SDPS-38 A BRAIN METASTASES PREDICTION MODEL IN BREASTCANCER WOMEN

Abstract BACKGROUND Breast cancer (BC) is a leading cause of mortality and the most frequent malignancy in women; deaths are due to metastatic disease, expressly brain metastases (BM). Currently, there no biomarker or prediction model accurately predict BM. OBJECTIVE To generate BM from variables acquired at BC diagnosis. METHODS A retrospective cohort women diagnosed 2009 2020 single center was divided into training (TD) validation datasets (VD). The TD used multivariable model. modeĹs performance measured applying area under curve (AUC), C-statistic, Akaike information criteria (AIC). RESULTS 5,009 patients were (n 3339) VD 1670). In TD, with best (lowest AIC) built following variables: Age, estrogen receptor status, tumor size, axillar adenopathy, AJCC anatomic clinical stage, Ki-67, Scarf-Bloom-Richardson score. This had an AIC 1241 AUC 0.793 (95%CI 0.761 – 0.825) p <0.0001 TD. 10-fold cross-validation showed good stability VD, = 0.812 (IC95% 0.774 0.850) P < 0.0001 644. Finally, we present online APP calculator for its use. CONCLUSION breast cancer, pathological diagnosis displayed robust measure individual odds currently considered external other institutions countries.