Validation of nomograms to predict non-sentinel lymph node metastasis after positive sentinel lymph node in breast cancer: Indian cohort background.

Abstract

568 Background: Axillary lymph node metastasis is still the important prognostic factor in the management of breast cancer (BC). Where we have moved towards axillary conservation in clinically node negative (cN0), the debate on what after 1-2 sentinel lymph nodes positive (SLN+ve) still continues. The ideal situation would be wherein we can accurately predict which patient has a risk of additional non SLN+ve. Several nomograms have been developed to predict the risk of NonSLN+ve. But in view of the differences in tumor size and nodal burden between our patients and the western data, we conducted a study to validate some of these nomograms in our cohort of early BC with positive LN on Low axillary sampling (LAS). Methods: Clinico-pathological data was collected for operable BC (OBC) with cN0 who underwent upfront SLNB or AS. This was entered into the various nomograms and the probability of the Non SLN+ve was calculated. Nomograms with AUC of greater than 0.7 were pre-defined as considerable discrimination. Results: From 2013 to 2018, 2350 women with cN0 OBC underwent LAS. Of which, 670 (28.5%) had a positive node on LAS. Median pT size was 3 cm with 327 (48%), LVI +ve 152 (77%) ENI +ve, 525 (78.4%) Hormone receptor +ve and 485 (72.4%) grade 3 tumors. Of 670, 239 (35.7%) had a NonSLN+ve on completion axillary dissection. The AUC values for nomograms included,ie. MSKCC, MDAnderson, Tenon, Cambridge, Shanghai, Mayo clinic and Turkish were 0.769, 0.77, 0.55, 0.74, 0.65, 0.529, 0.563 respectively. Only three nomograms, MDA, MSKCC and Cambridge had an AUC of more than 0.7. However, they were associated with poor sensitivity and specificity and high FNR (Table) making them clinically unreliable for this cohort. Conclusions: All 7 nomograms were not validated in our study. The larger T size and higher nodal burden of our cohort may be responsible for the same. We thus need to develop an Indian nomogram to predict the risk of non SLN+ve for our patients. [Table: see text]