Timing of PD-L1 Blockade with Durvalumab May Affect Outcomes of CD19 CAR-T Cell Therapy for Relapsed/Refractory Large B-Cell Lymphoma

Introduction CD19-targeted chimeric antigen receptor T (CAR-T) cell therapy has shown high overall response rates (ORRs) in patients (pts) with relapsed/refractory (R/R) large B-cell lymphoma (LBCL); however, CD19 CAR-T cells fail to elicit a complete response (CR) in a considerable proportion of pts. PD-L1 expression in the tumor microenvironment could impair the efficacy of CAR-T cell therapy. PD-1/PD-L1 pathway blockade may enhance the function and antitumor activity of CAR-T cells. Here, we report the analysis of a phase I dose-finding study (NCT02706405) of combination therapy with CD19-targeted 4-1BB-costimulated CAR-T cells (JCAR014) and escalating doses of durvalumab (durva), an anti-PD-L1 monoclonal antibody, in adults with R/R LBCL. Methods Pts were treated in 1 of 2 groups. All pts received lymphodepletion (LD) with cyclophosphamide (Cy) and fludarabine (Flu) followed by JCAR014 infusion. Pts in group 1 (Grp 1) received the first dose of durva (225 mg or 750 mg) after JCAR014 infusion. Pts in group 2 (Grp 2) received the first dose of durva (7.5 mg, 22.5 mg, 75 mg, 225 mg, or 750 mg) 1 day prior to JCAR014 infusion. Up to 10 doses of durva were administered at 4-week intervals after JCAR014 at the highest identified safe dose until toxicity or disease progression. Adverse events (AEs) were graded by CTCAE 4.03, except for cytokine release syndrome (CRS), which was graded according to Lee 2014 consensus criteria. Response was reported according to PET/CT 2014 Lugano criteria. Results Between 01/23/2017 and 07/02/2020, 34 pts were screened, 32 underwent leukapheresis, and 29 were enrolled and received LD and JCAR014. Eleven pts were treated in Grp 1 and 18 in Grp 2 (Table 1). All pts received 2 x 106 JCAR014 cells/kg, except for the first 2 pts treated on the study who received 7 x 105 cells/kg. The most common AEs of any grade related to LD and JCAR014 with or without durva were CRS (41%), neutropenia (21%), neurotoxicity (17%), and hypogammaglobulinemia (17%). The incidences of any grade and grade ≥ 3 CRS and neurotoxicity were similar between Grp 1 and Grp 2 (Table 2A). Patients in Grp 2 had later onset of CRS after JCAR014 infusion compared with Grp 1 (median of 6 vs. 4 days, P = .05) and shorter duration of CRS (median of 3 vs. 8 days, P = .08). Among the 27 pts evaluable for dose-limiting toxicity (DLT), 2 (7%) had DLT: 1 pt with grade 4 CRS in Grp 2 (durva 225 mg) and 1 pt with prolonged grade 3 neurotoxicity in Grp 2 (durva 750 mg). No maximum-tolerated dose of durva was identified. In pts evaluable for response (n = 26), the ORR and the CR rate at 3 months were 35% (95% CI, 17-56) and 27% (95% CI, 14-46). We retrospectively compared the response rates of LBCL pts treated with JCAR014 in combination with durva (NCT02706405) with LBCL pts who received the same Cy/Flu lymphodepletion regimen and JCAR014 dose (2 x 106 cells/kg) without durva (JCAR014 alone cohort) on our previous phase I/II clinical trial (NCT01865617). There was a trend towards lower ORR (P = .08) and CR rate (P = .09) in pts treated with JCAR014 in combination with durva compared with those treated in the JCAR014 alone cohort (Table 2B). There were no significant differences in ORR and CR rate between patients in Grp 1 and those who received JCAR014 alone. However, despite a trend towards lower tumor burden, pts in Grp 2 (first durva before JCAR014) had lower ORR (P = .07) and CR rate (P = .03) compared with those treated in the JCAR014 alone cohort, and a trend towards lower CR rate compared with those treated in Grp 1 (first durva after JCAR014; P = .16). Patients treated with JCAR014 alone had significantly shorter time to CAR-T cell peak counts in blood compared with patients treated with JCAR014 in combination with durva (median of 8 vs. 14 days, P = .01). We did not observe significant differences in peak CAR-T cell expansion, area under the curve (AUC) from day 0 to 28 after JCAR014 infusion, and day 28 CAR-T cell counts by qPCR between the cohorts. Conclusion The combination of JCAR014 with durva for the treatment of adult pts with LBCL was safe. Our findings suggest that the timing of initiation of durva therapy is a key variable that may affect outcomes. To our knowledge, this is the first report suggesting that PD-L1 blockade may impact toxicity and antitumor response in pts with LBCL undergoing CD19 CAR-T cell therapy. Additional studies will be required to determine the optimal approach for combining CD19 CAR-T cell therapy with PD-1/PD-L1 pathway blockade in LBCL. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal