Bridging therapy before axi-cel for lymphoma

In this issue of Blood Advances, Jain et al, on behalf of the US Lymphoma chimeric antigen receptor T-cell (CAR T) Consortium, return to the subject of bridging therapy (BT) for patients with large B-cell lymphoma receiving axicabtagene ciloleucel (axi-cel) as a third or later line of therapy and question the notion that BT itself increases the likelihood of treatment failure.¹ This issue is an important one. Decisions around BT are often difficult and highly individualized based on the patient’s disease and expected time from apheresis to infusion (vein-to-vein time). Furthermore, research questions around BT are likely only to be addressed retrospectively, given that the prospective registration studies for axi-cel have all prohibited BT.² 

Previously, the consortium had identified inferior outcomes for recipients of BT compared with those with no BT in a cohort of 298 patients treated in 2018.³ Since then, a number of other groups have observed similar findings, summarized in a systematic review in 2022.⁴ This has led to a difficult situation for the clinician addressing a patient with rapidly progressive disease, for whom bridging is clearly necessary for disease control during the manufacturing and transportation period, but fraught with concern regarding the impact on early toxicity, cytopenias, and disease response.

In this article, with longer follow-up, Jain et al re-examine this same 2018 cohort but go to additional lengths to account for baseline risk factors by performing both multivariate analysis and propensity score matching, the latter being intended to account for unmeasured variables that might have influenced clinician decision to prescribe BT. Once again, the patients receiving bridging (n = 143 [52%]) had inferior progression-free survival and overall survival (PFS and OS, respectively). However, on propensity score matching (and multivariate analysis), these outcomes were similar between the bridging and no BT groups (PFS: hazard ratio [HR], 1.25; P = .23; OS: HR, 1.39; P = .13). Of note, BT in the form of radiotherapy with or without corticosteroids was not associated with better outcomes than other BT categories. Although toxicity in the form of cytokine release syndrome and severe immune effector cell–associated neurological syndrome were similar in bridging and no BT groups on unadjusted analysis, the rate of intensive care unit admission (42% vs 23%), mean duration of hospital stay (1.5 vs 15.7 days), and use of granulocyte colony stimulating factor (48% vs 33%) were higher in the BT group. Once again, however, this difference no longer appeared significant upon propensity score matching.

For the clinician trying to decide whether and how to use BT, these data could be viewed as reassuring, in that bridging itself does not seem to add significantly to toxicity risk nor detract from efficacy over and above the risk factors inherent to the patient with bulky, fast-growing disease. Conversely, however, our clinician may feel somewhat helpless in knowing that no matter what choice is made regarding bridging, the high-risk features present at apheresis will still influence outcome after CAR T.

As Jain et al point out, this article comes with a few caveats that may be relevant to our clinician treating patients today. First, response to BT itself was not available for analysis. Secondly, at the time this cohort was treated, polatuzumab vedotin was not available in the United States as BT. These are important in the light of a recent UK registry study that identified complete or partial response to bridging to be associated with a 42% reduction in disease progression and death after CAR T.⁵ The rituximab bendamustine polatuzumab regimen was used in 38% of bridged patients in this cohort and was associated with double the response rate compared with other systemic BTs.

It should also be noted that none of these patients were receiving axi-cel as second-line therapy, which is now approved and becoming the most common indication for axi-cel in the United States after the reporting of the ZUMA-7 study, which, of course, did not allow BT either.⁶ It is possible, even likely, that the effect of bridging and responses to bridging will be different in a less heavily treated second-line population.

Finally, the time taken for apheresis, transportation, and manufacture of autologous CAR T products influences decisions around BT. This varies according to location: reported vein-to-vein times are shorter in the United States (26-28 days^1,7) than in Europe (35-41 days^8-10). Although additional manufacturing sites are likely to reduce transport and possibly manufacturing times in Europe, other regions of the world (eg, Australia and Japan) remain relatively distant from manufacturing sites.

In these regions, and in situations in which lymphoma is rapidly progressing, the article by Jain et al is likely to provide reassurance that using BT is not increasing the risk profile even further. When vein-to-vein time and disease kinetics are such that bridging is not necessary, this article suggests that withholding BT is not doing harm.

The authors deserve to be commended for returning to their previous analysis to further examine this issue with the aid of additional follow-up and statistical methods. Prospective, randomized evaluation of different bridging strategies is the next step to answer this question. Until then, we remain grateful for large research consortia and registries in addressing the key challenges of real-world practice.

Conflict-of-interest disclosure: D.P. has served on advisory boards for Kite Gilead and received honoraria, which were paid to the author’s institution for education and research. S.B. declares no competing financial interests.