To evaluate the efficacy of MRI in bladder or rectal invasion of cervical cancer.

Between January 2000 and December 2005, 106 cervical cancer patients who underwent cystoscopy or sigmoidoscopy examination retrospectively compared with results of the MRI finding. A 5-point invasion score was used to determine bladder or rectal invasion in MRI (1 = no invasion, 3 = serosal invasion, 5 = definite mucosal invasion). Score of 3 or above was suspicious for both rectal and bladder invasion.

Eighty one patients with negative for both rectal and bladder invasion in MRI were normal in cystoscopy and sigmoidoscopy. MRI identified 17 patients with suspected bladder invasion, 7 patients had confirmed bladder invasion in cystoscopy. MRI identified 11 patients with suspected rectal invasion, 1 patients had confirmed rectal invasion in sigmoidoscopy. The detection of rectal and bladder invasion by MRI had a negative predictive value of 100%.

The use of a 5-point scoring system for predicting rectal or bladder invasion in MRI is accurate in detecting cervical cancer patients with no evidence of rectal or bladder invasion. The cervical cancer patients with no evidence of rectal or bladder invasion in MRI can obviate invasive cystoscopy and sigmoidoscopy.