新疆医科大学学报

目的 构建间歇性糖皮质激素冲击治疗甲状腺相关眼病(Thyroid-associated ophthalmopathy, TAO)疗效的列线图预测模型，分析该治疗方法的安全性。方法 回顾性分析2018年11月-2024年2月内江市第一人民医院内分泌科收治的103例TAO患者的临床资料，采用间歇性糖皮质激素冲击治疗，根据治疗完成后的疗效将患者分为不良组、良好组。比较两组一般资料，采用多因素Logistic回归模型分析间歇性糖皮质激素冲击治疗TAO疗效的影响因素，建立对应的列线图预测模型。分别采用受试者工作特征曲线(Receiver operating characteristic,ROC)、校正曲线、决策曲线评估风险预测模型的准确性、校准度及临床适用性。统计治疗期间不良反应。结果 103例患者中30.10%效果不良；Logistic回归模型分析显示，吸烟史、C反应蛋白(CRP)、总胆固醇(TC)、促甲状腺激素受体抗体(TRAb)、病程、突眼度是间歇性糖皮质激素冲击治疗TAO疗效不良的危险因素(P<0.05)。建立对应的列线图预测模型，对各变量对应的数值进行评分，获取所有变量分数之和总分及间歇性糖皮质激素冲击治疗TAO疗效不良的风险。ROC曲线显示，该模型预测间歇性糖皮质激素冲击治疗TAO疗效不良的曲线下面积(AUC)为0.947,灵敏度为93.55%,特异度为94.44%。Hosmer-Lemeshow检验显示，该风险预测模型的预测概率与实际概率的校正曲线比较，差异无统计学意义(χ²=0.445,P=0.305)。决策曲线显示，该风险预测模型在风险阈值0～0.68范围内获取临床净收益。治疗期间，103例患者中不良反应总发生率为7.77%。结论 吸烟史、CRP、TC、TRAb、病程、突眼度是间歇性糖皮质激素冲击治疗TAO疗效不良的危险因素，对应的列线图模型经验证临床效能良好，且间歇性糖皮质激素冲击治疗TAO具有良好的安全性。

Objective To construct a nomogram prediction model for the efficacy of intermittent corticosteroid pulse therapy in thyroid-associated ophthalmopathy(TAO) and to analyze the safety of this treatment modality. Methods The clinical data of 103 TAO patients admitted to the hospital from November 2018 to February 2024 were retrospectively analyzed. All patients were treated with intermittent glucocorticoid shock therapy. The patients were divided into a poor outcome group and a good outcome group based on treatment efficacy. General clinical characteristics were compared between the groups, and multivariate Logistic regression was used to identify factors impacting the therapy′s success. A nomogram prediction model was developed, and its performance was assessed in terms of accuracy, calibration and clinical utility using the receiver operating characteristic(ROC) curve, calibration curve and decision curve analysis. Adverse reactions occurring during the treatment period were also recorded. Results 30.10% of the 103 patients demonstrated poor treatment outcomes. Logistic regression analysis revealed that smoking history, C-reactive protein(CRP), total cholesterol(TC), thyrotropin receptor antibody(TRAb), disease duration and exophthalmos were all significant risk factors for poor response to intermittent glucocorticoid pulse therapy in TAO(P<0.05). A nomogram model was developed, with each variable assigned a score based on its value. The total score was then used to estimate the risk of poor treatment outcomes. The ROC curve showed that the model predicted poor outcomes with an area under the curve(AUC) of 0.947, a sensitivity of 93.55%, and a specificity of 94.44%. The Hosmer-Lemeshow test indicated no significant differences between predicted and observed outcomes(χ²=0.445, P=0.305), and decision curve analysis revealed clinical net benefits within a risk threshold range of 0 to 0.68. Adverse reactions occurred in 7.77% of the patients. Conclusion Smoking history, CRP levels, TC levels, TRAb, disease duration and exophthalmos are significant predictors of poor outcomes in TAO patients receiving intermittent glucocorticoid pulse therapy. The nomogram model demonstrates strong predictive power and clinical utility. Additionally, the therapy itself appears to be safe with a low incidence of adverse effects.