Cross-Cohort Correlative Analysis of Clinically High-Risk Patients (pts) Treated with Idecabtagene Vicleucel (ide-cel) in KarMMa-2: Association between Progression-Free Survival (PFS) and Tumor Burden and Immune Status at Apheresis

Introduction

Despite improvements in outcomes with triplet and quadruplet combination regimens in front-line (1L) multiple myeloma (MM) treatment (Tx), pts with disease that relapses early after 1L Tx or responds inadequately to 1L autologous stem cell transplantation (ASCT) have poor prognosis. Therefore, Txs with novel mechanisms of action are needed. Ide-cel, a BCMA-directed CAR T cell therapy, resulted in deep and durable responses in pts with functional high-risk MM. KarMMa-2 (NCT03601078) is a phase 2, multicenter, multicohort trial of ide-cel: cohort 2a includes pts who progressed ≤ 18 mo of starting 1L therapy with ASCT, cohort 2b includes pts who progressed ≤ 18 mo of starting 1L therapy without ASCT, and cohort 2c includes pts who achieved less than very good partial response ~100 d after 1L ASCT. With a median follow-up of 21.5, 30.1, and 39.4 mo, respectively, median PFS (95% CI) for cohorts 2a, 2b, and 2c was 11.4 mo (5.6–19.6), not reached (NR; 12.2–NR), and NR (38.0–NR). Previous correlative analyses of ide-cel in relapsed and refractory MM have shown that pts with lower pre-Tx tumor burden and lower pre-Tx indicators of basal inflammation had better responses. This cross-cohort analysis evaluated tumor burden, baseline inflammatory status, and T-cell quality differences to develop hypotheses for biologic features underpinning robust PFS previously observed with ide-cel.

Methods

Levels of soluble BCMA (sBCMA) and 25 exploratory cytokines at infusion and post-Tx were measured by immunoassay. Apheresis and drug product material were characterized by high-parameter flow cytometry and immunoassay. Pharmacokinetics (PK) were evaluated by droplet digital polymerase chain reaction. These analyses were exploratory and post hoc in nature; statistical tests were utilized to identify relationships of interest to develop working hypotheses of the underlying biology.

Results

Baseline tumor burden, evaluated by median sBCMA (interquartile range [IQR]), was highest in cohort 2b at 156 ng/mL (83–300) compared with 2a at 91 ng/mL (41–169) and 2c at 27 ng/mL (19–48). Pts in cohort 2b had the lowest baseline levels of a subset of inflammatory cytokines, including interferon [IFN]-γ and interleukin-4. Phenotypic data from apheresis material showed that, among the 3 cohorts, pts in cohort 2b also had the lowest amount of effector memory T cells positive for CD45RA (T_EMRA; terminally differentiated T cells) and CD57+ T cells (senescent T cells), and the highest amount of CCR7+CD27+ T cells (early memory T cells). Cohorts 2a and 2c had similar phenotypic profiles in the apheresis material. Median time from ASCT to infusion (IQR) was 12.1 mo (9.7–14.3) in cohort 2a and 5.8 mo (5.2–6.2) in cohort 2c. Pts in cohort 2b also had a higher CD4:CD8 ratio in the apheresis material vs the other 2 cohorts. Profiling the drug products from pts in cohort 2b showed higher CAR+ T cell frequency and CAR-specific IFN-γ, and tumor necrosis factor-α and granzyme secretion upon stimulation vs 2a and 2c, with a consistent and comparable potency profile among all cohorts. While sBCMA level at infusion was lower in cohort 2c vs 2a, the similarity in inflammatory state and T-cell quality may reflect impacts from recent melphalan use before ASCT and bone marrow recovery after ASCT, as most pts in both cohorts were < 18 mo from ASCT. Robust cell expansion was observed across cohorts.

Conclusions

The encouraging median PFS outcome previously observed in cohort 2b relative to existing therapies in this high-risk pt population may be driven by the higher T-cell quality and better inflammatory state pre-Tx despite high baseline tumor burden (sBCMA). Pts in cohorts 2c and 2a had comparatively less optimal baseline immune status vs cohort 2b. The promising PFS observed in cohort 2c may be reflective of sufficient T-cell quality to drive optimal responses in the context of relatively low tumor burden at study entry. The median PFS in cohort 2a may reflect the combination of high tumor burden and less optimal baseline inflammatory status. These observations suggest a working model in which ide-cel may demonstrate robust PFS in pts with early-line, high-risk MM and high tumor burden with good baseline immune status and T-cell quality, and in pts with less favorable baseline immune status and T-cell quality when tumor burden is relatively low and stable.