This study aimed to develop and validate a prediction model for recurrent pulmonary embolism (PE) using clinical variables and CT-derived body composition parameters, including skeletal muscle area (SMA), pectoralis muscle area (PMA), and subcutaneous adipose tissue area (SATA). A retrospective cohort study was conducted among patients with confirmed PE. Demographic, clinical, and laboratory variables were collected, and body composition parameters—including SMA, PMA, and SATA—were quantified from the CT component of SPECT/CT examinations using Slice-O-Matic software. Predictors considered for model building comprised sex, age, PE type, serum uric acid, creatinine, white blood cell (WBC) count, body mass index (BMI), and CT-derived body composition indices. Patients were randomly allocated into training and validation cohorts at a 7:3 ratio. A multivariable logistic regression model was developed to predict recurrence and presented as a nomogram. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC). This retrospective study developed and validated a nomogram for predicting PE recurrence using clinical and CT-derived body composition parameters. Among 184 patients (70% training, 30% validation), the cohorts were well-balanced except for sPESI scores. From 27 candidate predictors, LASSO regression identified eight non-zero coefficients features for model construction. The nomogram demonstrated moderate discriminative ability, with AUCs of 0.757 (95% CI: 0.673–0.840) in the training cohort and 0.679 (95% CI: 0.531–0.826) in validation. Calibration curves showed good agreement between predicted and observed outcomes, and decision curve analysis confirmed clinical utility across most threshold probabilities. The integrated nomogram provides a practical tool for predicting pulmonary embolism recurrence and confirms the prognostic value of routine CT-derived body composition parameters. While this model shows potential for assisting in individualized patient management, its generalizability requires further external validation.
Cao et al. (Tue,) studied this question.