OBJECTIVE: External anal sphincter atrophy seen at endoanal MRI may predict poor outcome of surgical anal sphincter repair for an external anal sphincter defect. The purposes of this study were to compare external phased-array MRI to endoanal MRI for depicting external anal sphincter atrophy in patients with fecal incontinence and to evaluate observer reproducibility in detecting external anal sphincter atrophy with these techniques. SUBJECTS AND METHODS: Thirty patients with fecal incontinence (23 women, seven men; mean age, 58.7 years; age range, 37-78 years) underwent both endoanal and external phased-array MRI. Images were evaluated for external anal sphincter atrophy by three radiologists. Measures of differences and agreement between both MRI techniques and of interobserver and intraobserver agreement of both techniques were calculated. RESULTS: The MRI techniques did not significantly differ in their ability to depict external anal sphincter atrophy (p = 0.63) with good agreement (kappa = 0.72). Interobserver agreement was moderate (kappa = 0.53-0.56) for endoanal MRI and moderate to good (kappa = 0.55-0.8) for external phased-array MRI. Intraobserver agreement was moderate to very good (kappa = 0.57-0.86) for endoanal MRI and fair to very good (kappa = 0.31-0.86) for external phased-array MRI. CONCLUSION: External phased-array MRI is comparable to endoanal MRI in depicting external anal sphincter atrophy and, thereby, in selecting patients for anal sphincter repair. Because results among interpreters varied considerably depending on the experience level, both techniques can be recommended in the diagnostic workup of fecal incontinence only if sufficient experience is available.