PURPOSE: To determine whether gadolinium enhancement is helpful in rectal tumor staging determined by MRI and using an endorectal surface coil. MATERIALS AND METHODS: Between January 1995 and July 1996, we studied 17 MRI scans in which the scanning procedure had involved the use of an endorectal coil ; this was a prostate coil in six patients, and a colon coil in eleven. Eight patients were male and nine were female ; they were aged between 39 and 77 (mean, 59) years, and the tumors which had presented were adenocarcinoma (n=15), lymphoma (n=1) and villous adenoma (n=1). Precontrast scanning showing invasion of the rectal wall and perirectal fat were interpreted, and postcontrast T1WI and pathological findings were then compared. Fifteen patients underwent surgical resection but the other two (one adenocarcinoma and one lymphoma) underwent only an endoscopic biopsy. RESULTS: On precontrast scanning with the prostate coil, accurate staging was possible in three cases (one of stage T2, and two of stage T3) ; we overstaged two cases of stage T2 as stage T3. On postcontrast T1WI, however, we additionally understaged one case of stage T3 as stage T2. In a case of adenocarcinoma proven by biopsy, no definite difference was noted between pre- and postenhanced scan. On precontrast scan using a colon coil, accurate staging was possible in six cases (two of stage T1, one of stage T2 and three of stage T3). We overstaged a case of stage T2 as stage T3 and understaged three cases of stage T3 as stage T2. On postcontrast T1WI, however, we accurately diagnosed one additional case of stage T3, not diagnosed on precontrast scan. In one case of bioptically-proven lymphoma, no definite difference was noted between pre- and postenhanced scan. CONCLUSION: In rectal tumor staging, pre- and postenhanced scans are both 60% accurate. In MRI using an endorectal surface coil, gadolinium enhancement is not, therefore significantly helpful.