2 Clarke Drive
Suite 100
Cranbury, NJ 08512
© 2024 MJH Life Sciences™ and OncLive - Clinical Oncology News, Cancer Expert Insights. All rights reserved.
Steroid use while cytokine release syndrome and neurologic toxicities are at grade 1, instead of waiting until grade 3, reduces the rate of CAR T-cell treatment–related CRS and neurologic events.
Max S. Topp, MD
Steroid use while cytokine release syndrome (CRS) and neurologic toxicities are at grade 1, instead of waiting until grade 3, reduces the rate of CAR T-cell treatment—related CRS and neurologic events.
In a cohort of patients with relapsed/refractory (R/R) large B cell lymphoma (LBCL) who were treated with axicabtagene ciloleucel (axi-cel; Yescarta) who received early corticosteroid and tocilizumab as part of a nonrandomized expansion arm of the ZUMA-1 study, only 2% experienced grade ≥3 CRS and only 17% experienced grade ≥3 neurologic events, compared with 13% and 28%,1 respectively, in patients who started them later, said lead study author Max S. Topp, MD, at the 2019 ASH Annual Meeting.
“Earlier steroid intervention has the potential to improve the risk/benefit profile of CAR T-cell therapy with axi-cel,” said Topp, professor of oncology, University Hospital of Wuerzburg, Germany. “This intervention led to a lower total cumulative steroid usage in this cohort with no impact on expansion of T cells. It reduced markers of inflammatory response, had no impact on ongoing response with or without adjustment for tumor burden at a median follow-up of 8.7 months.”
ZUMA-1 was a pivotal multicenter single-arm phase I/II study in which axi-cel treatment in patients with R/R LBCL induced an objective response rate of 83%, with a 58% rate of complete response, a median overall survival (OS) of 25.8 months, and a 3-year OS rate of 47%.2 In cohorts 1 and 2 of ZUMA-1, grade ≥ 3 CRS and neurologic events occurred in 11% and 32% of patients, respectively.2 Cohort 3 evaluated the use of prophylactic tocilizumab on day 2, which appeared to decrease rates of grade ≥3 CRS but not rates of grade ≥3 neurologic events.3 An additional safety cohort (cohort 4) was added to evaluate the effect of earlier corticosteroid use on the rates of these adverse events.1
Cohorts 1 and 2 of the phase 2 portion of ZUMA-1 included 101 patients. Cohort 1 enrolled 77 patients with R/R diffuse LBCL; cohort 2 enrolled 24 patients with R/R primary mediastinal B cell lymphoma/transformed follicular lymphoma. Cohort 4 enrolled 41 patients with R/R LBCL. Optional bridging therapy was allowed in cohort 4.
In cohorts 1 plus 2, neither corticosteroids nor tocilizumab were given with grade-1 CRS or neurologic events; both were given with grade-2 CRS only if patients had comorbidities or were older. Tocilizumab was administered to all patients with grade ≥3 CRS. Patients with grade ≥2 neurologic events received tocilizumab but not corticosteroids. Corticosteroids were given with grade-3 neurologic events only if no improvement was observed with tocilizumab. Both tocilizumab and corticosteroids were used in the case of grade-4 CRS or neurologic events.
Patients in cohort 4 received early steroid intervention starting at grade 1 neurologic events and at grade 1 CRS when no improvement was observed after 3 days of supportive care.
At baseline, 57% of patients in cohorts 1 plus 2 had stage IV disease versus 46% in cohort 4. About two-thirds (65%) of patients in cohorts 1 and 2 had progressive disease to most recent chemotherapy, compared with 37% in cohort 4. Patients enrolled in cohort 4 had a lower median tumor burden by sum of product diameters and lower pretreatment serum LDH level compared with cohorts 1 plus 2. Median tumor burden was 3273 mm2 in cohort 1 plus 2 versus 2100 mm2 cohort 4, and the median levels of serum LDH were 356 and 262 U/L, respectively. The median ferritin level was 786 ng/mL in the combined cohort 1 plus 2 versus 393 ng/mL in cohort 4.
The rates of CRS of any grade were 93% in the combined cohorts 1 and 2 and in cohort 4. The median time to onset of CRS was 2 days in each cohort and the median duration of CRS was 8 days in cohort 1 plus 2 versus 7 days in cohort 4.
The rates of neurologic events of any grade were 64% in cohort 1 plus 2 and 61% in cohort 4. The median duration of neurologic events was 12 days and 9 days, respectively. In patients with tumor burden above the median, the rates of grade ≥3 CRS and neurologic events was reduced from 6% and 20%, respectively, in cohort 1 plus 2 to 1% and 4%, respectively, with earlier use of corticosteroids and tocilizumab in cohort 4.
The ORR in cohort 4 was 73%, with a CR rate of 51%. More than half (54%) of patients in cohort 4 remained in ongoing response with at least 6 months of follow-up, compared with a 44% ongoing response rate at the primary analysis of cohort 1 plus 2 (also with ≥6 months of follow-up). Responses were comparable between cohorts when evaluated by tumor burden. Median duration of response was 8.9 months in cohort 4, which is consistent with that observed at the primary analysis of cohorts 1 plus 2 (8.1 months).
“By using early onward steroids, we could lower the amount of steroids that were actually given over the clinical trial,” said Topp. Some 85% of patients in cohort 1 plus 2 received at least 5 doses of steroids compared with only 43% of patients in cohort 4. The median cumulative steroid dose was reduced from 5451 mg in cohort 1 plus 2 to 939 mg in cohort 4.
Early steroid use had no impact on CAR T-cell expansion; the overall peak number of CAR T cells and the area under the curve were similar between cohort 1 plus 2 and cohort 4, said Topp. “Altogether, giving steroids [early] had a positive impact on the inflammatory environment in these patients,” as inflammatory serum cytokines associated with CAR T-cell—related toxicity were reduced in cohort 4, he said.