A pilot assessment of whether external coil MRI is useful to assess evacuatory disorders

This study assessed the value of common surface coil magnetic resonance imaging (MRI) in patients with evacuatory disorders including fecal incontinence and constipation. These findings were then compared with those from other standard physiological examinations and/or surgical findings. From July 1996 to June 1997, 14 consecutive patients underwent surface coil MRI for evaluation of either fecal incontinence (n=5) or constipation (n=9). In patients with incontinence we compared the findings from endoanal ultrasound (EAUS), anal MRI, and surgery regarding morphopathological findings of the internal and external anal sphincter components. In constipated patients the findings of videoprography and dynamic pelvic MRI were compared regarding the presence of rectocele, rectoanal intussusception, and sigmoidocele as well as the measurements of anorectal angle and perineal descent. The five incontinent patients were all women, with a median age of 67 years (range 43-77). EAUS revealed an anterior sphincter defect in two patients, a posterior defect in one, and normal anal sphincter images in two. Surgical findings confirmed an anterior external anal sphincter scar in two patients, an internal anal sphincter defect in one, and an anatomically normal anal sphincter in two. In one patient, although anal MRI showed posterior external anal sphincter defect, EAUS and surgery revealed normal external anal sphincter appearance. The accuracy rate between EAUS and anal MRI was only 20%, that between surgery and anal MRI 40%, and that between surgery and EAUS 80%. Thus EAUS was more accurate than anal MRI in incontinent patients. The nine constipated patients were all women, with a mean age of 59 years (range 40-78). Videoproctography revealed an anterior rectocele in six patients, rectoanal intussusception in three, and sigmoidocele in five; no abnormalities were identified in two patients. On dynamic pelvic MRI anterior rectocele was seen in three patients and sigmoidocele in two, and five studies were interpreted as normal. One of the patients underwent sigmoidectomy for sigmoidocele, and five patients were treated by biofeedback. Thus the accuracy rate of dynamic pelvic MRI against videoproctography was 60% for anterior rectocele, 40% for sigmoidocele, and zero for rectoanal intussusception. In conclusion, neither MRI for the evaluation of patients with fecal incontinence nor for the evaluation of patients with constipation added any significant information that would warrant its continued use in these patient groups. Perhaps the more widespread availability of an endoanal coil will alter this conclusion; however, at the present time we cannot routinely endorse the expense, time, or inconvenience of these MRI investigations in patients with these diagnoses. Larger prospective comparative studies are required prior to endorsing the technique.