Background:

TILs involving tumor and/or its associated stroma may be indicative of an immune response that may either facilitate anti-tumor immunity and clearance or immune tolerance and evasion. In this study we performed a meta-analysis evaluating the relationship between TILs and pathologic response to neoadjuvant chemotherapy (NAC) since this is considered as a surrogate endpoint of disease outcomes in ER negative or Her2 positive tumors.

Methods:

We searched PubMed and Embase (1991-May 2013), and ASCO abstracts (2009-2012), using a combination of free text search and controlled vocabulary search. We identified 1147 reports which met our initial search criteria, and they were reviewed to identify those which met the following criteria: (1) evaluated the presence of TILs, defined as tumor and/or stromal lymphocytes (CD4, CD8 or FOXP3) identified by H&E, IHC or gene expression before NAC, (2) classified TIL's “high/low” or “positive/negative”, and (3) correlated TILs with pathological complete response (pCR) or near pCR after NAC. Standard anthracycline-containing regimens were used as NAC in most studies, and anti-HER2 therapy was not used in most studies with HER2-positive disease. Results are presented as pooled odds ratios (OR) with 95% confidence intervals (CI), based on random-effects (to account for between study variance and heterogeneity due to different cutoffs and subtypes of T lymphocytes). Sensitivity analysis was done and publication bias was investigated using a funnel plot. We employed the Chi(2) test and calculated the I(2) statistic to investigate study heterogeneity. Meta-analysis statistics were calculated using StatsDirect Version 2.7.9.

Results:

Seven studies including 1641 patients met our criteria for inclusion in this analysis. A TIL ratio classified as either high or positive was associated with a significantly higher likelihood of achieving a pCR/near pCR after NAC (OR 3.68; 95% CI 1.93–7.01. p<0.0001) [Table 1]. This effect was driven mainly by a difference in ER negative tumors (OR 4.04, 95% CI 2.16-7.57. p<0.0001) and Her2 positive tumors (OR 5.61, 95% CI 1.8–17.47, p = 0.0007); the association was present, but nonsignificant, in ER positive tumors (OR 2.17, 95% CI 0.95-4.98). Sensitivity analyses did not change the inference. Funnel plots suggested low likelihood of publication bias (Harbord Egger test, p = 0.604) for all studies, and the I(2) statistic was 67.5%.

Table 1. Characteristics and Odds ratio for individual subtypes

Subtype No of studies TIL low pCR% TIL high pCR% OR 95% CI 
All 1641 12.5% 28.6% 3.68 1.93 -7.01 
ER-/PR- and HER2- [except one study defined by ER-/PR-] 403 23.6% 41.3% 4.04 2.16 - 7.57 
Her2+ 326 16.9% 23.4% 5.61 1.80 - 17.47 
ER/PR+ 558 5.6% 11.5% 2.17 0.95 - 4.98 
Subtype No of studies TIL low pCR% TIL high pCR% OR 95% CI 
All 1641 12.5% 28.6% 3.68 1.93 -7.01 
ER-/PR- and HER2- [except one study defined by ER-/PR-] 403 23.6% 41.3% 4.04 2.16 - 7.57 
Her2+ 326 16.9% 23.4% 5.61 1.80 - 17.47 
ER/PR+ 558 5.6% 11.5% 2.17 0.95 - 4.98 

Conclusions:

In this systematic review and meta-analysis, high or positive TIL status before NAC was associated with a significantly better pathologic response to NAC (surrogate for DFS and OS), particularly in patients with ER/PR-negative or HER2-overexpressing disease. Patients with tumors characterized by low or absent TILs require novel therapeutic approaches, and may be candidates for immunotherapeutic approaches to enhance innate immunity or reverse immune tolerance.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-08.