利用机器学习整合CT影像组学和临床特征以预测新冠后肺纤维化

Integrating CT radiomics and clinical features using machine learning to predict post-COVID pulmonary fibrosis.

作者信息

Zhao Qianqian, Li Yijie, Zhao Chunliu, Dong Ran, Tian Jiaxin, Zhang Ze, Huang Lin, Huang Jingwen, Yan Junhai, Yang Zhitao, Ruan Jiangnan, Wang Ping, Yu Li, Qu Jieming, Zhou Min

机构信息

Department of Pulmonary and Critical Care Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, No.197, Ruijin No.2 Road, Shanghai, 200025, China.

Institute of Respiratory Diseases, Shanghai Jiao Tong University School of Medicine, No.197, Ruijin No.2 Road, Shanghai, 200025, China.

出版信息

Respir Res. 2025 Jul 2;26(1):227. doi: 10.1186/s12931-025-03305-7.

DOI:10.1186/s12931-025-03305-7

PMID:40604746

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC12225148/

Abstract

BACKGROUND

The lack of reliable biomarkers for the early detection and risk stratification of post-COVID-19 pulmonary fibrosis (PCPF) underscores the urgency advanced predictive tools. This study aimed to develop a machine learning-based predictive model integrating quantitative CT (qCT) radiomics and clinical features to assess the risk of lung fibrosis in COVID-19 patients.

METHODS

A total of 204 patients with confirmed COVID-19 pneumonia were included in the study. Of these, 93 patients were assigned to the development cohort (74 for training and 19 for internal validation), while 111 patients from three independent hospitals constituted the external validation cohort. Chest CT images were analyzed using qCT software. Clinical data and laboratory parameters were obtained from electronic health records. Least absolute shrinkage and selection operator (LASSO) regression with 5-fold cross-validation was used to select the most predictive features. Twelve machine learning algorithms were independently trained. Their performances were evaluated by receiver operating characteristic (ROC) curves, area under the curve (AUC) values, sensitivity, and specificity.

RESULTS

Seventy-eight features were extracted and reduced to ten features for model development. These included two qCT radiomics signatures: (1) whole lung_reticulation (%) interstitial lung disease (ILD) texture analysis, (2) interstitial lung abnormality (ILA)_Num of lung zones ≥ 5%_whole lung_ILA. Among 12 machine learning algorithms evaluated, the support vector machine (SVM) model demonstrated the best predictive performance, with AUCs of 0.836 (95% CI: 0.830-0.842) in the training cohort, 0.796 (95% CI: 0.777-0.816) in the internal validation cohort, and 0.797 (95% CI: 0.691-0.873) in the external validation cohort.

CONCLUSIONS

The integration of CT radiomics, clinical and laboratory variables using machine learning provides a robust tool for predicting pulmonary fibrosis progression in COVID-19 patients, facilitating early risk assessment and intervention.

摘要

背景

缺乏用于新冠后肺纤维化（PCPF）早期检测和风险分层的可靠生物标志物凸显了先进预测工具的紧迫性。本研究旨在开发一种基于机器学习的预测模型，整合定量CT（qCT）影像组学和临床特征，以评估新冠患者发生肺纤维化的风险。

方法

本研究共纳入204例确诊的新冠肺炎患者。其中，93例患者被分配到开发队列（74例用于训练，19例用于内部验证），而来自三家独立医院的111例患者构成外部验证队列。使用qCT软件分析胸部CT图像。从电子健康记录中获取临床数据和实验室参数。采用具有5折交叉验证的最小绝对收缩和选择算子（LASSO）回归来选择最具预测性的特征。独立训练12种机器学习算法。通过受试者工作特征（ROC）曲线、曲线下面积（AUC）值、敏感性和特异性评估它们的性能。

结果

提取了78个特征并将其减少到10个用于模型开发。这些特征包括两个qCT影像组学特征：（1）全肺网状影（%）间质性肺疾病（ILD）纹理分析，（2）间质性肺异常（ILA）_肺区≥5%的数量_全肺_ILA。在评估的12种机器学习算法中，支持向量机（SVM）模型表现出最佳的预测性能，在训练队列中的AUC为0.836（95%CI：0.830 - 0.842），在内部验证队列中为0.796（95%CI：0.777 - 0.816），在外部验证队列中为0.797（95%CI：0.691 - 0.873）。