The effect of non-oncology drugs on clinical and genomic risk in early luminal breast cancer

e12505 Background: In women with ER-positive, HER2-negative early stage breast cancer (BC), treatment decisions for adjuvant chemotherapy are based on genomic risk and clinical risk. Recently, an effect of non-oncology medications on cancer cell lines and cancer outcome have been suggested. In this study we aimed to systematically examine the impact of non- oncology drugs on clinical risk and genomic risk (based on OncotypeDx recurrence score [RS]) in early BC. Methods: We collected data from 1385 files of ER positive HER2 negative breast cancer patients regarding their clinical risk (stage and grade), genomic risk (OncotypeDx RS) as well as data regarding medications the patients received during the month before their surgery. Statistical analysis was applied to identify the influence of various medications on OncotypeDx RS. Results: Out of the various medications we examined, we found that Levothyroxine was significantly associated with high median OncotypeDx RS (RS median = 25 ;p < 0.0001) and Metformin was associated with low median OncotypeDx RS (RS median = 12 p < 0.0011) in comparison to patients not receiving these medications (RS median = 17). By contrast there were no differences in the clinical risk between patients who received Metformin or Levothyroxine. Notably, Levothyroxine and metformin did not impact proliferation marker (Ki67) levels but did impact progesterone-related features, suggesting they influence genomic risk through estrogen dependent modules. Indeed, scores of other genomic tests (PAM50, Mammaprint), which are determined largely by proliferative features, were not influenced by Levothyroxine or Metformin. Finally, by using contemporary guidelines to recommend adjuvant chemotherapy based on clinical risk and genomic risk (OncotypeDx ) we show that patients (Age > 50) who received Metformin treatment had 14.5% chance to be recommended adjuvant chemotherapy while patients who received Levothyroxine had 49% (p = 0.0001). Conclusions: The results of this study indicate significant impact of Metformin and Levothyroxine on clinical decisions with potential impact on early BC patients.