P5-13-12: Can the Ki67 Proliferation Index Predict the Oncotype DX Recurrence Score in Lymph Node Negative, ER Positive Breast Cancer?

Abstract

Abstract Background Oncotype DX is a 21-gene assay for node-negative, ER positive breast cancer. Results are expressed as a Recurrence Score (RS). This predicts a low, intermediate or high risk of distant relapse. Proliferation-related genes such as Ki67 impact heavily upon the calculation of the RS. The assay encompasses several histopathologic factors (ER, HER2) and is limited by cost, proprietary nature and turnaround time. We investigated whether Ki67 proliferation index (PI) and routine clinicopathologic parameters could be used to predict RS. Materials and Methods We obtained clinicopathologic details for 69 patients (diagnosed 2007–2010) who had RS available. Quantitative analysis of Ki67 immunohistochemistry was performed on primary tumour. Patients were classified as having a low (≤15%), intermediate (16-30%) or high (>30%) Ki67 PI. Risk of recurrence was calculated using Adjuvant! Online. Patients with categorized as having a poor, moderate, good or excellent prognosis as per the Nottingham prognostic index (NPI) and a low, intermediate or high risk of relapse, according to the St. Gallen criteria. Results Mean age at diagnosis was 51.4 years (range: 34–74 years). Mean tumour size was 1.9 cm (range: 0.8−4.1cm). Mean Ki67 PI was 13.4 % (range 0.2 to 52.6%). RS was low (0-17) in 34 cases, intermediate (18-30) in 26 cases and high (>30) in 9 cases. The RS was significantly correlated with Ki67 expression (Spearman's rank correlation, r = 0.472, p<0.001). RS was significantly higher in patients with a high Ki67 PI, when compared to patients with low and intermediate Ki67 PI (Mann Whitney U, p<0.05). RS was significantly higher in patients with grade 3 tumours, when compared to patients with grade 1 and 2 tumours (Mann Whitney U, p<0.05). Risk of recurrence as predicted by Adjuvant! Online, was significantly higher in patients with a high RS, when compared to patients with low and intermediate RS (Mann Whitney U, p<0.05). RS was significantly higher in patients with moderate prognosis, as determined by the NPI, when compared to patients classified as having a good/excellent prognosis (Mann Whitney U, p<0.05). No patients were classified by the NPI as having a poor prognosis. RS was significantly higher in patients deemed to be at high risk of relapse by the St. Gallen criteria, when compared to patients classified as low and intermediate risk (Mann Whitney U, p<0.05). The majority (66.7%) of cases in the high RS group had LVI present. Conclusions This pilot study indicates a trend whereby Ki67 and grade will predict whether the (Oncotype DX RS is low/intermediate or high. Adjuvant! Online, St. Gallen and NPI are also able to distinguish patients with a high RS from those with low/intermediate RS. In order to verify these findings, we are developing a statistical model using these 69 patients as a training set. As there is insufficient follow-up to correlate either RS or Ki67 with relapse, a potential test set comprised of an additional 50 LN negative, ER positive patients, with an average of 13 years follow-up, has been identified. Ki67 analysis of these cases is underway to address this issue. Acknowledgement This work is funded by SFI (SRC award, 08/SRC/B1410 to MTCI). Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-13-12.