Introduction: The function of tumor-infiltrating T cells in lymphoma is unknown. Previous studies have shown that an increased percentage of tumor-infiltrating CD4+ T cells was associated with a better overall survival in diffuse large B-cell lymphoma (Ansell S. et al. JCO 2001; Xu et al., British Journal of Haematology 2001). Furthermore, the CD4+ T cells observed in one study was associated with a CD3+/HLA-DR + phenotype, suggesting that they were activated T helper cells. Recently, a molecular predictor model for follicular lymphoma was reported based on microarray analysis of gene expression within the cell populations of biopsy specimens. Three groups of “predictor genes” were identified: one representing B-cell differentiation and two groups representing genes expressed by T cells and macrophages, respectively (ASH abstract, 2003). This raised the question of whether it was simply the presence of tumor-infiltrating T cells that correlated with clinical outcome as opposed to specific activation pathways within those cells.
Purpose: The purpose of this study was to examine whether the percentage of tumor-infiltrating T cells is correlated with prognosis of follicular lymphoma.
Patients and Methods: We identified 293 follicular lymphoma patients with clinical outcomes available within the Stanford lymphoma database, who had undergone an excisional biopsy, and whose tumor biopsy specimens had been analyzed by flow cytometry. These patients were diagnosed between 1977 and 2003. Among these patients, 167 (57%) had follicular small cleaved lymphoma, 112 (38%) had follicular mixed lymphoma, and 14 (5%) had follicular large cell lymphoma. A total of 66 deaths were documented. The overall survival was estimated from the time of diagnosis to the time of death or last follow-up. Biopsies had been analyzed for cell surface marker expression by single parameter flow cytometry. The percentage of CD3+, CD4+, CD8+, CD20+, and HLA−DR+ cells was determined directly. We also estimated activated T cells, defined by HLA−DR+ T cells, by subtracting the value for total B cells from the value for HLA + cells. Cell population percentages were treated as continuous variables and related to overall survival, each in a univariate analysis.
Results: We found no statistically significant correlation between the percentage of any cell population and overall survival. For each group (CD3+, CD4+, CD8+, and estimated HLA−DR+ T cells), we used the percentage of the cell population to divide patients into 3 subgroups of equal number of patients: low-, medium-, and high-percentage. We then compared Kaplan-Meier curves of the 3 subgroups for each cell population. We found no significant differences in survival among the 3 subgroups for each of the cell populations, CD3+, CD4+, CD8+, and the estimated activated T cells, after correcting for multiple hypotheses testing.
Conclusion: The percentage of tumor-infiltrating T cells or T cell subsets was not correlated with survival outcome in patients with follicular lymphoma. This finding provides the basis for building molecular predictive models based on gene expression analysis, which represents not only the presence of certain cells but their physiologic state as well.