| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiology (M.P.T., A.C.D., J.S.) and Clinical Epidemiology and Biostatistics (M.G.W.D., P.M.M.B.), Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands; Departments of Radiology (R.G.H.B.) and Surgery (C.G.M.I.B.), University Hospital Maastricht, Maastricht, the Netherlands; and Department of Radiology, Onze Lieve Vrouwe Gasthuis, Amsterdam, the Netherlands (V.P.M.v.d.H.). From the 2003 RSNA Annual Meeting. Supported by grant 945-01-013 of the Netherlands Organization for Health Research and Development. Received July 2, 2004; revision requested September 2; revision received October 6; accepted November 5. Address correspondence to M.P.T. (e-mail: m.p.terra{at}amc.uva.nl).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: The medical ethics committees of both participating hospitals approved the study, and informed consent was obtained. Thirty patients (23 women, seven men; mean age, 58.7 years; range, 37–78 years) with fecal incontinence underwent MR imaging with both endoanal and external phased-array coils. MR images were evaluated by three radiologists with different levels of experience for external and internal anal sphincter defects. Measures of inter- and intraobserver agreement of both MR imaging techniques and of differences between both imaging techniques were calculated.
RESULTS: Both MR imaging techniques did not significantly differ in the depiction of external (P > .99) and internal (P > .99) anal sphincter defects. The techniques corresponded in 25 (83%) of 30 patients for the depiction of external anal sphincter defects and in 28 (93%) of 30 patients for the depiction of internal anal sphincter defects. Interobserver agreement was moderate to good for endoanal MR imaging and poor to fair for external phased-array MR imaging. Intraobserver agreement ranged from fair to very good for both imaging techniques.
CONCLUSION: External phased-array MR imaging is comparable to endoanal MR imaging in the depiction of clinically relevant anal sphincter defects. Because of the weak interobserver agreement, both MR imaging techniques can be recommended in the diagnostic work-up of fecal incontinence only if sufficient experience is available.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Initially, most patients with fecal incontinence will be treated with conservative therapy, which includes medical treatment and physiotherapy (ie, electric stimulation and/or pelvic floor muscle training with biofeedback) (9–12). Patients in whom conservative treatment has failed may be candidates for surgery. Those with a defect of the external anal sphincter will be selected for repair of the sphincter ring (anterior anal sphincter repair) (13–16). No surgical option is available for patients with an isolated disruption of the internal anal sphincter.
Imaging has a central position in the assessment of the integrity of the external and internal anal sphincter muscles, since physical examination is not reliable in the detection of external and internal anal sphincter defects (17) and electromyography is too painful and precludes anatomic assessment of the entire sphincter complex (18,19). Further, with anal manometry one is not able to distinguish between sphincter dysfunction due to loss of sphincter integrity or neuropathy (20).
At present, two endoanal imaging techniques are used: endoanal ultrasonography (US) and endoanal magnetic resonance (MR) imaging. Both imaging techniques have been shown to be accurate in mapping defects of the external anal sphincter (21–28). Authors of several studies have investigated the diagnostic accuracy of both techniques in the depiction of external and internal anal sphincter defects. The reported results of these studies vary. Some of the variability can be attributed to the differences in study design and levels of experience (29–32). Both endoanal US and endoanal MR imaging have been shown to be useful in the preoperative identification of patients who may benefit from surgical treatment. Yet these two modalities have two important disadvantages. Both endoluminal techniques are used primarily at specialized centers because a dedicated device is necessary. In addition, the introduction of the endoluminal probe or coil leads to discomfort.
These two disadvantages of endoluminal techniques could be overcome with the use of external phased-array coils. Recent studies of MR imaging in depicting anal anatomy, perianal disease, and rectal tumors have demonstrated that MR imaging with external phased-array coils is accurate in demonstrating the anal anatomy and the presence and extent of anorectal disease (33–37). To our knowledge, no study has evaluated the diagnostic accuracy and reproducibility of MR imaging with external phased-array coils in detecting external and internal anal sphincter defects in patients with fecal incontinence.
If external phased-array MR imaging is comparable to endoanal MR imaging in the depiction of external and internal anal sphincter defects, external phased-array MR imaging is preferable in the diagnostic work-up of patients with fecal incontinence as this technique is more widely available and less invasive.
Thus, the purposes of our study were to prospectively compare external phased-array MR imaging with endoanal MR imaging in depicting defects of the external and internal anal sphincter in patients with fecal incontinence and to prospectively evaluate the observer reproducibility in the detection of the external and internal anal sphincter defects with MR imaging techniques.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Medical History and Endoanal US
Medical history was obtained from all patients and comprised the duration of fecal incontinence, the degree of fecal incontinence, and possible risk factors for anal sphincter trauma. The degree of fecal incontinence was assessed according to the scoring system of Vaizey et al (38). This scoring system contains items about the type (gas, fluid, or solid) and frequency of fecal incontinence and additional items that address alteration in lifestyle, need to wear a pad or plug, use of constipating medication, and presence of urge incontinence. The total Vaizey incontinence score ranges from 0 (complete continence) to 24 (complete incontinence).
Possible risk factors for anal sphincter trauma comprised prior anorectal surgery (surgery for anal fistulas, hemorrhoids, anal fissures, or anal sphincter repair) (6–8) and complicated vaginal delivery (high-birth-weight infant, long second stage of labor, instrumental delivery, episiotomy, or breech delivery) (3–5,39–42).
Endoanal US was performed by four residents of the gastroenterologic department at one hospital and by two residents of the surgical department at the other hospital. The residents were experienced in obtaining and evaluating endoanal US findings (each resident with approximately 1 year of experience) and at both hospitals were supervised by a surgeon with considerable experience in endoanal US (both surgeons with at least 10 years of experience). Endoanal US was performed in both hospitals according to a standard procedure. A US scanner (model 3535, Bruel and Kjaer, Gentfofte, Denmark; SDD-2000, Multiview Aloka, Tokyo, Japan) was used with a radial endoscopic probe and a 7.5-MHz transducer. The endoscopic probe was introduced into the anus to the level of the anorectal verge, with the patient in the left lateral position, and was slowly withdrawn. The integrity of the external and internal anal sphincter was assessed. A defect of the external or internal anal sphincter was defined by a discontinuity of the muscle ring (anatomic defect) and/or was characterized by the loss of normal architecture, with an area of amorphous texture that usually has low reflectiveness (functional defect, scar tissue).
After testing, all patients underwent physiotherapy of the pelvic floor. Patients in whom physiotherapy failed and with an external anal sphincter defect at endoanal US and/or endoanal MR imaging were referred for anterior anal sphincter repair.
MR Imaging
Endoanal and external phased-array MR imaging examinations were performed successively according to the standard procedure with the same system. At one hospital, endoanal MR imaging was performed first, followed by external phased-array MR imaging. At the other hospital, the order of the performance of MR imaging techniques was reversed. Imaging parameters were optimized for each MR imaging system and coil on the basis of extensive previous experience.
Endoanal MR imaging.—At both hospitals, endoanal MR imaging was performed with a 1.5-T MR unit (Horizon Echospeed, GE Medical Systems, Milwaukee, Wis; or Gyroscan ACS-NT, Philips Medical Systems, Best, the Netherlands) and a dedicated endoanal coil with a diameter of 19 mm. All patients were asked to fast 4 hours prior to MR imaging examinations to reduce artifacts from bowel peristalsis. Bowel relaxants (1 mL of butylscopolamine bromide [20 mg/mL, Buscopan; Boehringer, Ingelheim, Germany] or 1 mg of glucagon hydrochloride [Glucagen, Bagsvaerd, Denmark]) were used at one of the institutions. At both hospitals, the endoanal coil was covered with a condom, and after application of a lubricant, it was inserted into the anal canal in a left lateral position. After positioning of the coil, the patients were turned in the supine position, and supportive pads were used to stabilize the position of the endoanal coil.
At one institution, the following T2-weighted fast spin-echo sequence was used: repetition time msec/echo time msec of 2500/70, echo train length of 10, 10 x 10-cm (transverse) and 16 x 16-cm (coronal) field of view, 256 x 512 imaging matrix, 3-mm section thickness, 0.3-mm intersection gap, and two signals acquired. At the other institution, the following T2-weighted fast spin-echo sequence was used: 3000/80, echo train length of 25, 20 x 20-cm field of view (transverse and coronal), 256 x 512 imaging matrix (512 x 512 reconstruction matrix), 3-mm section thickness, 0.3-mm intersection gap, and four signals acquired. Transverse and coronal images with a section orientation perpendicular and parallel to the anal sphincter and the endoanal coil were obtained. Each of the 15 patients, within a hospital, underwent an identical MR imaging protocol. The total imaging time for endoanal MR imaging was 15 minutes.
External phased-array MR imaging.—At the hospital at which endoanal MR imaging was performed first, both the external phased-array coil and the endoanal coil were positioned at the beginning of the procedure. External phased-array MR imaging was performed after removal of the endoanal coil. At the other center, external phased-array MR imaging was performed before endoanal MR imaging, with subsequent introduction of the endoanal coil after completion of the external phased-array MR imaging sequences. All patients were placed in the supine position, with the pelvis centered at the proximal end of a posterior phased-array spine coil in the feet-first position. At one institution, a body phased-array coil was placed anteriorly; at the other institution, a quadrature phased-array coil was placed anteriorly.
At one institution, external phased-array MR imaging was performed by using the following T2-weighted fast spin-echo sequence: 2500/70, echo train length of 10, 30 x 30-cm field of view, 256 x 256 (transverse) and 256 x 512 (coronal) imaging matrix, 3-mm section thickness, 0.3-mm intersection gap, and two signals acquired. At the other institution, the imaging parameters were 3000/80, echo train length of 25, 20 x 20-cm field of view, 256 x 512 imaging matrix (512 x 512 reconstruction matrix, transverse and coronal), 3-mm section thickness, 0.3-mm intersection gap, and six signals acquired. Transverse and coronal images with section orientation perpendicular and parallel to the long axis of the anal canal were obtained. Each of the 15 patients, within a hospital, underwent an identical MR imaging protocol. The total imaging time for external phased-array MR imaging was 15 minutes.
Image Analysis
Three radiologists (observer A, R.G.H.B.; observer B, J.S.; and observer C, V.P.M.v.d.H.), working at different hospitals, read the MR images. All three radiologists have considerable experience with abdominal MR imaging (observer A, B, and C with 9, 11, and 7 years, respectively). Each observer had a different level of experience in evaluating the anorectum, pelvis, and pelvic floor with both MR imaging techniques. Observer A was more experienced with external phased-array MR imaging (at least 600 examinations) than with endoanal MR imaging (approximately 50 examinations). Observer B was familiar with external phased-array MR imaging (approximately 500 examinations) but was more experienced with endoanal MR imaging (approximately 1000 examinations). The experience level of observer C was similar for both MR imaging techniques (approximately 200 examinations with each technique).
To compare both MR imaging techniques and to evaluate interobserver agreement, all observers evaluated separately the images of the endoanal MR imaging examinations and of the external phased-array MR imaging examinations of all patients. The observers evaluated both MR imaging techniques with at least 6 weeks in between to avoid recall bias.
To evaluate the intraobserver agreement for both MR imaging techniques, all observers reevaluated in random order images from 15 examinations of each MR imaging technique. All observers performed the second reading after an interval of at least 4 weeks to avoid recall bias. All observers were blinded to their own and each other's results and were not aware of endoanal US findings and the medical history of the patients except for age, sex, and presence of fecal incontinence.
The MR imaging examinations were reviewed on a workstation (Impax SP4 SU4 DS3000, AGFA, Mortsel, Belgium; Easy Vision Workstation, Philips Medical Systems; or eFilm workstation, version 1.5.3, Merge eFilm, Toronto, Ontario, Canada).
The radiologists evaluated and recorded the quality of the MR images on the basis of the presence of artifacts and the identification of external and internal anal sphincter as either adequate (moderate to good) or inadequate (poor) for interpretation. Artifacts were classified as absent, motion (blurriness involving the complete image), peristalsis (blurriness in the phase-encoding direction), susceptibility artifacts, and artifacts related to the inner structure of the coil and/or suboptimal positioning of the coil. Identification of the external and internal anal sphincter was scored as adequate (moderate to good) or inadequate (poor) on the basis of the possibility to evaluate the internal structure of the external anal sphincter and the internal anal sphincter, respectively. The radiologists also noted the presence and location of the defects of the external and internal anal sphincter. A defect was defined as a discontinuity of the muscle ring (anatomic defect) and/or was recognized by a hypointense deformation of the normal pattern of the muscle layer owing to the replacement of muscle cells by fibrous tissue (functional defect, scar tissue) (28).
The extent of the defect was axially indicated in terms of clock positions (1- through 12-o'clock positions, with 12-o'clock position indicating anterior, 3-o'clock position indicating left lateral, 6-o'clock position indicating posterior, and 9-o'clock position indicating right lateral positions) and longitudinally in millimeters from the lower edge of the external anal sphincter (ie, lower edge of the anal sphincter). A defect, as defined by the surgeons working in both participating hospitals, was considered relevant when it comprised at least 30° ("1 hour") of the circumference of the sphincter ring and extended at least 5 mm in longitudinal direction. An independent research fellow (M.P.T.) compared findings of endoanal MR imaging with those of external phased-array MR imaging. To ensure that the same finding was depicted with both techniques, the axial and longitudinal extent of the defect was taken into account.
Statistical Analysis
A t test for independent samples (43) was used to test for the difference in age between men and women. P < .05 was considered to indicate a statistically significant difference. The external phased-array MR imaging readings were compared with endoanal MR imaging readings. A majority rule (at least two of three observers) was used for the data regarding the following parameters of interest: image quality, presence of artifacts, identification of the external and internal anal sphincter muscle, and presence of an external or internal anal sphincter defect.
We used the McNemar test (43) to test for differences between both MR imaging techniques in the depiction of anal sphincter defects. P < .05 was considered to indicate a significant difference. Corresponding and discrepant diagnoses of external and internal anal sphincter defects were noted. To check the plausibility of the existence of external and/or internal anal sphincter defects, MR imaging findings were compared to the available data from medical history (n = 30), endoanal US (n = 30), and surgery (n = 3). Only patients with a failure of pelvic floor physiotherapy and an external anal sphincter defect at endoanal US and/or endoanal MR imaging were candidates for surgery. Some of these patients refused surgery. In others, the surgeon believed that the patient would not benefit from surgical repair of the external anal sphincter.
Cohen 
statistics (43) with 95% confidence intervals were used to express inter- and intraobserver agreement for endoanal MR imaging and external phased-array MR imaging. values were interpreted as follows: < 0.20, poor agreement; = 0.20–0.40, fair; = 0.41–0.60, moderate; = 0.61–0.80, good; = 0.81–1.00, very good (43).
We used SPSS for Windows (version 11.5; SPSS, Chicago, Ill) and StatXact for Windows (version 3.0; Cytel Software, Cambridge, Mass) to perform statistical analysis of our data.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Anterior anal sphincter repair was performed in three patients (two women), and in these patients an external anal sphincter defect was identified by the surgeon.
Image Quality
Endoanal MR imaging.—Image quality of endoanal MR imaging was adequate in all patients. The identification of the external anal sphincter was adequate in all patients. The identification of the internal anal sphincter was adequate in 29 patients and inadequate in one patient.
Motion artifacts were present in 11 patients but did not lead to an inadequate identification of the external anal sphincter in any of these patients; in only one patient, they led to an inadequate identification of the internal anal sphincter. Artifacts related to the inner structure of endoanal coil and/or suboptimal positioning of the endoanal coil were depicted in four patients but did not lead to an inadequate identification of the sphincter muscles.
External phased-array MR imaging.—The quality of external phased-array MR imaging was scored as adequate in all patients. The identification of the external anal sphincter was adequate in all patients. The identification of the internal anal sphincter was adequate in 29 patients and inadequate in one patient. No artifacts were found with external phased-array MR imaging.
Depiction of Anal Sphincter Defects
Endoanal MR imaging.—With endoanal MR imaging, 11 external and two internal anal sphincter defects were depicted. In nine female patients, an isolated external anal sphincter defect was depicted, and in two patients (one woman) both the external and internal anal sphincter were disrupted. No isolated internal anal sphincter defects were found.
All isolated external anal sphincter defects were located at the anterior part of the external anal sphincter. Patients with an isolated external anal sphincter defect underwent prior anorectal surgery (n =1), had one or more risk factors during vaginal delivery (n = 6), underwent anorectal surgery and had one risk factor during vaginal delivery (n = 1), or had no known risk factors for anal sphincter damage (n = 1) (Table 2). Of the two combined external and internal anal sphincter defects, one was located at the anterior part and one was located at the posterior part of the anal sphincter complex. One of these patients had undergone prior anorectal surgery, and the other patient had three risk factors during vaginal delivery (Table 2).
|
External phased-array MR imaging.—With external phased-array MR imaging, 10 external anal sphincter defects and two internal anal sphincter defects were identified. In eight female patients, an isolated external anal sphincter defect was found; in two patients (one woman),both external and internal anal sphincters were disrupted. In none of the patients was an isolated internal anal sphincter defect depicted.
Isolated external anal sphincter defects were depicted at the anterior part in seven patients and in one patient at the posterior part of the external anal sphincter. Patients with an isolated external anal sphincter defect had one or more risk factors during vaginal delivery (n = 7) or underwent anorectal surgery and had one risk factor during vaginal delivery (n = 1) (Table 3).
|
Endoanal US helped confirm nine of the external anal sphincter defects and one of the internal anal sphincter defects depicted at external phased-array MR imaging (Table 3). In three patients who underwent surgery, an external anal sphincter defect was identified by the surgeon. Two of these defects were also depicted with external phased-array MR imaging.
External phased-array MR imaging compared with endoanal MR imaging.—External phased-array MR imaging and endoanal MR imaging did not differ significantly in the assessment of external (P > .99) or internal (P > .99) anal sphincter defects. In the assessment of external anal sphincter defects, external phased-array MR imaging results corresponded to those of endoanal MR imaging in 25 (83%) of 30 patients (Table 4) (Figs 1,2). In the assessment of internal anal sphincter defects, external phased-array MR imaging results corresponded to those of endoanal MR imaging in 28 (93%) of 30 patients.
|
|
|
|
|
All discrepant diagnoses regarding the presence of anal sphincter defects were scored at the anterior part of the external (Figs 3, 4) and internal anal sphincters. Among the discrepant diagnoses for external anal sphincter defects, endoanal US helped confirm two of three defects depicted with endoanal MR imaging and two depicted with external phased-array MR imaging. Among the two discrepant diagnosis for internal anal sphincter defects, endoanal US helped confirm only the defect depicted with endoanal MR imaging.
|
|
|
|
Reproducibility
Interobserver agreement for endoanal MR imaging.—For the detection of external anal sphincter defects, there was agreement between observers A and B in 23 (77%) of 30 patients, observers A and C in 22 (73%) of 30 patients, and observers B and C in 23 (77%) of 30 patients.
For the detection of internal anal sphincter defects, there was agreement between observers A and B in 26 (87%) of 30 patients, observers A and C in 28 (93%) of 30 patients, and observers B and C in 26 (87%) of 30 patients.
In all patients except one, discrepant diagnoses between observers regarding the presence of an anal sphincter defect were made at the anterior part of the anal sphincter complex. values with 95% confidence intervals for the detection of external and internal anal sphincter defects at endoanal MR imaging are shown in Table 5.
|
For the detection of internal anal sphincter defects, there was agreement between observers A and B in 25 (84%) of 30 patients, observers A and C in 27 (90%) of 30 patients, and observers B and C in 24 (80%) of 30 patients.
Discrepant diagnoses between the observers regarding the presence of an anal sphincter defect were made in all patients except one at the anterior part of the anal sphincter complex. values with 95% confidence intervals for the detection ofexternal and internal anal sphincter defects at external phased-array MR imaging are shown in Table 6.
|
For the detection of internal anal sphincter defects with endoanal MR imaging, there was agreement between the first and second reading for observers A, B, and C in 14 (93%) of 15 patients, 15 (100%) of 15 patients, and 13 (87%) of 15 patients, respectively.
Discrepant diagnoses within observations of the observers regarding the presence of external and internal anal sphincter defects were made in all patients at the anterior part of the anal sphincter complex. values with 95% confidence intervals for the detection of external and internal anal sphincter defects at endoanal MR imaging are shown in Table 7.
|
For the detection of internal anal sphincter defects with external phased-array MR imaging, there was agreement between the first and second reading for observers A, B, and C in 15 (100%) of 15 patients, 12 (80%) of 15 patients, and 14 (93%) of 15 patients, respectively.
Discrepant diagnoses within observations of the observers regarding the presence of external and internal anal sphincter defects were made in all patients except one at the anterior part of the anal sphincter complex. values with 95% confidence intervals for the detection of external and internal anal sphincter defects at external phased-array MR imaging are shown in Table 8.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
A number of potential limitations of this study should be taken into account. The prevalence of external and internal anal sphincter defects detected by the observers was not high in this series. We found a defect of the external anal sphincter in about one-third of the patients and that of the internal anal sphincter in only two patients. These findings reflect the daily clinical practice, where external and internal anal sphincter defects are present in only some of the patients with fecal incontinence (44), which limits the power of the study to detect differences.
Although all participating radiologists had considerable experience with abdominal MR imaging, each observer had a different level of experience in evaluating the anorectum, pelvis, and pelvic floor with each of the MR imaging techniques. Therefore, we used a majority rule for the depiction of external and internal anal sphincter defects with both MR imaging techniques.
We did not use surgery as the reference standard since surgery is not infallible. Surgery (anterior anal sphincter repair) is performed only for lesions of the anterior external anal sphincter, while in daily practice patients also present for imaging with lesions at other locations (although, only few patients in this study). Moreover, it would not have been possible, for ethical reasons, to verify all lesions during surgery, as patients with a false-positive external anal sphincter defect at MR imaging would have undergone unjustified surgical intervention.
Endoanal MR imaging and endoanal US have previously been shown to be accurate in the depiction of external and internal anal sphincter defects (21–28). However, we realized that both imaging techniques are not suitable as a reference standard in this study, since findings of previous comparison studies have shown disagreement between these techniques in some patients (29,30).
Since both surgery and imaging are not perfect as a reference standard, we performed a comparison study between an experimental technique (ie, external phased-array MR imaging) and an imaging technique (endoanal MR imaging) accepted in daily practice for the depiction of external and internal anal sphincter defects in patients with fecal incontinence.
Previous study findings have shown that women who become incontinent after childbirth have disruption of the external anal sphincter and often have internal anal sphincter injury as well (3–5), which indicates severe anal sphincter damage. In our study, the majority of external anal sphincter defects were depicted with both MR imaging techniques as an isolated external anal sphincter defect, and in only two patients they were accompanied by an internal anal sphincter defect. These findings also applied to the findings of endoanal US. This may lead one to conclude that the patients included in this study have a different spectrum of defects. Nevertheless, on the basis of medical history, most patients with an anal sphincter defect had one or more risk factors for anal sphincter trauma during vaginal delivery. Furthermore, the majority of anal sphincter defects were depicted with both MR imaging techniques in women and at the anterior part of the external anal sphincter. These findings are in concordance with those of previous studies, which reported that obstetric sphincter trauma frequently causes sphincter damage at the anterior part of the anal sphincter complex (3–5).
Previous study findings showed that the high intrinsic spatial resolution of endoanal MR imaging results in an accurate demonstration of external and internal anal sphincter defects (26–28). Because of the higher signal-to-noise ratio near the coil, endoanal MR imaging generated images at higher spatial resolution (in-plane resolution, 0.08–0.15 mm2) than did external phased-array MR imaging (in-plane resolution, 0.3–1.4 mm2). Our study findings show that the spatial resolution of external phased-array MR imaging was sufficient for adequate image quality and the identification of the anal sphincter muscles, which allowed depiction of external and internal anal sphincter defects, despite its lower spatial resolution. Although there were some differences between both hospitals in the T2-weighted sequences used for external phased-array MR imaging, with consequently slight differences in spatial resolution, this was not reflected in the differences in reader's assessment of image quality since image quality was scored as adequate in all patients.
In a large number of patients, both MR imaging techniques corresponded in the assessment of sphincter integrity. The patient's relevant medical history and other findings (ie, endoanal US and surgery) were comparable between patients with an anal sphincter defect at external phased-array MR imaging and endoanal MR imaging.
In concordance with a previous study (45), we found that the interobserver agreement for endoanal MR imaging was at least moderate for assessing anal sphincter integrity. The reproducibility of external phased-array MR imaging was weaker.
The weak interobserver agreement of both MR imaging techniques can most likely be explained by the differences in the experience level between observers, which lead to differences in the knowledge of anal anatomy and familiarity with the changes in sphincter morphology that occur in sphincter defects. This difference in the frame of reference might be related to the relative scarcity of articles about MR imaging of external and internal anal sphincter lesions in patients with fecal incontinence (26–28,46). This may result in a higher weight of personal experience and interpretation during reading. This viewpoint is endorsed by the observation that intraobserver agreement was stronger for each observer familiar with his/her own specific MR imaging technique. A better understanding of the spectrum of MR findings and of how they relate to endoanal US findings and outcome seems mandatory. Radiologists can then be trained in reading endoanal and external phased-array MR images of the anal sphincter complex, which can be expected to increase the reproducibility and reader performance for both MR imaging techniques.
With both MR imaging techniques, discrepant diagnoses between observers were made predominantly at the anterior part of the anal sphincter complex. The anatomy of the anterior part of the external anal sphincter is difficult to interpret because it is a complex structure of intersecting external anal and transverse perineal muscle fibers and longitudinal layer fibers that are closely aligned to each other and with identical signal intensity. Thereby, the posterior vaginal wall in women and the bulbospongiosus muscles are very close anteriorly, leaving only part of the anterior external anal sphincter surrounded by fatty tissue and consequently leading to less clear demarcation. In some patients, this may have resulted in an inaccurate delineation of the external anal sphincter and thereby in a different interpretation of sphincter integrity, despite an adequate identification of the external anal sphincter muscle. Internal anal sphincter defects are seldom isolated, as described earlier. Authors of a prior study reported that interobserver agreement of endoanal MR imaging between the observers was strongest if the anal sphincters were either both intact or both disrupted (45). However, the prevalence of combined external and internal anal sphincter defects was low in this study.
On the basis of findings of this study, we can conclude that external phased-arrayMR imaging is as suitable as endoanal MR imaging in the depiction of clinically relevant anal sphincter defects in patients with fecal incontinence. However, since the interobserver agreement is weak, and the intraobserver agreement is related to the relevant experience level of an observer with each MR imaging technique, both external phased-array MR imaging and endoanal MR imaging can be recommended in the diagnostic work-up of fecal incontinence only when sufficient experience is available.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, M.P.T.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, M.P.T.; clinical studies, M.P.T., R.G.H.B., C.G.M.I.B., J.S.; statistical analysis, M.P.T., M.G.W.D., P.M.M.B., A.C.D.; and manuscript editing, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. E. Lockhart, J. R. Fielding, H. E. Richter, L. Brubaker, C. G. Salomon, W. Ye, C. M. Hakim, C. Y. Wai, A. H. Stolpen, and A. M. Weber Reproducibility of Dynamic MR Imaging Pelvic Measurements: A Multi-institutional Study Radiology, November 1, 2008; 249(2): 534 - 540. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. F. El Sayed, S. El Mashed, A. Farag, M. M. Morsy, and M. S. Abdel Azim Pelvic Floor Dysfunction: Assessment with Combined Analysis of Static and Dynamic MR Imaging Findings Radiology, August 1, 2008; 248(2): 518 - 530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Dobben, R. J. F. Felt-Bersma, F. J. W. ten Kate, and J. Stoker Cross-Sectional Imaging of the Anal Sphincter in Fecal Incontinence Am. J. Roentgenol., March 1, 2008; 190(3): 671 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Berton, G. Gola, and S. R. Wilson Sonography of Benign Conditions of the Anal Canal: An Update Am. J. Roentgenol., October 1, 2007; 189(4): 765 - 773. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. C. Dobben, M. P. Terra, M. Deutekom, J. F. M. Slors, L. W. M. Janssen, P. M. M. Bossuyt, and J. Stoker The Role of Endoluminal Imaging in Clinical Outcome of Overlapping Anterior Anal Sphincter Repair in Patients with Fecal Incontinence Am. J. Roentgenol., August 1, 2007; 189(2): W70 - W77. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. Terra, R. G. H. Beets-Tan, V. P. M. van der Hulst, M. Deutekom, M. G. W. Dijkgraaf, P. M. M. Bossuyt, A. C. Dobben, C. G. M. I. Baeten, and J. Stoker MRI in Evaluating Atrophy of the External Anal Sphincter in Patients with Fecal Incontinence Am. J. Roentgenol., October 1, 2006; 187(4): 991 - 999. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
