Preoperative clinical scores compared with dual-energy computed tomography parameters in predicting complete resection of ovarian advanced high-grade serous carcinoma
The aim of this study was to compare clinical scores with dual-energy computed tomography (CT) parameters of ovarian advanced high-grade serous carcinoma (HGSC) in prediction of complete (R0) resection. Between November 2021 and July 2024, 56 advanced HGSC patients with complete dual-energy CT images and clinicopathological information were enrolled. Clinical scores, including the Suidan score, computed tomography-based peritoneal cancer index (CT-PCI), computed tomography-based Eisenkop (CT-Eisenkop) score, and Aletti score, were evaluated, and surgical outcomes were prospectively collected. Dual-energy CT parameters were measured from the solid primary tumour component in two-phase contrast-enhanced images. Receiver operating characteristic curves and logistic regression analysis were used to evaluate the diagnostic efficacy in predicting R0 resection. Of the 56 advanced HGSC patients, 32 (57.14%) had R0 resection and 24 (42.86%) had tumour residue. Virtual monoenergetic images at 40 keV (VMI40) were analysed using the ratio of venous phase (VP) to arterial phase (AP) enhancement, expressed as (VP-AP)/AP. The VMI40 (VP-AP)/AP ratio was identified as the only independent risk factor for R0 prediction (odds ratio: 0.019; P=0.022). The VMI40 (VP-AP)/AP ratio showed considerable area under the curve (AUC) performance (AUC of 0.724) for R0 resection, as did the CT-PCI and the CT-Eisenkop scores (both with an AUC of 0.722), compared to the Suidan score. Using CT-Eisenkop score + VMI40 (VP-AP)/AP showed the highest AUC for R0 resection (AUC: 0.853, sensitivity: 94.7%, and specificity: 69.0%). The VMI40 (VP-AP/AP) ratio, a dual-energy CT parameter, outperforms traditional CT-based clinical scores for predicting R0 resection in advanced HGSC. Predictive accuracy is further improved when combined with the CT-Eisenkop score.