Abstract P2-08-25: PD-L1 expression at the protein and RNA levels as prognostic factor in early breast cancer

Abstract

Abstract Introduction:PD-L1 expression as assessed by immunohistochemistry (IHC) is a clinically relevant biomarker in certain malignancies such as lung cancer, since it selects appropriate candidates for PD-1 blockade. Since these agents are under evaluation for breast cancer, discovering and validating predictive biomarkers is of outmost importance. However, the clinical utility of PD-L1 expression in breast cancer is questionable, in light of prior inconclusive reports which have used various IHC antibodies, scoring methods and cut-offs. Moreover, there are only few previous studies on comparing IHC and RNA data at the same cohort, not limited to a single subtype.Methods: Our cohort is derived from a nested case-control study consisting of 619 patients diagnosed with primary breast cancer between 1997-2005 in Stockholm health care region.Tissue microarrays from epithelial tumor areas have been constructed using duplicate cores from primary tumors and tissue sections were used for IHC with PD-L1 (Ventana; clone SP263) antibody. Positivity was defined as the presence of any single cell with membranous expression of PD-L1. Gene expression profiling was performed using DNA microarrays (GSE48091). Data on clinical and pathological tumor characteristics, survival, loco-regional and systemic treatments, and follow-up have been collected. Correlations between transcript and protein expression levels were estimated using Mann-Whitney test, while survival analyses were conducted using the Kaplan-Meier method. Furthermore, we associated an immune gene module score (IMS) –whose predictive power in neoadjuvant and metastatic settings has been previously demonstrated- with PD-L1 transcript levels by using Spearman's rank correlation coefficient.Results: IHC data were available for 87.4% (541/619) of the patients. PD-L1 was expressed on tumor cells in 9.6% (52/541) of the patients while it was also expressed by immune cells in 23.1% (125/541) of the patients. Any PD-L1 expression (tumor and/or immune cells) was noted in 24.2% (131/541) of the patients. PD-L1 transcript levels and protein expression on tumor, immune and/or both cell types were statistically significantly associated (p< 2.2e-16). In the whole cohort, patients with higher PD-L1 transcript levels were associated with better breast cancer-specific survival(BCSS) (p=0.0061). In addition, within intrinsic subtypes, high PD-L1 transcript expression was significantly associated with better BCSS only in basal-like (p=0.019) disease. There was no significant correlation between improved BCCS and PD-L1 protein expression by tumor (p=0.13), immune (p=0.12) or both types of cells (p=0.2). PD-L1 transcript levels were also positively associated with the IMS (Spearman's rho = 0.85). Conclusions:The prognostic value of PD-L IHC expression in breast cancer remains inconclusive. However, RNA expression of PD-L1 may be more informative as a prognostic factor, especially in basal-like disease and merits further validation. Citation Format: Zerdes I, Sifakis E, Matikas A, Tobin NP, Charlotte R, Rassidakis GZ, Bergh J, Foukakis T. PD-L1 expression at the protein and RNA levels as prognostic factor in early breast cancer [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 P2-08-25.