Abstract S1-07: Prognostic value of tumor-infiltrating lymphocytes (TILs) in two phase III randomized adjuvant breast cancer trials: ECOG 2197 and ECOG 1199

Abstract

Abstract Purpose: The immune system influences breast cancer prognosis. Analysis of the international BIG 02-98 trial suggests that TILs are associated with disease-free survival (DFS) in patients with operable triple negative breast cancer (TNBC, Loi et al, JCO 2013). Here we seek to validate the association of increased lymphocytic infiltration of primary breast tumors with an improved prognosis in TNBCs in two adjuvant Phase III trials conducted by the Eastern Cooperative Group. Methods: “Whole” H&E stained sections of tumors from E2197 and E1199 were evaluated by two independent pathologists blinded to outcome data for density of TILs in intratumoral (iTIL) and stromal compartments (sTIL). Cases from E2197 and E1199 were randomly selected based on TNBC status (E2197 central review, E1199 local determination) and availability of tumor sections. The association of DFS with TIL scores was determined by fitting proportional hazards models stratified on study. Nonrecurrences from E2197 were given a relative weight of 1.43, to reflect that they had been undersampled, and robust variance estimators were used. Results: 500 cases were obtained for which 481 were evaluable (190 for E2197, 291 for E1199). The median age of patients was 49 with 54% being premenopausal, 59% nodal involvement, 66% of tumors being > 2cm and all patients received anthracycline-based chemotherapy. 80% of tumors had at least 10% lymphocytic infiltration in the stroma (range 10-80%) but only 15% of cancers had at least 10% TILs intratumorally (range 10-50%). The lymphocyte-predominant breast cancer (LPBC) phenotype, defined as ≥50% stromal or intratumoral lymphocytic infiltration was seen in 4.4% of TNBC (21/481). At a median follow-up of 10.6 years, 190 DFS events were reported. Higher TIL scores were associated with a better prognosis: for every 10% increment in sTIL, a 14% reduction of risk for recurrence or death was observed (p = 0.02) and for every 10% increment in iTIL, a 28% risk reduction was seen (p = 0.06). Multivariate analysis confirmed sTIL to be an independent prognostic marker of DFS. Hazard Ratio (95% CI)p-valuesTIL (10% increase)0.84 (0.74,0.95)0.005Tumor Size < = 2.0 vs. 2.1-5.0cm0.59 (0.42,0.82)0.002Tumor Size < = 2.0 vs. >5.0cm0.46 (0.27,0.78)0.004Node-neg vs. 1-3 N+0.53 (0.37,0.78)0.001Node-neg vs. >3 N+0.23 (0.14,0.38)<0.001Age 41-50 vs. 24-401.32 (0.84,2.09)0.23Age 51-60 vs. 24-401.43 (0.88,2.30)0.14Age >60 vs. 24-402.12 (1.29,3.49)0.003Table: Estimated DFS Hazard Ratios, 95% CI and Individual Term P-values for the Multivariate Model for the Linear Effect of the Stromal TIL Score The presence versus the absence of stromal TILs as binary marker was also a strong prognostic marker with a 31% reduction of risk of recurrence or death for the group with sTILs (95% CI 2 to 51, p = 0.04). Conclusion: We confirm in two independent data sets from national randomized clinical trials using contemporary adjuvant chemotherapy that stromal lymphocytic infiltration constitutes a robust prognostic factor in TNBCs. The data provide strong evidence for the incorporation of this feature into routine histopathological assessment as a prognostic factor for patients with TNBCs. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S1-07.