Radiomics derived from HRCT can predict mortality in patients with acute exacerbation of idiopathic pulmonary fibrosis.

Background: There is a dismal prognosis for idiopathic pulmonary fibrosis (IPF), a progressive interstitial lung disease. During acute exacerbations, the survival outcomes are significantly worse. This study aimed to evaluate the efficacy of high-resolution computed tomography (HRCT)-based radiomics in predicting mortality in patients experiencing acute exacerbation of idiopathic pulmonary fibrosis (AE-IPF). Methods: From April 2012 to February 2023, 146 AE-IPF patients were retrospectively enrolled at West China Hospital of Sichuan University and divided at random in a 7:3 ratio between a training set (n=102) and an inside validation set (n=44). Chest HRCT was used to obtain radiomics characteristics. A nomogram that combined the radiomics risk score with specific clinical characteristics was produced by logistic regression classifiers, which also generated radiomics and clinical models. The C-index, area under the receiver operating characteristic curves, and decision curves were used to evaluate the predictive performance of the models. The nomogram-derived risk score was used to categorize patients into high-risk and low-risk categories for the training and validation datasets. The log-rank test and Kaplan-Meier curves were utilized in survival analysis. Results: All-cause mortality occurred in 72.6% of patients over a median follow-up of 17 months. The nomogram predicted 1-year overall survival (OS) with the area under the curve (AUC) of 0.739 and 0.717 for the training and validation datasets, respectively. The nomogram surpassed both clinical and radiomics models in predicting 3-year OS, with higher AUC values in training and validation datasets (0.789 Conclusions: The radiomics-based nomogram can predict the mortality of patients with AE-IPF.