Outcomes of Patients with Refractory and Relapsed Primary Mediastinal B Cell Lymphoma: A Canadian Single-Center Analysis

Background: Primary Mediastinal B- Cell Lymphoma (PMBCL) is an aggressive, rare non-Hodgkin Lymphoma. Despite improved outcomes with rituximab-based frontline therapy, approximately 10-15% of patients are either primary refractory or relapse after a short remission duration. Historical studies have shown inferior response rates and survival in chemorefractory patients (Kuruvilla, 2008) compared to chemosensitive disease. Data defining outcomes of relapsed or refractory PMBCL (rrPMBCL) are infrequently reported. We aimed to describe outcomes for patients with rrPMBCL who received salvage chemotherapy after frontline therapy with intent to transplant and/or subsequent treatments including PD1 inhibitors, or CAR T- cell therapy.

Methods: Using the REB-approved lymphoma database at Princess Margaret Cancer Center, we identified patients with PMBCL receiving 2^nd line therapy with ASCT intent between January 2012 and August 2023. Primary endpoint was event-free survival (EFS), defined as time from first salvage therapy to the earliest date of disease progression, commencement of new therapy or death from any cause.

Results: We identified 26 patients with rrPMBCL. Median age was 28 years (range 21-65) and 62% were female. At initial presentation, 43% of patients had advanced stage (Stage III/ IV) while 38% had extranodal involvement. Twenty-two (85%) patients were primary refractory, and 4 patients relapsed at a median of 14.7 (range-11-21) months after receiving rituximab and anthracycline-based frontline chemotherapy (RCHOP: 22, dose-adjusted EPOCH R: 4). At relapse, half (n=13) had stage IV disease; 65% (n=17) had extranodal involvement including CNS involvement in 3 patients. All patients received platinum-based salvage chemotherapy (GDP -24, DHAP-1, DICEP-1) with an ORR of 27 % (n=7; all PR). Five patients received a second line of salvage chemotherapy (MiniBEAM-3, ICE-1, DHAP-1) because of inadequate response or progressive disease. Ten (38%) patients ultimately proceeded to ASCT (7 in PR, 3 in SD). Although post-ASCT day-100 response rate was 60% (CMR -2, PR-4), 3 of 4 patients with PR pre-transplant relapsed at 5, 7 and 12 mos.

Thirteen patients received anti-CD19 CAR T cell therapy (Axicabtagene ciloleucel-12; Tisagenleluecel-1) as 3^rd line (n=10) or 4^th line (n=3) treatment at a median of 4 (range 2.7-16) months post their first progression/relapse. Bridging therapy was administered in 7 (54%); localised mediastinal radiation and steroids in 3, polatuzumab-BR in 2 and pembrolizumab in 2 patients. All but 2 patients had progressive disease before CAR T cell infusion. ORR rate at 3-months was 54% (n=7; CR-3, PR-4); 2 patients died due to progression prior to 3 months.

Nine patients received checkpoint inhibitors (8-pembrolizumab, 1- Nivolumab plus Brentuximab) in 3^rd-line (n=4) or 4^th line (post CAR T; n=5). ORR was achieved in 55% (n=5; CR-4, PR-1). All 4 patients who attained a CR received pembrolizumab post CAR T cell therapy relapse/ progression with durable responses. (DOR- 5, 20, 29 and 35 months)

At a median follow up of 14 (range-2-113) months from first salvage therapy, 15 patients had died [progressive disease- 12, infections post CAR T therapy-2 (COVID; CMV viremia + HLH), unknown cause-1]. 1-year EFS of our cohort was dismal at 8% (95%CI 2.2-31). Median PFS post ASCT was 9.95 months with 2-year PFS of 10%. Median PFS of patients receiving CAR T cell therapy was 12 months with 2-year PFS of 46.7%. Median OS of the whole cohort from first salvage was 17 months (95%CI 11.86- not reached) with 1-year and 2-year OS of 62% (95% CI 45-85) and 35% (95% CI 20-62) respectively. 2-year OS was significantly better (66.7%) in patients who underwent CAR T cell therapy vs those who did not (25%; p=0.01). We analysed factors predictive of OS after first relapse/progression using Cox-proportional hazard regression model: undergoing CAR T cell therapy was the only factor favouring better survival (adjusted HR 0.27 (95% CI 0.09-0.82; p=0.02) with no impact of stage, LDH, extranodal disease, IPI at relapse, and ASCT.

Conclusion: Outcomes in patients with rrPMBCL remain poor, with a minority of patients responding to salvage chemotherapy and proceeding with ASCT. CAR T-cell therapy appears promising in improving outcomes in rrPMBCL. Multicentric larger prospective studies are needed to explore the role of PD1 inhibitors post CAR T relapse, given the encouraging outcomes in this small subgroup of our study.