One hundred twenty-four patients with invasive carcinoma of the cervix were examined with transrectal ultrasound (TRUS) before treatment. Since surgery preceded any other kind of therapy, sonographic findings could be compared with the surgical pathologic stage. The accuracy of staging with TRUS was 83%, compared with an accuracy of 78% for clinical staging performed according to the criteria of the International Federation of Gynecology and Obstetrics (FIGO). For extent of parametrial involvement, the sensitivity of TRUS evaluation was 78%, with a specificity of 89% and a diagnostic accuracy of 87%. The same reliability parameters for clinical evaluation were 52%, 92%, and 84%, respectively. From these data and from analysis of the cases--in which a discrepancy between clinical and TRUS staging was observed--it can be concluded that TRUS could be usefully applied to routine pretreatment evaluation of patients with cervical carcinoma.