Head-to-Head Comparison of the Diagnostic Performance of FDG PET/CT and FDG PET/MR in Patients With Cancer: A Systematic Review and Meta-Analysis

Amit Singnurkar,Raymond Poon,Ur Metser
DOI: https://doi.org/10.2214/AJR.24.31519
2024-07-17
Abstract:Background: The available evidence on the use of FDG PET/MRI performed using an integrated system in patients with cancer has grown substantially. Objective: To perform a systematic review and meta-analysis comparing the diagnostic performance of FDG PET/CT and FDG PET/MRI in patients with cancer. Evidence Acquisition: MEDLINE, EMBASE, and Cochrane Database of Systematic Reviews were searched from July 1, 2015 to January 25, 2023 for studies reporting a head-to-head comparison of the diagnostic performance of FDG PET/CT and FDG PET/MRI in patients with cancer. The two modalities' diagnostic performance was summarized, stratified by performance endpoint. For endpoints with sufficient data, meta-analysis was performed using bivariate modeling to produce summary estimates of pooled sensitivity and specificity. For remaining endpoints, reported performance in individual studies was recorded. Evidence Synthesis: The systematic review included 29 studies with a total of 1656 patients. For patient-level detection of regional nodal metastases (5 studies), pooled sensitivity and specificity for PET/MRI were 88% (95% CI, 74-95%) and 92% (95% CI, 71-98%), respectively, and for PET/CT were 86% (95% CI, 70-94%) and 86% (95% CI, 68-95%), respectively. For lesion-level detection of recurrence and/or metastases (5 studies), pooled sensitivity and specificity for PET/MRI were 94% (95% CI, 78-99%) and 83% (95% CI, 76-88%), respectively, and for PET/CT were 91% (95% CI, 77-96%) and 81% (95% CI, 72-88%), respectively. In individual studies not included in meta-analysis, PET/MRI in comparison with PET/CT showed staging accuracy in breast cancer of 98.0% versus 74.5% and in colorectal cancer of 96.2% versus 69.2%, sensitivity for primary tumor detection in cervical cancer of 93.2% versus 66.2%, and sensitivity, specificity, and accuracy for lesion-level liver metastasis detection of 91.1-98.0% versus 42.3-71.1%, 100.0% versus 83.3-98.6%, and 96.5-98.2% versus 44.7-86.7%, respectively. In three studies, management was more commonly impacted by information from PET/MRI (5.2-11.1%) than PET/CT (0.0-2.6%). Conclusions: PET/MRI showed comparable or superior diagnostic performance versus PET/CT across a range of cancers and endpoints. Clinical Impact: The findings help to identify clinical settings where PET/MRI may provide particular clinical benefit for oncologic evaluation.
What problem does this paper attempt to address?