Comparison of 3-Tesla to 1.5-Tesla Magnetic Resonance Enterography to assess multifocal and multicentric bowel endometriosis: Results in routine practice

Study objective: To compare 1.5-Tesla (1.5T) to 3-Tesla (3T) Magnetic resonance enterography (MRE) in assessing multifocal (multiple lesions affecting the same segment) and multicentric (multiple lesions affecting several digestive segments) bowel endometriosis in patients with suspected-colorectal endometriosis, in routine practice. Study design: We conducted a retrospective, comparative study in Tenon University Hospital, Paris, France. We included patients scheduled for colorectal resection from April 2014 to February 2018. All patients underwent 1.5 T or 3 T MRE before surgery and endometriosis lesions were confirmed by surgery. The diagnostic performance of 1.5T and 3T MRE was evaluated by sensitivity, specificity, positive and negative predictive values (PPV and NPV), accuracy, positive and negative likelihood ratios (LR) and area under the curve of receiver operating curves (AUC ROC). Results: Ninety-eight patients were included. Fifty-two (53%) patients presented unifocal lesions, 31 (32%) multifocal lesions, and 23 (24%) multicentric lesions. In assessment of multifocal lesions, the sensitivity, specificity, positive and negative LRs were 0.57 (12/21), 0.89 (23/26), 4.95 and 0.58 for 1.5T MRE, and 0.10 (1/10), 0.95 (39/41), 2.05 and 0.95 for 3T MRE. For the diagnosis of multicentric lesions, 1.5 T MRE showed sensitivity, specificity, positive and negative LRs of 0.46 (5 of 11), 0.92 (33 of 36), 5.45 and 0.60 respectively and 3 T MRE showed sensitivity, specificity, and negative LRs of 0.33 (4/12),1.00 (39/39), and 0.67, respectively. 1.5 T MRE was more accurate than 3 T MRE for diagnosing multifocal lesions (p = 0.02), but there was no difference for multicentric lesion assessment (p = 0.66). Conclusion: In routine conditions, 3 T MRE and 1.5 T MRE are of similar low performance for diagnosing multifocal and multicentric bowel endometriosis. (C) 2018 Elsevier B.V. All rights reserved.