Head-to-Head comparison of 18F-FDG PET/MRI and 18F-FDG PET/CT for preoperative tumor staging and lymph node metastasis diagnosis in endometrial and cervical cancers: a meta-analysis

Background Accurate preoperative staging is critical for managing endometrial and cervical cancers. While MRI and PET/CT are current cornerstones, the emerging hybrid PET/MRI technology promises superior soft-tissue contrast and metabolic assessment. This study aimed to directly compare the diagnostic accuracy of 1 ⁸F-FDG PET/MRI versus 1 ⁸F-FDG PET/CT for preoperative tumor staging and lymph node metastasis (LNM) detection in these cancers. Methods We systematically searched PubMed, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials, followed by a meta-analysis of the eligible studies. Results Six studies ( n =419) were included, all of which compared lymph node metastasis detection, and four of which also compared tumor staging. For lymph node metastasis---the clinical question where PET contributes most---the primary analysis using a random-effects model including all six studies yielded pooled sensitivity of 83% (95% CI: 70–96%) and 78% (95% CI: 65–92%) for PET/MRI and PET/CT, respectively, and pooled specificity of 88% for both modalities, with substantial heterogeneity. A sensitivity analysis excluding one outlier study produced broadly similar estimates (sensitivity: 93% vs. 85%; specificity: 93% vs. 92%). No formal paired comparison was performed, and equivalence cannot be concluded. For local tumor staging, PET/MRI showed significantly higher accuracy than PET/CT (OR=3.02, 95% CI: 1.87-4.86, P<0.001), a finding likely attributable to MRI’s superior soft-tissue contrast. Conclusion This meta-analysis suggests that PET/MRI offers superior accuracy for local tumor staging compared to PET/CT in endometrial and cervical cancers, likely reflecting the inherent advantage of MRI over CT. For lymph node metastasis, the pooled point estimates for PET/MRI and PET/CT were numerically close; however, a formal paired comparison could not be performed due to the absence of discordant-pair data, and equivalence cannot be concluded. This finding nonetheless underscores the central role of the PET component for nodal detection. Given the limited number of included studies and the retrospective design of the primary studies, these findings should be interpreted cautiously and require confirmation through larger, prospective multicenter studies. Trial registration The protocol for this meta-analysis was registered on the PROSPERO platform (Registration ID: CRD420261387587).