Interest of axillary lymph node dissection in the case of micrometastatic invasion of the sentinel lymph node: Evaluation of three predicting models of risk of nonsentinel lymph node involvement.

Abstract

e11568 Background: To evaluate three predicting risk models of non sentinel lymph node (NSLN) involvement in case of micrometastatic sentinel node (SLN) involvement for breast cancer. Methods: This retrospective study included 72 successive patients with a micrometastatic SLN involvement who had surgery between March 1996 and October 2007. All patients had an immediately or delayed axillary lymph node dissection (ALND). The MSKCC nomogram, the Stanford nomogram and the Tenon score were applied to our population in order to calculate the probability of NSLN involvement. Statistical analysis with establishment of a ROC curve was applied to each model. Results: Concerning the MSKCC nomogram, w ith a threshold value of 10%, sensitivity was 50%, specificity was 70%, and negative predictive value (NPV) was 89%. The area under the ROC curve (AUC) was 0.6 (significant). ALND was avoided in 49 patients (68%) but it failed to detect 5 out of 10 (50%) patients with NSLN involvement. When using a threshold of 7%, the AUC was 0.69 with a sensitivity of 90% and a NPV of 97%. It avoided 31 out of 72 ALND (43%) while having only 3% chance of leaving metastasis when abstaining from ALND. Concerning the Tenon score, using a threshold of 3.5, the sensitivity lowered to 50% while the specificity was 72% and the AUC 0.62. It was not clinically applicable because 8 out of 10 patients (80%) with NSLN involved were spared from a necessary ALND. Concerning the Stanford nomogram, our results were not interpretable because the AUC was not significant. Conclusions: None of these three models was sufficiently reliable to be used in the daily practice. The MSKCC nomogram showed the most encouraging results especially for a threshold of 7% which is not validated in the literature. Completion axillary dissection should be performed as a standard of management in case of micrometastatic SLN involvement until more available data.