Analysis of clinical and pathological outcomes of sentinel lymph node biopsy in the SNAC Trial.

Abstract

Sentinel Lymph Node Biopsy (SLNB) has become the standard procedure to assess the status of the lymphatic drainage in early stages of breast cancer. Currently, SLNB has replaced Complete Axillary Dissection (CAD) when the lymph nodes are clinically negative. The detection of a Sentinel Lymph Node (SLN) is critical for the success of SLNB procedure, but the accurate contribution of various clinical and pathological factors to the detection of SLNs has remained a controversial issue. This study, which is based on the database of a randomised controlled surgical trial of Sentinel Node biopsy versus Axillary Clearance (SNAC), is aimed at delineating the factors that can influence the outcomes of the SLNB and determining the clinical and pathological issues can predict the pathological outcomes of both SLNs and Non Sentinel Lymph Nodes (NSLN). Furthermore, a comparison of the ability of Cambridge nomogram, Stanford nomogram and Tenon score to predict the status of non sentinel lymph nodes was performed. Method: Retrospective analysis was made of all patient data involved in the SNAC trial in 2006:1088 patients were randomised into two groups: 544 patients allocated to the first group underwent sentinel lymph node biopsy followed by CAD; the second group 544 patients underwent SLNB followed by CAD if positive SLNs were identified. A combination of three techniques including preoperative lymphoscintigraphy (LSG), intraoperative blue dye injection and gamma probe in the operation theatre were used for mapping of the SLNs. Retrieved nodes were examined via Haematoxylin and Eosin and immunohistochemistry. Results: SLNs were identified in 1024 (94.6%) patients involved in the SNAC trial. Our analysis revealed that the highest Identification Rate (IR) was achieved by using the combination of three detection techniques: SLNs were detected in 96.3% (905 out of 940). The identification rates of preoperative LSG, intraoperative gamma probe and blue dye were 81.4%, 91.8% and 83% respectively. Patients‘ weights and the mode of primary tumour presentation have a significant impact on the overall outcomes of the sentinel lymph node detection. There were variations in the outcomes of multivariate analysis of factors influencing the detection of SLNs for each technique. The False Negative Rate (FNR), which was calculated only for patients who had SLNB followed by immediate CAD, was 8.3% and no significant correlation was found between various clinicopathological factors and the FNR. The involved SLNs were detected in 291 (28.4%) patients and involved non sentinel lymph nodes were identified in 118 out of 291 patients with positive SLNs. Regarding the prediction of the SLN status, the multivariate analysis models demonstrated that the size of the primary tumour (p = 0.000), the presence of Peritumoural Vascular Invasion (PVI) (p = 0.000), the primary tumour palpability (p = 0.036) and the site of the primary tumour (p = 0.038) were the most significant factors to predict the histopathological status of SLNs. The primary tumour size (p=0.015) and the diameter of the metastatic lesion (p=0.009) are the most significant predictors of the non sentinel lymph node status. Our validation of three nomograms to predict the positivity of NSLN revealed that the areas under the Receiver operating characteristics (ROC) curves were 0.697, 0.714 and 0.724 for Cambridge nomogram, Stanford nomogram and Tenon score respectively. Conclusion: the combined technique is superior to the single technique for SLN detection. The appropriate selection of patients to undergo SLNB procedure can minimise the failure of the sentinel lymph node detection. The presence of involved SLNs can be predicted from the primary tumour characteristics. The diameter of the primary tumour and the size of the metastatic lesion are independent predictors of NSLN involvement.