|Objectives:To assess the accuracy of magnetic resonance imaging (MRI) and transrectal ultrasound (TRUS) in the evaluation of tumor size and in thedetection of residual tumor following neoadjuvant chemotherapy (NACT) in patients with cervical cancer. Methods:This was a prospective study involving 42 women with locally advanced histologically confirmed cervical cancer referred for NACT. Clinical examination, TRUS and MRI were performed before and after NACT. The tumor volume was calculated using three standardized diameters (anteroposterior, laterolateral and craniocaudal) that were measured using both TRUS and MRI. Thereafter patients underwent surgical treatment and the same tumor measurements were taken by a pathologist using a fixed surgical specimen. Tumor volumes were calculated from tumor dimensions using the ellipsoid formula, and data obtained from both imaging methods were compared with pathological results as the gold standard. Results:Twelve cases were excluded from the study owing to disease progression (these patients were referred for primary radiotherapy) or inability to perform MRI, leaving data from 30 patients for the final analysis. On average, tumor volume decreased after NACT by 84.6 and 87.1% as measured by MRI and TRUS, respectively. The agreement between measurements obtained by MRI and histology did not reach significance (intraclass correlation coefficient, 0.344 (95% CI, -0.013 to 0.610), P=0.059), while agreement between TRUS and histology was statistically significant (intraclass correlation coefficient, 0.795 (95% CI, 0.569-0.902), P<0.001). The accuracy of residual tumor detection (for non-microscopic tumors >5mm(3) in volume) reached 77% for both MRI and TRUS. The sensitivity of TRUS was, however, lower than that of MRI (83 vs 96%). The positive predictive values were similar for the two methods. Conclusions:TRUS should be considered as an accurate diagnostic method in the evaluation of tumor volume after NACT in patients with cervical cancer and may constitute a reliable alternative imaging method to MRI.