To compare diagnostic performance of preoperative transvaginal ultrasound (TVS) and intraoperative macroscopic examination for determining myometrial infiltration in women with low-risk endometrial cancer, and to estimate the agreement between the two methods.
This was a single-center observational study comprising women with preoperative diagnosis of well- or moderately differentiated endometrioid carcinoma of the endometrium. All women underwent preoperative TVS by a single examiner. According to the examiner's subjective impression, myometrial infiltration was stated as¿¿¿50% or¿<¿50%. Surgical staging was performed in all cases. Intraoperative macroscopic examination of the removed uterus was performed by pathologists who were unaware of the ultrasound findings, and myometrial infiltration was stated as¿¿¿50% or¿<¿50%. Definitive histological diagnosis of myometrial infiltration was made by frozen section analysis and was used as the gold standard. Sensitivity and specificity with 95% CIs were calculated for TVS and intraoperative macroscopic inspection and compared using McNemar's test. Agreement between TVS and intraoperative macroscopic inspection was estimated using Cohen's kappa index (¿) and percentage of agreement.
Of 209 eligible women, 152 were ultimately included. Mean (±¿SD) age was 60.9¿±¿10.2¿years, with a range of 32-91¿years. Definitive histological diagnosis revealed that myometrial infiltration was <¿50% in 114 women and ¿¿50% in 38 women. Sensitivity and specificity of TVS for detecting deep myometrial infiltration were 81.6% and 89.5%, respectively, whereas the respective values for intraoperative macroscopic examination were 78.9% and 90.4% (McNemar's test, P¿>¿0.05 when comparing TVS and intraoperative macroscopic examination). Agreement between methods was moderate with ¿¿=¿0.54 (95%¿CI, 0.39-0.69) and percentage of agreement of 82%.
Although the agreement between preoperative TVS and intraoperative macroscopic examination for detecting deep myometrial infiltration was only moderate, both methods had similar accuracy when compared with frozen section histology. Preoperative TVS might reasonably be proposed as a method for assessing myometrial infiltration as an alternative to intraoperative macroscopic examination, especially when performed by an experienced examiner and image quality is not poor.