Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Beets-Tan, R. G. H.
Right arrow Articles by van Engelshoven, J. M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Beets-Tan, R. G. H.
Right arrow Articles by van Engelshoven, J. M. A.
(Radiology. 2001;218:75-84.)
© RSNA, 2001

Gastrointestinal Imaging

Preoperative MR Imaging of Anal Fistulas: Does It Really Help the Surgeon?1

Regina G. H. Beets-Tan, MD, Geerard L. Beets, MD, PhD, Arthur Gerritsen van der Hoop, MD, PhD, Alfons G. H. Kessels, MD, MSc, Roy F. A. Vliegen, MD, Cor G. M. I. Baeten, MD, PhD and Jos M. A. van Engelshoven, MD, PhD

1 From the Departments of Radiology (R.G.H.B.T., R.F.A.V., J.M.A.v.E.), Surgery (G.L.B., A.G.v.d.H., C.G.M.I.B.), and Clinical Epidemiology (A.G.H.K.), University Hospital of Maastricht, P. Debyelaan 25, 6229 HX Maastricht, the Netherlands. From the 1999 RSNA scientific assembly. Received December 2, 1999; revision requested December 30; final revision received April 24, 2000; accepted May 8. Address correspondence to R.G.H.B.T. (e-mail: rbe@rdia.azm.nl).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
PURPOSE: To evaluate the accuracy of magnetic resonance (MR) imaging with a quadrature phased-array coil for the detection of anal fistulas and to evaluate the additional clinical value of preoperative MR imaging, as compared with surgery alone.

MATERIALS AND METHODS: Fifty-six patients with anal fistulas underwent high-spatial-resolution MR imaging. Twenty-four had a primary fistula; 17, a recurrent fistula; and 15, a fistula associated with Crohn disease. MR imaging findings were withheld from the surgeon until surgery ended and verified, and surgery continued when required.

RESULTS: MR imaging provided important additional information in 12 (21%) of 56 patients. In patients with Crohn disease, the benefit was 40% (six of 15); in patients with recurrent fistulas, 24% (four of 17); and in patients with primary fistulas, 8% (two of 24). The difference between patients with or without Crohn disease and between patients with a simple fistula versus the rest was significant (P < .05). The sensitivity and specificity for detecting fistula tracks were 100% and 86%, respectively; abscesses, 96% and 97%, respectively; horseshoe fistulas, 100% and 100%, respectively; and internal openings, 96% and 90%, respectively.

CONCLUSION: High-spatial-resolution MR imaging is accurate for detecting anal fistulas. It provides important additional information in patients with Crohn disease–related and recurrent anal fistulas and is recommended in their preoperative work-up.

Index terms: Anus, abnormalities, 757.245 • Crohn disease, 74.262, 75.262 • Intestines, abnormalities, 74.262, 75.262, 757.245 • Magnetic resonance (MR), coils, 70.121411


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A majority of anal fistulas have a single simple fistula track that is easily identified during surgery, and surgical treatment is generally successful (1). However, 5%–15% of anal fistula tracks have a more complicated course, with secondary extensions outside the anal sphincter, often with horseshoe fistulas and ischiorectal and supralevator abscesses. These so-called complex fistulas are often associated with recurrent fistulas and fistulas associated with underlying Crohn disease. Failure in accurate assessment of the secondary extensions during surgery may be responsible for the high rate of recurrence (2). Diagnostic studies that allow the acquisition of accurate preoperative information on the course of the primary track and its secondary extensions may improve the surgical treatment of these complex fistulas. Contrast material–enhanced fistulography is correct in only 16% of patients (3), and computed tomography usually fails to depict subtle fistula tracks and abscesses because of the inherently low soft-tissue contrast resolution (4,5). Anal and transrectal endosonographic studies show better resolution of fistulas and their relation to the anal sphincter muscles (6). The limited field of view, however, is a considerable disadvantage, and endosonography is reported to be no more accurate than examination under anesthesia (7). In general, the imaging of perianal fistulas was disappointing until the introduction of magnetic resonance (MR) imaging.

To our knowledge, the first reports on the accuracy of MR imaging for the detection and classification of fistulas (8,9) are from 1992 and 1994. In these publications, Lunniss et al reported a concordance rate of 86%–88% between MR imaging and surgical findings. Since then, authors of some articles (10–15) have reported high accuracy values for MR imaging in the detection of fistula tracks and secondary extensions. It is, however, unclear whether these high accuracy values also lead to better surgery. In our opinion, a preoperative MR image has no clinical value when it depicts only lesions that are easily found at standard surgical exploration. Some authors (8,9,11,15) have suggested that MR imaging can depict extensions that can be missed at surgery, but none of their studies were designed to evaluate the additional clinical value of MR imaging. To our knowledge, investigators in only two studies (16,17) have specifically addressed this issue, and they found little benefit of MR imaging in the treatment of patients with perianal fistulas. In the study by Scholefield and colleagues (16), most patients had primary simple fistulas; therefore, their conclusion may not be valid for patients with more complex fistulas. Zbar and deSouza (17) found important additional information in only one of 11 patients with complex fistulas who were examined by using an endoanal MR imaging technique. This technique, however, has the inherent disadvantage of a limited field of view. The clinical effect of an MR imaging technique that offers a wide field of view is still unclear.

The purpose of this prospective study was to evaluate the accuracy of preoperative high-spatial-resolution MR imaging with a phased-array coil for the detection of anal fistulas and, more important, to evaluate its additional clinical value, as compared with surgery alone, in a group of patients with a high prevalence of complex fistulas.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
The study was approved by the local institutional review board. Between December 1997 and August 1999, all patients with the clinical diagnosis of anal fistula who were scheduled for surgical exploration were considered candidates for inclusion. Fifty-six patients (32 men, 24 women) with a mean age of 42 years (age range, 13–79 years) entered the study. After informed consent was obtained, the patients underwent preoperative high-spatial-resolution MR imaging. There was no other preoperative imaging.

The study population represented the spectrum of patients treated in a specialized colorectal surgery unit in our academic referral center. Fifteen patients had a primary or recurrent perianal fistula with associated Crohn disease (mean age, 34 years; age range, 13–61 years). Of the remaining 41 patients without Crohn disease (mean age, 45 years; age range, 19–79 years), 24 had a primary fistula at presentation, nine of whom had previously undergone perianal abscess drainage. Seventeen patients had undergone previous fistula surgery and had a recurrence at presentation. Therefore, three groups of patients were defined: 15 patients with Crohn disease, 17 with a recurrent fistula, and 24 with a primary fistula not associated with Crohn disease. Patients in the first two groups were referred to as having a "complex" fistula, as opposed to patients in the last group, who were referred to as having a "simple" fistula.

Imaging Techniques
MR imaging (Gyroscan, Powertrak 6000, NT Release 6.2.1; Philips Medical Systems, Best, the Netherlands) was performed at 1.5 T (23.0 mT/m; rise time, 0.2 msec; slew rate, 105 T/m/sec). A quadrature phased-array spine coil was used. All subjects were placed in a supine feet-first position, with the pelvis centered on the proximal end of the coil. T1-weighted, two-dimensional turbo spin-echo (SE) (repetition time, 656 msec; echo time, 10 msec [656/10]; echo train length, five; 8-mm section thickness; 0.8-mm gap; four signals acquired; 166 x 256 matrix; 25-cm field of view; and 1.4-minute acquisition time) and T2-weighted, two-dimensional turbo SE ([3,427/150] echo train length, twenty-five; 3-mm section thickness; 0.3-mm gap; eight signals acquired; 175 x 256 matrix; 20-cm field of view; 2.64-mm3 voxel size; and 6.5-minute acquisition time) sequences were used. T2-weighted imaging was performed in sagittal, coronal, and transverse planes. T1-weighted turbo SE imaging was performed in the transverse plane and served for accurate planning of coronal and transverse T2-weighted imaging, because it was important to angle these planes exactly parallel and perpendicular to the long axis of the anal canal. Patients did not receive bowel preparation, and no catheters were inserted into the anal canal during the procedure. The total imaging time was 20 minutes.

Image Evaluation
The MR images were prospectively evaluated by a radiologist who was experienced in reading pelvic MR images (R.G.H.B.T.). The images were evaluated for the presence and site of the primary fistula track, of any external and internal opening, and of any abscess or horseshoe extension; a yes or no score was given. The fistula classification according to Parks and colleagues (18) was also given. Criteria for fistulas and fluid collections were based on shape and signal intensities, as described in the literature (14,19). Mucin-containing fistulas were recognized as tubular structures with a hyperintense signal surrounded by a hypointense rim on T2-weighted two-dimensional turbo SE images. Non–mucin-containing fistulas were recognized as tubular structures with a hypointense signal on T2-weighted images. Fluid-filled cavities were hyperintense on T2-weighted images and surrounded by a border of hypointense signal. All findings were recorded on a standardized fistula surgery form (Fig 1), which was put into an envelope that was then sealed without disclosure of the information to the surgeon.



View larger version (50K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. St Mark’s Hospital fistula surgery form, based on the Parks classification (18). (Reprinted, with permission, from reference 9.)

 
Surgery
Surgery was performed by one of the three colorectal surgeons with experience in complicated fistula surgery (including C.G.M.I.B. and A.G.v.d.H.) Surgical assessment was performed with the patient under general or local-regional anesthesia and in the lithotomy position. The extent of disease was established by cannulating the fistulas with probes and by laying open all primary tracks, extensions, and abscesses. The presence and site of a primary fistula track, the site of any external and internal opening, the presence and site of any abscess or horseshoe fistula, and the Parks classification were recorded on a separate fistula surgery form. While the patient was still under anesthesia and before reconstructive surgery (ie, mucosal advancement) was performed, the corresponding envelope containing the MR imaging findings was opened, and the surgical results were compared with the MR imaging results, as recorded on the forms. When MR imaging findings suggested more extensive disease, the surgeon went back to verify these additional MR imaging findings. These findings were considered true-positive additional MR imaging findings only when they were confirmed and led to additional surgical drainage.

Interobserver and Intraobserver Variability
To determine the inter- and intraobserver variability and to elucidate the sensitivity and specificity profile of the MR images, all images were again evaluated, now retrospectively. This was performed by three observers: the radiologist who had evaluated the images prospectively (observer 1) (R.G.H.B.T.), a radiology resident (observer 2) (R.F.A.V.), and a surgeon who was experienced in reading pelvic MR images and who was not involved in the surgical treatment of any of the 56 patients (observer 3) (G.L.B.). The images were evaluated for the same items as in prospective evaluation, but now a confidence level scoring system was used instead of a simple yes or no score. The following confidence levels were used: definitely absent, probably absent, possibly present, probably present, or definitely present.

Analysis and Statistics
The final surgical findings after correction with MR imaging were accepted as the reference standard against which the MR imaging findings were compared. A primary track, abscess, or horseshoe fistula was considered as correctly depicted by MR imaging when both the classification and location were in agreement with the findings at final surgery. An internal opening was considered as correctly identified when it was at the correct level in the anal canal and was within the correct quadrant. The sensitivity, specificity, positive predictive value, and negative predictive value were calculated for high-spatial-resolution MR imaging in predicting the presence and exact location of primary tracks, abscesses, horseshoe fistulas, and internal openings. Additional information from preoperative MR imaging for each of the three groups of patients and the three observers was compared by performing the Pearson {chi}2 test.

Receiver operating characteristic (ROC) analyses for the detection of primary tracks, abscesses, horseshoe fistulas, and internal openings were performed on the data generated by using the confidence level scoring by the three observers. For each observer, the accuracy of high-spatial-resolution MR imaging for the detection of primary tracks, abscesses, horseshoe fistulas, and internal openings was measured by calculating the area under the ROC curve (Az). Differences in accuracies among the three observers were evaluated by performing a pairwise comparison of the Az values (20,21).

Interobserver agreement was measured by using the linear-weighted {kappa} statistic for pairwise comparison of the three observers for the detection of primary tracks, abscesses, horseshoe fistulas, and internal openings (22). Intraobserver agreement was studied for only observer 1. For the first reading, a yes or no score was used, whereas for the second, the graded confidence level score was used. To compare both readings, the second was dichotomized, with the cutoff point between "probably absent" and "possibly present." A nonweighted {kappa} statistic was used for this intraobserver agreement and the agreement on classification of fistulas. {kappa} values ranged from 0 (no agreement) to +1 (perfect agreement) and were interpreted as poor (0.00), slight (0.01–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), or almost perfect (0.81–1.00) (23).

Statistical analysis was performed by using SPSS for Windows Release 8.0 (SPSS, Chicago, Ill). P values less than .05 were considered to indicate a significant difference.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
All patients tolerated MR imaging, and there were no motion artifacts. Overall, MR imaging showed important additional information in 12 (21%) of 56 patients. MR imaging showed additional information in six (40%) of the 15 patients with Crohn fistulas, in four (24%) of the 17 patients with recurrent fistulas, and in two (8%) of the 24 patients with only primary fistulas. The difference in benefit from MR imaging between the group with primary simple fistulas and the group with complex fistulas was significant (P < .05), as was the difference in benefit between the group with Crohn fistulas versus the group with non–Crohn fistulas (P < .05).

The correlation of the MR imaging findings, the initial surgical findings, and the final surgical findings after correction by using MR imaging are given in Table 1. Additional information from preoperative MR imaging is found in the false-negative surgical findings column. MR imaging correctly depicted all 56 primary tracks but resulted in the misclassification of one transsphincteric fistula as a sinus. MR imaging showed eight false-positive findings of a primary track. Two intersphincteric fistulas were missed at initial surgery but were correctly predicted at MR imaging. MR imaging depicted 25 of 26 abscess collections. Seven supralevator abscess collections were missed at initial surgery but were correctly predicted by using MR imaging and were confirmed at extended surgery. MR imaging led to the false prediction of three abscesses. All 14 horseshoe fistulas were detected by using MR imaging and included the four collections that the surgeons initially missed. MR imaging was used to correctly identify 52 of the 54 internal openings. There were six false-positive predictions of an enteric entry site. Three times, MR imaging correctly revealed an internal opening that was missed at initial surgical exploration. All three openings were in the lower part of the rectum.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Correlation of MR Imaging Findings with Surgical Findings for the Detection and Classification of Abnormalities
 
The sensitivities, specificities, and positive and negative predictive values of high-spatial-resolution MR imaging for the detection of primary tracks, abscesses, horseshoe fistulas, and internal openings are given in Table 2.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Accuracy of High-Spatial-Resolution MR Imaging for the Detection of a Primary Track, Abscess, Horseshoe Fistula, or Internal Opening
 
The ROC curves and Az values for the detection and classification of anal fistulas, abscesses, horseshoe fistulas, and internal openings are shown in Figure 2. The high Az values (>0.90) indicated that the presence or absence of fistula tracks, abscesses, and horseshoe fistulas could be predicted with accuracy. For the internal opening, the Az value was lower, which indicated that these were more difficult to predict. There was no significant difference in Az values between the three observers.



View larger version (64K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Chart shows ROC curves and area under the curve (AUC), or Az values, for the three observers in the detection of fistula tracks, abscesses, horseshoe fistulas, and internal openings. The 95% CIs of the estimated Az values for fistula tracks were as follows: Az1: 0.97, 1.00; Az2: 0.92, 1.00; and Az3: 0.94, 1.00. For abscesses, Az1: 0.92, 1.00; Az2: 0.88, 1.00; and Az3: 0.88, 1.00. For horseshoe fistulas, Az1: 1.00, 1.00; Az2: 0.88, 1.00; and Az3: 0.82, 1.00. For internal openings, Az1: 0.89, 0.99; Az2: 0.85, 0.97; and Az3: 0.86, 0.97.

 
Table 3 shows the analysis of interobserver and intraobserver agreement for the detection and classification of a fistula track in the detection of abscesses, horseshoe fistulas, and internal openings. In general, the interobserver agreement for the detection of fistula tracks, abscesses, horseshoe fistulas, and internal openings was good to very good. There was better agreement for the detection of abscesses and horseshoe fistulas than for the detection of fistula tracks and internal openings. The intraobserver agreement of the experienced radiologist (observer 1) was very good, with the best agreement achieved for the detection of horseshoe fistulas and abscess collections. For the classification of fistula tracks, there was only moderate agreement between the observers but better intraobserver agreement for the radiologist.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Inter- and Intraobserver Agreement for the Detection and Classification of Abnormalities with MR Imaging
 
The radiologist’s reading provided important additional information in 12 (21%) of 56 patients. The resident provided additional information in 10 (18%) patients; the surgeon, in 11 (20%) patients. This difference was not significant.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
In this study, the additional clinical value and accuracy of preoperative high-spatial-resolution MR imaging, as compared with surgery alone, in the treatment of patients with anal fistulas were evaluated. Overall, preoperative MR imaging revealed important additional information in 12 (21%) of 56 patients. The benefit was most obvious in patients with Crohn disease (40%) and in patients with recurrent fistulas (24%). In patients with a primary simple fistula, the benefit was only 8%.

The findings of our study seem to be in contrast with the findings of a 27-patient study by Scholefield and colleagues (16), who found that preoperative MR imaging was of little use in the surgical treatment of perianal fistulas. This difference in outcome can be explained by the type of patients enrolled in both studies. The study by Scholefield and colleagues (16) consisted mainly of patients with a primary simple fistula, whereas in the present study, the percentage of complex fistulas was higher (57%). The 8% additional information from MR imaging in the group of patients with a primary fistula in our study was in agreement with the 7.4% (two of 27) in their study, and the conclusions are therefore the same: Preoperative diagnostic imaging is rarely needed in patients with a primary simple fistula (Fig 3).



View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. Transsphincteric fistula. (a, b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense fistula track (white arrowhead) posterior to the anal canal (black arrowhead). The track crosses the sphincter muscles (white arrow in a) posteriorly. (c, d) Transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show the track (black arrow in c) first coursing laterodorsally from the left external sphincter muscles (arrowhead in c), then crossing the external sphincter at a midanal level (black arrow in d). A transsphincteric fistula track was diagnosed at MR imaging and confirmed at surgery. The internal opening was correctly predicted as being at the five o’clock position dorsally, at midanal level (white arrow in d). Note that the internal opening is not directly visualized but can be inferred only from the course and proximity of the fistula track in the sphincter muscle compartments. The arrowhead in d indicates the internal sphincter muscle.

 


View larger version (162K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. Transsphincteric fistula. (a, b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense fistula track (white arrowhead) posterior to the anal canal (black arrowhead). The track crosses the sphincter muscles (white arrow in a) posteriorly. (c, d) Transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show the track (black arrow in c) first coursing laterodorsally from the left external sphincter muscles (arrowhead in c), then crossing the external sphincter at a midanal level (black arrow in d). A transsphincteric fistula track was diagnosed at MR imaging and confirmed at surgery. The internal opening was correctly predicted as being at the five o’clock position dorsally, at midanal level (white arrow in d). Note that the internal opening is not directly visualized but can be inferred only from the course and proximity of the fistula track in the sphincter muscle compartments. The arrowhead in d indicates the internal sphincter muscle.

 


View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3c. Transsphincteric fistula. (a, b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense fistula track (white arrowhead) posterior to the anal canal (black arrowhead). The track crosses the sphincter muscles (white arrow in a) posteriorly. (c, d) Transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show the track (black arrow in c) first coursing laterodorsally from the left external sphincter muscles (arrowhead in c), then crossing the external sphincter at a midanal level (black arrow in d). A transsphincteric fistula track was diagnosed at MR imaging and confirmed at surgery. The internal opening was correctly predicted as being at the five o’clock position dorsally, at midanal level (white arrow in d). Note that the internal opening is not directly visualized but can be inferred only from the course and proximity of the fistula track in the sphincter muscle compartments. The arrowhead in d indicates the internal sphincter muscle.

 


View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3d. Transsphincteric fistula. (a, b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense fistula track (white arrowhead) posterior to the anal canal (black arrowhead). The track crosses the sphincter muscles (white arrow in a) posteriorly. (c, d) Transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show the track (black arrow in c) first coursing laterodorsally from the left external sphincter muscles (arrowhead in c), then crossing the external sphincter at a midanal level (black arrow in d). A transsphincteric fistula track was diagnosed at MR imaging and confirmed at surgery. The internal opening was correctly predicted as being at the five o’clock position dorsally, at midanal level (white arrow in d). Note that the internal opening is not directly visualized but can be inferred only from the course and proximity of the fistula track in the sphincter muscle compartments. The arrowhead in d indicates the internal sphincter muscle.

 
Fistulas associated with Crohn disease and recurrent fistulas, the complex fistulas, are often more difficult to evaluate during surgery because of the extensive secondary extensions and multiple perianal and perirectal abscess collections. When not identified and properly treated, these extensions and abscesses may lead to recurrences (2). The recurrent nature of Crohn perianal disease is well known, and a conservative treatment policy is often advocated (24). It is, however, debatable whether these recurrences are an inevitable consequence of the nature of Crohn disease. It may be that many recurrences in patients with Crohn fistulas are simply due to overlooked abscesses and secondary extensions (25). Results of the first study by Lunniss et al (8) suggested that MR imaging could depict more extensions than could surgical exploration. Our study results showed that preoperative MR imaging helps surgeons to identify all secondary extensions of a complex fistula. The largest additional value was seen in the detection of supralevator abscesses and horseshoe fistulas (Figs 4, 5). Secondary extensions are easily missed above the pelvic floor and in patients with multiple extensions. It is likely that the additional information obtained from MR imaging will improve the surgical results in patients with complex fistulas, although this can be reliably established in only a prospective randomized trial of surgery with and without preoperative MR imaging.



View larger version (166K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a. Abscess and fistula track missed at surgery. (a) Coronal and (b) transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense lesion with a hypointense rim (arrowhead in a) above the pelvic floor, which suggests a small abscess in the supralevator space. Note the layering of debris within the fluid collection (arrow in b). Also note the hyperintense structure between the sphincter muscles (arrow in a) that extends toward the abscess collection. MR imaging was used to correctly predict the presence of a left intersphincteric blind-ending fistula track leading toward a supralevator abscess collection on the left side. Both the track and abscess were missed at initial surgical exploration because the patient was thought to have only a sinus on the contralateral side (not shown).

 


View larger version (183K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b. Abscess and fistula track missed at surgery. (a) Coronal and (b) transverse T2-weighted two-dimensional turbo SE (3,427/150) MR images show a hyperintense lesion with a hypointense rim (arrowhead in a) above the pelvic floor, which suggests a small abscess in the supralevator space. Note the layering of debris within the fluid collection (arrow in b). Also note the hyperintense structure between the sphincter muscles (arrow in a) that extends toward the abscess collection. MR imaging was used to correctly predict the presence of a left intersphincteric blind-ending fistula track leading toward a supralevator abscess collection on the left side. Both the track and abscess were missed at initial surgical exploration because the patient was thought to have only a sinus on the contralateral side (not shown).

 


View larger version (176K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a. Horseshoe fistula and rectal opening missed at surgery. (a-d) Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR images from posterior (a) to anterior (d) show two hyperintense fluid collections that cross the midline, one located infralevatorly (arrow in a and b) and the other at the level of the pelvic floor (arrowhead in a and b). A secondary track (arrow in c) ascends from the highest collection toward the rectal lumen. The rectal opening (arrowhead in d) is clearly visualized. MR imaging findings suggested the presence of two horseshoe fistulas at different levels, with connection to the rectal lumen. The surgeons identified the lower horseshoe fistulas but missed the higher extensions because the problem was thought to have been solved by draining the lower extension. Because of the additional MR imaging information, surgical exploration was extended, and the higher extensions were found and properly treated.

 


View larger version (173K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b. Horseshoe fistula and rectal opening missed at surgery. (a-d) Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR images from posterior (a) to anterior (d) show two hyperintense fluid collections that cross the midline, one located infralevatorly (arrow in a and b) and the other at the level of the pelvic floor (arrowhead in a and b). A secondary track (arrow in c) ascends from the highest collection toward the rectal lumen. The rectal opening (arrowhead in d) is clearly visualized. MR imaging findings suggested the presence of two horseshoe fistulas at different levels, with connection to the rectal lumen. The surgeons identified the lower horseshoe fistulas but missed the higher extensions because the problem was thought to have been solved by draining the lower extension. Because of the additional MR imaging information, surgical exploration was extended, and the higher extensions were found and properly treated.

 


View larger version (178K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5c. Horseshoe fistula and rectal opening missed at surgery. (a-d) Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR images from posterior (a) to anterior (d) show two hyperintense fluid collections that cross the midline, one located infralevatorly (arrow in a and b) and the other at the level of the pelvic floor (arrowhead in a and b). A secondary track (arrow in c) ascends from the highest collection toward the rectal lumen. The rectal opening (arrowhead in d) is clearly visualized. MR imaging findings suggested the presence of two horseshoe fistulas at different levels, with connection to the rectal lumen. The surgeons identified the lower horseshoe fistulas but missed the higher extensions because the problem was thought to have been solved by draining the lower extension. Because of the additional MR imaging information, surgical exploration was extended, and the higher extensions were found and properly treated.

 


View larger version (178K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5d. Horseshoe fistula and rectal opening missed at surgery. (a-d) Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR images from posterior (a) to anterior (d) show two hyperintense fluid collections that cross the midline, one located infralevatorly (arrow in a and b) and the other at the level of the pelvic floor (arrowhead in a and b). A secondary track (arrow in c) ascends from the highest collection toward the rectal lumen. The rectal opening (arrowhead in d) is clearly visualized. MR imaging findings suggested the presence of two horseshoe fistulas at different levels, with connection to the rectal lumen. The surgeons identified the lower horseshoe fistulas but missed the higher extensions because the problem was thought to have been solved by draining the lower extension. Because of the additional MR imaging information, surgical exploration was extended, and the higher extensions were found and properly treated.

 
Accuracy
In general, the sensitivity and specificity figures for the detection of fistula tracks, abscesses, horseshoe fistulas, and internal openings are in agreement with those in other studies (9–12,14). Abscesses and horseshoe fistulas were identified with a high accuracy. Although the sensitivity for the detection of a primary track (100%) was high, the specificity (86%) was lower. Seven of eight false-positive predictions of fistula tracks were found at surgical exploration to be healed fibrotic tracks. In retrospect, the MR imaging appearance of a healed fibrotic track typically lacked the hyperintense signal of fluid inside the hypointense fistula track (Fig 6). It will, however, remain difficult to differentiate a healed fibrotic track from an open inactive track. The internal fistula opening is not always directly visualized at MR imaging (10,14,16). The location of the internal opening often must be inferred from the course of the fistula track in the sphincter muscles (Fig 3). With this method, a high sensitivity was obtained at the expense of some false-positive predictions and therefore a lower specificity.



View larger version (172K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6a. (a) True-positive diagnosis of a fistula track. Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a hyperintense tubular structure in the left ischiorectal fossa that is surrounded by a hypointense rim (black arrow) and was correctly diagnosed as an extrasphincteric fistula track. The track ends in a small abscess (white arrow) in the left side of the pelvic floor. (b) False-positive diagnosis of a fistula track. Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a hypointense track (large arrow) that courses to the left sphincter muscle complex. This hypointense structure was mistaken for an inactive fistula track. At surgery, only fibrosis was found. Note the detailed anatomy of the anal sphincter complex: the external sphincter muscle (arrowhead a), the puborectalis muscle (arrowhead b), the pelvic floor muscles (arrowhead c), and the longitudinal muscle in the intersphincteric space (small arrow).

 


View larger version (165K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6b. (a) True-positive diagnosis of a fistula track. Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a hyperintense tubular structure in the left ischiorectal fossa that is surrounded by a hypointense rim (black arrow) and was correctly diagnosed as an extrasphincteric fistula track. The track ends in a small abscess (white arrow) in the left side of the pelvic floor. (b) False-positive diagnosis of a fistula track. Coronal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a hypointense track (large arrow) that courses to the left sphincter muscle complex. This hypointense structure was mistaken for an inactive fistula track. At surgery, only fibrosis was found. Note the detailed anatomy of the anal sphincter complex: the external sphincter muscle (arrowhead a), the puborectalis muscle (arrowhead b), the pelvic floor muscles (arrowhead c), and the longitudinal muscle in the intersphincteric space (small arrow).

 
These findings were also reflected in the ROC curves. The highest Az values and therefore the highest accuracies were obtained for the detection of horseshoe fistulas and abscesses. Accuracy was lowest for the detection of the internal opening.

The sensitivity and specificity figures must be interpreted with caution. Nearly all of the patients in our study had a fistula because they were selected for surgery only when there was a proved or strongly suspected fistula. Such selection bias leads to methodologic difficulties. Awareness of the fact that patients will undergo surgery anyway and almost certainly have a fistula leads to overreading of the MR images. In the clinical setting of the current study, it was more important to indicate all possible tracks and extensions than to avoid a false-positive reading. This can produce a high sensitivity at the expense of a lower specificity.

Another problem associated with the selection bias of this study was the difficulty in defining the exact number of true-negative results. Because many patients had bilateral extensions, the decision was made to analyze the accuracy not per patient but per side. The downside of this may be artificially elevated specificity figures. It becomes even more complicated when one realizes that the reference standard of surgery itself is not infallible and that MR imaging is sometimes more accurate than surgical exploration (8). In our opinion, one should therefore be cautious about attaching too much value to sensitivity and specificity figures.

Interobserver and Intraobserver Variability
For the detection of abscesses and horseshoe fistulas, there was good correlation between the radiologist, the radiology resident, and the surgeon and therefore little difference among the readers in providing important additional information. There was less agreement on the detection of primary tracks and internal openings and only moderate agreement on the fistula classification. Although one could argue that classifying a fistula is important, it is, in our opinion, more important to detect all secondary extensions, because these are easily overlooked at surgery. The ROC curves again showed no significant difference in performance between the experienced radiologist, the less experienced resident, and the surgeon with experience in reading pelvic MR images. The high intraobserver agreement indicated consistency in image interpretation by the same reader. All of these findings show that the results of our study are reproducible and can be generalized to other centers. Surgeons who treat these patients should familiarize themselves with reading pelvic MR images to obtain the maximum benefit from preoperative MR imaging.

MR Imaging Technique
MR imaging can be improved by enhancing contrast resolution and increasing spatial resolution. Most MR imaging studies in patients with perianal sepsis (8–11, 14,26–30) have involved the use of a body coil with an inherent low spatial resolution. In these studies, contrast resolution was improved by using fat-suppression techniques, dynamic gadolinium-enhanced imaging, and MR fistulography. Fat-suppression sequences such as short inversion time inversion recovery (STIR) (8–10,14,30) and frequency-specific spectral presaturation enhance fistula tracks but have longer acquisition times than do SE sequences, and the anatomy of the pelvic floor is not well visualized (29,30). Dynamic gadolinium-enhanced imaging has been described as superior to STIR for detecting active sepsis, with the main advantage of being faster but with the additional cost of gadolinium use (11,14). MR fistulography with instillation of saline can facilitate the detection of fistula tracks, but the technique is cumbersome and depends on the existence of an external opening (28).

With the MR imaging technique used in the present study, we aimed to improve the spatial resolution by using a phased-array coil, which resulted in a completely noninvasive high-spatial-resolution MR imaging technique (31). The multiple-coil arrangement in a phased-array coil increases the signal-to-noise ratio and allows us to obtain images with smaller voxel sizes and higher spatial resolution than those obtained with a body coil. The specific coil used in our study was a quadrature phased-array spine coil. Unlike the linear arrangement of the coil components in a standard torso or pelvic phased-array coil, the coil components in a synergy spine coil are arranged in quadrature. This quadrature arrangement further improves the signal-to-noise ratio, and even smaller voxel sizes can be generated. The 2.6-mm3 voxel size of the T2-weighted two-dimensional turbo SE sequence in our study was substantially smaller than the 11–31 mm3 voxel size in other techniques in which a body coil (9,14) or even a phased-array coil (12,32, 33) was used.

MR imaging with an endoanal coil can also generate images with a high spatial resolution because of the very high signal-to-noise ratio near the coil (34). However, the results of fistula imaging have been disappointing. Investigators in a large study (13) in which endoanal MR imaging was compared with body-coil MR imaging found a surgical concordance rate of 68% for endoanal MR imaging, as compared with 96% for body-coil MR imaging. Investigators in another study (34) showed a similar disappointing concordance rate of 64%. The main drawback of the endoanal MR imaging technique is that it fails to show many secondary extensions that lie beyond the range of the coil (12,13). DeSouza et al (12) suggested that this problem can be overcome by combining the endoanal coil with a phased-array coil. Our study findings show that the combination of a high spatial resolution and a wide field of view can also be achieved by using a quadrature phased-array coil. The visualization of the perianal region and the entire pelvic region in a single investigation can be of particular value in patients with Crohn anal fistulas, who often present with simultaneous intrapelvic disease, as illustrated in Figure 7.



View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7a. Intrapelvic disease in two patients with Crohn perianal fistulas. (a) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a presacral fluid collection (arrow) with connection to the bladder (not shown) and sigmoid colon (arrowhead) that was correctly diagnosed as a presacral abscess that drained to the bladder and sigmoid colon. (b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR image depicts a fistula track (arrowhead) that connects the rectal lumen (bullet arrow) with the vaginal lumen (arrow), which is suggestive of a rectovaginal fistula and was confirmed at surgery.

 


View larger version (153K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7b. Intrapelvic disease in two patients with Crohn perianal fistulas. (a) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR image shows a presacral fluid collection (arrow) with connection to the bladder (not shown) and sigmoid colon (arrowhead) that was correctly diagnosed as a presacral abscess that drained to the bladder and sigmoid colon. (b) Sagittal T2-weighted two-dimensional turbo SE (3,427/150) MR image depicts a fistula track (arrowhead) that connects the rectal lumen (bullet arrow) with the vaginal lumen (arrow), which is suggestive of a rectovaginal fistula and was confirmed at surgery.

 
In conclusion, high-spatial-resolution MR imaging with a quadrature phased-array coil is accurate for the detection of perianal fistulas and provides important additional information on secondary extensions in patients with complex fistulas. The largest additional value from preoperative MR imaging was obtained in patients with complex fistulas that were associated with Crohn disease and recurrences. High-spatial-resolution MR imaging is therefore recommended in the preoperative work-up of patients with Crohn and recurrent anal fistulas. In our opinion, there is little place for MR imaging in the treatment of patients with primary simple fistulas.


   ACKNOWLEDGMENTS
 
We acknowledge the technical assistance of Etienne Lemaire and Henk Schoenmakers, the secretarial assistance of Ine Kengen, and the assistance of Geert-Jan van Zonneveld in preparing the figures.


   FOOTNOTES
 
Abbreviations: Az = area under the ROC curve, ROC = receiver operating characteristic, SE = spin echo, STIR = short inversion time inversion recovery

Author contributions: Guarantor of integrity of entire study, R.G.H.B.T.; study concepts, R.G.H.B.T., G.L.B., A.G.v.d.H., C.G.M.I.B., J.M.A.v.E.; study design, R.G.H.B.T., G.L.B., A.G.v.d.H., J.M.A.v.E.; definition of intellectual content, R.G.H.B.T., G.L.B., A.G.v.d.H., J.M.A.v.E.; literature research, R.G.H.B.T.; clinical studies, R.G.H.B.T., A.G.v.d.H., C.G.M.I.B.; data acquisition, R.G.H.B.T.; data analysis, R.G.H.B.T., G.L.B., R.F.A.V.; statistical analysis, R.G.H.B.T., G.L.B., A.G.H.K.; manuscript preparation, R.G.H.B.T.; manuscript editing, R.G.H.B.T., G.L.B., J.M.A.v.E.; manuscript review, all authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Shouler PJ, Grimley RP, Keighley MR, Alexander-Williams J. Fistula-in-ano is usually simple to manage surgically. Int J Colorectal Dis 1986; 1:113-115.[Medline]
  2. Seow C, Phillips RK. Insights gained from the management of problematical anal fistulae at St Mark’s Hospital, 1984–88. Br J Surg 1991; 78:539-541.[Medline]
  3. Kuijpers HC/MIDDLE>, Schulpen T. Fistulography for fistula-in-ano: is it useful?. Dis Colon Rectum 1985; 28:103-104.[Medline]
  4. Yousem DM, Fishman EK, Jones B. Crohn disease: perirectal and perianal findings at CT. Radiology 1988; 167:331-334.[Abstract/Free Full Text]
  5. Guillaumin E, Jeffrey RB, Jr, Shea WJ, Asling CW, Goldberg HI. Perirectal inflammatory disease: CT findings. Radiology 1986; 161:153-157.[Abstract/Free Full Text]
  6. Van Outryve MJ, Pelckmans PA, Michielsen PP, Van Maercke YM. Value of transrectal ultrasonography in Crohn’s disease. Gastroenterology 1991; 101:1171- 1177.[Medline]
  7. Choen S, Burnett S, Bartram CI, Nicholls RJ. Comparison between anal endosonography and digital examination in the evaluation of anal fistulae. Br J Surg 1991; 78:445-447.[Medline]
  8. Lunniss PJ, Armstrong P, Barker PG, Reznek RH, Phillips RK. Magnetic resonance imaging of anal fistulae. Lancet 1992; 340:394-396.[Medline]
  9. Lunniss PJ, Barker PG, Sultan AH, et al. Magnetic resonance imaging of fistula-in-ano. Dis Colon Rectum 1994; 37:708- 718.[Medline]
  10. Barker PG, Lunniss PJ, Armstrong P, Reznek RH, Cottam K, Phillips RK. Magnetic resonance imaging of fistula-in-ano: technique, interpretation and accuracy. Clin Radiol 1994; 49:7-13.[Medline]
  11. Beckingham IJ, Spencer JA, Ward J, Dyke GW, Adams C, Ambrose NS. Prospective evaluation of dynamic contrast enhanced magnetic resonance imaging in the evaluation of fistula in ano. Br J Surg 1996; 83:1396-1398.[Medline]
  12. deSouza NM, Gilderdale DJ, Coutts GA, Puni R, Steiner RE. MRI of fistula-in-ano: a comparison of endoanal coil with external phased array coil techniques. J Comput Assist Tomogr 1998; 22:357-363.[Medline]
  13. Halligan S, Bartram CI. MR imaging of fistula in ano: are endoanal coils the gold standard?. AJR Am J Roentgenol 1998; 171:407-412.[Abstract/Free Full Text]
  14. Spencer JA, Ward J, Beckingham IJ, Adams C, Ambrose NS. Dynamic contrast-enhanced MR imaging of perianal fistulas. AJR Am J Roentgenol 1996; 167:735-741.[Abstract/Free Full Text]
  15. Spencer JA, Chapple K, Wilson D, Ward J, Windsor AC, Ambrose NS. Outcome after surgery for perianal fistula: predictive value of MR imaging. AJR Am J Roentgenol 1998; 171:403-406.[Abstract/Free Full Text]
  16. Scholefield JH, Berry DP, Armitage NC, Wastie ML. Magnetic resonance imaging in the management of fistula in ano. Int J Colorectal Dis 1997; 12:276-279.[Medline]
  17. Zbar AP, deSouza NM. Prospective comparison of endosonography, magnetic resonance imaging and surgical findings in anorectal fistula and abscess complicating Crohn’s disease. Br J Surg 1999; 86:1093-1094.[Medline]
  18. Parks AG, Gordon PH, Hardcastle JD. A classification of fistula-in-ano. Br J Surg 1976; 63:1-12.[Medline]
  19. Hussain SM, Stoker J, Schouten WR, Hop WC, Lameris JS. Fistula in ano: endoanal sonography versus endoanal MR imaging in classification. Radiology 1996; 200:475-481.[Abstract/Free Full Text]
  20. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148:839- 843.[Abstract/Free Full Text]
  21. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29-36.[Abstract/Free Full Text]
  22. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull 1968; 70:213-230.
  23. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159-174.[Medline]
  24. Keighley MR, Allan RN. Current status and influence of operation on perianal Crohn’s disease. Int J Colorectal Dis 1986; 1:104-107.[Medline]
  25. Gruwez JA, Christiaens MR. Anal-perineal lesions in Crohn’s disease. Ann Gastroenterol Hepatol (Paris) 1991; 27:251-255[French].[Medline]
  26. Van Beers B, Grandin C, Kartheuser A, et al. MRI of complicated anal fistulae: comparison with digital examination. J Comput Assist Tomogr 1994; 18:87-90.[Medline]
  27. Orsoni P, Barthet M, Portier F, Panuel M, Desjeux A, Grimaud JC. Prospective comparison of endosonography, magnetic resonance imaging and surgical findings in anorectal fistula and abscess complicating Crohn’s disease. Br J Surg 1999; 86:360-364.[Medline]
  28. Myhr GE, Myrvold HE, Nilsen G, Thoresen JE, Rinck PA. Perianal fistulas: use of MR imaging for diagnosis. Radiology 1994; 191:545-549.[Abstract/Free Full Text]
  29. Haggett PJ, Moore NR, Shearman JD, Travis SP, Jewell DP, Mortensen NJ. Pelvic and perineal complications of Crohn’s disease: assessment using magnetic resonance imaging. Gut 1995; 36:407-410.[Abstract/Free Full Text]
  30. Halligan S, Healy JC, Bartram CI. Magnetic resonance imaging of fistula-in-ano: STIR or SPIR?. Br J Radiol 1998; 71:141-145.[Abstract]
  31. Beets-Tan RG, Beets GL, van der Hoop AG, et al. High-resolution magnetic resonance imaging of the anorectal region without an endocoil. Abdom Imaging 1999; 24:576-581.[Medline]
  32. Laniado M, Makowiec F, Dammann F, Jehle EC, Claussen CD, Starlinger M. Perianal complications of Crohn disease: MR imaging findings. Eur Radiol 1997; 7:1035-1042.[Medline]
  33. Semelka RC, Hricak H, Kim B, et al. Pelvic fistulas: appearances on MR images. Abdom Imaging 1997; 22:91-95.[Medline]
  34. Hussain SM, Stoker J, Schutte HE, Lameris JS. Imaging of the anorectal region. Eur J Radiol 1996; 22:116-122.[Medline]



This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
V. R. Muthusamy and K. J. Chang
Optimal Methods for Staging Rectal Cancer
Clin. Cancer Res., November 15, 2007; 13(22): 6877s - 6884s.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
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 page
 J Ultrasound MedHome page
G. Engin
Endosonographic imaging of anorectal diseases.
J. Ultrasound Med., January 1, 2006; 25(1): 57 - 73.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
G. N. Buchanan, S. Halligan, C. I. Bartram, A. B. Williams, D. Tarroni, and C. R. G. Cohen
Clinical Examination, Endosonography, and MR Imaging in Preoperative Assessment of Fistula in Ano: Comparison with Outcome-based Reference Standard
Radiology, December 1, 2004; 233(3): 674 - 681.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
R. G. H. Beets-Tan and G. L. Beets
Rectal Cancer: Review with Emphasis on MR Imaging
Radiology, August 1, 2004; 232(2): 335 - 346.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
G. N. Buchanan, S. Halligan, S. Taylor, A. Williams, R. Cohen, and C. Bartram
MRI of Fistula In Ano: Inter- and Intraobserver Agreement and Effects of Directed Education
Am. J. Roentgenol., July 1, 2004; 183(1): 135 - 140.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
S. Dwarkasing, S. M. Hussain, W. C. J. Hop, and G. P. Krestin
Anovaginal Fistulas: Evaluation with Endoanal MR Imaging
Radiology, April 1, 2004; 231(1): 123 - 128.
[Abstract] [Full Text] [PDF]

Home page
 J Ultrasound MedHome page
A. Mallouhi, H. Bonatti, S. Peer, P. Lugger, F. Conrad, and G. Bodner
Detection and Characterization of Perianal Inflammatory Disease: Accuracy of Transperineal Combined Gray Scale and Color Doppler Sonography
J. Ultrasound Med., January 1, 2004; 23(1): 19 - 27.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
O. Schaefer, C. Lohrmann, and M. Langer
Digital Subtraction MR Fistulography: New Diagnostic Tool for the Detection of Fistula In Ano
Am. J. Roentgenol., December 1, 2003; 181(6): 1611 - 1613.
[Full Text] [PDF]

Home page
 RadiologyHome page
P. J. Pickhardt, S. Bhalla, and D. M. Balfe
Acquired Gastrointestinal Fis