Durable Responses to JCAR017 Observed in High-Risk DLBCL

Jeremy S. Abramson, MD

Responses to lisocabtagene maraleucel (liso-cel; JCAR017) have been potent and durable in the treatment of patients with relapsed/refractory diffuse large B-cell lymphoma (DLBCL). Eight of the patients who received liso-cel in the TRANSCEND trial have been treated as outpatients, requiring fewer hospitalization days with no instances of severe cytokine release syndrome (CRS), reported Jeremy S. Abramson, MD, at the 2018 ASCO-SITC Clinical Immuno-Oncology Symposium.

Separate exploratory analyses of the DLBCL population treated with liso-cel, presented by Tanya Siddiqi, MD, found that high tumor burden and a series of inflammatory biomarkers were associated with high chimeric antigen receptor (CAR) T-cell expansion and higher rates of CRS and neurotoxicity.

The updated results presented by Abramson, the clinical director of the Lymphoma Program at the Massachusetts General Hospital Cancer Center, show an 81% (95% CI, 62%-94%) objective response rate (ORR) and a 63% (95% CI, 42%-81%) complete remission (CR) rate in patients with relapsed/refractory DLBCL.¹Early overall survival (OS) data are encouraging among the patients with high-risk DLBCL, with a 6-month OS rate of 86% (95% CI, 73%-93%), which improves to 94% (95% CI, 78%-98%) among patients in CR.

TRANSCEND is a multicenter seamless design pivotal trial with 3 dosending cohorts. Patients in the study were treated with liso-cel, a CD19-directed CAR T-cell product with a 4-1BB co-stimulatory domain that is formulated into a precise 1:1 ratio of CD4-positive and CD8-positive CAR T cells.

Two dose levels (DLs) were used: 5 x 107 cells (DL1) as a single or double dose and liso-cel at 1 x 108 cells (DL2) as a single infusion. Based on the dose findings, 2 dose expansion cohorts were studied: DL1 and DL2 as single infusions, after which DL2 moved forward into the pivotal cohort focused on DLBCL. Data from all 91 patients in the phase I study were presented by Abramson, as well as that on 67 patients treated in the core DLBCL subset.

At study entry, patients underwent leukapheresis followed by manufacturing of the cellular product. Prior to receipt of liso-cel, patients underwent lymphodepleting chemotherapy with udarabine and cyclophosphamide for 3 days, followed 2 to 7 days later by liso-cel infusion.

The pivotal population had either de novo DLBCL, transformed from follicular lymphoma (FL), or high-grade (double/triple hit) B-cell lymphoma. Prior stem cell transplant (SCT) was permitted. Patients with prior allotransplant were not included in the pivotal cohort. Of the 140 patients leukapheresed, “there have been only 2 cases where we were unable to manufacture the liso-cel product,” he said. A total of 108 patients have been treated with liso-cel, of which 91 were evaluable (45 at DL1 single dose, 6 at DL1 double dose, and 40 at DL2).

In the DLBCL cohort (full and core dataset), the median age was about 60 years (range, 20-82), with 36% ≥65 years. Twenty percent in the full dataset and 24% in the core dataset had double-/ triple-hit cytogenetics. In the core population, 76% had de novo DLBCL and 25% had transformed from FL. In the full dataset, patients had a median of 3, and up to 12, prior lines of therapy. Two-thirds of patients were de ned as chemorefractory, and 51% had never achieved CR to prior therapy. Forty-three percent in the full dataset had prior SCT, most of which were autologous.

At all dose levels, the ORR was 74% (95% CI, 63%-83%), with a CR rate of 52% (95% CI, 41%-63%). Within the core DLBCL population, patients treated at DL2 had an 81% ORR and 63% achieved CR compared with 76% (95% CI, 59%-89%) and 47% (95% CI, 30%-65%), respectively, at DL1. For the entire core DLBCL population, the median duration of response was 9.2 months, “but has not been reached in the patients who achieved CR,” Abramson said.

Responses at 3 months occurred across all high-risk groups, including double- and triple-hit disease, patients who had never achieved CR previously, and patients with chemorefractory disease.

“Our optimized commercial-ready product is being turned around from apheresis to product release in less than 21 days,” he said.

ONLY 1 CASE OF SEVERE CRS

There were few grade 3/4 toxicities. The rate of any-grade CRS in the full cohort was 35%, with only 1 case of severe CRS among the 91 evaluable patients. Neurotoxicity was observed in 19% of patients; 12% had grade 3/4. There was no increase in either CRS or neurotoxicity between DL1 and DL2, “validating the selection of this dose level for our pivotal population,” said Abramson.

No grade 5 events of CRS or neurotoxicity were observed; however, there were 2 grade 5 treatment-emergent adverse events, one of septic shock, which wasn’t related to treatment, and the other was diffuse alveolar damage.

The median time to CRS was 5 days, and the median time of onset of neurotoxicity was 10 days. “The low overall incidence of these toxicities, as well as the timing of onset, we think supports exploration of outpatient administration of this product,” he said.

At the time of the data cut-off, 8 patients have been treated as outpatients. One patient remained an outpatient throughout the duration of the dose-limiting toxicity period. The other 7 were admitted with fever, and in 6 of them, the fever was attributed to grade 1/2 CRS. Only 1 patient with CRS required treatment with tocilizumab (Actemra). When liso-cel was given on an outpatient basis, the mean number of hospitalization days was 9.3 compared with 15.6 days when administered on an inpatient basis.

Baseline (pre-lymphodepletion) characteristics that appeared to increase the risks of CRS and neurotoxicity in the core population were lactate dehydrogenase level >500 U/L and a high tumor burden (sum of the products of diameters ≥50 cm²),²said Siddiqi, a hematologist/oncologist at City of Hope National Medical Center. Among patients with both of these characteristics, the rate of CRS was 77% compared with 19% among those with neither, and the rates of neurotoxicity were 54% and 11%, respectively.

Increased CAR T-cell expansion was observed with DL2 compared with DL1, which corresponded to better response without increased toxicity. There was a trend toward greater CAR T-cell expansion with higher baseline tumor burden and level of in ammatory cytokine levels, she said. Similarly, baseline levels of the cytokines interleukin (IL)-7, IL-15, macrophage in ammatory protein-1-alpha, and tumor necrosis factor-alpha were higher in those with greater T-cell expansion.

“Not just baseline tumor burden but also [a] higher in ammatory state at baseline seems to affect patients in terms of getting any-grade CRS or any-grade neurotoxicity,” said Siddiqi.

Preliminary modeling data suggested an ideal therapeutic window that may exist for CAR T-cell expansion in which toxicity is limited but ef cacy is optimized. If patients at risk for lower or higher CAR T-cell expansion could be identified, perhaps strategies could be created to drive patients into the optimal window of intermediate T-cell expansion.

References:

1. Abramson JS, Siddiqi T, Palomba ML, et al. High durable CR rates and preliminary safety pro le for JCAR017 in R/R aggressive b-NHL (TRANSCEND NHL 001 Study): A de ned composition CD19-di- rected CAR T-cell product with potential for outpatient administration.J Clin Oncol.2018;36(suppl 5S; abstr 120). meetinglibrary.asco.org/record/156443/abstract.
2. Siddiqi T, Abramson JS, Palomba ML, et al. Correlation of patient characteristics and biomarkers with clinical outcomes of JCAR017 in R/R aggressive B-NHL (TRANSCEND NHL 001 study).J Clin Oncol.2018;36(suppl 5S; abstr 122). meetinglibrary.asco.org/record/156501/abstract.