Comparison of 18F-FDG PET/MR and PET/CT for pretreatment TNM staging of hilar cholangiocarcinoma

Purpose 18 F-FDG PET/MR has been applied to the diagnosis and preoperative staging in various tumor types; however, reports using PET/MR in hilar cholangiocarcinoma (HCCA) are rare. We investigated the value of PET/MR for preoperative staging and compared it with PET/CT in HCCA. Methods Fifty-eight patients with HCCA confirmed by pathology were retrospectively analyzed. 18 F-FDG PET/CT imaging was performed first, followed with whole-body PET/MR imaging. SUV max of tumor and normal liver tissue were measured. Paired T test was used to compare SUV max of tumor and normal liver tissue of PET/CT and PET/MR. In addition, McNemar test was used to compare the accuracy of TNM staging and Bismuth-Corlette typing between PET/CT and PET/MR. Results There was no significant difference in SUV max between PET/CT and PET/MR in primary tumor lesions (6.6 ± 5.5 vs. 6.8 ± 6.2, P = 0.439). SUV max of PET/CT and PET/MR in normal liver parenchyma was significantly different (3.0 ± 0.5 vs. 2.1 ± 0.5, P < 0.001). The accuracy of PET/MR in diagnosing T staging and N staging was significantly higher than those of PET/CT (72.4% vs. 58.6%, P = 0.022 and 84.5% vs. 67.2%, P = 0.002). There was no significant difference between PET/CT and PET/MR in M staging (94.8% vs. 98.3%, P = 0.5). The classification accuracy of PET/MR in Bismuth-Corlette was significantly higher than that of PET/CT (89.7% vs. 79.3%), P = 0.031. Conclusions The diagnostic accuracy of 18 F-FDG PET/MR was superior to that of PET/CT in preoperative T staging, N staging, and Bismuth–Corlette classification of HCCA. In M staging, the diagnostic accuracy of PET/MR was similar to that of PET/CT.