Abstract
Abstract 3953
Poster Board III-889
AIMS
Several studies concur that the microenvironment determines outcome in follicular lymphoma (FL), but they disagree regarding which components thereof are important. Our hypothesis was that several immune cell subsets are important for disease outcome and their individual prognostic importance should be demonstrable in the same analysis and in competition with clinical factors. Specifically, we hypothesized that (1) CD8+ cells are associated with good prognosis (presumably due to tumor cell killing), as are (2) cells positive for programmed death-1 (PD-1) or FOXP3 (due to diminished B-cell stimulation), while (3) CD4+ cells are associated with poor prognosis (due to B-cell stimulation).
PATIENTS AND METHODS
Seventy FL patients with extreme clinical outcome (“poor” and “good” cases) were identified in a cohort of 197 patients. The criterion for poor outcome was death from lymphoma <5 years after diagnosis (n=33). The general criteria for good outcome were absence of a lymphoma-related death and/or transplantation and one of the following three statements had to be true: (1) never treated against lymphoma and followed for ≥5 years (n=11); (2) never relapsed after first-line anti-lymphoma treatment and followed for ≥8 years (n=14); (3) relapsed but never received intensive or frequent (≥3 years between) treatments and followed for ≥10 years (n=12), rendering totally 37 good-outcome patients. A tissue microarray was constructed from diagnostic and relapse biopsies of these 70 patients. Sections of the microarray were stained for CD3, CD7, CD4, FOXP3, PD-1, CD8, TIA-1, granzyme B, perforin, CD57, CD56, CD68, and tryptase. The number of positive cells for each staining were quantified using computerized image analysis, separating cells inside and outside the follicles (follicular and interfollicular compartments).
RESULTS
Between the two clinical extreme groups there were great differences in the FL International Prognostic Index (FLIPI), as expected (P<0.0001). In univariate analysis, the amounts of several immune subsets were different between the two groups with borderline or stronger significance. These subsets were taken to multivariate analysis together with the FLIPI. Independently of the FLIPI, CD4+ cells were associated with poor (Odds Ratio [OR] 1.26; P=0.025) but PD-1+ (OR 0.58; P=0.020) and CD8+ (OR 0.94; P=0.024) cells with good outcome. In a second multivariate analysis, where the subsets in the follicular and interfollicular comparments were analyzed (again in competition with the FLIPI), the prognostic values of CD4+ and PD-1+ cells were accentuated when they were follicular (OR 2.16; P=0.010 and OR 0.34; P=0.019, respectively), and that of CD8+ cells when interfollicular (OR 0.86; P=0.014). Follicular FOXP3+ cells were also associated with good outcome (OR 0.09; P=0.018) and interfollicular CD68+ cells with poor (OR 1.36; P=0.040).
CONCLUSION
We conclude that there are many important immune cell subsets in the microenvironment of FL. Independently of the FLIPI, and of each other, PD-1+, FOXP3+, CD4+, CD8+, and CD68+ cells correlate with outcome. This suggests several different but not mutually exclusive mechanisms which all affect the course of the disease.
Disclosures:
No relevant conflicts of interest to declare.
Local glucocorticoid production in the mouse lung is induced by immune cell stimulation