Evaluating the liver fibrosis and waist-to-height ratio(LFWHR) as a gallstone disease predictor: a nomogram model and SHAP analysis.

Gallstone disease (GD) is a prevalent gastrointestinal disorder worldwide, closely associated with obesity, metabolic diseases, and liver fibrosis. This study aimed to investigate the relationship between liver fibrosis and waist circumference to height ratio (LFWHR) with GD and to construct a nomogram model to predict gallstones. A total of 8694 participants from the 2017-2023 NHANES database were included in this study. Multivariate logistic regression was used to assess the relationship between LFWHR and GD, and subgroup analyses and interaction tests were performed. A predictive model was established using the absolute shrinkage and selection operator (LASSO) and logistic regression for multivariate feature selection. Nomograms were created to demonstrate the predictive model. The model was evaluated using the area under the ROC curve, calibration curve, and decision curve. Finally, we performed interpretability analysis by calculating SHAP values and plotting force diagrams and swarm diagrams. In the fully adjusted multivariate logistic regression model (Model 3), each unit increase in LFWHR was associated with a 51% higher likelihood of gallstone formation [OR 1.51 (95% CI 1.30-1.75)]. After stratifying by LFWHR, the Q3 level was associated with a higher risk of gallstones [OR 1.74 (95% CI 1.35-2.26)], and the Q4 level was also associated with a higher risk of gallstones [OR 1.75 (95% CI 1.33-2.32)] compared to the Q1 level. This correlation was stronger in people under 60 years of age. After feature screening, nomograms and individual nomograms were constructed for the predictive model of gallstones, yielding an AUC of 0.767 (95% CI, 0.747-0.787). The DCA analysis of the present model indicated a net benefit in the high-risk threshold range of 2-93%. The blue bars in the important figure indicate the mean size of the SHAP value, which was 0.385 for LFWHR. The swarm plot demonstrates the direction and size of the sample contribution to gallstones for each variable. We found for the first time that elevated levels of LFWHR were significantly associated with a high incidence of gallstones, and the nomogram prediction model constructed using LFWHR has potential clinical predictive value. Clinicians can utilize this tool to identify high-risk factors for gallstones at an early stage and reduce the incidence of gallstones.