Predictive nomogram for preoperative differential diagnosis of benign and malignant gallbladder lesions. Nomogram for diagnosis of gallbladder cancer

Abstract

Background Gallbladder cancer is a rare and fatal malignant tumor, and difficult to be found in time due to the atypical symptoms. Early detection and treatment of gallbladder cancer is essential. Methods By evaluating the relationship between clinical features and contrast-enhanced ultrasound nature of 237 cases of gallbladder lesions, the preoperative predictors with differential diagnosis value for malignancy and benignity were identified and integrated into the nomogram by stepwise multivariate logistic regression analysis. The predictive performance of the nomogram was assessed by receiver operating characteristic curve analysis, calibration curve analysis and decision curve analysis, and compared with the prediction model combining neutrophil-to-lymphocyte ratio and CA19-9. Bootstrap analysis was used for the interval validation. Results The nomogram predicting benign and malignant gallbladder lesions was constructed by the predictors with preoperative diagnostic value (L, DBil, gallbladder wall thickness and features of gallbladder lesions by CEUS). The C-statistic of the nomogram is 0.981 and superior than the C-statistic of the combination of neutrophil-to-lymphocyte ratio and CA19-9. The prediction accuracy, reliability and clinical utility were considerable in the performance evaluation. Internal validation of the nomogram was performed using Bootstrap with 1000 resamplings, yielding a Kappa value of 0.800 and an Accuracy of 0.911. Conclusions The predictive nomogram is conducive to the differentiation of benign and malignant gallbladder lesions and play an important guiding role in the clinical treatment decision-making process.