RESUMO
OBJECTIVES: To determine the radiological modalities that provide at the initial workout the most accurate information regarding the operability and the type of operation for patients with primary cervical, endometrial, and ovarian carcinomas. PATIENTS AND METHODS: The medical records of 611 patients with gynaecological cancer were reviewed. The preoperative radiological findings were compared with the intraoperative and pathological ones. The diagnostic accuracy of CT and MRI at various sites was evaluated for all three kinds of carcinoma in combination. RESULTS: MRI was more accurate than CT in determining cervical enlargement (82% vs 73%), parametrial invasion (91% vs 74%) and the only examination that could evaluate cervical tumour size as well as cervical stromal and myometrial infiltration. Regarding lymph node involvement their results were similar (86% vs 88%). Both methods were comparably accurate in evaluating ovarian tumours (82% vs 84%), ascites (82% vs 81%), omental (73% vs 77%) and mesenterial infiltration (88% vs 93%). They also proved to be highly accurate (100% vs 98%) in the evaluation of solid abdominal organs. CONCLUSIONS: Non-enchanced MRI should only be used for the preoperative evaluation of a patient with cervical carcinoma, while CT with intravenous and per os contrast media for one with ovarian cancer. Regarding patients with endometrial cancer, no high-resolution method is required for endometrioid grade I tumours, while contrast-enchanced MRI should be employed for all other cases.