Postoperative CT-derived spleen volumetry for recurrence risk stratification after curative-intent surgery for hepatic alveolar echinococcosis.
Abstract
Purpose
To evaluate automated CT-derived spleen volumetry as an opportunistic imaging biomarker for postoperative risk stratification in patients with surgically treated hepatic alveolar echinococcosis (AE) and to compare its prognostic performance with clinical and laboratory parameters.
Methods
This retrospective single-center cohort study included consecutive patients undergoing curative-intent R0 resection for hepatic AE. Baseline was defined as the immediate postoperative CT. Automated spleen segmentation was performed on baseline and 12-month CT using TotalSegmentator, and relative spleen volume change was calculated. The primary endpoint was imaging-defined disease recurrence within 24 months. Clinical, laboratory, serological (indirect hemagglutination assay [IHA], total IgE), and imaging variables were analyzed using univariate logistic regression; significant predictors entered multivariate analysis. ROC analysis assessed discriminatory performance.
Results
62 patients were included; 21 developed recurrence within 24 months. In univariate analysis, 4 features were significant, with relative spleen volume change showing the strongest association with recurrence (OR 1.158; 95% CI 1.082-1.239; p < 0.001). In multivariate analysis, only relative spleen volume change remained independently associated with recurrence (OR 1.152; 95% CI 1.056-1.255; p = 0.001). Relative spleen volume change showed excellent discrimination (AUC 0.926; 95% CI 0.851-0.978); a 24.3% increase yielded sensitivity of 0.81 and specificity of 0.90. Total IgE and IHA titers, including longitudinal changes, were not associated with recurrence.
Conclusions
Automated postoperative CT-derived spleen volumetry may provide an objective imaging biomarker for postoperative risk stratification after curative-intent surgery for hepatic AE, independently predicting recurrence, whereas routinely available clinical, serological and laboratory markers showed no significant association in this cohort.