The goal of this study was to determine the sensitivity, specificity, and accuracy of transrectal ultrasound (TRUS) in comparison to magnetic resonance imaging (MRI) in the evaluation of tumor volume, early parametrial infiltration, and identification of residual tumor in early-stage cervical cancer. Patients in whom an early-stage cervical cancer was diagnosed by clinical examination were enrolled in the study. Only those patients who were examined by both MRI and TRUS with following surgical treatment were included. Imaging results were compared with pathology findings. Altogether, 120 patients were consecutively enrolled from January 2004 to February 2006. Data from 95 patients were evaluated. Correlation coefficient for TRUS- and MRI-derived volumes versus volumes at pathology reached R = 0.996 and R = 0.980, respectively. The accuracy for detecting tumor in 95 patients was 93.7% for TRUS and 83.2% for MRI (P <or= 0.006). In small tumors (<or=1 cm(3)), the accuracy of tumor detection by TRUS was 90.5% and 81.1% by MRI (P <or= 0.049). The accuracy of parametrial infiltration detection by TRUS and MRI was 98.9% and 94.7%, respectively (P <or= 0.219). The accuracy was not influenced by body mass index values. In conclusion, our results showed the accuracy of TRUS as being comparable to the more commonly used MRI in staging of early-stage cervical cancer. TRUS and the pathology-derived volumes correlated tightly; accuracy of TRUS was superior to MRI, especially in detection of residual tumors following conization.