Barium Enema Evaluation of Colonic Involvement in Endometriosis

Transcription

1 Gastrointestinal Imaging Original Research Faccioli et al. Barium Enema of Endometriosis Gastrointestinal Imaging Original Research Niccolò Faccioli 1 Riccardo Manfredi 1 Paride Mainardi 2 Emiliano Dalla Chiara 2 Elide Spoto 2 Luca Minelli 3 Roberto Pozzi Mucelli 1 Faccioli N, Manfredi R, Mainardi P, et al. Keywords: bowel endometriosis, double-contrast barium enema, endometriosis, intestinal endometriosis DOI: /AJR Received August 24, 2007; accepted after revision October 10, Department of Radiology, University of Verona, Policlinico G. B. Rossi, Piazzale Scuro 10, Verona, Italy. Address correspondence to N. Faccioli (nfaccioli@sirm.org). 2 Department of Radiology, Sacro Cuore Hospital, Negrar, Verona, Italy. 3 Department of Gynecology, Sacro Cuore Hospital, Negrar, Verona, Italy. AJR 2008; 190: X/08/ American Roentgen Ray Society Barium Enema Evaluation of Colonic Involvement in Endometriosis OBJECTIVE. The purpose of our study was to define the role of double-contrast barium enema (DCBE) compared with laparoscopy in the diagnosis and local staging of intestinal endometriosis. MATERIALS AND METHODS. A search of our radiology database revealed the cases of 234 women who underwent surgical resection for pelvic endometriosis with associated intestinal surgery for intestinal endometriosis. We retrospectively evaluated all preoperative DCBE images for the presence of bowel endometriosis and the number, site (rectum, sigmoid, cecum), and size of the lesions. The radiographic findings at DCBE were retrospectively correlated with those at surgical pathologic examination. RESULTS. DCBE revealed 211 intestinal lesions of bowel endometriosis in 168 (71.8%) of 234 patients with pelvic endometriosis clinically enrolled. Forty (23.8%) of the 168 women had more than one endometriotic bowel nodule (two nodules in 37 cases, three in three cases). Laparoscopy revealed 233 intestinal lesions in 174 (74.3%) of the patients. Fifty-four (31.0%) of 174 women had more than one endometriotic bowel nodule (two nodules in 49 cases, three in five cases). There was 100% correlation between the DCBE and histologic findings as far as site and size of the lesions were concerned. DCBE had a sensitivity of 88.4%, specificity of 93.0%, positive predictive value of 97.5%, negative predictive of 71.0%, and accuracy of 89.5% in the identification of bowel endometriosis. CONCLUSION. DCBE is helpful in discerning bowel wall involvement in endometriosis, enabling proper surgical planning. DCBE also appears to have a role in the management of endometriosis. P elvic endometriosis is a disease characterized by the presence of functional endometrial glands and stroma in locations outside the uterine cavity. Five percent to 27% of women with pelvic endometriosis present with endometriotic implants on the intestinal wall. Bowel endometriosis affects the rectum and the rectosigmoid junction in 70 85% of cases. Less frequent sites are the appendix, cecum, and distal ileum [1 4]. The clinical presentation of bowel endometriosis can mimic irritable bowel syndrome and, in the most severe cases, obstructive bowel disease. Symptoms are related to the site and extent of bowel implants. The implants can be single and discrete or multifocal and diffuse. The lesions also can be superficial, localized to the bowel serosa, or they can invade the subserosa and the muscularis propria, causing bowel-wall thickening due to fibrosis. Most lesions, however, are asymptomatic and are often diagnosed at surgery [5, 6]. Complete laparoscopic excision of bowel endometriosis offers good long-term symptomatic relief. To ensure complete removal of all endometriotic implants, intestinal surgery with or without intestinal resection is needed. A thorough preoperative evaluation should be performed to correctly ascertain the presence, anatomic distribution, and extent of intestinal endometriosis so that intestinal surgery can be planned. The preoperative decision to perform intestinal surgery greatly influences the type of bowel preparation, the need for a colorectal surgeon, the patient s psychological acceptance of proctectomy if needed, and the provision of signed informed consent [5]. The diagnosis of bowel endometriosis is challenging because gastrointestinal symptoms are not specific. Various imaging techniques have been proposed for the diagnosis 1050 AJR:190, April 2008

2 Barium Enema of Endometriosis of bowel endometriosis, but all have limitations. CT [7 9] and MRI [10 12] have been described, but the importance of doublecontrast barium enema (DCBE) in the preoperative evaluation of patients with suspected intestinal endometriosis has been underplayed [13 17]. DCBE has advantages such as ease of performance, low cost, and ready availability. The only disadvantage is limited luminal perspective for the examination. The purpose of this study was to define the role of DCBE, particularly in the diagnosis and local staging of intestinal endometriosis, by comparing the results with those at laparoscopy. Materials and Methods Patient Population This retrospective study was performed in the period January 2002 December 2006 (overall duration, 60 months). The study was approved by the local institutional review board, and written informed consent was obtained from all patients. Inclusion criteria for the study were clinical symptoms indicative of serious pelvic endometriosis (i.e., dyspareunia, dysmenorrhea, or chronic pelvic pain), transvaginal sonographic findings of endometriosis, gastrointestinal symptoms suggestive of bowel endometriosis (cramping abdominal pain before or during passage of stools; rectal pain coincident with menses; symptoms of acute, chronic, or intermittent bowel occlusion), and negative results of serum tests for carcino embryonic antigen, CA-125, and CA All of the women who underwent DCBE were believed to have intestinal endometriosis. DCBE was performed before surgical intervention for pelvic endometriosis and for preoperative localization before intestinal resection. Exclusion criteria were previous radiologic examination of the bowel, surgery for endometriosis, and bowel surgery other than appendectomy (39 of 234 patients). Before surgery, all patients were informed that DCBE was being evaluated for the diagnosis of bowel endometriosis. Independently of the findings at DCBE, the patients signed written informed consent forms that included consent for serosal shave excision or segmental bowel resection if endometriotic intestinal lesions were found. The study population consisted of 234 women (mean age, 31.6 years; range, years). All patients had clinical symptoms of pelvic endometriosis: 197 (84.1%) had dyspareunia; 226 (96.5%), dysmenorrhea; and 232 (99.1%), chronic pelvic pain. All patients had transvaginal sonographic findings of pelvic endometriosis. All patients also had symptoms indicative of bowel endometriosis: 211 (90.2%) had cramping abdominal pain before or during passage of stools; 207 (88.5%), rectal pain coincident with menses; and 99 (42.3%), symptoms of acute, chronic, or intermittent bowel obstruction. All women had negative results of serum tests for carcinoembryonic antigen, CA-125, and CA Surgery All patients underwent surgery with a mean interval between DCBE and surgery of 33 days (range, days). Independently from the findings at DCBE, all women underwent operative laparoscopy because of the presence of pelvic endometriosis. During laparoscopic surgery, laparo scopic segmental resection of the involved part of the intestine was scheduled for endometriotic nodules larger than 2.5 cm detected at DCBE and for multiple bowel nodules. In all the other patients, superficial excision, serosal shaving, or full-thickness disk excision was first attempted, and laparoscopic segmental resection was considered a second choice. All of the surgical procedures were performed by the same surgeon. All visible endometriotic lesions were excised in all patients; when intestinal resection was needed, it was performed by an intestinal surgeon. Histopathology Histopathologic criteria for the diagnosis of colorectal endometriosis were the presence of ectopic endometrial and stromal tissues penetrating through at least the serosa of the bowel wall. Intestinal endometriosis was confirmed with pathologic examination of all resected specimens that were histologically evaluated (serosa, muscularis propria, mucosa). Imaging Because of the chronic character of endometriosis, no attempt was made to schedule DCBE at a particular phase of the menstrual cycle. All patients suspended medical therapy with progestins or gonadotropin-releasing hormone at least 3 months before the procedure. All DCBE procedures were performed with a fluoroscopic system (Sireskop SX 40, Siemens Medical Solutions) with motorized table tilt equipped with an imaging system (Fluorospot TOP, Siemens Medical Solutions). Patient preparation included a low-residue diet for 3 days before the examination. On the day before surgery, the patients ingested 13 tablets of glycosides of senna (Pursennid, Novartis Farma) and 15 g of magnesium sulfate and followed them with 2 L of liquids to minimize the dehydration caused by the preparation. Barium in a 100% weight-to-volume ratio (Prontobario Colon, Bracco) was instilled into the rectum while the patient lay in the leftside-down lateral position. A first lateral view of the rectum was obtained. Once barium reached the hepatic flexure, the colon was drained by gravity of as much barium as possible to empty the rectal ampulla while not completely clearing the entire rectosigmoid colon. The anti cholinergic agent hyoscine N-butylbro mide (Buscopan, Boehringer Ingelheim) was then used to induce colonic hypotonia. Room air was then gently and intermittently insufflated into the colon with an insufflator (Blue Air Bulb Insufflator, Bracco). Each colonic segment was viewed in detail on spot radiographs and mid- to high-magnification digital images. The whole procedure lasted an average of 10 minutes with a maximum of 3 minutes of fluoroscopy. The procedure was well tolerated by 210 (89.7%) of the patients, and 24 (10.2%) of the patients reported mild discomfort. In no case was the procedure terminated owing to patient intolerance. Image Analysis The DCBE images were reviewed by a gastrointestinal radiologist with more than 15 years of experience. The radiologist was not aware of the clinical findings or the patient s history. Image analysis included presence of bowel endometriosis on the basis of extrinsic mass effect on the bowel wall, shortening or flattening of the bowel wall, crenulation of the mucosa, or a combination of these factors [5, 14, 17]; number of lesions; site of the lesions (rectum, sigmoid, cecum); and size of lesions. The radiographic findings on DCBE were retrospectively correlated with those at surgical pathologic examination. Statistical Analysis Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated. The Pearson correlation co efficient was used to evaluate the correlation be tween the diameter of the endometriotic nodules estimated at DCBE and the diameter measured at pathologic evaluation. Data were analyzed with the SPSS software package (release , SPSS). A value of p < 0.05 was considered statistically significant. Results Bowel endometriosis was diagnosed with DCBE in the cases of 168 of 234 clinically enrolled patients with pelvic endometriosis. At laparoscopy, abnormal findings suggestive of bowel endometriotic nodules were detected in 174 patients, and the endometriotic nature of all removed bowel lesions was confirmed at histologic examination. Bowel lesions were not recognized in 66 patients with DCBE, and 60 patients did not have bowel lesions at surgery. DCBE depicted 211 intestinal lesions in 168 (71.8%) of the patients. Forty (23.8%) of AJR:190, April

3 Faccioli et al. the 168 women had more than one endometriotic bowel nodule (two nodules in 37 cases, three nodules in three cases) (Fig. 1). At laparoscopy, however, 233 intestinal lesions were identified in 174 (74.3%) of the patients. Fifty-four (31.0%) of the 174 women had more than one endometriotic bowel nodule (two nodules in 49 cases, three nodules in five cases). At DCBE, 113 lesions were found at the rectum (Fig. 2), 85 at the sigmoid, and 13 at the cecum. No nodules were found at the terminal ileum. At laparoscopy, 117 nodules were found at the rectum, 92 at the sigmoid, and 24 at the cecum. No nodules were found at the terminal ileum. Two hundred six (97.6%) of the 211 lesions found at DCBE were confirmed at surgery and histologic examination; 194 of 198 (98.0%) of these lesions were at the rectum and sigmoid. Four lesions were not identified: two adnexal inflammatory masses originating from the ovary and involving the bowel wall (Fig. 3) and two suspected double lesions. Fourteen of 194 lesions involved the mucosa, 94 the submucosa, and 86 the serosa. Twelve (92%) of 13 lesions at the cecum were confirmed at surgery and histologic examination (Fig. 4). One lesion was a falsepositive finding. Six lesions involved the muscle wall, five the submucosa, and one lesion only the serosa. Twenty-seven lesions not preoperatively recognized were identified at surgery. Six of these lesions were in patients who had normal findings at DCBE. Fifteen of the 27 lesions were at the sigmoid and 12 at the cecum. Eleven of the 27 lesions were strictly adjacent to other lesions correctly identified at DCBE. Seven of the 27 lesions were superficial, not infiltrating the muscle wall. Five lesions were not recognized because of difficulty depicting them in tricky locations, such as the appendix and near the ileocecal valve. In one case residual feces covered a lesion at the cecum. In another case nonocclusive stenosis caused poor passage of contrast medium at the cecum, limiting the evaluation. Two lesions, at the rectum and the rectosigmoid junction, were not identified at DCBE owing to diagnostic errors. At DCBE the mean largest diameter of the endometriotic nodule was estimated to be 2.5 ± 1.2 [SD] cm (range, cm). At pathologic evaluation the mean largest diameter of the endometriotic nodule was 2.5 ± 1.1 cm (range, cm). A statistically significant positive correlation was observed for the diameter of the endometriotic nodules A Fig year-old woman with suspected intestinal implants of endometriosis. A and B, Lateral (A) and oblique (B) spot images show three endometriotic lesions exhibiting extrinsic mass effect with crenulation of contour and spiculation that are direct signs of infiltration of bowel wall (arrows). Small polypoid lesion (arrowhead) is benign tubular adenoma confirmed at surgery. Fig year-old woman with suspected intestinal implants of endometriosis and finding of rectal localization of intestinal endometriosis. Doublecontrast barium enema image shows extrinsic mass effect and spiculation (arrow) of rectal wall that appears infiltrated. Surgical and histologic findings confirmed intestinal implants of endometriosis. estimated at DCBE and the diameter measured by the pathologist (Pearson s correlation coefficient, r = 0.966; p < 0.001). Ninety-six nodules were removed with partial-thickness nodulectomy (superficial excision or serosal shaving), 120 with fullthickness nodulectomy, and 15 with bowel resection. Fourteen women underwent bowel resection. One patient underwent two bowel resections because she had two endometriotic nodules, one on the rectum and one on the cecum. A small polypoid lesion at the sigmoid recognized with DCBE was resected and at histologic examination proved to be benign tubular adenoma (Fig. 1B). Fig year-old woman with suspected intestinal implants of endometriosis. Double-contrast barium enema examination showed pathologic pelvic process involving bowel serosa at rectosigmoid junction. Finding of extrinsic mass effect and spiculation (arrows) owing to poor wall distention after air insufflation suggested endometriotic wall infiltration. Histologic examination of these alterations showed adnexal inflammation originating from ovary. The findings were true-positive in 206 cases, true-negative in 66, false-negative in 27, and false-positive in five of the cases. DCBE had a sensitivity of 88.4%, specificity of 93.0%, positive predictive value of 97.5%, negative predictive value of 71.0%, and accuracy of 89.5% in the identification of bowel endometriosis. In all women who underwent bowel surgery, there was 100% correlation between DCBE and histologic findings as far as site and size of the lesions were concerned. Discussion Implants of intestinal endometriosis are typically located on the antimesenteric edge of B 1052 AJR:190, April 2008

4 Barium Enema of Endometriosis A B Fig year-old woman with suspected intestinal implants of endometriosis and finding of cecal localization of intestinal endometriosis. A, Double-contrast barium enema image of cecum shows endometriotic localization as endoluminal filling defect (arrow) due to wide extrinsic mass effect confirmed at surgery. B, Photograph of surgical specimen shows endometrial glands and stroma infiltrating fibromuscular tissue, smooth-muscle proliferation, and fibrous reaction, resulting in thickening of wall with endometriosis. Multiple cysts (arrow) are present in perivisceral fat tissue. Mucosal sparing is typical of this disease. the bowel, and their appearance is variable. Sometimes they appear as small pigmented nodules on the peritoneum, and at other times as larger lesions with puckering of the serosa. The infiltration of the muscularis propria can determine the presence of mural thickening and associated luminal stenosis. At histologic examination, endometrial glands and stroma are found invading the bowel wall from the serosa inward. The submucosa may be involved, but an important feature is mucosa free of disease [1, 13]. Small endometriotic nodules do not directly cause symptoms but can become symptomatic with growth. Widespread nodules infiltrating the muscular layer cause pain and a range of gastrointestinal symptoms. Bowel involvement may be suspected on the basis of the clinical history and the findings at physical examination but often is a late diagnosis, sometimes found only in the course of surgery. A precise diagnosis regarding the presence, location, and extent of bowel endometriosis is useful for preoperative evaluation and surgical planning [5]. The preoperative decision to perform intestinal surgery greatly influences the type of bowel preparation, the need for a colorectal surgeon, the patient s psychological acceptance of proctectomy if needed, and provision of informed signed consent [5]. In our study, we included only patients who consented to intestinal surgery, but the rejection rate at our institution is 16%. The guidelines for preoperative assessment of patients with suspected intestinal endometriosis are not well established. Most scientific reports have considered only rectal endometriosis, not endometriosis of the entire large bowel. The evaluation is usually performed with several radiologic techniques, but the ideal technique for the diagnosis of endometriosis should include not only the rectum but also the whole colon, the cecum, and the terminal ileum, even though this location is extremely rare [18]. Only DCBE, MRI, and CT enteroclysis can depict the entire large bowel, but MRI data are inconclusive. One of the most cited articles [12] reported a sensitivity of 76.5% and specificity of 97.9% in the diagnosis of rectal involvement in women with deep infiltrating endometriosis, but the authors did not investigate the whole colon. MRI is limited in the diagnosis of bowel endometriosis because the lesions often have histologic components (e.g., fibrosis) that can alter the signal intensity pattern. Other limitations are the time required to complete each imaging study and the presence of motion artifacts. CT enteroclysis, which can depict nonspecific wall thickening, is a new and promising technique despite its limitations of radiation exposure and administration of iodinated contrast medium to women of reproductive age [7 9]. DCBE has been used in the diagnosis of endometriotic lesions of the bowel, but some authors deny its utility or do not mention it [13, 16]. At DCBE the presence of an extrinsic mass effect in association with mucosal fine crenulation is highly suggestive of the presence of endometriotic lesions involving the bowel, as previously reported and observed in all patients of our study. Appreciation of the en face appearance of the crenulated surface pattern is extremely important, because this abnormality may be the sole finding. At times, crenulation is quite striking; it can also be extremely subtle and easily overlooked by an untrained radiologist. The differential diagnoses have been reported to be many, including inflammatory lesions such as diverticulitis, pelvic inflammatory disease, and benign and malignant colonic neoplasm. In our experience, only metastatic carcinoma and pelvic abscesses mimicked the appearance of endometriosis [5]. Many surgeons [12, 19] are concerned about performing both radiologic and endoscopic investigations because these studies cannot clearly show the depth of parietal involvement, as does CT enteroclysis. That may be true, but we believe that what really matters is the preoperative choice to perform intestinal surgery in accordance with lesion size (less or more than 2.5 cm). In our study population, we recognized lesions from 0.7 to 7.2 cm. Unlike Biscaldi et al. [8, 9], we did not find it important to evaluate the infiltration of the lesion, because the surgeon had to resect all the foci in evaluating the extent of disease and intraoperatively determining the correct surgical approach based on the findings. Nevertheless, DCBE clearly depicts preoperatively whether stenosis of the lumen is present, a finding that can change the surgical approach. In our clinical practice DCBE has a role in the diagnosis of intestinal involvement because it depicts the classic signs of extrinsic involvement of the visceral wall with satisfactory accuracy. These morphologic features, together with clinical and anamnestic data, aid in formulation of a differential diagnosis from other abdominal diseases with parietal involvement. Of the 27 lesions not identified at DCBE, 12 were at the cecum, a site particularly difficult to evaluate, but a report [5] in the literature indicates that DCBE has the highest accuracy for lesions at this site. The false-negative findings at this site were due to incomplete distention, residual feces, and stenosis by muscular contraction that did not allow passage of contrast medium. Eleven endometriotic foci at the rectum and sigmoid were not correctly recognized owing to proximity to other correctly identified lesions with common fibrotic involvement of the bowel wall. In nine cases not identified at DCBE, seven lesions had superficial implants that were inapparent. Two cases were not AJR:190, April

5 Faccioli et al. effectively recognized at DCBE despite involvement of the muscle wall. Our results might have been biased because the population was selected. The findings may not apply to other patients, particularly to young otherwise healthy women with only dysmenorrhea who would be at extremely low risk of having bowel involvement. Although DCBE is not expensive and is easy to perform, there are limitations to its use. The experience of the radiologist in the diagnosis of bowel endometriosis is most important. Another limitation is that the whole abdomen cannot be examined with DCBE, because only pathologic changes in the mucosal surface are depicted. Furthermore, the origin of an extrinsic mass altering the profile of a bowel loop can be difficult to determine. We believe, however, that young women selected because of clinical suspicion of intestinal endometriosis usually do not have other pathologic conditions. This supposition was confirmed in our series because only two (0.8%) of 234 clinically selected women had other pathologic conditions that necessitated additional CT or MRI. Nevertheless, DCBE has to be preferred to CT because of its lower radiation dose and over MRI because of its lower cost. In our practice, the mean time for fluoroscopy is 3 minutes, and the effective dose is approximately 6 msv, one-third the dose of CT of the abdomen [20]. We evaluated the contribution of DCBE to the detection and staging of bowel endometriosis, assessing presence, site, and size of lesions and comparing the results with those of laparoscopy. To this aim, we selected only patients with endometriosis, which limits the statistical value of the data. The data confirm, however, that DCBE is highly accurate in the detection of intestinal endometriosis necessitating subsequent intestinal surgery. Our data show that in expert hands, findings at DCBE are predictive of bowel wall involvement in endometriosis and allow proper surgical planning. The reliability of DCBE is related to rigorous technique and operator skill. The technique is inexpensive and simple to perform and appears to have a role in the management of endometriosis. References 1. Woodward PJ, Sohaey R, Mezzetti TP. Endometriosis: radiologic pathologic correlation. RadioGraphics 2001; 21: Olive DL, Schwartz LB. Endometriosis. N Engl J Med 1993; 328: Lu PY, Ory SJ. Endometriosis: current management. Mayo Clin Proc 1995; 70: Jubanyik KJ, Comite F. Extra-pelvic endometriosis. Obstet Gynecol Clin North Am 1997; 24: Landi S, Barbieri F, Fiaccavento A, et al. Preoperative double-contrast barium enema in patients with suspected intestinal endometriosis. J Am Assoc Gynecol Laparosc 2004; 11: Prystowsky JB, Stryker SJ, Ujiki GT, et al. Gastrointestinal endometriosis: incidence and indications for resection. Arch Surg 1988; 123: La Seta F, Buccellato A, Tese L, et al. Multidetector-row CT enteroclysis: indications and clinical applications. Radiol Med 2006; 111: Biscaldi E, Ferrero S, Fulcheri E, et al. Multislice CT enteroclysis in the diagnosis of bowel endometriosis. Eur Radiol 2007; 17: Biscaldi E, Ferrero S, Remorgida V, et al. Bowel endometriosis: CT-enteroclysis. Abdom Imaging 2007;32: Zanardi R, Del Frate C, Zuiani C, et al. Staging of pelvic endometriosis using magnetic resonance imaging compared with the laparoscopic classification of the American Fertility Society: a prospective study. Radiol Med 2003; 105: Del Frate C, Girometti R, Pittino M, et al. Deep retroperitoneal pelvic endometriosis: MR imaging appearance with laparoscopic correlation. RadioGraphics 2006; 26: Chapron C, Vieira M, Chopin N, et al. Accuracy of rectal endoscopic ultrasonography and magnetic resonance imaging in the diagnosis of rectal involvement for patients presenting with deeply infiltrating endometriosis. Ultrasound Obstet Gynecol 2004; 24: Remorgida V, Ferrero S, Fulcheri E, et al. Bowel endometriosis: presentation, diagnosis, and treatment. Obstet Gynecol Surv 2007; 62: Gedgaudas RK, Kelvin FM, Thompson WM, et al. The value of the preoperative barium-enema examination in the assessment of pelvic masses. Radiology 1983; 146: Rollandi GA, Biscaldi E, DeCicco E. Double contrast barium enema: technique, indications, results and limitations of a conventional imaging methodology in the MDCT virtual endoscopy era. Eur J Radiol 2007; 61: Kinkel K, Frei KA, Balleyguier C, et al. Diagnosis of endometriosis with imaging: a review. Eur Radiol 2006; 16: Roseau G, Dumontier I, Palazzo L, et al. Rectosigmoid endometriosis: endoscopic ultrasound features and clinical implications. Endoscopy 2000; 32: Scarmato VJ, Levine MS, Herlinger H, et al. Ileal endometriosis: radiographic findings in five cases. Radiology 2000; 214: Gordon RL, Evers K, Kressel HY. Double-contrast enema in pelvic endometriosis. AJR 1982; 138: Levatter RE. Radiation risk of body CT: what to tell our patients and other questions. Radiology 2005; 234: AJR:190, April 2008