Abstract P5-06-14: Integrating CD8, FOXP3 and PD-L1 expression in prognostic models for triple negative breast cancer (TNBC): An analysis of 265 patients treated with standard therapy for stage I-III disease

Abstract

Abstract Background: Tumor infiltrating lymphocytes (TILs) are strong prognostic biomarkers for early TNBC. We evaluated the role of CD8, FOXP3 and PD-L1 expression in refining prognostic models for non-metastatic TNBC in a large cohort of patients treated with standard therapy. Methods: Consecutive patients diagnosed with stage I-III TNBC (ER/PgR <10%, HER2 0/1+ or ISH non amplified) between May 2012 and December 2015 were included. All patients received treatment with surgery, chemotherapy (neoadjuvant or adjuvant) and radiotherapy (when indicated). For each case, three FFPE tumor slides were stained for CD8 (Clone C8/144B, Dako Cytomation), FOXP3 (Clone 236A/E7, dilution 1:200, Abcam) and PD-L1 (IHC 73-10 Research Use Only assay developed by Agilent Technologies). One slide was stained for cytokeratins with MNF116. Digital slides were evaluated by a Visiopharm® software application, following alignment of the CD8, FOXP3 and PD-L1 slides with the MNF116 slide. The density of CD8 and FOXP3 expression was calculated as the number of cells/mm2 of stroma area. For PD-L1, the % of positive stromal cells over the total of stromal cells was analyzed. Disease-free survival (DFS) was calculated from diagnosis to relapse or death. The Harrell’s c-index was used to determine the cut-offs for CD8, FOXP3 and PD-L1 to be used in survival analyses. Results: 265 TNBC pts were evaluated. Median TILs was 10% (Q1-Q3 3%-25%), median CD8 was 249 (Q1-Q3 109-568), median FOXP3 was 57 (Q1-Q3 21-134), median PD-L1 was 5.2% (Q1-Q3 0.2%-25.4%). TILs, CD8, FOXP3 and PD-L1 were positively correlated with each other (p<0.001): CD8 showed strong correlation with TILs (Spearman’s coefficient 0.753), FOXP3 and PD-L1 showed moderate correlation with TILs (Spearman’s coefficient 0.535 and 0.587). Higher TILs, CD8, FOXP3 and PD-L1 were associated with age ≤50yrs (p=0.002, p=0.004, p=0.065, p=0.011), Grade 3 (p=0.001, p=0.006, p=0.006, p=0.003) and Ki67 ≥30% (p=0.056, p=0.024, p=0.004, p=0.005). There was no association between immune markers and stage. Among classic clinicopathologic factors, TILs (10% increments) and stage at diagnosis were independent prognostic parameters in multivariate analysis (HR 0.81, 95% CI 0.69-0.94 p=0.005 for TILs, HR 2.01 95% ci 1.01-4.23 P=0.047 for stage II vs stage I and HR 5.31 95% CI 2.54-11.11 p<0.001 for stage III vs stage I). In univariate analysis, high CD8 (≥443), high FOXP3 (≥57) and high PD-L1 (>20%) were all significantly associated with improved DFS (HR 0.36 95%CI 0.18-0.72, p=0.004 for CD8; HR 0.48 95%CI 0.28-0.80, p=0.005 for FOXP3; HR 0.52 95%CI 0.28-0.97, p=0.039 for PD-L1). FOXP3 and PD-L1 provided significant additional prognostic information beyond a model containing TILs and stage: likelihood ratio χ2 5.12, p=0.024 for FOXP3; likelihood ratio χ2 5.52, p=0.019 for PD-L1. CD8 did not add relevant prognostic information beyond TILs and stage (likelihood ratio χ2 2.76, p=0.097). Including both FOXP3 and PD-L1 did not add further prognostic information to models already containing TILs, stage and either FOXP3 or PD-L1. Conclusions: FOXP3 and PD-L1 expression evaluated with a software-assisted method were prognostic for stage I-III TNBC pts treated with standard therapy and may contribute to refine the prognostic stratification beyond stage and TILs. This study was supported by a grant from Merck KGaA. Citation Format: Maria Vittoria Dieci, Vassilena Tsvetkova, Gaia Griguolo, Federica Miglietta, Deborah Bacchin, Giulia Tasca, Carlo Alberto Giorgi, Enrico Cumerlato, Enrico Orvieto, Valentina Guarneri, Pierfranco Conte. Integrating CD8, FOXP3 and PD-L1 expression in prognostic models for triple negative breast cancer (TNBC): An analysis of 265 patients treated with standard therapy for stage I-III disease [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-06-14.