Abstract PD2-07: mRNA signatures predict response to durvalumab therapy in triple negative breast cancer (TNBC)– Results of the translational biomarker programme of the neoadjuvant double-blind placebo controlled GeparNuevo trial

Abstract

Abstract Background: The GeparNuevo trial showed a numerical increase in the pCR rate to 53% vs 44%; p=0.281 compared to placebo in TNBC with the addition of the anti-PD-L1 antibody durvalumab to a neoadjuvant anthracycline-taxane containing chemotherapy (Loibl S et al. ASCO 2018). In a predefined subgroup analysis, a significant increase of the pCR rate was observed for patients that received durvalumab for 2 weeks alone prior to the start of chemotherapy in a window phase (61% vs 41%, p interaction=0.048), while the pCR rate was not increased for the subset of patients that did start durvalumab together with chemotherapy. Here we report the main results of the translational programme for GeparNuevo with a focus on mRNA signatures predictive for pCR in pretherapeutic core biopsies. Methods: A total of 162 baseline FFPE core biopsies were evaluable for expression of 2560 genes using the HTG EdgeSeq® system that combines a modified nuclease protection assay with next generation sequencing. Data was processed as recommended by the HTG and median transformed for further analyses. For differential gene expression analyses, the data was scale-normalized (TMM normalization; EdgeR package) and linear models were fit (limma package). Prior to these analyses, genes were filtered based on minimal expression (> 4) and variability (IQR > 1). As a first step, predefined immune-genes signature (TILs signature) (Denkert et al. JCO 2016) as well as IFN-gamma signatures were evaluated for correlation with pCR in logistic regression models. Subsequently, we performed a differential gene expression analysis according to therapy response for the durvalumab-arm and the placebo arm using the pre-filtered candidate genes. Gene names are not included in this abstract to allow filing of IP, but full gene names will be presented at the SABCS meeting. Results: The predefined TIL- and IFN-gamma signatures were associated with increased pCR rates in the complete cohort (TIL-signature: OR 1.44, 95% CI 1.15-1.82, p=0.002; IFN-Gamma-signature: OR 1.63, 95% CI 1.22-2.24, p=0.002) as well as in the durvalumab arm (p=0.012 and 0.042) and the placebo arm (p=0.050 and 0.011). These signatures were general pCR predictors without specificity for durvalumab response. Additional 44 genes were significantly (p<0.05) correlated with pCR in the durvalumab arm. Of those, 21 genes were upregulated and 23 genes were downregulated in pCR patients. 14 of the 21 upregulated genes are related to tumorbiologically relevant immune cell functions. A total of 6 of the 44 genes had a positive test for interaction (interaction p<0.05) with the therapy arm (durvalumab + NACT vs. placebo + NACT), suggesting that these genes might specifically predict response to durvalumab. Additional analyses investigating the role of molecular tumor subtypes, additional immune gene signatures and other subgroup analyses will be presented at the meeting. Conclusion: Our results show that specific immune-related gene expression signatures predict response to durvalumab in primary triple negative breast cancer. The trial was financially supported by Astra Zeneca and Celgene Citation Format: Loibl S, Sinn BV, Karn T, Untch M, Treue D, Sinn H-P, Weber KE, Hanusch CA, Fasching PA, Huober J, Zahm D-M, Jackisch C, Thomalla J, Blohmer J-U, Marmé F, Klauschen F, Rhiem K, Felder B, von Minckwitz G, Burchardi N, Schneeweiss A, Denkert C. mRNA signatures predict response to durvalumab therapy in triple negative breast cancer (TNBC)– Results of the translational biomarker programme of the neoadjuvant double-blind placebo controlled GeparNuevo trial [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr PD2-07.