Could a 2D/3D US based model be helpful in the assessment of myometrial invasion at time of intraoperative frozen section? A pilot study.

BACKGROUND The assessment of myometrial invasion is a pivotal step in the preoperative staging of endometrial cancer. Intraoperative frozen section (FS) represents a reliable tool in directing surgeon's choices. Preoperative transvaginal ultrasound (US) showed high accuracy in evaluating myometrial invasion. This study aimed to understand if the application of a standardized ultrasonographic protocol for the pre-operative evaluation of myometrial invasion can help pathologists in improving the accuracy of FS. Furthermore, the agreement between US and FS in the assessment of myometrial invasion was assessed. METHODS Sixty-six patients who underwent surgery for endometrial cancer were analyzed. Preoperative 2D/3D ultrasound was performed in all the patients. Myometrial invasion was estimated by subjective assessment and objective measurement techniques. Data from US were reported to pathologists through a prefilled form with depth and site of the maximum myometrial invasion. Diagnostic performance of US and FS were compared having the definitive histological examination as the gold standard. RESULTS Influenced by the information given by our 3D US-model, FS showed a 90% sensitivity and a 93% specificity, with a 93% PPV and an 89% NPV. The agreement with histology was strong (K=0.824). Myometrial invasion was missed at the level of the isthmus by FS just in one case. Subjective assessment was confirmed as the most reliable ultrasonographic technique in assessing myometrial invasion, with 90% sensitivity, 78% specificity, 80% PPV and 89% NPV. The agreement with histology was substantial (K=0.68). CONCLUSIONS The application of a preoperative 2D/3D US assessment would seem to help pathologists in detecting myometrial invasion in difficult areas of the uterus such as the isthmus, reducing downstaging and overtreatment.