Using MR imaging with a body coil parametrial invasion was determined prospectively in 169 consecutive patients considered on the basis of clinical examination to have carcinoma confined to the cervix. After radical hysterectomy correlation with histologic examination was performed for the left and right parametrium separately. The criterion for parametrial invasion was a high-signal-intensity lesion with disruption of the full thickness of the cervical stroma combined with areas of abnormal signal intensity within the parametrial region on T2-weighted images. Histologic examination showed that 18 parametria in 13 patients were invaded by tumor. MR had an overall accuracy of 93%, a sensitivity of 89%, and a specificity of 93% in demonstrating parametrial involvement. Positive and negative predictive values were 43% and 99%. The main weakness of MR was 21 false-positive tests. This represents a limitation when MR is performed with a body coil.