Prognostic assessment of sepsis-induced acute respiratory distress syndrome in older patients using clinical and CT-based radiomic features

Abstract

Background

Sepsis-induced acute respiratory distress syndrome (SI-ARDS) is associated with high mortality rates, necessitating early risk stratification. This study aimed to develop and validate a radiomics-based nomogram integrating computed tomography (CT) features and clinical parameters to predict 28-day mortality in older patients with SI-ARDS.

Methods

In this retrospective cohort study, 302 older patients (≥60 years) diagnosed with SI-ARDS between January 2019 and December 2023 were enrolled. Radiomic features were extracted from admission chest CT images. Patients were randomly allocated to training (n = 242) and validation (n = 60) cohorts. Three predictive models—radiomic, clinical, and combined—were constructed using Maximum Relevance Minimum Redundancy (MRMR) algorithm and Least Absolute Shrinkage and Selection Operator (LASSO) regression. Model performance was assessed using the concordance index (C-index), calibration curves, and decision curve analysis. A nomogram was developed based on the optimal model for clinical application.

Results

The fusion model achieved superior discrimination compared with the radiomic model, clinical model, and Sequential Organ Failure Assessment score in both cohorts (C-index: training, 0.850 vs. 0.798, 0.781, and 0.654; validation, 0.839 vs. 0.768, 0.779, and 0.696; all p < 0.001). The model demonstrated excellent calibration and provided greater net clinical benefit across threshold probabilities of 10 %–90 %. Risk stratification using the nomogram identified distinct prognostic groups with significantly different 28-day survival (log-rank p < 0.001).

Conclusion

The nomogram developed from the fusion model demonstrated superior predictive performance for 28-day mortality in older patients with SI-ARDS compared to conventional scoring systems, though multicenter validation is required to confirm clinical utility.