Abstract
RATIONALE AND OBJECTIVE: To evaluate the predictive value of CT-based thrombi texture analysis for assessing external iliac vein thrombosis stability. MATERIALS AND METHODS: A total of 108 patients with external iliac vein thrombosis diagnosed through clinical examination and ultrasound at the First Hospital Affiliated with Hebei North University between May 2020 and October 2024 were enrolled. Patients were divided into external iliac vein thrombosis with acute pulmonary embolism (APE) (n = 58) and external iliac vein thrombosis without APE (n = 50) groups. The region of interest (region of interest, ROI) was manually delineated on the CT images of external iliac vein thrombosis patients using 3D-Slicer software. Texture features were extracted from CT thrombi images using the Least Absolute Shrinkage and Selection Operator algorithm. Imaging, clinical, and combined models were developed through logistic regression analyses and evaluated using receiver operating characteristic (ROC) curves, calibration curves, and clinical decision curves, with the areas under the curves (AUCs) compared using the DeLong test. RESULTS: External iliac vein thrombosis textural features (gray-level run-length matrix run entropy, gray-level co-occurrence matrix (GLCM) cluster shade, GLCM-maximum cross-correlation), and clinical parameters (smoking, sex, D-dimer), were identified as independent predictors of external iliac vein thrombosis stability. The AUCs for the imaging, clinical, and combined models were 0.792 (0.705–0.880), 0.832 (0.753–0.910), and 0.890 (0.825–0.954), respectively. The combined model outperformed the clinical and imaging models (P < 0.05). Clinical decision curve analysis revealed superior predictive accuracy (0.03–0.75) compared to the clinical (0.13–0.87) and imaging (0.12–0.83) models. It achieved the highest net benefit at a risk probability > 0.03, offering optimal efficacy and the greatest clinical value in predicting external iliac vein thrombosis stability. CONCLUSION: Integrating CT texture features with clinical parameters enhances the prediction of external iliac vein thrombosis stability, underscoring its potential clinical applications.