Impact of Bridging Chemotherapy on Clinical Outcomes of CD19 CAR T Therapy in Relapse/Refractory Diffuse Large B- Cell Lymphoma in Real World Experience

Background: Autologous chimeric antigen receptor (CAR) T cell therapy has shown to be effective in relapsed or refractory (R/R) diffuse large B cell lymphoma (DLBCL) but requires in real world and in Europe at least a 4 weeks period of cell processing. During this "bridging period", patients are vulnerable to disease progression and complications. We sought to characterize bridging therapy strategies.

Methods: We performed a retrospective review of patients (pts) with R/R DLBCL treated with commercialized CD19 CAR T cells, either Axicabtagene Ciloleucel (Yescarta) or tisagenlecleucel (Kymriah). Bridging therapy (BT) was defined by any therapy given from enrolment to cell infusion. We divided bridging therapy in 2 groups: high intensity BT (HI, including chemotherapy+/-immunotherapy) and low intensity BT (LI, including monoclonal antibodies rituximab, brentuximab, dexamethasone, lenalidomide). We evaluated toxicity (cytokine releasing syndrome (CRS), CAR-related encephalopathy syndrome (CRES), infections) and efficacy (OS, PFS) after infusion according to the intensity of the BT.

Results: 46 pts were enrolled for commercialized CAR T cells (Kymriah n=25, Yescarta n=21) including R/R DLBCL (n=35), TFL (n=5) or PMBL (n=6) between June 2018 and April 2019. The median age was 57 (IQR, 42 to 64). Number of pts with aaIPI > 2 was 20. The median number of previous lines was 3 (range 3-5), including 15 autologous stem cell transplants and 2 allotransplants. Median time between enrollment and infusion, and between collection and infusion were 59 (35 to 118) and 40 (32 to 90) days, respectively.

Overall, 30 (65%) pts received a HI and 16 (35%) a LI or no bridging therapy. Median number of BT cycles was 2 (range, 1-4). 13 patients had various regimen because of uncontrolled disease. Pts receiving HI presented at enrollment with more elevated LDH (60% vs 31%), more involved extranodal sites (n> 2, 32% vs 0%), higher IPI (>2, 77% vs 43%, p=.049), and higher total metabolic tumor volume (TMTV) measured on FDG PET/CT (median, 90 vs 10, p=.002).

After BT, 44 pts were infused. 2 pts died before infusion because of a pulmonary infection (n=1) and a lymphoma progression (n=1), both in the HI group. At infusion response evaluation showed that 25 (57%) pts progressed during BT, without difference in the HI or the LI groups (61 % vs 50%). Median TMTV in HI and LI was 95 (10 -256) vs 18 (8 - 44), respectively. After CAR T infusion, 14 pts (32%) developed an infectious disease (7 in the LI group, 7 in the HI group, p=0.31). Median duration of neutropenia (PNN < 1000) was 8 days, 7 in the LI and 9 in the HI (p= 0.21).

In the LI and HI groups, 11 and 19 pts developed a CRS (0 grade 3-4), 1 and 11 pts developed a CRES (1 and 3 grade 3-4), respectively. The HI/LI did not impact the occurrence of CRS (p=1.00) but that of CRES (p=0.03). Pts with abnormal LDH (p=0.02), high number of courses (p=0.03) or PS>0 (p=0.049) had a higher risk of CRS, whereas patients with high IPI (p=0.03), and high values of CRP (P=0.02) or decreased albumin levels (p=0.008) had a higher risk of CRES.

After CAR T cells infusion, with a median follow up of 5.7 months, the median PFS was not reached in the LI group, and was of 3 months in the HI group (p=0.06). The median OS was not reached in the LI group, and was of 12.5 months in the HI group (p=0.09). In multivariate analyses, IPI was predictor of better PFS, whereas low IPI and TMTV were predictors of better OS. No evidence of any effect of intensity of the BT on OS or PFS was observed in these multivariate analyses.

Conclusion: The type of bridging therapies was very heterogeneous. Its intensity was closely related to tumor burden at enrolment. There was no significant difference in terms of efficacy between the HI and LI groups, but a higher frequency of CRES in the HI group.