Sixty-three consecutive patients with nonseminomatous testicular carcinoma were prospectively examined for paraaortic metastases before undergoing a staging laparotomy with lymphadenectomies or biopsies. Fifty patients had computed tomographic examinations and paraaortic metastases were correctly predicted in 25 of the 38 patients (sensitivity 66%) with paraaortic tumor. In the 56 patients having the study, lymphography was more accurate as 34 of the 44 patients with metastases (sensitivity 77%) were identified. Using inferior vena cavography in 55 patients, only 24 of 41 patients with metastatic tumor (sensitivity 59%) were diagnoses. All three methods had specificities of 100% in those patients who were free of paraaortic metastases. Overall accuracies were 74% for computed tomography, 82% for lymphography, and 69% for inferior vena cava cavography. Although lymphography was the most accurate method for nonbulky tumor, computed tomography was most useful for defining the extent of bulk disease. Inferior vena cavography did not contribute any new information and was responsible for one false-positive result which was explained by computed tomography. A combination of lymphography followed by computed tomography provided the most accurate assessment of paraaortic metastases.