Comparison of the prognostic value of 18F-fluorodeoxyglucose PET/MRI and PET/computed tomography in recurrent ovarian cancer

PURPOSE: Peritoneal metastases significantly affect prognosis in ovarian cancer. We aimed to compare the prognostic value of 18F-fluorodeoxyglucose PET/computed tomography (18F-FDG PET/CT) and 18F-FDG PET/MRI performed for restaging in patients with suspected recurrence and elevated CA-125. MATERIALS AND METHODS: Forty-five recurrent ovarian cancer patients who underwent same-session 18F-FDG PET/CT followed by whole-abdomen 18F-FDG PET/MRI for restaging were included from a previously published prospective cohort; patients were followed for overall survival (OS; median follow-up: 61.9 months). Kaplan-Meier and Cox regression analyses were performed. RESULTS: 18F-FDG PET/MRI detected peritoneal involvement more frequently than 18F-FDG PET/CT (78% vs. 58%, P = 0.012) and yielded higher peritoneal cancer index (PCI) scores (P < 0.001). During follow-up, 35 patients (77.8%) died, and the median OS was 27.7 months. On 18F-FDG PET/CT, the number of peritoneal lesions (P = 0.002), PCI score (P = 0.004), and distant organ metastases (P = 0.017) were significantly associated with OS. Peritoneal involvement showed clearer survival separation on 18F-FDG PET/CT than on 18F-FDG PET/MRI (P = 0.046 vs. P = 0.222). For 18F-FDG PET/MRI, the number of peritoneal lesions (P = 0.038) and PCI score (P = 0.005) were significantly associated with OS, whereas peritoneal involvement on 18F-FDG PET/MRI was not associated with OS. Thus, binary PET/MRI positivity and quantitative PET/MRI-derived disease burden showed different prognostic associations. In addition, among 18F-FDG PET/MRI-positive patients, OS did not differ significantly between those positive only on 18F-FDG PET/MRI and those positive on both 18F-FDG PET/MRI and 18F-FDG PET/CT (P = 0.246). In the multivariable Cox model including age and both imaging modalities simultaneously, only 18F-FDG PET/CT-defined peritoneal lesion burden (>5 vs. ≤5 lesions) remained independently associated with OS [hazard ratio = 3.80, 95% confidence interval (CI) = 1.35-10.68, P = 0.012], whereas 18F-FDG PET/MRI-defined lesion burden (>5 vs. ≤5 lesions) did not retain statistical significance (hazard ratio = 1.31, 95% CI = 0.53-3.25, P = 0.555). CONCLUSION: Although 18F-FDG PET/MRI detected peritoneal metastases more frequently and yielded higher PCI scores, 18F-FDG PET/CT-derived peritoneal lesion burden demonstrated stronger independent prognostic value. 18F-FDG PET/CT appears sufficient for baseline prognostic stratification, whereas 18F-FDG PET/MRI may still provide complementary value for anatomical localization, detection of small-volume disease, and treatment planning.