Prognostic Significance of Bone Marrow Involvement At Diagnosis and At the Time of Autologous Stem Cell Transplantation for Lymphoma Relapse,

Abstract 4122

Bone marrow (BM) involvement at the time of initial diagnosis occurs in 4%-14% of patients (pts) with Hodgkin (HL) and 18%-39% with non-Hodgkin (NHL) lymphomas. Both clinical and prognostic implications of BM involvement are largely related to upstaging of lymphoma. Despite formal guidelines on selecting appropriate candidates for initial staging BM biopsy, it is performed frequently as a part of diagnostic work up and almost routinely prior to autologous stem cell transplantation (ASCT) for all lymphomas. We investigated the prognostic impact of sequential BM involvement (i.e negative or positive at diagnosis [Dx-, Dx+] and pre-transplant [Tx-, T+]) on post-transplant disease relapse in a cohort of 788 consecutive lymphoma pts who underwent ASCT from 1/1998 to 12/2010 at our institution. All pts (median age=52 yrs, 63% males, 92% Caucasians) received Bu/Cy/VP-16 preparation for their ASCT. A total of 244 pts (31%) were Dx+ and 544 (69%) were Dx-. 203 pts (25.8%) were Dx+/Tx-; 41 (5.2%) were Dx+/Tx+; 533 (67.6%) were Dx-/Tx-; and 11 (1.4%) were Dx-/Tx+. Only 2% of those with negative BM at diagnosis (Dx-) converted to positive (Dx-/Tx+) prior to ASCT (i.e. negative predictive value [NPV]=98%). Dx-/Tx+ group was not associated with any particular lymphoma subtype. 83.2% of pts cleared their BM involvement by the time of ASCT. Out of 52 pts with positive BM involvement at the time of ASCT, 20% had prior negative BM as opposed to 80% with prior positive BM. 642 (81.5%) pts had NHL (40.6% DLBCL, 18.4% FL, 10.8% MCL, 7.9% T-cell NHL, 4% other B-cell NHL) and 18.5% had HL. 79.2% had stage III-IV and 132 (16.8%) had bulky (>10 cm) disease at the time of diagnosis. Median time from lymphoma diagnosis to ASCT was 17 months. At the time of ASCT, 551 pts (69.9%) were in their 2^ndcomplete or partial remission (CR2/PR2). 559 pts (70.9%) had their CD34+ stem cells mobilized with G-CSF+VP-16, whereas 155 (19.7%) received G-CSF alone. 330 pts (41.9%) experienced disease relapse post-transplant. Median time-to-relapse was 29.1 months for Dx+/Tx+, 23.2 for Dx-/Tx+, 37.5 for Dx+/Tx-, and it was not reached for Dx-/Tx- (p =0.03). Prognostic factors for lymphoma relapse after ASCT in the univariate Cox proportional hazards analysis included BM involvement (HR=1.28, 95% CI, 1.01–1.63 for Dx+/Tx-; HR=1.69, 95% CI, 1.12–2.54 for Dx+/Tx+; Dx-/Tx-=reference), male gender (HR=1.43, 95% CI, 1.13–1.8), number of prior chemotherapies (HR=1.14, 95% CI, 1.05–1.23 per 1 regimen increase), disease remission status at transplant (HR=1.48, 95% CI, 1.1–1.99 for CR2/PR2 and HR=2.25, 95% CI, 1.5–3.38 for relapsed/refractory lymphoma, CR1/PR1=reference), and type of lymphoma (HR=1.74, 95% CI, 1.15–2.62 for T-cell NHL, FL=reference). Bone marrow involvement remained prognostic for lymphoma relapse along with gender, disease remission status, and type of lymphoma in the multivariable Cox analysis (Table). In conclusion, our study demonstrated independent prognostic significance of BM involvement in the context of ASCT for NHL and HL. It also revealed superior NPV of BM involvement at diagnosis with no impact of positive BM conversion on post-transplant lymphoma relapse which calls into question the routine practice of screening pre-transplant BM examinations.