Obstructed defecation syndrome is the inability to evacuate stools from the rectum, caused by anatomic (rectocele, enterocele, intussusception) or functional (paradoxic pelvic floor contraction, eg, anismus) conditions of the posterior pelvic floor. Symptoms occur in 13% of the general female population and are highly prevalent in urogynecologic patients.1,2 Investigations are required because clinical examination alone may not reveal the underlying condition.3 Underdiagnosis can lead to ineffective treatment, increase in complications, and recurrence of symptoms after surgery.
Evacuation proctography, also known as video defecography, is considered the “reference standard” because it was the first proposed imaging technique for the assessment of posterior pelvic floor disorders.4–11 However, evacuation proctography is known to overdiagnose conditions by identifying these in asymptomatic volunteers and has significant interobserver variability so should be interpreted with caution.12–15 Furthermore, it requires ionizing radiation, bowel preparation, and is embarrassing, because the patient has to evacuate contrast in a nonprivate setting.
Magnetic resonance imaging (MRI) and ultrasonography overcome these disadvantages. Comparative studies to assess test accuracy of MRI and transperineal and endovaginal ultrasonography have been conducted using evacuation proctography as the reference standard5–11,16–24; however, we hypothesize that MRI or ultrasonography has an equal or better test accuracy than evacuation proctography.
The objective of this study was to establish the diagnostic test accuracy of evacuation proctography, MRI, and transperineal and endovaginal ultrasonography for the detection of posterior pelvic floor disorders in women with obstructed defecation syndrome to evaluate which is the best available imaging technique for diagnosis of rectocele, enterocele, intussusception, and anismus. Secondarily we aimed to identify the most patient-friendly imaging technique.
MATERIALS AND METHODS
In this prospective cohort study, between January 2014 and January 2015, consecutive women with symptoms of obstructed defecation syndrome were recruited from tertiary urogynecology or colorectal clinics in Croydon University Hospital. Obstructed defecation syndrome was defined as the inability to evacuate stools from the rectum, causing the feeling of incomplete emptying, excessive straining, or the need to digitally assist evacuation.25 Evacuation proctography and MRI proctogram were requested as part of the hospital's protocol and additional transperineal and endovaginal ultrasonography were performed as part of this study. Exclusion criteria were age younger than 18 years, inability to understand English, and lacking mental capacity. Ethical approval was obtained from the National Research Ethics Service (REC 13/LO/1665). The study was registered with clinical trials in September 2014 as a result of an oversight on our part. All women gave signed informed consent. Demographic data were collected including age, ethnicity, parity, body mass index (calculated as weight (kg)/[height (m)]2), and previous pelvic floor surgery. Time among evacuation proctography, MRI, and ultrasonography was kept as short as possible.
Evacuation proctography was performed by an experienced radiologist with a special interest in pelvic floor (A.S.). The small bowel was opacified with oral diluted barium 1 hour before the procedure and the rectum was prepared with glycerin suppositories. The rectum was filled with 120 mL barium paste (barium sulphate mixed with potato powder). The patient was sitting on a radiolucent commode with a metal ruler placed adjacent to the patient to calibrate the images for analysis. Images were recorded in the sagittal plane at rest, during contraction, straining, and evacuation of contrast.
Magnetic resonance imaging was performed with a closed 1.5-T magnet MRI scanner by specifically trained radiographers. The rectum was filled with 120 mL contrast (ultrasound gel). The patient was scanned in the supine position with the knees and hips flexed to facilitate evacuation of contrast. T2-weighted fast acquisition images were obtained simultaneously in the midsagittal and coronal planes to evaluate pelvic organ descent and pelvic floor muscles motion while the patient was instructed via headphones to perform the rest–squeeze–relaxation–strain–evacuation maneuver and to empty the rectum as completely as possible.
Pelvic floor ultrasonography, consisting of transperineal and endovaginal ultrasonography, was performed by an experienced ultrasonographer (I.M.A.v.G.) using of Profocus ultrasound scanner with the patient in a supine position with hips and knees semiflexed. No vaginal or rectal contrast was used. For transperineal ultrasonography, a convex transducer (3.5–6.0 MHz, focal range 10–135 mm) was gently placed on the perineum in a vertical position. For endovaginal ultrasonography, a high-resolution linear array transducer (6–12 MHz, focal range 3–60 mm, contact surface 65 mm) was placed in the vagina. For both transperineal and endovaginal ultrasonography, images were acquired at rest, squeeze, and maximum Valsalva. Three Valsalva maneuvers were recorded as a cineloop and the best cineloop was used for analysis. After each imaging technique, the patients were asked to give a visual analog scale (VAS) score, ranging from 0 to 10, for the level of embarrassment and discomfort of the imaging technique to assess patient's preference.
Offline analysis of images obtained by each technique was performed by two observers blinded to clinical and other imaging findings (I.M.A.v.G. and A.S. for evacuation proctography and MRI and I.M.A.v.G. and K.K. for transperineal and endovaginal ultrasonography). In case of discrepancies, final diagnosis was made by a third observer [for evacuation proctography and MRI, a radiologist (H.B.) with more than 30 years' experience in pelvic floor imaging and for transperineal and endovaginal ultrasonography, an urogynecologist (R.T.) with more than 10 years of experience in pelvic floor ultrasonography].
Predefined cutoff values were used based on existing literature. A rectocele is a bulging of the anterior rectal wall into the vagina. Its depth was measured as the maximum depth perpendicular to the expected contour of the anterior rectal wall.17,26 The presence of a rectocele on evacuation proctography and MRI was defined as a bulge of more than 20 mm. Rectoceles of less than 20 mm are often found in asymptomatic women and are therefore assumed not to be of clinical relevance.12,20 A rectocele on transperineal ultrasonography was present if its depth was more than 10 mm.6–8 The presence of a rectocele on endovaginal ultrasonography was defined similar to transperineal ultrasonography, because currently no cutoff value for rectocele using endovaginal ultrasonography has been defined. An example of a rectocele present on all four imaging techniques is provided in Appendix 1 Case 1 (available online at http://links.lww.com/AOG/A989). An enterocele is the descent of small bowel loops or rectosigmoid between the rectum and vagina on Valsalva maneuver.26 An enterocele was present on evacuation proctography and MRI when the small bowel was descending below the pubococcygeal line in the rectovaginal space.17 On transperineal ultrasonography, an enterocele was present if small bowel was visible below the superoinferior border of the symphysis pubis2 and on endovaginal ultrasonography when small bowel was visible in the region of the rectovaginal septum. An example of a enterocele present on all four imaging techniques is provided in Appendix 1 Case 2 (available online at http://links.lww.com/AOG/A989). An intussusception is an intraluminal folding of the rectal wall that extends into the rectum (grade 1), anal canal (grade 2), or externally (grade 3).26 An intussusception was defined as present on all imaging techniques if there was a full-thickness circumferential invagination of the rectal wall during straining.5 Partial and mucosal intussusceptions were classified as no intussusception, because these are often seen in asymptomatic patients and are therefore assumed not to be of clinical relevance.13 An example of an intussusception present on all four imaging techniques is provided in Appendix 1 Case 3 (available online at http://links.lww.com/AOG/A989). Anismus is a paradoxic pelvic floor contraction during attempts to evacuate.8,10,11,21 It was defined as present on all techniques if a paradoxic contraction of the puborectalis muscle during straining was visualized or as a persistent impression of the puborectalis muscle on the posterior rectal wall. An example of anismus present on all four imaging techniques is provided in Appendix 1 Case 4 (available online at http://links.lww.com/AOG/A989).
For statistical analysis, the posterior pelvic floor disorders were dichotomized into presence or absence. Because evacuation proctography has not been validated, we do not regard it as a reference standard. In the absence of a reference standard, and when no a priori consensus exists about what combination of tests would be a suitable reference standard, the method to evaluate the accuracy of multiple diagnostic tests would be latent class analysis.27 Latent class analysis combines the results of the diagnostic techniques through a statistical model to identify the true patient status. It assumes that the actual results of the techniques are imperfect observations of the true unobserved status: latent classes “healthy” and “diseased.” To handle possible conditional dependence, random effects were applied.28 The resulting model was used to estimate sensitivity, specificity, and area under the curve (AUC) with 95% bootstrap CIs.29 Only women who underwent all four imaging techniques were included in the latent class analysis, because missing values are not permitted in this type of analysis. Test accuracy is defined as follows: AUC 0.9–1.0 excellent, 0.8–0.9 good, 0.7–0.8 moderate, 0.6–0.7 fair, 0.5–0.6 poor, and below 0.5 as a useless test. In all included women, interobserver agreement was analyzed using the intraclass correlation coefficient for continuous variables and Cohen's κ for dichotomous variables both ranging from 0 to 1: 0.0–0.2 poor, 0.2–0.4 fair, 0.4–0.6 moderate, 0.6–0.8 good, 0.8–1.0 excellent agreement. Differences in measurements among the imaging techniques were assessed using the paired sample t test. Optimum cutoff values for rectocele depth were established by maximizing the Youden's index, that is, by finding the point of maximum sensitivity and specificity. To assess patient acceptability, VAS scores were compared using the Wilcoxon rank test. Sample size calculation was performed based on the expected prevalence, sensitivity, and specificity with a precision (half width of 95% CI) of 0.1 and the expected κ with a precision of 0.2 and a power of greater than 80% to detect a κ >0. Statistical analyses were performed using SPSS 22 and R 3.3.1 (package randomLCA) and a P value of <.05 was considered statistically significant.
A total of 146 consecutive women with obstructed defecation syndrome were referred for evacuation proctography. Of those, four women were not eligible for inclusion and 11 did not want to participate in the study, leaving 131 women for inclusion (Fig. 1). Mean age was 54 years (range 25–90 years), body mass index 27 (±4.9 SD), and parity 2.2 (±1.3 SD). Ethnicity was 77% Caucasian, 8% Asian, and 15% black. Previous pelvic organ prolapse surgery had been performed in 24 (18%) women, 39 (30%) had a hysterectomy, and 12 (9%) had previous surgery for obstructed defecation syndrome (six stapled transanal rectal resection, four rectopexy, one with both). One hundred fourteen women (87%) had the feeling of incomplete emptying on a weekly basis, 62 (47%) had to digitally assist evacuation either vaginally or anally at least weekly, and 62 (47%) had to strain excessively at least 50% of the time. All women underwent evacuation proctography and endovaginal and transperineal ultrasonography. In four women, MRI was contraindicated and five women had no MRI for other reasons. Pelvic floor ultrasonography, consisting of transperineal and endovaginal ultrasonography, was performed at the same time. The time difference (median) between evacuation proctography and MRI was 11.5 days (range 0–92 days), between evacuation proctography and ultrasonography 3.0 days (range 0–58 days), and between MRI and ultrasonography 8.5 days (range 0–89 days). No adverse events from performing these imaging techniques were observed during this clinical trial. The prevalence of the posterior compartment disorders diagnosed on each of the four imaging techniques is presented in Table 1.
For the diagnosis of rectocele, no significant difference in sensitivity, specificity, or AUC was found among all four imaging techniques. Estimates of diagnostic test accuracy were highest for MRI (AUC 0.79) and transperineal ultrasonography (0.85) (Table 2). Cohen's κ for interobserver agreement was highest for transperineal ultrasonography (0.72) and MRI (0.68) with no significant differences among all techniques. Intraclass correlation coefficient for interobserver reliability for rectocele depth measurements was significantly higher for transperineal ultrasonography (0.89) and MRI (0.82) as compared with evacuation proctography (0.55, P<.001) (Table 3). Mean rectocele depth was 17.1 mm (±6.3 SD), 25.6 mm (±9.8 SD), 19.5 mm (±7.9 SD), and 10.7 mm (±2.7 SD) on evacuation proctography, MRI, and transperineal and endovaginal ultrasonography, respectively. Rectocele depth measured significantly larger on MRI as compared with evacuation proctography (mean difference 9.5 mm; P<.001) and significantly smaller on endovaginal ultrasonography as compared with evacuation proctography (mean difference 7.4 mm; P<.001); however, rectocele depth on transperineal ultrasonography was comparable with evacuation proctography (mean difference 1.1 mm; P=.311). Based on the maximization of the Youden's index (highest sensitivity and specificity), when taking the 20-mm cutoff on evacuation proctography as the reference standard, the optimum cutoff value for MRI is 26.5 mm, for transperineal ultrasonography 16.3 mm, and for endovaginal ultrasonography 9.1 mm.
For the diagnosis of enterocele, no significant difference in sensitivity, specificity, or AUC was found among all four imaging techniques. Estimates of diagnostic test accuracy were highest for transperineal (AUC 0.73) and endovaginal ultrasonography (AUC 0.87) (Table 2). Cohen's κ for interobserver agreement was significantly higher for transperineal ultrasonography (0.94) and MRI (0.80) as compared with evacuation proctography (0.54, P=.01 and P<.001). Intraclass correlation coefficients for interobserver reliability for enterocele depth measurements were highest for evacuation proctography (0.61) and MRI (0.85) with no significant differences among all techniques (Table 3). The mean enterocele depth below the pubococcygeal line was 26.9 mm (±14.9 SD) on evacuation proctography, 48.3 mm (±24.6 SD) on MRI, and 35.5 mm (±20.8 SD) below the pubic bone on transperineal ultrasonography. There was no significant difference in enterocele depth between evacuation proctography and MRI (mean difference 15.8 mm; P=.067) and between evacuation proctography and transperineal ultrasonography (mean difference 0.36 mm; P=.974).
For diagnosis of intussusception, no significant difference in sensitivity, specificity, or AUC was found among all four imaging techniques. Estimates of diagnostic test accuracy were highest for evacuation proctography (AUC 0.76) and endovaginal ultrasonography (0.77) (Table 2). Cohen's κ for interobserver agreement was fair for MRI (0.37) and transperineal ultrasonography (0.22) and poor for endovaginal ultrasonography (0.10) and evacuation proctography (−0.03) with no significant differences among all techniques (Table 3).
For diagnosis of anismus, no significant difference in sensitivity, specificity, or AUC was found among all four imaging techniques. Estimates of diagnostic test accuracy were highest for endovaginal (AUC 0.95) and transperineal ultrasonography (AUC 0.78) (Table 2). Cohen's κ for interobserver agreement was highest for MRI (0.81) with transperineal ultrasonography being significantly worse (0.43, P=.02) (Table 3).
Evacuation proctography was significantly more uncomfortable (VAS 5.4) and embarrassing (VAS 6.8) compared with the other three imaging techniques (P<.001–.022) (Table 4). Transperineal and endovaginal ultrasonography were described as less embarrassing (VAS 1.8–2.1) and uncomfortable (VAS 0.7–1.8) compared with MRI and evacuation proctography (P<.001).
There is no one optimal test for the diagnosis of all conditions. On comparison of four imaging techniques in a noninferiority setting, all were shown to have similar diagnostic test accuracy. Evacuation proctography is not the best available imaging technique. Evacuation proctography had good accuracy for diagnosis of intussusception; MRI for rectocele; transperineal ultrasonography for rectocele, enterocele, and anismus; and endovaginal ultrasonography for enterocele, intussusception, and anismus. Transperineal and endovaginal ultrasonography were better tolerated than MRI and evacuation proctography.
Sensitivity of transperineal ultrasonography and MRI for diagnosis of rectocele might have been artificially increased because of the difference in predefined and optimum cutoff values. We recommend the use of our newly established optimum cutoff value for transperineal ultrasonography (15 mm), which is in agreement with Dietz et al,30 and endovaginal ultrasonography (10 mm) because the presence of the probe in the vagina restricts the rectocele bulge. Magnetic resonance imaging's excellent sensitivity for rectocele and interobserver agreement of rectocele depth are probably the result of precise tissue discrimination facilitating measurements. Magnetic resonance imaging had the lowest sensitivity for diagnosis of enterocele, which may be missed as a result of restricted pushing caused by operator dependence, supine positioning, or embarrassment (Appendix 1 Case 5, available online at http://links.lww.com/AOG/A989). However, MRI demonstrated excellent interobserver agreement for enterocele and was best at distinguishing hernia sac content, for example, peritoneocele cannot be identified on evacuation proctography because it does not contain contrast-filled bowel. Diagnosis of intussusception is enhanced on evacuation proctography by barium contrast providing good discrimination between the rectal wall and surrounding tissue and on endovaginal ultrasonography by the probe being close to the area of interest. Intussusceptions could be missed as a result of obscuration by the presence of enterocele, incomplete emptying, or absence of evacuation (Appendix 1 Case 6, available online at http://links.lww.com/AOG/A989). Ultrasonography enables repetition of dynamic maneuvers, which makes it easier to detect anismus. Interobserver agreement of well-trained operators ranged from poor to excellent for all techniques, indicating extensive training is required for all four imaging methods to enhance diagnostic test accuracy.
Magnetic resonance imaging showed similar high accuracy for diagnosis of rectocele but lower accuracy for enterocele, intussusception, and anismus as compared with other studies.11,17,19–23 Magnetic resonance imaging is better at distinguishing between mucosal and full-thickness intussusception than evacuation proctography5; therefore, evacuation proctography may have picked up mucosal intussusception as full thickness, decreasing the sensitivity of MRI. A similarly good accuracy of transperineal ultrasonography for the diagnosis of rectocele, enterocele, and anismus and equally poor sensitivity for intussusception was found as compared with other studies.6–10,24 Our study showed endovaginal ultrasonography to have equally good accuracy for diagnosis of rectocele and intussusception,24 but better accuracy for enterocele and anismus as compared with Hainsworth et al.24 This could be explained by the use of latent class analysis. These conditions being missed by evacuation proctography did not reduce sensitivity of endovaginal ultrasonography, but more accurately reduced sensitivity of evacuation proctography. Our results regarding evacuation proctography cannot be compared with other studies because in most evacuation, proctography is used as reference standard, hence 100% test accuracy. Similar to Steensma et al,7 ultrasonography was the patient's preferred technique. Evacuation proctography was significantly more embarrassing than the other techniques probably as a result of lack of privacy during evacuation. In summary, other studies assessing test accuracy in a similar population using similar cutoff values found either equal or better test accuracy compared with our findings and similar patient acceptability, meaning our results are generalizable. Our results are in agreement with studies reporting MRI and transperineal ultrasonography to be equally good as evacuation proctography.8,17–19 This study adds to the evidence that evacuation proctography should no longer be used as the reference standard for assessment of posterior pelvic floor disorders.
Strengths of this study are its prospective design, adequate sample size, recruitment of consecutive women, comparison of four different imaging techniques in the same patient, minimal time difference among techniques, blinded analysis of all images by two observers, and the use of predefined cutoff values. Limitations include different patient position, use of contrast, evacuation phase, and cutoff values used for the four imaging techniques. The Valsalva maneuver was not standardized; therefore, patients’ effort could have varied among the techniques and might have caused differences in diagnostic outcomes.
In this study, no significant differences in test accuracy were found; therefore, logical choices from clinical reasoning (risk of radiation, availability, costs, expertise, and patient acceptability) may lead our way in decisions for or against one or another imaging technique. Pelvic floor ultrasonography (the combination of transperineal and endovaginal ultrasonography) could be performed at the time of consultation, enabling correlation to symptoms and examination. We suggest ultrasonography might be used for the initial assessment of women with obstructed defecation syndrome, because it has good accuracy for all conditions and good patient acceptability, but should only be performed by an experienced ultrasonographer after extensive training. This is in agreement with the recent report on the terminology for female anorectal dysfunction.25 If ultrasonography is inconclusive and symptoms persist despite conservative treatment, evacuation proctography or MRI could be considered before embarking on surgery. We recommend evacuation proctography if intussusception is suspected and MRI in women of childbearing age to avoid radiation.
1. Whitcomb EL, Lukacz ES, Lawrence JM, Nager CW, Luber KM. Prevalence of defecatory dysfunction in women with and without pelvic floor disorders. J Pelvic Med Surg 2009;15:179–87.
2. Guzman Rojas R, Kamisan Atan I, Shek KL, Dietz HP. The prevalence of abnormal posterior compartment anatomy and its association with obstructed defecation symptoms in urogynecological patients. Int Urogynecol J 2016;27:939–44.
3. Kelvin FM, Hale DS, Maglinte DD, Patten BJ, Benson JT. Female pelvic organ prolapse: diagnostic contribution of dynamic cystoproctography and comparison with physical examination. AJR Am J Roentgenol 1999;173:31–7.
4. Mahieu P, Pringot J, Bodart P. Defecography: I. Description of a new procedure and results in normal patients. Gastrointest Radiol 1984;9:247–51.
5. Dvorkin LS, Hetzer F, Scott SM, Williams NS, Gedroyc W, Lunniss PJ. Open-magnet MR defaecography compared with evacuation proctography in the diagnosis and management of patients with rectal intussusception. Colorectal Dis 2004;6:45–53.
6. Perniola G, Shek C, Chong CC, Chew S, Cartmill J, Dietz HP. Defecation proctography and translabial ultrasound in the investigation of defecatory disorders. Ultrasound Obstet Gynecol 2008;31:567–71.
7. Steensma AB, Oom DM, Burger CW, Schouten WR. Assessment of posterior compartment prolapse: a comparison of evacuation proctography and 3D transperineal ultrasound. Colorectal Dis 2010;12:533–9.
8. Martellucci J, Naldini G. Clinical relevance of transperineal ultrasound compared with evacuation proctography for the evaluation of patients with obstructed defaecation. Colorectal Dis 2011;13:1167–72.
9. Weemhoff M, Kluivers KB, Govaert B, Evers JL, Kessels AG, Baeten CG. Transperineal ultrasound compared to evacuation proctography for diagnosing enteroceles and intussusceptions. Int J Colorectal Dis 2013;28:359–63.
10. Beer-Gabel M, Carter D. Comparison of dynamic transperineal ultrasound and defecography for the evaluation of pelvic floor disorders. Int J Colorectal Dis 2015;30:835–41.
11. Foti PV, Farina R, Riva G, Coronella M, Fisichella E, Palmucci S, et al. Pelvic floor imaging: comparison between magnetic resonance imaging and conventional defecography in studying outlet obstruction syndrome. Radiol Med 2013;118:23–39.
12. Shorvon PJ, McHugh S, Diamant NE, Somers S, Stevenson GW. Defecography in normal volunteers: results and implications. Gut 1989;30:1737–49.
13. Freimanis MG, Wald A, Caruana B, Bauman DH. Evacuation proctography in normal volunteers. Invest Radiol 1991;26:581–5.
14. Müller-Lissner SA, Bartolo DC, Christiansen J, Ekberg O, Goei R, Höpfner W, et al. Interobserver agreement in defecography—an international study. Z Gastroenterol 1998;36:273–9.
15. Goei R, van Engelshoven J, Schouten H, Baeten C, Stassen C. Anorectal function: defecographic measurement in asymptomatic subjects. Radiology 1989;173:137–41.
16. Healy JC, Halligan S, Reznek RH, Watson S, Bartram CI, Phillips R, et al. Dynamic MR imaging compared with evacuation proctography when evaluating anorectal configuration and pelvic floor movement. AJR Am J Roentgenol 1997;169:775–9.
17. Lienemann A, Anthuber C, Baron A, Kohz P, Reiser M. Dynamic MR colpocystorectography assessing pelvic-floor descent. Eur Radiol 1997;7:1309–17.
18. Gufler H, Laubenberger J, DeGregorio G, Dohnicht S, Langer M. Pelvic floor descent: dynamic MR imaging using a half-Fourier RARE sequence. J Magn Reson Imaging 1999;9:378–83.
19. Kelvin FM, Maglinte DD, Hale DS, Benson JT. Female pelvic organ prolapse: a comparison of triphasic dynamic MR imaging and triphasic fluoroscopic cystocolpoproctography. AJR Am J Roentgenol 2000;174:81–8.
20. Vitton V, Vignally P, Barthet M, Cohen V, Durieux O, Bouvier M, et al. Dynamic anal endosonography and MRI defecography in diagnosis of pelvic floor disorders: comparison with conventional defecography. Dis Colon Rectum 2011;54:1398–404.
21. Pilkington SA, Nugent KP, Brenner J, Harris S, Clarke A, Lamparelli M, et al. Barium proctography vs magnetic resonance proctography for pelvic floor disorders: a comparative study. Colorectal Dis 2012;14:1224–30.
22. Fiaschetti V, Pastorelli D, Squillaci E, Funel V, Rascioni M, Meschini A, et al. Static and dynamic evaluation of pelvic floor disorders with an open low-field tilting magnet. Clin Radiol 2013;68:e293–300.
23. Faucheron JL, Barot S, Collomb D, Hohn N, Anglade D, Dubreuil A. Dynamic cystocolpoproctography is superior to functional pelvic MRI in the diagnosis of posterior pelvic floor disorders: results of a prospective study. Colorectal Dis 2014;16:O240–7.
24. Hainsworth AJ, Solanki D, Hamad A, Morris SJ, Schizas AM, Williams AB. Integrated total pelvic floor ultrasound in pelvic floor defaecatory dysfunction. Colorectal Dis 2017;19:O54–65.
25. Sultan AH, Monga A, Lee J, Emmanuel A, Norton C, Santoro G, et al. An International Urogynecological Association (IUGA)/International Continence Society (ICS) joint report on the terminology for female anorectal dysfunction. Int Urogynecol J 2017;28:5–31.
26. Mellgren A, Bremmer S, Johansson C, Dolk A, Udén R, Ahlbäck SO, et al. Defecography. Results of investigations in 2,816 patients. Dis Colon Rectum 1994;37:1133–41.
27. Rutjes AW, Reitsma JB, Coomarasamy A, Khan KS, Bossuyt PM. Evaluation of diagnostic tests when there is no gold standard. A review of methods. Health Technol Assess 2007;11:iii, ix–51.
28. Qu Y, Tan M, Kutner MH. Random effects models in latent class analysis for evaluating accuracy of diagnostic tests. Biometrics 1996;52:797–810.
29. van Smeden M, Naaktgeboren CA, Reitsma JB, Moons KG, de Groot JA. Latent class models in diagnostic studies when there is no reference standard—a systematic review. Am J Epidemiol 2014;179:423–31.
30. Dietz HP, Zhang X, Shek KL, Guzman RR. How large does a rectocele have to be to cause symptoms? A 3D/4D ultrasound study. Int Urogynecol J 2015;26:1355–9.
Supplemental Digital Content
© 2017 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.