivity, specificity, PPV, NPV and accuracy of 95.97%, 83.96%, 90.43%, 87.5%, 94.68% was observed.

Conclusions EUS can detect AF in patients with UCBDD. ‘G-CAL score’ provides simple validated clinical screening test, which can help in selecting patients for EUS examination.

Publication History

Article published online:
27 March 2025