Abstract
Introduction: Based on efficacy of CD22-targeted CAR T-cells in a phase I trial in adults with relapsed/refractory (r/r) large B-cell lymphoma (LBCL)—including in those with relapse following CD19 CAR T-cells (CAR19) (Frank M, et al. Lancet 2024), firicabtagene autoleucel (firi-cel/CRG-022), an autologous CD22-directed chimeric antigen receptor (CAR) T-cell therapy was tested in a Phase 2 study.
Methods: The primary objective of this phase 2, open-label multicenter study was to evaluate efficacy of firi-cel in adults with r/r LBCL who progressed after CAR19 (cohort 1). The primary endpoint was overall response rate (ORR) as determined by blinded central review using Lugano Response Criteria. Cohort 2 was for those with a non-conforming product; cohort 3 was for those who received prior CAR19 and bispecific T-cell engagers. In contrast to the phase I study, CD22 expression was not required. Secondary objectives were to evaluate toxicity and additional efficacy endpoints, including ORR, complete response (CR), duration of response (DOR), progression-free survival (PFS) and overall survival (OS) as determined by individual investigators. Firi-cel was centrally manufactured using the Miltenyi Prodigy and utilized a different, more rapid in-culture timeline (5-9 days) compared to the phase I trial (7-12 days). Enrollment started Aug 2023 and data cut-off was 4/8/2025.
Results: 138 patients underwent screening and 101 proceeded to leukapheresis (37 screen failures). 93 products were manufactured and 84 patients were infused. 15 patients dropped out due to manufacturing failures (7), disease progression (3), PI decision (4), or other (1). The median vein-to-vein time was 35 days (range, 24-86 days). For those infused, the median age was 65.5 years (19-87 years), 56 (67%) were male, 6 (7%) were non-white and 12 (14%) were Hispanic. The median number of prior lines of therapy was 3 (2-8) and the median time from prior CAR19 to apheresis was 7.4 months (1.6-66 months). 12 (14%) had bulky disease and 46 (55%) had elevated LDH. The median H-score for tumor CD22 expression was 65 (range 0-300). Among 52 patients with an available sample, 10 had undetectable CD22.
At data cut-off (median follow-up of 4.8 months) and using investigator-determined responses (as central review was not available after trial closure), for cohort 1 (n=69), the ORR and CR rates were 73% (95% CI, 60-83%) and 46% (34-59%), respectively. The median DOR was 2.4 months (95% CI, 2.0-5.1). The median PFS and OS were 3.1 months (2.6-4.9) and 12.4 months (7.4-non-estimable (NE)). Results for cohort 2 (n=3) were comparable and will be reported later.
For cohort 3 (n=12), the ORR and CR rates were 50.0% (21-79%) and 25% (6-57%) respectively. The median DOR was 2.2 months (1.1-NE), while the median PFS and OS were 2.8 months (0.9-3.4) and 5.9 months (2.8-NE), respectively.
Any and grade 3+ CRS occurred in 66 (79%) and 4 (5%) patients. 8 (10%) had any grade ICANS with no Gr3+ seen. Immune effector cell associated hemophagocytic lymphohistiocytosis (IEC-HS) occurred in 21 (25%) subjects and was grade 3+ in 11 (13%). 5 (6%) had a fatal treatment-emergent adverse event, all related to IEC-HS or complications thereof.
Firi-cel harvested after 5 (n=23), 7 (n=31), or 9 (n=15) days associated with 3-month CR rates of 32%, 15%, and 7%, respectively, and circulating firi-cel detectable in the peripheral blood in 91%, 58%, and 0% of patients at 3 months, respectively. Interestingly, day 5 products were enriched with central memory T-cells; while day 9 products were enriched with effector memory T-cells.
CD22 tumor expression did not correlate with ORR or IEC-HS incidence or severity. Longer follow up is needed to determine whether CD22 tumor expression significantly correlates with DOR/PFS; so far patients with undetectable CD22 tumor expression trended toward shorter DOR/PFS.
The trial was terminated due to limited durable responses and higher than anticipated toxicity.Conclusions: Firi-cel demonstrated high ORR in patients with r/r LBCL after prior CAR19. However, DOR was low likely due to altered manufacturing protocols and enrollment of patients with undetectable CD22 disease. Altered manufacturing may also have contributed to higher incidence and severity of IEC-HS. Further exploration of firi-cel pharmacokinetics and biology in relationship to toxicity will be essential to better understand the continued potential of CD22-targeting in LBCL.
