Development and Validation of a Predictive Model for 6-Month All-Cause Mortality in Pulmonary Embolism

Background/Objectives: Pulmonary embolism (PE) remains a major cause of cardiovascular mortality, yet existing risk stratification tools predominantly focus on short-term outcomes. In this study, we set out to develop and externally validate a predictive model for 6-month all-cause mortality in PE patients using routinely available clinical parameters. Methods: We conducted a retrospective analysis of patients with PE from two independent cohorts. The MIMIC-IV database served as the training set for variable selection using least absolute shrinkage and selection operator (LASSO) regression. External validation was performed using data from the First Affiliated Hospital of Fujian Medical University. Model performance was assessed through discrimination (area under the receiver operating characteristic curve [AUC]), calibration, and decision curve analysis. Results: The training cohort comprised 814 patients from the MIMIC-IV database, with a 6-month all-cause mortality rate of 21.3% (n = 174). LASSO regression initially selected 13 predictors. External validation in the local cohort (n = 217, 6-month all-cause mortality: 16.5%) refined these to four core predictors: respiratory rate, oxygen saturation (SpO2), and the systemic inflammation response index (SIRI). The parsimonious model demonstrated good discrimination (AUC = 0.81, 95% confidence interval [CI]: 0.74–0.88), excellent calibration (Hosmer–Lemeshow p = 0.367), and positive net clinical benefit across a range of threshold probabilities. After standardizing SIRI to absolute counts, SIRI was retained as a pre-specified inflammatory predictor. Conclusions: This simplified model incorporating four readily available clinical variables accurately predicts 6-month all-cause mortality in patients with PE.