The Prognostic Value of Molecular Bone Marrow Involvement in Aggressive Lymphoma

Introduction: Morphological bone marrow (BM) involvement in diffuse Large B-cell lymphoma (DLBCL) and primary T-cell lymphoma (PTCL) has prognostic and therapeutic implications. The prognostic role of molecular BM involvement (B and T cell gene rearrangement (GR) studies), in patients with no evidence for involvement is uncertain.

Objective: To evaluate the prognostic role of molecular BM involvement in patients with DLBCL and PTCL, and its significance in the rituximab era.

Methods: This is a retrospective study of 416 patients with DLBCL (358) or PTCL (58) treated 1995-2010. All patients underwent BM biopsy for morphology and GR studies on BM aspirates at diagnosis and/or at end of treatment (EOT). Most patients were treated with CHOP or R-CHOP based protocols after diagnosis. Medical records were reviewed for details of initial disease characteristics, type and response to treatment, relapse if occurred and the date of death or last follow-up. Patients were classified into one of three groups according to BM results: morphology positive (MorphPos); morphology negative/GR positive (MorphNeg/GRpos); morphology negative/GR negative (MorphNeg/GRneg).

Main outcome measures: Overall survival (OS) and progression free survival (PFS) among the groups of patients at diagnosis and at EOT.

Results: At diagnosis there were 89 MorphPos, 49 MorphNeg/GRpos and 278 MorphNeg/GRneg patients. GRpos patients had lower complete remission rate (73.9%) than GRneg patients (84.8%; p=0.087). The median OS differed significantly among groups: MorphPos patients 66.4 months (95%CI: 50.7-82); MorphNeg/GRpos patients 89 months (95%CI: 66-111); and MorphNeg/GRneg patients 125 months (95%CI: 113-136), p<0.001. MorphNeg status improved PFS (p<0.001), however GRneg status did not (p=0.09). The hazard ratio for GRpos at diagnosis in MorphNeg patients was 1.52 (95% CI: 0.97-2.4; p=0.068). Among the subgroup of DLBCL patients, the median OS was significantly different: MorphPos patients 75.6 months (95%CI: 57.7-94.1); MorphNeg/GRpos patients 95.5 months (95%CI: 66.9-124.1); and MorphNeg/GRneg patients 129.6 months (95%CI: 118.2-141.1), p=0.036. 73% of DLBCL patients received rituximab. Rituximab improved OS (p<0.001) in all DLBCL patients, yet did not overcome the GR-positivity negative effect on OS. For PTCL patients, BM morphology and/or GR status did not affect survival. 72 patients were tested at the EOT, 5 were MorphPos and 17 GRpos. EOT MorphPos patients had marginally worse OS (p=0.07). EOT GR status in MorphNeg patients did not affect survival (p=0.5)

Conclusions: BM GR positivity at diagnosis, independent of BM morphological involvement, predicts worse OS in DLBCL even in the rituximab era, but not in PTCL. EOT morphologic but not molecular involvement, had marginally negative effect on survival.