Rapid Molecular RT-PCR Assay To Detect Metastases in Sentinel Lymph Nodes of Breast Cancer Patients May Accurately Predict Further Axillary Positivity.

Abstract

Abstract Background: Data on the clinical cutoff and validation trials of the Breast Lymph Node (BLN) Assay were presented at the San Antonio Breast Cancer Symposia in 2006 and 2007. These trials led to FDA approval on July 17, 2007 as a stand alone test for intraoperative decision-making for complete axillary node dissection (cALND) on patients with invasive breast cancer. The goal of intraoperative assays is to reduce the 2nd surgery cALND, but also not to perform cALND on non-clinically significant disease by current staging (N0 (i+) or ITCs). This report is the initial post-market clinical data on 391 consecutive cases of invasive breast cancer from a single institution, single surgeon experience.Methods: The sentinel lymph nodes (SLN) were examined using the FDA approved cutting scheme. 50% of each node (alternating slices) was homogenized & analyzed by real-time RT-PCR markers: cytokeratin 19 & mammaglobin. Typically the BLN assay gives a qualitative result (pos/neg), but can provide quantitative values, currently for research use only. Only the on-label qualitative result was used for patient management. Remaining slices were submitted for frozen section (FS) & permanent section H&E (PS). All patients that were positive by FS, PS or the BLN assay had cALND.Results: The BLN assay performance in clinical use compared to PS on SLNs is similar to trial data: sensitivity 92% & specificity 92%. FS performance compared to PS is sensitivity 68% & specificity 99%; low sensitivity is in part due to an increase of micromets. Assay turn around time is 36 min. 89 patients had cALNDs. BLN Assay qualitative results have similar PPV to PS but greater NPV. Looking at factors commonly in nomograms (age, tumor size, grade, PS results and ER/PR/HER2 status) to predict cALND positivity none had statistical significance, but the ROC curve for all these factors together (baseline) had a 0.69 AUC. Comparatively BLN assay quantitative results were statistically significant & produce a ROC curve with 0.72 AUC. When BLN assay is added to baseline, no variables are statistically significant, mostly due to small sample size, but AUC increases (0.75). However the ROC curve seems most promising with BLN assay alone (Figure 1). More data is needed for statistical significance.Discussion: Although there are many nomograms & clinical prediction rules published, often all the data points are not available intraoperatively & have limited reliability. The alternative is to wait for PS & then decide whom to bring back for further surgery. This leads to increased cost, stress on the patient & staff if another procedure is recommended. Conversely, a prediction rule that can help decide when not to proceed to a cALND even with SLN positivity would be extremely useful. The BLN Assay offers a good prediction of ALND positivity that adds value when added to other risk factors & may add the same or more value alone intraoperatively. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 1016.