常规胸部CT多域表型预测COPD急性加重风险的价值

Value of Multidomain Phenotyping on Routine Chest CT in Predicting the Risk of Acute Exacerbation of COPD

摘要: 目的：探讨常规胸部CT多域表型预测慢性阻塞性肺疾病(COPD) 12个月中重度急性加重的价值。方法：回顾性纳入684例COPD患者，按7:3比例分为建模队列和验证队列。收集临床资料，并在常规胸部CT上评价肺气肿评分、黏液栓评分、PA/A比值、胸大肌面积指数和胸大肌密度。采用LASSO和多因素Logistic回归构建临床模型、影像模型和联合模型，并评价区分度、校准度和临床净获益。结果：252例患者在CT检查后12个月内发生中重度急性加重。联合模型在建模队列和验证队列中的AUC分别为0.884和0.851，均高于临床模型和影像模型；校准表现及决策曲线亦显示联合模型具有较好的预测效能。结论：常规胸部CT多域表型可辅助识别COPD急性加重高风险患者，具有一定临床应用价值。

Abstract: Objective: To explore the value of multidomain routine chest CT phenotyping for predicting 12-month moderate-to-severe acute exacerbation in patients with chronic obstructive pulmonary disease (COPD). Methods: This retrospective study included 684 patients with COPD, who were randomly divided into derivation and validation cohorts at a ratio of 7:3. Clinical variables and routine chest CT features, including emphysema score, mucus plug score, pulmonary artery-to-aorta ratio, pectoralis muscle area index and pectoralis muscle attenuation, were collected. LASSO and multivariable logistic regression were used to develop clinical, imaging and combined models. Results: A total of 252 patients developed 12-month moderate-to-severe exacerbation. The combined model achieved AUCs of 0.884 and 0.851 in the derivation and validation cohorts, respectively, and showed better calibration and clinical net benefit than single-domain models. Conclusion: Multidomain routine chest CT phenotyping may help identify patients with COPD at high risk of exacerbation and provide an explainable basis for follow-up management.