Pelvic MRI staging of endometriosis at 3 T without patient preparation or anti-peristaltic: Diagnostic performance outcomes

PURPOSE: Endometriosis is a disease of significant burden among pre-menopausal women characterised by the appearance of functional endometrial tissue in locations outside the uterus. Deep infiltrating endometriosis (DIE) is an invasion of the endometriotic lesion that exceeds 5 mm in depth into the peritoneum. In most cases, pelvic MRI is the imaging modality of choice for the pre-operative workup for DIE to guide surgery. The recommended standard for pelvic MRI images is at 1.5 T with patient preparations in the form of laxatives, anti-spasmodics and/or rectal contrast. However, there appears to be equally diagnostic quality imaging obtainable with a 3 T system using fast imaging acquisition without any patient preparation. This reduces imaging time, movement artefacts, is more comfortable for the patient and is easier for workflow coordination. This study aims to confirm that this approach reaches diagnostic performance outcomes comparable to reported international standards, particularly in the detection of endometriotic bowel lesions, in order to guide surgical pre-operative planning. METHODS AND MATERIALS: Pre-operative diagnostic radiology reports were identified by a search of the Radiology Information System (RIS) for all pelvic MRI studies performed at a tertiary referral centre, King Edward Memorial Hospital for Women, between January 2015 to April 2017 that contained the keyword "endometriosis". Reported sites of endometriotic deposits at MRI findings were tallied based on anatomical location and correlated to laparoscopic surgical report findings and/or pathology report as a reference standard. Pooled sensitivities and specificities were then calculated and compared with established studies. RESULTS: Ninety-eight MRI studies were identified, of which 76 identified DIE and 22 were normal studies. Sixty-one patients did not have a surgical or pathology record. Of the remainder who underwent laparoscopy, operative and/or pathology reports were obtainable in 37 female patients, with a median age of 35 years (range: 24 to 49 years). The average time interval from MRI report to surgical operation was 195 days (range: 5 to 563 days). Middle compartment estimated sensitivity was 79.4% (95% CI = (69.4%, 89.4%)), specificity 95.1% (95% CI = (91.2%, 98.9%)). Posterior compartment estimated sensitivity was 76.5% (95% CI = (66.4%, 86.6%)), specificity 99.4% (95% CI = (98.1%, 100%)). Overall sensitivity 76.9% (95% CI = (69.7%, 84.0%)), specificity 98.5% (95% CI = (97.3%, 99.6%)). Sensitivity and specificity of detecting bowel endometriosis were estimated to be 94.4% (95% CI = (83.9%, 100%) and 94.7% (84.7%, 100%)) respectively. CONCLUSION: Using a 3 T MRI system without patient preparation is feasible and achieves benchmark diagnostic performance outcomes in the pre-operative assessment for DIE, especially in detecting posterior compartment lesions and bowel DIE to correctly guide surgical planning.