Respiratory medicine · 2026
Predicting Progressive Pulmonary Fibrosis in Idiopathic Inflammatory Myopathy-associated Interstitial Lung Disease: A Machine Learning Approach Integrating Quantitative CT and Clinical Features.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Sun Chao, Wu Wenxin, Li Yijie, Wang Yuheng, Wang Shuaibo, Zhang Ze, Teng Jialin, Cao Hua, Wang Ping, Zhou Min
Аннотация
BACKGROUND: The prognosis of progressive pulmonary fibrosis (PPF) in idiopathic inflammatory myopathy-associated interstitial lung disease (IIM-ILD) is poor, and current strategies for predicting its development are limited. This study aimed to use machine learning to integrate quantitative computed tomography (QCT) and clinical features to predict PPF in patients with IIM-ILD.
METHODS: We retrospectively included 180 IIM-ILD patients, of whom 86 developed PPF. We randomly assigned patients to training and validation sets in a 7:3 ratio. Least absolute shrinkage and selection operator regression and univariate analysis were employed to select the most predictive features. We independently trained 8 machine learning algorithms. Model performance was evaluated by receiver operating characteristic curves, area under the curve (AUC) values, and decision curve analysis.
RESULTS: 16 key clinical features and 7 QCT variables were selected for model development. The 10 variables with the highest contribution values were subsequently selected for the PPF prediction model. The clinical variables included Krebs von den Lungen-6, interleukin-6, percent-predicted forced expiratory volume in 1 second, carcinoembryonic antigen, ferritin, CD3+ T cells, natural killer cells, and CD19+ B cells. The QCT signatures included ILD Honeycomb Volume (%) Lt Lower + Rt Lower and ILD Consolidation Volume (%) Whole Lung. Among the 8 machine learning algorithms, the Random Forest model demonstrated the best predictive performance, with AUCs of 0.821 (95% CI: 0.680-0.925) in the training cohort and 0.943 (95% CI: 0.866-0.982) in the validation cohort.
CONCLUSIONS: The integration of QCT and clinical features using machine learning offers a potential non-invasive approach for predicting PPF in patients with IIM-ILD. This strategy may facilitate the early identification of individuals at high risk of PPF and potentially support timely clinical intervention.
METHODS: We retrospectively included 180 IIM-ILD patients, of whom 86 developed PPF. We randomly assigned patients to training and validation sets in a 7:3 ratio. Least absolute shrinkage and selection operator regression and univariate analysis were employed to select the most predictive features. We independently trained 8 machine learning algorithms. Model performance was evaluated by receiver operating characteristic curves, area under the curve (AUC) values, and decision curve analysis.
RESULTS: 16 key clinical features and 7 QCT variables were selected for model development. The 10 variables with the highest contribution values were subsequently selected for the PPF prediction model. The clinical variables included Krebs von den Lungen-6, interleukin-6, percent-predicted forced expiratory volume in 1 second, carcinoembryonic antigen, ferritin, CD3+ T cells, natural killer cells, and CD19+ B cells. The QCT signatures included ILD Honeycomb Volume (%) Lt Lower + Rt Lower and ILD Consolidation Volume (%) Whole Lung. Among the 8 machine learning algorithms, the Random Forest model demonstrated the best predictive performance, with AUCs of 0.821 (95% CI: 0.680-0.925) in the training cohort and 0.943 (95% CI: 0.866-0.982) in the validation cohort.
CONCLUSIONS: The integration of QCT and clinical features using machine learning offers a potential non-invasive approach for predicting PPF in patients with IIM-ILD. This strategy may facilitate the early identification of individuals at high risk of PPF and potentially support timely clinical intervention.