Magnetic resonance imaging classification of deep pelvic endometriosis: Description and impact on surgical management

Abstract

STUDY QUESTION: Is an MRI classification of deep pelvic endometriosis (DE) able to correctly predict the risk of DE surgery. SUMMARY ANSWER: A new radiological classification, that we have called the deep pelvic endometriosis index (dPEI) classification, is accurate and reproducible to assess the extension of the disease in central and lateral compartments and well correlated with operating time, hospital stay duration and the risk of voiding dysfunction. WHAT IS KNOWN ALREADY: Few imaging classifications are currently available to predict the extent of DE to help preoperative assessment of surgical outcomes and provide the patient with objective information about the risk of surgical complications. STUDY DESIGN, SIZE, DURATION: Retrospective monocentric observational study was conducted between 01 January 2017 and 31 December 2018 and included 150 women (mean age ¼ 34.5 years, 20-52 years) with DE on MRI and who subsequently underwent surgery. PARTICIPANTS/MATERIALS, SETTING, METHODS: Two radiologists independently graded the disease according to MRI-based Enzian classification (A (rectovaginal septum and vagina locations), B (uterosacral locations) and C (rectosigmoid locations)) and a new radiological classification, that we have called the dPEI, that grades the severity of DE as: mild, moderate or severe DE. MRI findings and classification were correlated with surgical and pathological results. MAIN RESULTS AND THE ROLE OF CHANCE: MRI-based and surgical Enzian classifications were concordant for A lesions in 78.7% (118/150), for B lesions in 34.7% (52/150) and for C lesions in 82.7% (124/150). Operating time and hospital stays were longer in Group A2 (rectovaginal septum and vagina, 1-3 cm) compared to A0, B2 (uterosacral, 1-3 cm) compared to B0, C3 (rectosigmoid >3 cm) compared to C2 (rectosigmoid 1-3 cm) or C0 (P < 0.001), in severe compared to moderate DE patients, and in moderate compared to mild extensive patients (P < 0.01). Patients with vaginal or rectosigmoid involvement were respectively six and three times more likely to experience high-grade complications according to Clavien-Dindo classification than patients without vaginal or rectosigmoid disease (P < 0.001). Postoperative voiding dysfunction was correlated with A lesions (odds ratio (OR) ¼ 6.82, 95% CI 2.34-20.5), moderate or severe DE (OR ¼ 4.15, 95% CI 1.26-17.9), the presence of at least unilateral lateral pelvic involvement (OR ¼ 3.6, 95% CI 1.14-11.2, P ¼ 0.03) and C lesions (OR ¼ 2.6, 95% CI 1.03-6.8, P < 0.01). LIMITATIONS, REASONS FOR CAUTION: The study was conducted in an expert center and needs to be validated in a multicenter study. There is a limited number of patients with lateral pelvic endometriosis beyond the parietal fascia, probably due to the low prevalence of this disease presentation. WIDER IMPLICATIONS OF THE FINDINGS: MRI imaging can be used to accurately predict postoperative complications for women with DE. This may help the clinician to preoperatively inform a patient about the risks of surgery. Larger clinical studies are required to validate these results.