Objective To construct a nomogram prediction model for hypersplenism in patients with hepatolenticular degeneration (HLD) based on ultrasonic elastography and clinical indicators, and evaluate its clinical application value.

Methods The clinical and ultrasonic data of 155 HLD patients admitted to the First Affiliated Hospital of Anhui University of Chinese Medicine from January 2022 to January 2024 were retrospectively analyzed. The patients were divided into the hypersplenism group (41 cases) and non-hypersplenism group (114 cases) according to the presence or absence of complicated hypersplenism. Univariate and multivariate logistic regression analyses were conducted using R language to screen out the independent risk factors of HLD combined with hypersplenism, and a nomogram prediction model was established based on the screened risk factors. The receiver operating characteristic (ROC) curve and its area under the curve were used to evaluate the discrimination of the model. Drawing the calibration curve and conducting the Hosmer-Lemeshow test verified the calibration of model; The clinical effectiveness was analyzed using the decision curve.

Results The results of univariate logistic regression analysis revealed that the age, portal vein diameter, liver shear wave velocity, total bilirubin, type IV collagen, hyaluronic acid and prothrombin time were the risk factors of hypersplenism in HLD patients (P ＜ 0.05 to P ＜ 0.01）. The multivariate logistic regression demonstrated that the portal vein diameter (OR = 1.646, 95%CI: 1.290–2.101, P ＜ 0.01), liver shear wave velocity (OR = 4.831, 95%CI: 1.675–13.932, P ＜ 0.01) and prothrombin time (OR = 1.531, 95%CI: 1.059–2.212, P ＜ 0.05) were the independent risk factors od the development of hypersplenism. Based on the above independent risk factors, a nomogram model for predicting HLD combined with hypersplenism was successfully constructed. The ROC curve AUC of this model was 0.858 (95%CI: 0.798–0.917, P ＜ 0.05), with a sensitivity of 71% and a specificity of 88%, indicating that this model could well distinguish whether HLD patients had hypersplenism or not. The results of Hosmer-Lemeshow test showed that the model had a good fit (P = 0.32, ＞0.05), indicating that the model could accurately predict the probability of hypersplenism in HLD patients. The decision curve analysis showed that the nomogram model yielded clinical net benefit within the risk threshold range of 8% to 91%.

Conclusions Portal vein diameter, liver shear wave velocity and prothrombin time are the independent risk factors for hypersplenism in patients with HLD. The nomogram prediction model constructed based on these factors can help clinicians accurately screen the high-risk patients with HLD who may develop hypersplenism.