CARs in pole position: ready for the race? Free

In this issue of Blood, Chavez et al¹ present the long-term results of the phase 2, single-arm ZUMA-12 study, which evaluated axicabtagene ciloleucel (axi-cel) as part of first-line therapy in patients with high-risk large B-cell lymphoma (LBCL). The primary analysis of ZUMA-12² (n = 37; 15.9 months’ median follow-up) previously reported a complete response rate of 78% with manageable toxicity. In this updated analysis (47.0 months’ median follow-up), the complete response rate increased to 86% (due to complete response after data cutoff or reclassified after external review) with an objective response rate of 92% and an estimated progression-free survival (PFS) of 75.1% at 3 years. No patients relapsed after 18 months postinfusion, and efficacy outcomes remained consistent across various high-risk characteristics, with no new safety signals emerging.

Despite the generally favorable outcome of LBCL, 30% to 40% of patients experience refractory disease or relapse after first-line chemoimmunotherapy, with some ultimately dying despite salvage therapy. Optimizing first-line therapy in these high-risk patients could obviate the need for salvage therapy and improve long-term outcome. Chimeric antigen receptor (CAR) T-cell therapies are promising candidates to address this unmet need given their success in later treatment lines.

In the ZUMA-12 study, high-risk LBCL was defined as an International Prognostic Index (IPI) score ≥3 or a double- or triple-hit lymphoma, with a positive interim (Deauville score 4-5) after 2 cycles of anti-CD20 antibody plus anthracycline-containing regimen (PET2). Although the prognostic value of IPI is well established, the adverse impact of double- or triple-hit lymphoma and interim PET2 remains controversial. A large international study³ revealed that among patients with double- or triple-hit lymphoma, only those whose MYC gene was rearranged with an immunoglobulin (IG) partner experienced significantly worse PFS and overall survival (OS). In contrast, patients with MYC rearranged to a non-IG partner (accounting for 45% of MYC rearrangements) showed outcomes comparable to those without MYC rearrangement. Also, several studies have evaluated the prognostic value of interim PET2, with contradictory results.^4-6 In these studies, the definition of interim PET positivity was inconsistent due to use of different methods (including visual Deauville score and dynamic semiquantitative analysis ΔSUVmax) and cutoffs. The prognostic value of interim PET2 is even more questionable after excluding patients with progressive disease (40% of the ZUMA-12 patients), who would have been eligible for second-line CAR T cells, and considering that PET2-positive patients may exhibit different prognoses depending on whether they achieve negativity after 2 additional cycles (PET4). In the GAINED study,⁷ nearly half of the PET2-positive patients (defined as having a ΔSUVmax ≤66%) achieved a negative PET4 result. For all these reasons, it is challenging to determine how standard chemoimmunotherapy would have performed in this population compared with the current results.

Despite these potential concerns, the encouraging results of ZUMA-12 prompted the initiation of the ongoing randomized, phase 3, ZUMA-23 trial, comparing frontline axi-cel to standard chemoimmunotherapy in patients with high-risk LBCL. Notably, the definition of high-risk LBCL has been modified in ZUMA-23: interim PET2 positivity is no longer required, double- or triple-hit lymphomas have been excluded from the “high-risk” definition, and the minimal IPI score cutoff has been raised from 3 to 4. The ZUMA-12 and ZUMA-23 trials pose a critical question: is the frontline use of CAR T-cell therapy in high-risk patients (some of whom might achieve cure with standard chemoimmunotherapy) more effective than reserving CAR T cells for second-line treatment in relapsed or refractory patients, thereby reducing financial and logistical burdens? Correlative evidence suggests that earlier CAR T-cell therapy may enhance the fitness of CAR T-cell product. Indeed, patients receiving fewer prior lines of therapy have higher proportions of naive-like T-cells (CCR7⁺ CD45RA⁺), which are associated with improved CAR T-cell expansion and, consequently, greater clinical efficacy.⁸ In ZUMA-7, second-line axi-cel demonstrated superior OS compared with the standard of care in transplant-eligible patients. A post hoc analysis⁹ further compared outcomes of second-line axi-cel with third-line CAR T-cell therapy in CAR T cell–naive patients, revealing better outcomes when therapy was administered in the second line. In the ongoing ZUMA-23 trial, it will be essential to determine whether frontline CAR T-cell therapy improves OS, bearing in mind that patients in the standard arm who experience progression or relapse may receive second-line CAR T-cell therapy.

In addition to ZUMA-23, over 15 phase 3 clinical trials are currently underway with the same goal of optimizing frontline treatment for LBCL, as reviewed by Qualls et al.¹⁰ Many of these trials evaluate combinations of R-CHOP (rituximab, cyclophosphamide, doxorubicin, Oncovin [vincristine], and prednisone) or R-CHOP-like regimens with novel agents (“R-CHOP + X”), such as bispecific antibodies, immunomodulatory drugs, or targeted therapies. Like ZUMA-23, these studies are often restricted to high-risk LBCL patients based on IPI score. Should these ongoing phase 3 trials yield positive results, a key question will emerge: what should be the optimal frontline therapy for LBCL, CAR T-cell therapy or R-CHOP + X? A direct, prospective, head-to-head comparison is unlikely to be conducted, leaving the choice between these strategies dependent on factors beyond efficacy. These include toxicity (particularly nonrelapse mortality), logistical considerations, cost-effectiveness, and impacts on quality of life. Additionally, predictive tools and biomarkers will be essential to guide personalized treatment decisions. Variability in patient responses to cell-based therapies and R-CHOP + X regimens underscores the need for advanced theranostic models to identify patients most likely to benefit from early CAR T-cell intervention. The race to optimize frontline LBCL treatment is on, and time will reveal whether CAR T cells will be part of the future therapeutic armamentarium!

Conflict-of-interest disclosure: The authors declare no competing financial interests.