IntroductionThe total tumour load (TTL) obtained from OSNA study in each of the sentinel lymph nodes has been identified as the most powerful predictor of axillary non-sentinel lymph node metastasis. In addition, the distinct molecular subtypes (MS) of breast cancer differ significantly not only in terms of incidence, prognosis and treatment but also in terms of the pattern of axillary metastatic involvement. We hypothesised that the prediction of metastatic disease in axillary lymphadenectomy could be enhanced by applying a predictive model based on the TTL and the intrinsic tumour subtype. ObjectiveTo evaluate the impact of the MS identified by immunohistochemistry on prediction of metastatic disease in axillary non-sentinel lymph nodes based on TTL. Material and methodsRetrospective, European multicenter study including 683 patients from 9 hospitals. ResultsUnivariate analysis identified 6 variables that were significantly correlated with axillary non-sentinel metastasis. Of these, the variables logarithmic value of the TTL, tumour diameter and MS diagnosed by immunohistochemistry were selected for multivariate analysis. The odds ratio estimated by the model were: logarithmic value of the TTL 1.527 (95% CI: 1.299-1.796), tumour diameter 1.503 (95% CI: 1.062-2.129) and MS 2.195 (95% CI: 1.246-3.867). ConclusionsThe strongest predictors of axillary involvement were MS, TTL and tumour diameter. These variables should be included in diagnostic algorithms as essential parameters for therapeutic decision-making on the axilla.