Development and Validation of a Nomogram Incorporating Ultrasonic and Elastic Findings for the Preoperative Diagnosis of Biliary Atresia

Abstract

To develop and validate a nomogram that incorporates the gallbladder morphology, hepatic elasticity, and demographic information for the prediction of biliary atresia (BA) in children. A total of 294 consecutive patients under the age of 70 days with cholestasis and suspected symptoms of BA were enrolled in this study, who were divided into a training cohort (150 patients) and a validation cohort (144 patients). Ultrasonography and two-dimensional shear wave elastography were performed for each patient prior to knowing the final diagnosis. Multivariate logistic regression was used to analyze the gallbladder morphologic feature in the sonogram (absence of gallbladder, small gallbladder, lower postprandial gallbladder contractibility, or abnormal gallbladder wall), hepatic elasticity and clinical data from the training cohort, and a diagnostic nomogram for BA was subsequently developed. The performance of the nomogram was respectively evaluated with respect to the discrimination and calibration in every cohort. The multivariate analysis showed that the factors of age (p = 0.009), gallbladder morphology (p = 0.001) and hepatic elasticity (p < 0.001) could serve as independent predictive factors to differentiate between BA and other causes of cholestasis. The nomogram incorporating these three parameters showed good discrimination and satisfactory calibration, indicating a better performance compared to using only the gallbladder morphologic features and hepatic elasticity. The observed area under the receiver operator characteristic curve in the training cohort and validation cohort was 0.939 (p < 0.001) and 0.942 (p < 0.001), respectively, with a sensitivity of 95.5% and a specificity of 83.4% in the combined cohort. The established nomogram shows a favored and improved predictive value for the diagnosis of BA.