Purpose: The aim of this study was to determine an optimal threshold method for the segmentation of malignant lesions from (18)F-FDG PET/CT images and to evaluate the prognostic value of the total lesion glycolysis in post-surgical patients with epithelial ovarian cancer.
Methods: We retrospectively reviewed 47 patients with pathologically proven epithelial ovarian cancer who underwent (18)F-FDG PET/CT imaging after surgery. The follow-up time was 26.6 ± 19.8 months (ranged from 4 to 89 months). For each patient, every lesion was segmented by 2 thresholds with 3D-area growing algorithm, standard uptake value (SUV) 2.5, and background method. The detection rates were compared. The optimal threshold method was then used to calculate whole-body metabolic tumor volume (WBMTV) and whole-body total lesion glycolysis (WBTLG). The prognostic significance of SUV(max), WBMTV, WBTLG, and other pathological variables for overall survival were assessed by Cox proportional hazards regression analysis and Kaplan-Meier survival analysis.
Results: A total of 142 metastatic lesions of 47 patients were confirmed by long-term clinical follow-up or pathological findings. The detection rates of the threshold SUV 2.5 and background methods were 37.32% (53/142) and 96.48% (137/142), respectively, which showed significant difference between the 2 methods (P < 0.005). In multivariate analysis, WBTLG, obtained from the background method, was an independent predictive factor associated with the prognosis (HR 1.043, 95% CI 1.01-1.078, P = 0.011), and none of the other factors had statistical association. Survival analysis also showed that the survival time was clearly shortened with WBTLG increasing (P < 0.001).
Conclusions: In this group of post-surgery patients with epithelial ovarian cancer, the background method could segment much more malignant lesions than SUV = 2.5 method, and WBTLG, obtained from this method, could be used as an independent prognostic factor.