Abstract P2-05-04: Evaluation of intra-tumor heterogeneity, test reproducibility and their impact in breast cancer samples assessed by Prosigna™: Results from a decision impact prospective study and a matched case-control study

Background: Recent molecular biology technologies reveals insight into tumor heterogeneity but quantification and impact on reproducibility of tests is not well known. The objective of this study was to assess the extent to which tumor heterogeneity may affect the prognosis of patients assessed by Prosigna™ (PAM50) gene signature assay compared to test reproducibility. Methods: Reproducibility was measured by testing replicate tissue sections from 186 FFPE breast tumor blocks across 2 sites (Institut Curie, Centre Jean Perrin) following independent pathology review at each site. Consecutive slides came from blocks of patients included in the Decision Impact prospective study which examined whether the Prosigna™ test influences adjuvant treatment decision (Clinical trial information: NCT02395575). To evaluate heterogeneity and its impact in terms of outcome, we selected among T1N0 patients treated in Institut Curie between 2003 and 2008, 32 patients who did recur and 28 matched control group who did not (2 9controls9 recurred during the study). Analyses were performed on two parts of each tumor. NanoString9s Prosigna™ outputs (risk of recurrence (ROR) score, 10 year probability of distant recurrence, risk category, and intrinsic subtype (Luminal A/B, HER2-enriched, Basal-like)) were measured and compared to evaluate heterogeneity (defined as difference in terms of subtype and/or risk category between the two parts) and reproducibility. Correlation between heterogeneity and outcome was performed. Impact was assessed by tumor board analysis. Results: Pearson correlation coefficients for ROR score and probability of distant recurrence predicted were .95 and .97, respectively in the reproducibility study and .82 and .86, in the tumor heterogeneity study. The measured standard deviation (SD) was 5.4 and 8.1 ROR units corresponding to 1.6% and 2.8% in terms of risk of distant metastasis free survival within the reproducibility study and the tumor heterogeneity study, respectively. Kappa coefficients for intrinsic subtype and risk category agreement were 0.88 and 0.87 in the reproducibility study, and 0.67 and 0.58 in the tumor heterogeneity study. Tumor board analysis of discordant cases showed that the impact, in terms of decision of chemotherapy administration, concerns 3% of patients because of reproducibility and 8% because of tumor heterogeneity, comparing favorably with the discordance between Prosigna™ and immunohistochemistry (27%). Probability of distant recurrence was higher in the cases (15%) compared to control (9%) (p=.001) in the tumor heterogeneity study confirming the performance of the Prosigna™ test. Conclusion: We validated in the prospective Decision Impact study the analytical performance of NanoString9s Prosigna™ assay across multiple clinical testing laboratories. We showed in these two studies that tumor heterogeneity has more impact than reproducibility performance. The clinical impact on the decision making based on tumor heterogeneity is however limited, since it does not correlate to outcomes, whereas the Prosigna™ ROR score has been shown to correlate very well to outcomes. Citation Format: Rouzier R, Bonneau C, Cayre A, Hequet D, Gentien D, Bonhomme A, Mouret-Reynier M-A, Dubot C, Cottu P, Roulot A, Morel P, Salomon A, Callens C, Guinebretiere J-M, Penault-Llorca F. Evaluation of intra-tumor heterogeneity, test reproducibility and their impact in breast cancer samples assessed by Prosigna™: Results from a decision impact prospective study and a matched case-control study [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-05-04.