Endoluminal MR imaging of diseases of the anus and rectum

Endoluminal MR Imaging of Diseases of the Anus and Rectum Jaap Stoker and Elena Rociu Endoluminal MRI of the rectum and anus was introduced in the first half of this decade to overcome the limitations of endoluminal sonography and body coil MRI. Endoluminal MRI is the imaging method of choice for fecal incontinence and anal tumors, whereas it is a competitive imaging method to phased array coil MRI in patients with perianal fistulas or rectal tumor. The purpose of this article is to describe the technique and major indications of endoluminal MR imaging of the anus and rectum.

Copyright© 1999by W.B. Saunders Company

HE ROLE OF THE radiologist in the diagnostic work-up of patients with anorectal diseases is expanding. Initially, barium studies and fistulographies were used. The subsequent introduction during the last two decades of computer tomography (CT), defecography, endoluminal ultrasound, magnetic resonance imaging (MRI), and endoluminal MRI has broadened the armamentarium significantly. Currently, MRI is the preferred imaging technique for many anorectal diseases. The widespread use of MRI and, to a lesser extent endoluminal sonography, has diminished the role of CT to relatively sparse indications. This role of MRI has resulted from the multiplanar capabilities and inherent high contrast resolution of MRI. The use of an endoluminal coil combines this high inherent contrast with high spatial resolution, resulting in an excellent demonstration of anorectal anatomy and pathology. 1-28 In this article, the methods to perform endoluminal MR imaging as well as the role of endoluminal MR imaging in patients with anorectal diseases are discussed. Emphasis will be on the major indications for endoluminal MR imaging: anorectal tumors, perianal fistulas, and fecal incontinence.

T

TECHNIQUE

For endoluminal MRI of the rectum, an endoluminal coil with a balloon can be used. 1-6,17,19,20,21 These coils were developed for MRI of the prostate and are adequate coils for imaging rectal pathology. However, this type of coil is not adequate for endoluminal MRI of the anal region. Dedicated rigid, cylindrical anal coils have been developed and have proven to give a detailed demonstration of the anal region. 7-16,18,2a,24-28Because dedicated endoanal coils can also be used for endorectal imaging, this type of coil is preferred at our institutions for both anal and rectal imaging. Endoluminal MRI of the anus and rectum is performed with off-axis sequences perpendicular or parallel to the anorectum. The axial plane is the

most important diagnostically, but longitudinal planes are mandatory to determine the superiorinferior extent of tumor and to minimize partial volume effects. Coronal or sagittal planes can also be used. The use of a radial sequence is a more time-efficient alternative to the combination of coronal and sagittal sequences? 6,27 The optimal imaging sequences have not been established. Besides data from the literature, the choice is influenced by experience and personal preferences. The anorectal anatomy and pathology are accurately demonstrated by T2w turbo spin-echo (TSE) or proton density (PD)w gradient-echo (GRE), and also by Tlw TSE with intravenous contrast medium. Fat suppression techniques can be valuable in patients with perianal fistulas to facilitate fistula identification. For rectal tumors, T2w TSE is preferred for imaging the relationship of the tumor to the muscular part of the rectal wall, whereas the spread of tumor through the rectal wall into the perirectal fat is somewhat more conspicuous with Tlw sequences. 27Axial and longitudinal sequences are mandatory to reduce partial volume effects. The use of an endoluminal coil and a small pixel size make the endoluminal technique relatively sensitive for motion artefacts. Patient motion can be reduced by careful positioning and adequate patient instruction. 27 Rectal motion is the other major reason for motion artefacts. This can be reduced by the use of butyl scopalamine bromide (Buscopan, Boehringer, Ingelheim, Germany) or glucagon, administered before imaging. The authors prefer intramuscular injection because this

From the Department of Radiology, Academic Medical Center, Amsterdam, and the Department of Radiology, University Hospital Rotterdam Dijkzigt, Rotterdam, The Netherlands. Address reprint requests to Jaap Stoker, MD, PhD, Department of Radiology, Academic Medical Center, PO Box 22700, 1100 DE Amsterdam, The Netherlands. Copyright © 1999 by W.B. Saunders Company 0887-2171/99/2001-0001510.00/0

Seminars in Ultrasound, CT, and MRI, Vol 20, No 1 (February), 1999: pp 47-55

47

48

STOKER AND ROCIU

endoluminal coil compresses the sphincter components, it does not significantly alter the relationship between the sphincter structures. Therefore, the MRI concept of anal anatomy is not caused by the presence of the coil, but by the detailed multiplanar demonstration of anal anatomy. The upper part of the anal sphincter is continuous with the rectum. The rectal wall comprises the mucosa, submucosa, and the circular and longitudinal muscular layers. The levator ani is connected to the anal sphincter complex at the level of the anorectal junction. ANORECTAL TUMORS

Fig 1, Rectal tumor. Radial T2w TSE of the anorectum demonstrates a tumor (T) in the distal rectum, The tumor does not extend into the perirectal fat and is close (arrow) to the anorectal junction.

Tumors of the rectum are relatively common, whereas anal tumors are relatively rare. The major role for imaging in anorectal tumors is staging. Staging is important because the spread of the disease will influence the type of therapy. Important items to be addressed are extension of tumor in or through the wall, the tumor-anus distance, the presence of enlarged lymph nodes, and distant metastases. Developments in treatment have made the tumor-anus distance relevant because anus-

results in a longer effective reduction of the rectal contractions than the intravenous route. ANATOMY

The anal sphincter is formed by several muscles. The innermost muscle is the internal sphincter, which is the continuation of the circular muscle of the rectum. Outside the internal sphincter is the fat-containing intersphincteric space. In this space is the longitudinal muscle, which is the continuation of the longitudinal muscle of the rectum. The intersphincteric space is bordered by the external sphincter and the puborectal muscle. The external sphincter surrounds the lower half of the intersphincteric space, and the puborectal muscle surrounds the upper half of the intersphincteric space. This concept of the anal anatomy, based on endoluminal MRI, is different from previous concepts with regard to the external sphincter. 7,16 In the previous concepts, based on surgery or dissection, the external sphincter was thought to comprise almost the complete outer part of the sphincter. The current concept results from the detailed demonstration of the anal anatomy and the multiplanar capabilities of endoanal MRI, which facilitates the identification of the external sphincter and puborectal muscle more accurately. Although the presence of an

Fig 2, Rectal tumor. Axial T2w TSE reveals a tumor (T) in the rectum not extending into the muscular part of the rectal wall (M). The window setting is optimized for demonstrating the tumor, with high signal intensity close to the coil.

ENDOLUMINAL MRI OF ANORECTAL DISEASES

saving (and thus continence-saving) surgery can be performed. Prognosis is influenced by the presence or spread of the disease through the wall and the presence of lymph node and distant metastases. Recent developments in medical imaging have had major impact on the diagnostic work-up of patients with anorectal tumor. Currently, endolumi2 nal sonography or MRI are used for local staging, whereas either MRI or CT are used for the detection of distant metastases. Local staging with MRI can be performed with body coil MRI, endoluminal MRI, or phased array MRI. Endoluminal MRI has the advantage of combining high intrinsic contrast resolution with high spatial resolution provided by an endoluminal coil. The detailed demonstration of the tumor with endoluminal MRI proved to be accurate in de,termining the local tumor spread 2,3,19,21 (Figs 1 through 3). The accuracy is approximately 85%. 3-6'17'19-21 The disadvantage of endoluminal MRI is the need for additional sequences with the body coil or phased array coil for the detection of distant metastases. Endoluminal MRI is not possible in patients with tight strictures and those with

49

relatively high tumors. Phased array coil MRI is not hindered by tight or high strictures, and the use of a large field of view with high spatial resolution allows the complete staging with the use of one coil. Comparative studies of endoluminal MRI and phased array MRI are sparse. 19 The rectal wall is more accurately demonstrated with endoluminal MRI, but better resolution does not result in improved tumor staging. These data indicate that phased array coil MRI may be the preferred imaging technique for staging rectal tumors. 28 PERIANAL FISTULAS

Perianal fistulas can occur without predisposing disease, or in patients with Crohn's disease. Inflammation of the cryptal glands in the anal submucosa results in the development of a track. In general, one track will develop, but multiple tracks may occur, especially in patients with Crohn's disease. Commonly, the track will only extend downward toward an external opening at the skin, but upward extension to or above the levator ani muscle also may occur.

Fig 3. Rectal tumor. (A) Axial T2w TSE reveals a tumor (T) in the rectum. The tumor does not seem to invade the muscular layer, but some doubts remain, probably caused by partial volume effects. An additional sequence in a longitudinal plane is necessary to fully appreciate the relation of the tumor to the wall. (B) Radial coronal oblique T2w TSE demonstrates that the tumor (T) in the distal rectum visible in (A), does not invade into the muscular layer. No enlarged lymph nodes are visible and the distance between tumor and anorectal junction (arrow) is sufficient for sphincter saving surgery.

50

STOKER AND ROCIU

Fig 4. Perianal fistulas. (A) Axial T2w TSE in a patient with Crohn's disease reveals multiple hyperintense tracks (T). Both intersphincteric tracks (tracks confined to the intersphincteric space) and transsphincteric tracks (tracks extending through the external sphincter) are present. (B) Axial SPIR (fat sat) T2w TSE at the same level. The hyperintense tracks are more conspicuous with the use of fat suppression, but the relationship of the tracts to the parts of the anal sphincter is more difficult to appreciate. (C) Radial coronal T2w TSE shows that the tracks do not extend into the supralevator space. At the right, parts of a complex transsphincteric track (Tr) is visible and at the left an intersphincteric track (IT) is seen. A, anal sphincter.

The role of imaging in perianal fistula is in describing the relation between the course of the track and the several components of the anal sphincter. This is important for proper manage1

ment, because the type of track determines the type of surgery. Commonly, the classification according to Parks et a129is used. In this classification, a track confined to the internal sphincter and intersphinc-

ENDOLUMINAL MRI OF ANORECTAL DISEASES

51

Fig 5. Perianal abscess. (A) Coronal T2w TSE reveals an abscess (A) within the right puborectal muscle and levator ani muscle and extending into the supralevator space (SLS). Compare with the left side for normal anatomy of the puborectal muscle (PRM) and levator ani muscle (LAM): (B) Axial SPIR (fat sat) T2w TSE demonstrates the large internal opening of this abscess (A) at right posteriorly.

teric space is an intersphincteric fistula. Extension of a track through the external sphincter is classified as a transsphincteric fistula. A suprasphincteric track extends above the levator ani muscle, and an extrasphincteric fistula has a course outside the anal sphincter to the rectum: A track can be simple, complex, or may have a horse-shoe shape. Initially, fistulography was used for classification. However, this procedure is often cumbersome and provides only limited results.30 A major drawback of fistulography is that the classification is indirect because the anal sphincter is not visualized. In the 1980s, endoluminal sonography was introduced and resulted in improved classification. 31 A limitation of endoluminal sonography is the troublesome identification of the external sphincter. The relation of a track to the internal sphincter can be accurately shown, but the relation to the external sphinct e r - o f major importance for classification--often has to be inferred. MRI, using a body coil, was introduced and proved to be a valuable technique. 32-34 However, the spatial resolution was limited, which resulted in incorrect identification

and classification of s o m e f i s t u l a s . 11 Endoluminal MRI results in a detailed demonstration of the anal anatomy, including the external sphincter. The relation of a track to the anal sphincter is more accurately identified with endoluminal MRI than with body coil MRI or endoluminal sonography. 11,12 This advantage results in better fistula identification and classification (Figs 4 through 7). The axial plane is the preferred imaging plane, whereas additional longitudinal imaging planes are used to improve classification 26 (Figs 4, 5). Coronal and sagittal imaging planes, or alternatively a radial sequence, can be used. With a radial sequence the slices have a spook wheel-like orientation. H,26T2w sequences, fat suppression, or T l w sequences after intravenous contrast medium can be used. The authors prefer an axial T2w sequence combined with an axial fat suppression sequence (STIR-TSE or fat saturation T2w TSE) and an additional T2w sequence in a longitudinal plane. 27 The fat suppression technique may facilitate the identification of tracks, but classification of tracks is not improved with fat suppression because the relationship of the

52

STOKER AND ROCIU

Fig 6. perianal abscess. (A) Axial T2w TSE shows a large horse-shoe shaped abscess (A) in the intersphincteric space and extending into the puborectal muscle (PRM) at the right. (B) On a SPIR (fat sat) T2w TSE image at the same level, the abscess is somewhat more conspicuou s, but the relationship of the abscess to the muscular sphincter parts is more difficult to appreciate.

track to the sphincter is more difficult to appreciate when fat suppression is applied (Figs 4 through 7). Tracks are often easily detected. Veins in the submucosal space, intersphincteric space, or ischioanal space may be mistaken for a track, and hemorrhoids can be mistaken for an abscess. 27 The former pitfall may be prevented by identifying the tortuous course of the thin-walled veins. A hemor, rhoid can be detected by identifying the continuity of the thin-walled hemorrhoid with surrounding vessels. FECAL INCONTINENCE

Fecal incontinence occurs in 2% of the population at age 45 years and older, and results in considerable discomfort that may lead to social isolation. 35 Anal sphincter defects and pudendal neuropathy are major causes of fecal incontinence. Treatment can be conservative (eg, rectal cleaning by enema, drugs, biofeedback), or surgical. The surgical treatment, in the majority of patients, is directed toward repair of defects of the external sphincter. Pudendal neuropathy, therefore, can negatively influence the outcome of treatment. The role

for imaging is in identifying anal sphincter defects, which may be treated by surgery. Until this decade, digital examination, anal manometry, and electromyography were used in the detection of anal sphincter defects. Digital examination and anal manometry proved to be too insensitive for the detection of sphincter defects. Electromyography has the advantage that it gives direct functional information, but major drawbacks are that the method is blind and invasive. Electrodes are commonly placed at four standard positions, and therefore defects between these points will be missed. Endoanal sonography was the first imaging technique to visualize the anal sphincter muscles and had significant impact on patient management. The technique is now part of the routine diagnostic work-up of patients with fecal incontinence. However, there are limitations caused by the limited contrast resolution. The differentiation between the external sphincter and the surrounding fat is often imprecise, resulting only in visualization of a hyperechoic region outside the internal sphincter. This region includes the intersphincteric space, the longitudinal muscle, the external sphincter, and the

ENDOLUMINAL MRI OF ANORECTAL DI~;EASES

53

Fig 8. Fecal incontinence, Axial PDW GRE reveals an anterior defect (arrows) of the internal sphincter (IS), longitudinal muscle (LM), and external sphincter (ES),

body coil MRI is too limited for adequate visualization of the sphincter complex. 36With the introduction of endoanal MRI, the accurate demonstration of anal anatomy, including the external sphincter, has become possible. Recent studies using endoanal MRI have led to new insight concerning the anal anatomy (see anatomy section).7,16This superior visualization of the anal anatomy results in the improved detection of external sphincter lesions (Fig 8). In a study of 22 patients, comparing

Fig 7. Perianal fistula. (A) Axial T2w TSE demonstrates a track (T) at the right. Posteriorly, a second hyperintense structure is visible, possibly a second track (arrow). (B) Axial SPIR (fat sat) T2w TSE confirms that the hyperintense structure at the right in (A) is a track, and reveals that the second structure is not a track but is fat, because it is not visible using a fat suppression technique.

ischioanal space. Lesions of the external sphincter are defined as disturbance of this hyperechoic region, but the external sphincter itself is often not directly visualized. For optimal treatment, the accurate visualization of the external sphincter is mandatory. MRI has inherent contrast resolution that is superior to sonography. The spatial resolution of

Fig 9. Fecal incontinence. Axial PD W GRE reveals a large anterior defect (arrows) of the external sphincter (ES) a s well as atrophy of the external sphincter. The internal sphincter is intact.

54

STOKER AND ROCIU

i m a g i n g results to surgical findings, endoanal M R I was more accurate than endoanal sonography in detecting sphincter lesions. 24 M R I detected 20 of the 22 external sphincter lesions, whereas sonography detected 16 of the 22 external sphincter lesions. E n d o a n a l M R I was also m o r e accurate in detecting the type of sphincter lesion (eg, defect, scar). A n o t h e r advantage of e n d o a n a l M R I is the detection of external sphincter atrophy, which is not possible with e n d o a n a l sonography (Fig 9). In a study of 15 patients, endoanal M R I proved to be accurate in the detection of external sphincter atrophy as compared with histopathology. 25 In a retrospective study, the presence of external atrophy proved to be an important predictor for the negative o u t c o m e of anal repair. 22

techniques in demonstrating the anal a n a t o m y and lesions of the anal sphincter in patiems with fecal incontinence. For perianal fistulas and anorectal tumors, e n d o l u m i n a l M R I is superior to e n d o l u m i hal sonography and body coil MRI. W h e t h e r endol u m i n a l M R I or phased array coil M R I is the optimal technique for i m a g i n g perianal fistulas has to be established. Phased array coil M R I has the advantage of a large field of view facilitating the evaluation of extensive tracks. E n d o l u m i n a l M R I has the advantage of the highest spatial resolution, which facilitates the detection of subtle, albeit important; tracks. For rectal tumors the use of phased array coil M R I seems to be the preferred i m a g i n g technique, b u t m o r e comparative studies of e n d o l u m i n a l M R I and phased array M R I are needed.

CONCLUSION

E n d o l u m i n a l M R I is an i m p o r t a n t part of the diagnostic work-up of patients with anorectal diseases. E n d o l u m i n a l M R I is superior to all other

ACKNOWLEDGMENT

The authors acknowledge Teun Rijsdijk and Andries Zwamborn for preparing the photographs.

REFERENCES 1. Imai Y, Kressel HY, Saul SH, et al: Colorectal tumors: An in vitro study of high resolution MR Imaging. Radiology 177:695-701, 1990 2. Chart TW, Kressel HY, Milestone B, et al: Rectal carcinoma: Staging at MR imaging with endorectal surface coil. Radiology 181:461-467, 1991 3. Schnall MD, Furth EE, Rosato EF, et al: Rectal tumor stage: Correlation of endorectal MR imaging and pathologic findings. Radiology 190:709-714, 1994 4. Murano A, Sasaki F, Kido C, et al: Endoscopic MRI using 3D-spoiled GRASS (SPGR) sequence for local staging of rectal carcinoma. J Comput Assist Tomogr 19:586-591, 1995 5. Masi A, Oimastroni M, Lascialfari L, et al: Endorectal magnetic resonance imaging for the locoregional staging of rectal carcinoma. Radiol Med 90:431-437, 1995 6. Joosten FBM, Jansen JBMJ, Joosten HJM, et al: Staging of rectal carcinoma using MR double balloon surface coil, MR endorectal coil, and intrarectal ultrasound: Correlation with lfistopathologic findings, J Comput Assist Tomogr 19:752-758, 1995 7. Hussain SM, Stoker J, La.m6ris JS: Anal sphincter complex: Endoanal MR imaging of normal anatomy. Radiology 197:671-677, 1995 8. deSouza NM, Kmiot WA, Puni R, et al: High resolution magnetic resonance imaging of the anal sphincter using an internal coil. Gut 37:284-287, 1995 9. deSouza NM, Puni R, Gilderdale DJ, et al: Magnetic resonance imaging of the anal sphincter using an internal coil. Magn Reson Q 11:45-56, 1995 10. deSouza NM, Puni R, Kmiot WA, et al: MRI of the anal sphincter. J Comput Assist Tomogr 19:745-751, 1995 11. Stoker J, Hussain SM, van Kempen D, et al; Endoanal coil in MR imaging of anal fistulas. AJR 166:360-362, 1996 12. Hussain SM, Stoker J, Schouten WR, et al: Fistulo-in-

ann: Endoanal sonography vs endoanal MR imaging in classification . Radiology 200:475-481, 1996 13. Hussain SM, Stoker J, Schtitte HE, et al: Imaging of the anorectal region. Enr J Radio122:116-122, 1996 14. Stoker J, Hussain SM, Lam&is JS: Endoanal magnetic resonance imaging versus sonography. Radiol Med 92:738-741, 1996 15. deSonza NM, Hall AS, Puni R, et al: High resolution magnetic resonance imaging of the anal sphincter using a dedicated endoanal coil. Comparison of magnetic resonance imaging with surgical findings. Dis Colon Rectum 39:926-934, 1996 16. Hussain SM, Stoker J, Zwarnborn AW, et al: Endoanal MR/ of the anal sphincter complex: Correlation with crosssectional anatomy and histology. J Anat 189:677-682, 1996 17. Indinnimeo M, Grasso RF, Cicchini C, et al: Endorectal magnetic resonance imaging in the preoperative staging of rectal tumors. Int Surg 81:419-422, 1996 18. deSouza NM, Gilderdale DJ, MacIver DK, et al: High resolution MR imaging of the anal sphincter in children: A pilot study using endoanal receiver coils. AJR 169:201-206, 1997 19. Blomqvist L, Holm T, Rubio C, et al: Rectal tumours. MR imaging with endorectal and/or phased array coils, and histopathological staging on giant sections. Acta Radiol 38:437444, 1997 20. Zagoria RJ, Schlarb CA, Ott D, et al: Assessment of rectal tumor infiltration utilizing endorectal MR imaging and comparison with endoscopic rectal sonography. J Surg Oncol 64:312-317, 1997 21. VoglTJ, Pegios W, Mack MG, et al: Accuracy of staging rectal tumors with contrast-enhanced transrectal MR imaging. AJR 168:1427-1434, 1997 22. Rociu E, Stoker J, Briel JW, et al: Endoanal MR imaging evaluation of sphincter atrophy. Radiology 205(P):453, 1997

ENDOLUMINAL MRI OF ANORECTAL DISEASES

23. Paley MR, Ros PR: MRI of the rectum: non neoplastic disease. Eur Radiol 8:3-8, 1998 24. Rociu E, Stoker J, Eijkemans MJC, et al: Endoanal sonography versus endoanal MR imaging in fecal incontinence. Radiology (accepted for publication) 25. Stoker J, Rociu E, Briel JW: External sphincter atrophy on endoanal MRI: Correlation with histopathology. Proceedings ISMRM Sixth Annual Meeting, Sydney, Australia, 18-24 April 1998, p 1025 26. Stoker J, Jong Tjien Fa VE, Eijkemans MJC, et al: Endoanal MRI of perianal fistulas: The optimal imaging planes. Eur Radiol 8:1212-1216, 1998 27. Stoker J, Rociu E, Zwamborn AW, et al: Endoluminal MR imaging of the rectum and anus: Technique, applications and pitfalls. Radiographics 1999 (accepted for publication) 28. Stoker J, Rociu E: Endoluminal MR imaging of anorectal diseases. JMRI 1999 (accepted for publication) 29. Parks AG, Gordon PH, Hardcastle JD: A classification of fistulo-in-ano. B r J Surg 63:1-12, 1976

55

30. Kuijpers HC, Schulpen T: Fistulography for fistula-inano. Dis Colon Rectum 28:103-104, 1985 31. Choen S, Burnett S, Bartram CI, et al: Comparison between anal endosonography and digital examination in the evaluation of anal fistulae. Br J Surg 78:445-447, 1991 32. Lunniss PJ, Armstrong P, Barker PG, et al: Magnetic resonance imaging of anal fistulae . Lancet 340:394-396, 1992 33. Barker PG, Lunniss PJ, Armstrong P, et ai: Magnetic resonance imaging of fistula-in-ano; Technique, interpretation and accuracy. Clin Radio149:7-!3, !994 34. Myhr GE, Myrvold HE, Nilsen G, et al: Perianal fistulas: Use of MR imaging for diagnosis. Radiology 19k545-554 , 1994 35. Gordon PH, Nivatvongs S: Principles and practice of surgery of the colon, rectum and anus. St. Louis, MO, Quality Medical Publishing Inc., 1992 36. Sch~fer A, Enck P, Ft~rst G, et al: Anatomy of the ai!al sphincters. Comparison of anal sonography to magnetic resonance imaging. Dis Colon Rectum 37:777-781, 1994