Abstract P3-06-15: Baseline CD4/CD8 tumor infiltrating lymphocytes (TIL) ratio predicts pathologic response to neoadjuvant chemotherapy (NC) in breast cancer

Abstract

Abstract Background: Tumor microenvironment immune response has been correlated with chemotherapy efficacy. The aim of this study was to determine whether baseline and post-treatment CD4 and CD8 TIL predict tumor response in women with breast cancer (BC) receiving NC. Methods: We analyzed TIL CD4 and CD8 subpopulations in BC patients treated with sequential anthracyclines and taxanes NC. A tissue microarray with paired pre- and post-NC biopsies was built, and immunohistochemically stained for CD4 and CD8. Tumor area-adjusted morphometric analysis was performed with Image J software after slide scanning and digitalization (results expressed as TIL count/mm2). Statistical analysis was done with SPSS 15.0 software. Results: We included 121 consecutive patients. Clinical stages: 15,7% IIA, 28% IIB, 33,3% IIIA, 6,6% IIIB, 16,5% IIIC. Histology: 93% ductal infiltrating carcinomas. Phenotypic classification by immunohistochemistry (IHC): 50,4% RE+ and/or RP+ and HER2NEU negative; 13,2% RE+ and/or RP+ and HER2NEU positive; 9,9% RE- and RP- and HER2NEU positive; 21,5% RE- and RP- and HER2NEU negative; 5% non-classifiable. Treatment: 80.2% AC × 4 courses followed by docetaxel × 4 courses. Pathological response: complete pathologic response (pCR) rate was 17,4% (primary tumor pCR: 20,7%; axillary pCR: 35,5%). After a median follow up of 52 months neither overall survival (OS) nor disease free survival (DFS) has been reached. Mean pre-chemotherapy CD4 count was 59,98 (± 107,53, SD). We did not found any association with clinical or pathologic tumor characteristic at diagnosis (stage, grade, hormonosensitivity, HER2NEU overexpression or IHC tumor classification). Mean CD4 count was significantly higher in patients achieving pCR than in patients without pCR (112,43 vs 49,91 CD4/mm2; p 0,009). Mean pre-chemotherapy CD8 count was 16,23 (± 56,96, SD) and was higher in grade 3 carcinomas (32,83 vs 22,37; p < 0,03); no other association with tumor characteristics or with pCR was found. Mean variation in CD4 after NC (=postNC — preNC count) was 47,35 (± 104,49) CD4/mm2, and it was larger in patients with pCR (102,01 vs 36,85 CD4/mm2; p < 0,010). An opposite non-significant trend was observed for CD8 variation after NC (pCR: 39,42 vs non-pCR: 20,22, p < 0,14). To integrate the influence of both CD4 and CD8 on response to chemotherapy, the predictive value of pre-treatment ratio between CD4 and CD8 counts (CD4/CD8) was analyzed. We found a highly significant difference for CD4/CD8 between patients with and without pCR to NC (103 vs 17; p 0.002). Multivariate logistic regression analysis demonstrated that only a high pretreatment CD4/CD8 ratio (p < 0,006), hormone sensitivity (p < 0,03) and Her2 overexpression (p < 0.01) independently predicted pCR after NC. No differences in overall or disease-free survival were demonstrated for high vs. low CD4/CD8 pre-NC ratio. Conclusion: A high pretreatment CD4/CD8 ratio independently predicted pathological complete response in patients with invasive breast cancer receiving anthracyclines and taxanes NC. Our results support the contribution of tumor microenvironment immunologic response to chemotherapy effects on breast cancer. [Supported by GEICAM-Beca Ana Ballil] Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-06-15.