Three-dimensional sonorectography: a new transvaginal ultrasound technique with intrarectal contrast to assess colorectal endometriosis

Deep infiltrating endometriosis (DIE) is difficult to diagnose. Although there are symptoms associated with DIE1, they are not specific for severity or location of DIE lesions2. Some studies have shown that the performance of transvaginal sonography (TVS) could be similar, if not superior, to that of magnetic resonance imaging (MRI), which is the reference technique for diagnosing endometriosis3-5. Here we describe the technique of three-dimensional sonorectography (or rectosonography) (3D-SRG), a new TVS technique with intrarectal contrast to assess bowel endometriosis, and report on the preliminary findings of the first 50 cases that we examined using 3D-SRG. Images were obtained with a Voluson E8 (GE Healthcare Ultrasound, Milwaukee, WI, USA) equipped with a 3D transvaginal multifrequency transducer (2.9–10 MHz). Each patient underwent colorectal preparation with enema (130 mL; Normacol® adult rectal solution, Norgine, Amsterdam, The Netherlands) before the procedure. The patients slowly injected 120 mL of warm water into their rectum with a conical-tip syringe. Using distention of the bowel water to increase the contrast and improve the quality of visualization of the rectosigmoid, TVS was performed. The various layers of the intestinal wall (of rectum and sigmoid) were examined, starting from the outermost and proceeding to the innermost layer. The serosal layer appeared as a thin hyperechoic line, the muscularis propria as two hypoechoic lines and the submucosal and mucosal layers as a distinct hyperechoic line. Bowel involvement was suspected when a solid hypoechoic nodule was seen to adhere to the serosal layer and infiltrate the intestinal muscularis (Figure 1). In these cases, multiple 3D-SRG acquisitions were performed to provide better characterization of the lesions, including measurement of their diameter in three planes and determination of their volume, their anatomic extension and whether they caused bowel occlusion (Figure 2). All patients also underwent an MRI examination following preparation with enema, during which intravaginal and intrarectal contrasts (ultrasonic gel and water) were used. The practicing radiologist was blinded to the results of 3D-SRG. Subjectively, all procedures were well tolerated. Two procedures were discontinued by the examiner for technical reasons: one (2%) because bowel preparation was inadequate and one (2%) because of excess water leakage to the upper sigmoid resulting in the absence of rectal contrast. Both occurred among the first 10 procedures, which suggests a learning curve associated with performance of the procedures. Eighteen of the 20 intestinal nodules (90%) were identified among 19 patients with both MRI and 3D-SRG. One nodule (5%) was seen only with MRI and one (5%) only with 3D-SRG. The two techniques concurred in revealing no intestinal lesions in 31 examinations. In addition to providing diagnostic images and accurate information on the characteristics of intestinal DIE, these results show a strong concordance between 3D-SRG and MRI in our first 50 examinations. Further studies using DIE pathology as the gold standard are needed to assess the performance of 3D-SRG compared with TVS without contrast, 2D-TVS and MRI. C. A. Philip†, C. Bisch†, A. Coulon‡, E. Maissiat‡, P. de Saint-Hilaire†, C. Huissoud†, R. Rudigoz† and G. Dubernard*† †Department of Obstetrics and Gynecology, Croix-Rousse University Hospital of Lyon, Lyon, France; ‡Department of Radiology, Croix-Rousse University Hospital of Lyon, Lyon, France *Correspondence. (e-mail: [email protected])