MR staging of clinical stage I and IIa cervical carcinoma: a reappraisal of efficacy and pitfalls.

The purpose of this study was to evaluate the diagnostic efficacy and pitfalls of magnetic resonance (MR) imaging in preoperative staging of cervical cancer. MR imaging was performed to determine the tumor staging for 31 patients with cervical carcinoma emphasizing tumor size, parametrial invasion, vaginal invasion and lymph node metastases. Tumor size was 3.23+/-1.75 cm (mean+/-standard deviation) at MR imaging compared with 2.79+/-1.76 cm at surgical-pathologic evaluation. The discrepancy between the tumor size determined by MR imaging and the measured surgical specimens was consistent in tumors larger than 1 cm. In assessing parametrial invasion, vaginal invasion and lymph node metastases, MR imaging had an accuracy of 96.7 and 87%. In determining stage of disease and differentiating operable (< or =stage IIA) from advanced disease (> or =stage IIB), MR imaging had an accuracy of 83.8 and 96.7%. Pitfalls leading to staging errors included difficulties in differentiating cancer foci from surrounding tissue edema and excluding vaginal invasion in the presence of large cervical cancer. In conclusion, MR imaging is accurate in the evaluation of parametrial invasion and useful in the differentiation of operable from advanced disease. The ability of MR imaging to exclude vaginal invasion in the presence of large cervical cancer and differentiate cancer foci from surrounding tissue edema is not as reliable.