系统性硬化症相关间质性肺病风险预测模型:基于LASSO-logistic回归OA

Objective To establish a prediction model for systemic sclerosis(SSc)-related in-terstitial lung disease(ILD)using LASSO-regularized logistic regression.Methods Patients di-agnosed with systemic sclerosis at the First Affiliated Hospital of Zhengzhou University from Janu-ary 2019 to June 2024 were included as study subjects.They were divided into a case group(99 cases,SSc-ILD group)and a control group(62 cases,SSc-non-ILD group)based on the pres-ence or absence of interstitial lung disease.Univariate analysis and LASSO regression were first used for variable screening,with the selected variables serving as independent variables in multi-variate logistic regression.Based on the final independently identified risk factors,a predictive model was established using R software,and its discrimination ability,calibration level,and clini-cal efficacy were evaluated.Results A prediction model established based on LASSO-logistic re-gression included five predictors:the dcSSc subtype,cough,digital ulcers,positivity for anti-Scl-70 antibodies,and lower serum albumin levels.Based on these factors,a diagnostic prediction model was established.The ROC curve demonstrated an AUC of 0.815(95％CI:0.748～0.882),indicating good discriminative ability.The calibration curve closely approximated the ref-erence line,showing excellent calibration performance.The DCA results revealed that this clinical prediction model exhibited high net benefit across a broad range of thresholds,demonstrating strong clinical utility.Conclusion The prediction model developed in this study to predict inter-stitial lung disease(ILD)in patients with SSc demonstrates good clinical applicability.