A novel nomogram based on routine clinical indicators for screening for Wilson's disease

Abstract

Background and aims: There is currently no single model for predicting Wilson's disease (WD). We aimed to create a nomogram using daily clinical parameters to improve the accuracy of WD diagnosis in patients with abnormal liver function. Methods: Between July 2016 and December 2020, we identified 90 WD patients with abnormal liver function who had homozygous or compound heterozygous mutations in the ATP7B gene. The control group included 128 patients with similar liver function but no WD during the same time period. To create a nomogram, we screened potential predictive variables using the least absolute shrinkage and selection operator model and multivariate logistic regression. Results: We developed a nomogram for screening for WD based on six predictive factors: serum copper, direct bilirubin, uric acid, cholinesterase, prealbumin, and reticulocyte percentage. In the training cohort, the area under curve (AUC) of the nomogram reached 0.967 (95% confidence interval (CI) 0.946–0.988), while the area under the precision-recall curve was 0.961. Based on the optimal cutpoint of 213.55, our nomogram performed well, with a sensitivity of 96% and a specificity of 87%. In the validation cohort, the AUC of the nomogram was as high as 0.991 (95% CI 0.970–1.000). Conclusions: We developed a nomogram that can predict the risk of WD prior to the detection of serum ceruloplasmin or urinary copper, greatly increasing screening efficiency for patients with abnormal liver function.

Keywords

• Wilson's disease (WD)
• Diagnosis
• Nomogram
• Probabilistic model
• Predictor