Despite its presence, it is sometimes impossible to detect endometriosis with certainty on an MRI scan. In trained teams, the sensitivity of MRI for diagnosing endometriosis is around 90% [1Bazot M., Darai E., Hourani R., Thomassin I., Cortez A., Uzan S., et al. Deep pelvic endometriosis: MR imaging for diagnosis and prediction of extension of disease Radiology 2004 ;  232 (2) : 379-389 [cross-ref]

Cliquez ici pour aller à la section Références, 2Bazot M., Nassar J., Daraï E., Thomassin I., Cortez A., Buy J.N., et al. Value of sonography and MR imaging for the evaluation of deep pelvic endometriosis J Radiol 2005 ;  86 : 461-467 [inter-ref]

Cliquez ici pour aller à la section Références]. We do not currently know whether surgical treatment of endometriosis that is not visible on MRI can improve symptoms. We wanted to find out whether complete excision of endometriosis improves pain symptoms depending on the MRI diagnosis (invisible, suspicious or typical endometriosis).

An observational prospective cohort study was conducted on 161 surgical records of women who underwent laparoscopic surgical treatment for suspected endometriosis between January 2017 and May 2021 in two French university centers. These patients underwent surgery for suspected deep endometriosis based on their symptoms and/or clinical criteria and in a situation of failure or refusal of hormonal treatment. Inclusion criteria were symptoms and/or clinical examination suggestive of endometriosis, complete surgical excision by laparoscopy, pre- and post-operative questionnaires, pelvic MRI report, and histological analysis if specimen was obtained. Exclusion criteria were as follows : more than 50% missing data for the calculation of the primary endpoint (ENDOPAIN-4D score). Each woman underwent a preoperative pelvic MRI. They were divided into three groups according to the MRI findings : typical form corresponds to an endometriosis nodule or an implant with a compatible location or an endometrioma; suspected endometriosis on MRI corresponds to indirect, non-specific signs of endometriosis (pelvic adhesions, non-nodular thickening of a utero-sacral ligament, uterine laterodeviation, retroverted uterus); no endometriosis or visible benign lesions corresponds to MRIs deemed normal by the radiologist. Each woman completed a pre- and post-operative questionnaire (ENDOPAIN-4D, at least one year after surgery) to assess pain symptoms. ENDOPAIN-4D is a 4-dimensional score for the assessment of pain, dyspareunia, digestive and urinary disorders. 3 It has been previously published and validated in this population [3Puchar A., Panel P., Oppenheimer A., Du Cheyron J., Fritel X., Fauconnier A. The ENDOPAIN 4D questionnaire: a new validated tool for assessing pain in endometriosis J Clin Med 2021 ;  10 : 3216 [cross-ref]

Cliquez ici pour aller à la section Références, 4Ahmadpour P., Jahangiry L., Bani S., Iravani M., Mirghafourvand M. Validation of the Iranian version of the ENDOPAIN-4D questionnaire for measurement of painful symptoms of endometriosis J Obstet Gynaecol 2022 ;  42 (6) : 2341-2348 [cross-ref]

Cliquez ici pour aller à la section Références]. The primary outcome was assessed using analysis of covariance (ANCOVA), with MRI group and baseline values for the dependent variable entered as independent variables in a model. In this analysis, we examine the effect of MRI classification on ENDOPAIN-4D score independent of the variables of age, BMI and baseline symptom intensity. To analyse the variation in the ENDOPAIN-4D score, we used the MCID (minimal clinically important difference), which is the smallest difference in the scores of an instrument that is considered useful by patients. The MCID for the ENDOPAIN-4D score was set at 10.9. [3Puchar A., Panel P., Oppenheimer A., Du Cheyron J., Fritel X., Fauconnier A. The ENDOPAIN 4D questionnaire: a new validated tool for assessing pain in endometriosis J Clin Med 2021 ;  10 : 3216 [cross-ref]

Cliquez ici pour aller à la section Références] This was an observational, noninterventional study. Written informed consent was not required. However, all patients received information about the study and were free to participate or not. Confidentiality of patient data was guaranteed. The Comité de Protection des Personnes (Ethics Committee) of the IVème Sud-Est in France approved this study (#18/002).

Of the 161 patients included, 26 had negative, 21 suspicious and 114 typical MRI scans. Histological analysis confirmed the presence of endometriosis in 83% of women in the group without visible endometriosis on MRI, 95% in women with suspected endometriosis on MRI and 98% in women with typical MRI. (Table, significant difference). The intensity of pain symptoms was similar in each groups before surgery (Table), and significantly decreased in the typical group after surgery compared to the other two groups (Figure). The adjusted mean difference between T0 and T1 was significantly greater for patients with typical MRI than for the other two groups. For the MRI-negative group, this difference just reached the MCID, but its confidence interval included 0 (Figure), meaning that there was no significant effect of surgery in this group. The results are unchanged when patients with negative histology are excluded.

Our results show that response to surgery is proportional to MRI signal intensity. Endometriosis can take a hidden form that is not visible on MRI scans tests and for which surgical treatment may not provide effective pain relief.