Retreatment with R-CHOP-like Therapy in Patients with Late Relapse of Diffuse Large B-Cell Lymphoma (DLBCL)

Background: Most patients (pts) with DLBCL who experience relapsed/refractory disease following R-CHOP will do so within 24 months of diagnosis, but late relapses (> 2 years [y]) can occur and are typically associated with better outcomes. Pts with late relapse have lymphomas that share few mutations with the original tumor and may represent de novo DLBCL arising from a common precursor cell (Hilton et al, JCO 2023). It has been hypothesized that these pts have chemotherapy-naive disease and may be effectively treated with a second round of R-CHOP-like therapy. This approach has been adopted for many pts managed in British Columbia (BC), particularly those who are unsuitable for more intensive options. The objective of this retrospective study was to evaluate outcomes in pts treated with R-CHOP-like therapy for late relapse DLBCL in BC.

Methods: The BC Cancer Centre for Lymphoid Cancer Database was used to identify all pts with de novo DLBCL treated with an initial R-CHOP-like regimen who developed late relapse (> 2 y from initial diagnosis) of DLBCL from January 2005 to April 2024 and received second-line R-CHOP-like therapy with curative-intent (typically R-CHOP or R-CEOP, in which etoposide is substituted for doxorubicin once maximum lifetime anthracycline exposure has been reached [Moccia et al, Blood Advances 2021]). Transformed lymphoma, PTLD, PMBCL, and CNS lymphoma were excluded. Primary outcome was time-to-progression (TTP). Other outcomes included overall response rate (ORR); complete response (CR) rate; progression-free survival (PFS); overall survival (OS); and disease-specific survival (DSS). Survival endpoints were calculated from the time of relapse to: progression of lymphoma or treatment/disease-related death (TTP), progression of lymphoma or death from any cause (PFS); death from any cause (OS); or treatment/disease-related death (DSS). Pts with ‘intent-to-transplant’ were evaluated for response, but time-to-event outcomes were censored at time of transplant.

Results: 53 pts meeting all inclusion criteria were identified with a median age of 77 y (range: 52-89) at time of relapse. 30 (57%) were male. Median time from initial diagnosis to relapse was 7.4 y (range: 2.5 – 15.9). 47 pts had biopsy confirmation of DLBCL at relapse. Cell of origin (COO) was GCB in 24/37 (65%), ABC/non-GCB in 12/37 (32%), or unclassifiable (1/37; 3%), according to gene expression profile (11/37) or Hans algorithm (26/37). COO concordance with initial diagnostic biopsy was 24/26 (92%). At time of relapse, in pts with available information, 80% had stage III-IV disease; 53% had poor performance status (ECOG 2-4) and 80% had high IPI (3-5). 15 pts (28%) started R-CHOP, with a median of 3 cycles (range 1-6) before switching to R-CEOP (n = 7), stopping for transplant (n = 4), limited stage (n = 2), or toxicity (n = 2). The remainder (n = 38) were started directly on R-CEOP. The median number of total cycles (R-CHOP and/or R-CEOP) was 5 (range 1-6), with 5 pts receiving consolidative radiation therapy. Four pts were treated with a limited stage approach. 19 pts (36%) did not complete their planned treatment, with toxicity or intolerance (n = 12; 63%) being the most common reason for early discontinuation. Most pts were not considered for transplant due to age or comorbidities. Six pts were intended for autologous stem cell transplant, with 5/6 proceeding to transplant.

Response rates were as follows: CR 60%, PR 11%, SD 4%, PD 9%, non-evaluable/missing 15%. With a median follow-up of 29 months (m), 2-y TTP was 55%, 2-y PFS was 46%, 2-y OS was 55% and 2-y DSS was 67%. Median TTP was 45 m (95% CI 0 – 97), median PFS was 20 m (95% CI 0 – 45), median OS was 39 m (95% CI 13 – 65), and median DSS was 59 m (95% CI 41 – 78). Outcomes were also assessed according to timing of relapse following initial therapy. Pts who relapsed more than 5 y from initial diagnosis (n=41) had a significantly better TTP than pts who relapsed between 2-5 y (n = 12) with 2-y TTP 68% vs 11%, respectively (HR 0.27 [95%CI 0.11-0.64]; p = 0.003), despite comparable baseline characteristics. No difference in TTP was observed between COO subgroups (HR 1.14 [95%CI 0.82-1.60]).

Conclusion: In pts with late relapse DLBCL, a retreatment strategy with R-CHOP-like therapy is a reasonable option. This approach results in durable remission in a high proportion of pts, thereby avoiding the need for more intensive therapies, particularly in pts who experience relapse more than 5 y from initial diagnosis.