Proportions of Bone Marrow Lymphocyte Subsets Predict Survival in Patients with Newly Diagnosed Diffuse Large B-Cell Lymphoma Independent of International Prognostic Index

Background: Previous studies on lymphocyte subsets in blood and lymph node suggest that host immune status at diagnosis may predict a better outcome in patients with diffuse large B-cell lymphoma (DLBCL). Since bone marrow is a major part of the host immune system and bone marrow biopsy is routinely performed as part of staging work-up in DLBCL, we studied the prognostic significance of bone marrow lymphocyte subsets at diagnosis as determined by flow cytometry in patients with DLBCL.

Methods: All DLBCL patients diagnosed and treated at our institution from 2000– 07 were identified through our cancer registry. We included all patients who had flow cytometric analyses of their bone marrow aspirates at the time of diagnosis. Patients with HIV, concurrent low grade lymphoma, Richter’s transformation, and DLBCL arising from the central nervous system and skin were excluded. We examined the significance of the following variables in predicting overall and disease-specific survival: International Prognostic Index (IPI) score (0–1,2, 3, or 4–5), chemotherapy received [rituximab/cyclophosphamide/doxorubicin/prednisone (R-CHOP), CHOP, or others], absolute lymphocyte count in the blood, and proportions total lymphocyte and lymphocyte subsets [B-cells (CD19+), total T-cells (CD3+), helper-T cells (CD3+/CD4+), cytotoxic-T cells (CD3+/CD8+), and NK-cells (CD16+/CD56+)] in the bone marrow. We used the method of Contal and O’Quigley to determine cut-off points for lymphocyte subsets.

Results: Eighty five patients met the study criteria and were included. The mean age at diagnosis was 65 years (range, 31–97) and 60% were males. IPI distribution was 12%, 25%, 22%, 26% and 15% for scores of 0, 1, 2, 3, and 4, respectively. Most patients received either RCHOP (62%) or CHOP (28%) chemotherapy. Only 7% had bone marrow involvement by DLBCL. Univariate analysis showed that lower IPI scores, treatment with CHOP±R, and higher proportions of bone marrow total lymphocytes, B-cells, T-cells (total and subsets), and NK-cells, but not absolute lymphocyte count in the blood, were associated with better overall and disease specific survival. When evaluated in a multivariate model (Table), higher proportion of total T-cells, lower IPI scores, and treatment with CHOP±R were significant independent predictors of better overall and disease specific survival. Higher proportion of NK-cells was a significant predictor of better disease-specific but not overall survival.

Conclusion: The presence of higher proportions of T-cells and NK-cells in the bone marrow predicts a better prognosis in patients with DLBCL independent of IPI score and treatment received. These findings support the importance of the role of host immune system in modulating the clinical course of DLBCL even in the era of rituximab-based treatment. If our findings are confirmed, routine assessment of bone marrow lymphocyte subsets at diagnosis in conjunction with IPI may provide a better prediction of clinical outcome in patients with DLBCL.