CAR-T for large B-cell lymphomas: the clock is ticking

In this issue of Blood Advances, Ghobadi et al¹ report that patients with relapsed or refractory large B-cell lymphomas (LBCL) who receive anti-CD19–directed autologous chimeric antigen receptor–modified T cells (CAR-T) as second-line therapy have better outcomes than patients who receive third-line CAR-T.

For some patients with LBCL, the pivotal phase 3 trials ZUMA-7² and TRANSFORM³ changed the second-line standard of care from salvage chemoimmunotherapy with autologous hematopoietic cell transplantation (HCT) to anti-CD19 CAR-T. In ZUMA-7, patients eligible for autologous HCT with primary refractory LBCL or LBCL relapsing within 1 year of frontline chemotherapy were randomized to second-line therapy with either axicabtagene ciloleucel (axi-cel), an autologous anti-CD19 CAR-T, or salvage chemotherapy followed by autologous HCT. Improved progression-free survival (PFS)^2,3 and overall survival (OS)⁴ were observed for patients who received second-line CAR-T. Thus, second-line CAR-T therapy is the new treatment paradigm for this challenging patient population.

In this article, the authors describe a post hoc analysis of third-line therapy outcomes for those patients with relapsed or refractory LBCL after participation in ZUMA-7.² The authors also compare previously reported second-line ZUMA-7 axi-cel² outcomes with third-line CAR-T therapy outcomes in those patients who were CAR-T naïve.

This research addresses 2 important questions for the management of LBCL. First, is there a benefit to earlier-line CAR-T? Second, what is the optimal treatment approach for third-line LBCL relapsing after or refractory to second-line CAR-T? Because of the recent approval of second-line CAR-T, there are limited data describing third-line therapy outcomes in LBCL after CAR-T.

The findings of Ghobadi et al support the principle that earlier CAR-T therapy is likely more effective and may help influence future approaches. Although no formal statistical comparisons are performed, the data suggest that PFS and OS are longer when axi-cel is administered in the second line⁴ rather than in third or later lines of therapy.¹ By comparison, the initial third-line CAR-T approval for LBCL was based on single-arm trials that reported 12-month PFS rates of 40% to 44%.^5-7 Ghobadi et al observe strikingly similar outcomes; the patients who received second-line chemotherapy followed by third-line CAR-T therapy had a 12-month PFS of 41%. Dismal PFS for those patients who received third-line chemotherapy was also noted, as is expected for patients who met the criteria for ZUMA-7 eligibility. Moreover, third-line and later CAR-T therapy appear more effective than the other therapies studied.

There is also correlative evidence suggesting that earlier CAR-T therapy may result in enhanced CAR-T product fitness. Patients who receive fewer lines of therapy have higher proportions of naïve-like T cells (CCR7⁺ CD45RA⁺),^8,9 which correlates with CAR-T expansion.⁹ Moreover, clinical trial data for axi-cel in LBCL with suboptimal response during first-line therapy are encouraging.⁸ 

The reported patients’ third-line and later treatment histories¹ also reflect the rapid rate of innovation in LBCL. CD20/CD3 bispecific antibodies are now approved for relapsed/refractory LBCL and have demonstrated monotherapy activity in patients who relapsed after or are refractory to CAR-T therapy.^10,11 Due to the timing of bispecific antibody approvals, few patients in this study received bispecific antibody therapy after CAR-T. The dearth of bispecific antibody use in this study limits the real-world applicability of these third-line LBCL observations.

The ZUMA-7 trial also required patients to be eligible for autologous HCT, have LBCL relapsed within 12 months of frontline chemoimmunotherapy, and not require bridging therapy. Therefore, the findings of Ghobadi et al may not apply to patients ineligible for ZUMA-7.

Otherwise, the major challenge in interpreting these data is the lack of baseline patient characteristics at the time of third-line or later therapy. Although the ZUMA-7 arms were initially balanced for patient characteristics at second-line therapy, we do not know whether the characteristics of those patients receiving third-line CAR-T vs non–CAR-T therapies were similar at the onset of treatment. Different clinical characteristics could affect outcomes.

The authors also report excellent outcomes for a small group of 8 patients who received third-line retreatment with axi-cel. Most patients with LBCL previously responsive to CAR-T therapy again responded, with an overall response rate for axi-cel retreatment of 75% (6/8). Although these results are interesting, their implication for practice is less clear. Three patients with response proceeded immediately to allogeneic HCT; 3 patients had complete responses only after later lines of therapy followed by autologous or allogeneic HCT. Thus, these noteworthy outcomes may be the result of HCT. Furthermore, it is possible that these outcomes are achieved because these patients have biologically more responsive disease than those with lymphomas refractory to CAR-T. Nevertheless, some of these patients appear to have benefited from disease control with CAR-T before HCT.

Another small, but notable, group received third-line autologous (9/60 [15%]) or allogeneic HCT (1/60) after CAR-T. As the authors observe, the precise efficacy of third-line HCT after second-line CAR-T therapy is unclear because it is unknown how many patients previously received salvage chemotherapy with intent for transplantation. Regardless, for this patient population, third-line HCT was successful in a portion, albeit minority, of patients. A priori identification of those patients with chemotherapy-sensitive LBCL is a challenge.

New third-line treatment strategies after CAR-T therapy are clearly needed based on the outcomes reported by Ghobadi et al. We eagerly await the results of ongoing clinical trials of CAR-T targeting alternative tumor-associated antigens administered after progression of standard-of-care anti-CD19 CAR-T therapy. Efforts are also needed to better define which patients will benefit from autologous or allogeneic HCT after CAR-T. Finally, the optimal sequencing of CAR-T and bispecific antibodies remains a key unanswered question. Regardless, immunotherapy for LBCL clearly has a promising future.

Conflict-of-interest disclosure: E.A.C. reports research funding from AbbVie, Genentech/Roche, and AstraZeneca; advisory board participation with AstraZeneca and BeiGene; and is supported by a Lymphoma Research Foundation Clinical Investigator Career Development Award.