Correlation between Oncotype DX recurrence score and classical risk factors in early breast cancer.
e12011 Background: OncotypeDx(ODX) predicts the likelihood of estrogen receptor (ER) positive breast cancer (BC) recurrence and assesses the likely benefit from both hormonal therapy and chemotherapy. Many clinical scores that estimate the risk category of ODX are being tested. Ki67 is frequently incorporated into these assessments, although there is no standard cut-off for its use. Methods: We retrospectively reviewed the electronic medical files of 190 patients with early stage ER+ BC for whom ODX recurrence score (RS) was available from 2014 to 2016. Our objective was to find out the degree to which classical clinicopathological variables -as defined by St. Gallen(SG) 2015- could predict ODX risk category, also to determine an optimal Ki67 cut-off in order to establish an accurate classification. Chi square test was used. Results: Mean age at diagnosis was 59 years (28-89). Mean tumor diameter was 15mm, 84.2% were intermediate grade. Mean expression of ER, PR and Ki67 were 87%, 53% and 22%, respectively. According to ODX 62.1% patients were low risk, 30.5% intermediate risk and 7.4% high risk. An overall concordance of 46.8% (73/190) was found between SG 2015 and the risk category of ODX (75.7% for low, 33.3% for intermediate and 23.9% for high RS). When changing SG Ki67 cutoffs to ≤20% (for low Ki67) and ≥30% (for high Ki67), an overall concordance of 56.3% (107/190) was found (69.6% for low, 47.3% for intermediate and 23.9% for high RS, with p=0.00) (Table 1). Conclusions: In best-case scenario, SG classical clinicopathological variables correctly classified 56.3% of patients of our series. Despite being a specialized center, the utility of classical clinicopathological variables for predicting ODX risk category is limited. [Table: see text]