Objective: The purpose of this study was to compare fat-suppressed T1-weighted with conventional MR images for the diagnosis of endometriosis, focusing on the detectability of peritoneal implants, and to evaluate the usefulness of MR imaging in predicting the severity of disease.
Subjects and methods: We analyzed the MR images of 31 patients with clinically suspected endometriosis. Laparoscopy confirmed no evidence of disease in two and endometriosis in 29. Conventional T1- and T2-weighted images and supplementary axial fat-suppressed images were obtained in all patients. In both groups, detection of endometriotic lesions with MR imaging was compared with detection at laparoscopy. Only the findings on fat-suppressed images were used to evaluate the usefulness of MR imaging in predicting the severity of disease. Early stages of disease included stages I and II; advanced stages of disease included stages III and IV.
Results: Diagnostic accuracy was higher with fat-suppressed imaging (77%) than with conventional imaging (55%) (p = .06). The overall sensitivity in detecting peritoneal implants was significantly higher with fat-suppressed imaging (61%) than with conventional imaging (27%) (p < .01). The disease was likely to be at an early stage when MR images showed peritoneal implants that were 5 mm or less in maximum diameter. The disease was likely to be at an advanced stage when MR images showed an endometrial cyst that was larger than 1.5 cm.
Conclusion: Our results show that fat-suppressed MR imaging is more accurate in the diagnosis of pelvic endometriosis and is better for predicting the severity of disease than conventional MR imaging is.