Tumor infiltrating lymphocytes and programmed death ligand-1 can identify different prognostic profiles in breast cancer.

Abstract

e12622 Background: Immunologic biomarkers such as tumor infiltrating lymphocytes (TILs) and programmed death ligand-1 (PD-L1) can provide prognostic information in breast cancer (BC) patients. Specifically, a high TILs level has been associated with pathologic complete response and enhanced disease-free survival (DFS). Furthermore, PD-L1 positivity has been associated with worse DFS and overall survival (OS). However, the interaction of both biomarkers as well as its association with survival outcomes in specific BC subtypes is still a subject of ongoing research. Methods: Medical records of women diagnosed with primary BC between 2013 and 2015 in a center in Monterrey, Mexico were reviewed. Eligible patients had at least 1 year of follow-up, stages I-III at diagnosis, and available tissue for TILs and PD-L1 assessment. PD-L1 positivity was defined as the presence of PD-L1 in ≥1% of tumor-infiltrating immune cells using the VENTANA SP142 assay. Classification of TILs into low ( < 30%) and high (≥30%) levels was performed for analytical purposes. Clinicopathological features were compared with Fisher’s exact tests and logistic regression models, as appropriate. The Kaplan-Meier method was used to calculate recurrence-free survival (RFS), and associations between variables were explored with log-rank or Cox regressions. Results: A total of 195 patients were included. Overall, 12.3% of BC biopsy specimens showed positivity to PD-L1 (8.3% [11/132] in HR+/HER2-, 5.9% [2/34] in HER2+, and 38% [11/29] in TNBC; p < 0.001). PD-L1 positivity was significantly associated with ER-negative status (OR 3.1; p = 0.013), high TILs (OR 5.7; p < 0.001), and high Ki67 expression (HR 12.5; p < 0.001). The median follow-up for the entire cohort was 61 months (95%CI 58-63). RFS in the PD-L1- group was significantly superior to the PD-L1+ group (85% vs. 66% at 5 years; p = 0.021). In a multivariate Cox model, PD-L1 positivity (HR 3.3), low TILs (HR 3.4), advanced stage (HR 2.8), and high histological grade (HR 2.5) were all independent prognostic factors for worse RFS. Based on these results, we further classified our cohort into four prognostic profiles based on PD-L1 and TILs status. The low TILs/PD-L1+ group experienced an inferior RFS than low TILs/PD-L1- patients (45% vs. 85% at 5-years; p-value < 0.001), while no difference was observed in the high TILs groups. Notably, when analyzing HR+/HER2- patients only, a similar statistical difference was observed between low TILs/PD-L1+ and low TILs/PD-L1- groups (HR 5.4, p = 0.009). Conclusions: In our cohort, PD-L1 positivity together with low TILs identified a subset of BC patients with a worse RFS. Interestingly, patients with HR+/HER2- BC fared worse in the low TILs/PD-L1+ subcategory than in the low TILs/PD-L1- group. Future studies are warranted to explore the prognostic value of categorizing patients based on TILs and PD-L1 status in different BC subtypes.