CLINICAL ARTICLE Accuracy of magnetic resonance imaging for diagnosis and preoperative assessment of deeply infiltrating endometriosis Luciana P. Chamié a, ⁎, Roberto Blasbalg a,b , Manoel O.C. Gonçalves c , Filomena M. Carvalho d , Maurício S. Abrão e , Ilka S. de Oliveira b,c a Fleury Medicina e Saúde, São Paulo, Brazil b Instituto de Radiologia da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil c Digimagem Medicina Diagnóstica, São Paulo, Brazil d Departamento de Patologia da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil e Departamento de Ginecologia da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil abstract article info Article history: Received 11 February 2009 Received in revised 17 March 2009 Accepted 16 April 2009 Keywords: Accuracy Deeply infiltrating endometriosis Diagnosis Magnetic resonance imaging Objective: To evaluate the accuracy of preoperative magnetic resonance imaging (MRI) findings relative to surgical presence of deeply infiltrating endometriosis (DIE). Methods: This prospective study included 92 women with clinical suspicion of DIE. The MR images were compared with laparoscopy and pathology findings. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MRI for diagnosis of DIE were assessed. Results: DIE was confirmed at histopathology in 77 of the 92 patients (83.7%). Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MRI to diagnose DIE at each of the specific sites evaluated were as follows: retrocervical space (89.4%, 92.3%, 96.7%, 77.4%, 90.2%); rectosigmoid (86.0%, 92.9%, 93.5%, 84.8%, 89.1%); bladder (23.1%, 100%, 100%, 88.8%, 89.1%); ureters (50.0%, 100%, 95.5%, 95.7%); and vagina (72.7%, 100%, 100%, 96.4%, 96.7%). Conclusion: MRI demonstrates high accuracy in diagnosing DIE in the retrocervical region, rectosigmoid, bladder, ureters, and vagina. © 2009 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. 1. Introduction Deeply infiltrating endometriosis (DIE) is defined as the presence of invasive endometriotic lesions that extend more than 5 mm from the peritoneal surface into adjacent structures, and is associated with fibrosis and muscular hyperplasia. The condition has gained increased attention from gynecologists, surgeons, and radiologists [1]. An ac- curate preoperative diagnosis of DIE, including the degree of extension into the pelvic tissues, provides invaluable information to patients and specialist physicians. Comprehensive preoperative counseling and thorough surgical planning have been shown to improve patient outcome (restitution of fertility) and well-being (re-establishment of anatomy leading to lesser or painless states) [2–4]. Diagnostic difficulties posed by the multifaceted presentation of pelvic endometriosis are highlighted by the delayed interval of 7–11 years from the beginning of symptoms to diagnosis [5–7]. Pelvic exams may be reported as normal in up to 40% of evaluations, despite disclosure of painful nodularities in the posterior cul-de-sac [6]. Diagnosis of DIE remains a major challenge in the clinical con- text and has become an area of focus for pelvic radiologists in the last two decades. Precise preoperative diagnosis of DIE requires radiologists to have meticulous knowledge of pelvic anatomy and the natural history of DIE, as well as expertise in the imaging technique employed prior to surgical and pathologic confirmation. Transvagi- nal, transrectal, and rectal endoscopic ultrasound used to diagnose DIE have shown variable accuracy depending on the anatomical sites affected and the experience of the investigators [7–10]. Despite noted limitations, magnetic resonance imaging (MRI) remains the best noninvasive method to evaluate the locations affected by pelvic endometriosis [11–15]. However, it has not been possible to general- ize the findings of these studies because some were retrospective, others included a small patient cohort, and some included women with a high prevalence of DIE. There are few data on the diagnostic accuracy of MRI for pelvic endometriosis; in addition, there have been few prospective studies and surveys do not represent all ethnic groups. Moreover, there have been some conflicting results. The aim of the present study was to evaluate the accuracy of MRI findings for the diagnosis of DIE at multiple sites, such as the retrocervical space, rectosigmoid, bladder, ureters, and vagina. 2. Materials and methods This prospective, observational, cross-sectional study was carried out in women who had a history and findings from a physical exam that were consistent with endometriosis. All women were seen at the International Journal of Gynecology and Obstetrics 106 (2009) 198–201 ⁎ Corresponding author. Rua Cincinato Braga, 232, São Paulo, Paraiso, SP 01333-910, Brazil. Tel.: +55 11 5014 6823. E-mail address: luciana.chamie@fleury.com.br (L.P. Chamié). 0020-7292/$ – see front matter © 2009 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijgo.2009.04.013 Contents lists available at ScienceDirect International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo