Sentinel lymph node detection in breast cancer patients by real-time virtual sonography constructed with three-dimensional computed tomography-lymphography.

Ultrasonography (US) is one tool for preoperative diagnosis of lymph node metastases in breast cancer. However, US cannot detect true sentinel lymph nodes (SLNs). We identified SLNs in 60 clinically node-negative breast cancer patients using a real-time virtual sonography (RVS) system to display in real time a virtual multi-planar reconstruction obtained from computed tomography (CT) volume data corresponding to the same cross-sectional image from US. CT volume data were obtained from our original three-dimensional CT lymphography (3DCT-LG), which accurately detects SLNs in breast cancer. SLN metastases were assessed by shape and visibility of the hilum. All patients underwent SLN biopsy and SLN metastases were examined pathologically. In all 60 patients, we were able to detect the same SLNs visualized by 3DCT-LG. Suspicious SLN metastases were identified in seven of the 60 patients, and four of seven patients were pathologically positive. Positive predictive value was 57%. The remaining 53 patients displayed non-suspect SLNs in which absence of metastasis from the SLN was confirmed histologically. Overall accuracy was 95%. This is a first attempt at preoperatively identifying SLNs using US guided by the RVS system in breast cancer patients. Although evaluation of SLN metastases was unsatisfactory, this method may be useful for preoperative fine-needle aspiration cytology for diagnosis of SLN metastases.