Sequencing bispecific antibodies and CAR T cells for FL

Visual Abstract

Abstract

Treatment for relapsed/refractory (R/R) follicular lymphoma (FL) has evolved over recent years with the introduction of multiple novel immunotherapies: anti-CD3  ×  CD20 bispecific antibody (BsAb) T-cell engagers and anti-CD19 chimeric antigen receptor T cells (CAR T). Both drug classes are highly active, and their adverse event profiles overlap considerably, with cytokine release syndrome, cytopenias, and infections being most common. However, key differences include accessibility and logistical considerations as well as distinct neurologic toxicities, which make recommending a BsAb or CAR T a nuanced decision for each patient with R/R FL. Notably, patients could receive both classes of therapies in sequence; however, data guiding this decision are sparse. Considering the 3 most advanced agents in each class, we generally favor BsAbs before CAR T as the standard-of-care third-line treatment for the typical patient with R/R FL without concern for aggressive histologic transformation (HT). This is based on a 3-year follow-up of the mosunetuzumab phase 2 trial in R/R FL highlighting durable complete responses after a time-limited therapy with an acceptable safety profile for patients of all ages and reasonable performance status. We generally prioritize CAR T before BsAbs for patients with proven or suspected HT given the curative-potential of this approach based on trial data from R/R diffuse large B-cell lymphoma; it is unknown whether BsAbs offer the same long-term benefit in transformed FL. Overall, with the ability to personalize the sequencing of BsAbs and CAR T, the recently expanding portfolio of highly effective immunotherapies for R/R FL is poised to offer considerable benefit to this patient population.

Learning Objectives

• Compare the efficacy of anti-CD3  ×  CD20 bispecific antibodies and anti-CD19 autologous CAR T cells in R/R follicular lymphoma

• Differentiate the safety of anti-CD3  ×  CD20 bispecific antibodies and anti-CD19 autologous CAR T cells in R/R follicular lymphoma

• Develop a framework for choosing between a bispecific antibody and anti-CD19 CAR T cell in R/R follicular lymphoma

Introduction

While follicular lymphoma (FL) is typically highly responsive to frontline therapy, the usual FL disease course after recurrence entails progressively shorter responses to therapy, with a risk of aggressive histologic transformation (HT) impacting prognosis and management.^1,2 Until recently, the treatment landscape for relapsed/refractory (R/R) FL included predominantly traditional chemoimmunotherapy regimens, lenalidomide, and tazemetostat. However, the approval of 2 promising immunotherapy classes (Table 1), bispecific antibody (BsAb) T-cell engagers and chimeric antigen receptor (CAR) T cells, meaningfully expands the options in this setting. Both classes exploit well- established B-cell lineage markers to target autologous T cells to lymphoma cells. Multiple agents in each class have shown impressive efficacy and reasonable safety profiles and have received accelerated approval from the US Food and Drug Administration for treatment of R/R FL after 2 or more prior therapies (Tables 2 and 3).^3-7 Importantly, confirmatory phase 3 trials are exploring these agents in either the first or second line, potentially affecting choices in later lines of therapy. We discuss the data for each class of agents and highlight key considerations for recommending and sequencing therapy for patients with R/R FL (Figure 1).

Table 1.

BsAb vs CAR-T design

Class	Agent	Structure	Target(s)	Clones
Bispecific antibody	Mosunetuzumab	IgG1	CD3  ×  CD20	Anti-CD3: UCHT1 (CD3ε) Anti-CD20: 2H7 (epitope shared w/rituximab)
Bispecific antibody	Epcoritamab	IgG1	CD3  ×  CD20	Anti-CD3: SP34-der (CD3ε) Anti-CD20: 7D8 (epitope shared w/ofatumumab)
Bispecific antibody	Odronextamab	IgG4	CD3  ×  CD20	Anti-CD3: REG1250 (CD3δε) Anti-CD20: 3B9-10 (epitope shared w/ofatumumab)
CAR T cell	Axicabtagene-ciloleucel	CD28 hinge CD28 transmembrane CD28 costimulatory CD3ζ signaling	CD19	Anti-CD19: FMC63 (murine scFv)
CAR T cell	Tisagenlecleucel	CD8α hinge CD8α transmembrane 4-1BB costimulatory CD3ζ signaling	CD19	Anti-CD19: FMC63 (murine scFv)
CAR T cell	Lisocabtagene-maraleucel	IgG4 hinge CD28 transmembrane 4-1BB costimulatory CD3ζ signaling	CD19	Anti-CD19: FMC63 (murine scFv)

scFv, single-chain variable fragments.

Table 2.

BsAb vs CAR-T efficacy

Class	Agent	Trial	N	Demographics		Follow-up (months)	ORR (%)	CRR (%)	Survival outcomes	Duration of response
				POD24 (%)	Refractory to prior therapy
Bispecific antibody	Mosunetuzumab	NCT02500407 (GO29781)	90	52	69	37	78	60	36-mo PFS: 43%	30-mo DOCR: 72%
Bispecific antibody	Epcoritamab	NCT03625037 (EPCORE NHL-1)	128a	42	69	17	82	63	Median PFS: 15.4 mo	18-mo DOCR: 72%
Bispecific antibody	Odronextamab	NCT02290951 (ELM-1)	128b	49	72	27	81	73	24-mo PFS: 45%	24-mo DOCR: 48%
CAR T cell	Axicabtagene-ciloleucel	NCT03105336 (ZUMA-5)	127	55	69	42	94	79	36-mo PFS: 54%	36-mo DOCR: 62%
CAR T cell	Tisagenlecleucel	NCT03568461 (ELARA)	97	63	78	29	86	68	24-mo PFS: 57%	24-mo DOCR: 78%
CAR T cell	Lisocabtagene-maraleucel	NCT04245839 (TRANSCEND FL)	107c	43	38	19	97	94	12-mo PFS: 81%	12-mo DOCR: 82%

^a EPCORE NHL-1 FL study: N  =  128 for efficacy and safety except for CRS and ICANS (N  =  86 optimization cohort).

^b ELM-2 study: N  =  128 for efficacy and safety except for CRS and ICANS (N  =  60 optimization cohort).

^c TRANSCEND FL study: third-line or beyond FL for efficacy (N  =  107), second-line or beyond FL for safety (N  =  130).

CRR, complete response rate.

Table 3.

BsAb vs CAR-T toxicity

Agent	CRS rate		Neurotoxicity rate		Treatment discontinuation due to AEs	Grade 3+ infections	Tocilizumab required
	Any grade	Grade 3+	Any grade	Grade 3+
Mosunetuzumab	44%	2%	5%	2%	4%	13%	8%
Epcoritamaba	49%	0%	0%	0%	19%	NR	12%
Odronextamabb	57%	1%	1%	0%	16%	41%	17%
Axicabtagene-ciloleucel	78%	6%	56%	15%	NA	15%	50%
Tisagenlecleucel	49%	0%	4%	1%	NA	9%	34%
Lisocabtagene-maraleucelc	58%	1%	15%	2%	NA	5%	25%

^a EPCORE NHL-1 FL study: N  =  128 for efficacy and safety except for CRS and ICANS (N  =  86 optimization cohort).

^b ELM-2 study: N  =  128 for efficacy and safety except for CRS and ICANS (N  =  60 optimization cohort).

^c TRANSCEND FL study: third-line or beyond FL for efficacy (N  =  107), second-line or beyond FL for safety (N  =  130).

NA, not applicable; NR, not reported.

Figure 1.

Pros and cons of bispecific antibodies vs CAR T cells for R/R FL. Made with BioRender.

CLINICAL CASE

At 45 years of age, our patient was originally diagnosed with stage IVA, low-grade FL with extensive adenopathy and pleural effusions. Frontline therapy was 6 cycles of rituximab plus bendamustine, and he achieved a complete response (CR). His disease progressed within 24 months of initial therapy (POD24),⁸ with biopsy-proven FL grade 3A and diffuse adenopathy. He received second-line ibrutinib on a clinical trial with a partial response after cycle 5 and progressive disease after cycle 11, followed by third-line rituximab plus 4-1BB agonist on a clinical trial with no response. He received fourth-line rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) and achieved a CR, followed by 2 years of maintenance rituximab. Four years later (9 years after initial diagnosis), his disease relapsed with diffuse lymphadenopathy, pleural effusions, and osseous involvement. A restaging positron emission tomographic scan showed standardized uptake values of 7 to 13, and 2 nodal biopsies confirmed FL grade 1 to 2. He is now inquiring about fifth-line therapy at 54 years of age.

Bispecific antibodies for relapsed FL

Bispecific antibodies (BsAbs) are engineered immunoglobulin (typically IgG)-based agents that recognize 2 distinct targets aiming to induce T-cell redirection. The first-in-class agent, mosunetuzumab, is an anti-CD3  ×  CD20 BsAb sharing 1 Fab with the anti-CD20 monoclonal antibody (mAb) rituximab to accomplish FL targeting while the second Fab achieves T-cell redirection via replacement of 1 heavy and light chain to recognize CD3. Beyond their Fabs targeting 2 different antigens, BsAb T-cell engagers also differ from mAbs in their Fc region, which is disabled to prevent the counterproductive killing of redirected T cells. Further, while experimental models show that redirected CD3+ T cells can directly kill CD20+ tumor B cells via cross-linking by mosunetuzumab,⁹ the true underlying immunology is complex, considering differential BsAb-induced tumor killing by circulating vs tumor-resident T cells, the ability of BsAbs to trigger both CD4+ and CD8+ T-cell proliferation and cytotoxicity, and the potentially immunosuppressive tumor microenvironment.^10-14

Early BsAb clinical trials revealed remarkable efficacy, with improved safety profiles, after the introduction of step-up dosing, as well as guidance on the earlier use of corticosteroids and tocilizumab for cytokine release syndrome (CRS). The original mosunetuzumab phase 1 study enrolled patients with CD20+ R/R B-cell non-Hodgkin's lymphoma (NHL).¹⁵ All patients received fixed-duration treatment: 8 cycles if achieving CR or 17 cycles for PR or stable disease after cycle 8. Step-up dosing in cycle 1 shifted CRS from predominantly grade 2 to grade 1. Histology-specific expansion cohorts led to the selection of a recommended phase 2 dose of 1/2/60  mg intravenously (IV; step-up dosing cycle 1, days 1, 8, and 15), 60-mg IV (cycle 2, day 1), and 30-mg IV (cycle 3, day 1 and beyond) for further study in FL.

The pivotal phase 2 study of fixed-duration, IV mosunetuzumab enrolled 90 patients with R/R FL after 2 or more prior lines of therapy including an anti-CD20 mAb and alkylator.⁵ Corticosteroid pretreatment included dexamethasone at 20  mg before all cycle 1 and 2 doses. Hospitalization was not required. The overall response rate (ORR) was 78% with a CR rate of 60%, which did not appear to be meaningfully impacted by POD24 status or high-risk mutations such as TP53 alterations.¹⁶ The 30-month duration of complete remission (DOCR) was 73%, and the 3-year progression-free survival (PFS) was 43%.¹⁷ CRS of any grade occurred in 44% of patients (grade 3+, 2%), mostly commonly after C1D1 and C1D15 dosing, with 11% and 8% of patients receiving corticosteroids and tocilizumab, respectively. Neurological adverse events (AEs) consistent with immune effector cell–associated neurotoxicity syndrome (ICANS) occurred in 2 patients (all grade 1-2). Grade 3+ infections occurred in 14% of patients. Of the 8 patient deaths, 6 were due to progressive lymphoma, and 2 were due to other causes. Based on these results, fixed-duration, IV mosunetuzumab received approval by the European Commission in June 2022 and accelerated approval by the US Food and Drug Administration (FDA) in December 2022 for the treatment of patients with R/R FL after 2 or more prior lines of therapy. We await additional data on the subcutaneous (SC) dosing of mosunetuzumab, which will hopefully enhance accessibility and ease of administration while further decreasing CRS severity.

Two other anti-CD3  ×  CD20 BsAb agents being studied in R/R FL include epcoritamab and odronextamab. In contrast to fixed-duration mosunetuzumab, both of these agents are administered until progression or unacceptable toxicity. Epcoritamab has been studied in R/R FL in 28-day cycles, administered SC weekly for 3 step-up doses during cycle 1, followed by full doses weekly during cycles 2 and 3, every 2 weeks during cycles 4 through 9, and every 4 weeks starting with cycle 10. Of the 128 patients treated at the 48-mg target dose,⁶ the ORR was 82% with a CR rate of 63%. Median DOCR was not reached after a median follow-up of 17.4 months, with a median PFS of 15.4 months. In the pivotal cohort, CRS of any grade occurred in 65% of patients (grade 3+, 2%), with 24% requiring tocilizumab. ICANS occurred in 6% of patients, with the original CRS mitigation of prednisone at 100  mg/d for 4 days with each cycle 1 dose. No ICANS was seen in a subsequent 86-patient optimization cohort utilizing dexamethasone at 15  mg before each cycle 1 dose. Treatment discontinuation due to AEs occurred in 24 (19%) of patients, while 10% of patients suffered fatal AEs, half due to COVID-19 infection. Given this efficacy and safety profile, epcoritamab was granted accelerated approval by the FDA in June 2024 and approval by the European Commission in August 2024 for treating patients with R/R FL after 2 or more prior lines of therapy.

The phase 2 trial of odronextamab enrolled 128 patients after 2 or more prior lines of therapy, including an anti-CD20 mAb and alkylator.^18-20 Odronextamab was administered on 21-day cycles, with step-up doses during cycle 1, intermediate doses during cycles 2 to 4, and full doses from cycle 5 and beyond. Due to unacceptable CRS rates, odronextamab underwent further dose optimization during the phase 2 study. The final schedule included IV administration with split dosing during cycle 1 step-up (0.7  mg divided over days 1 and 2, 4  mg divided over days 8 and 9, 20  mg divided over days 15 and 16), followed by 80  mg on days 1, 8, and 15 of cycles 2 to 4 and 160  mg every 2 weeks thereafter.¹⁹ The ORR was 81%, with a CR rate of 73%. The median DOCR was 23.7 months, and the 2-year PFS was 45%. CRS mitigation entailed dexamethasone at 20  mg on the day before, day of, and day after all cycle 1 and 2 treatments (21 total doses of dexamethasone). Of patients receiving the final 0.7-, 4-, and 20-mg split-dosing schema, 57% experienced CRS of any grade (grade 3+, 2%), with 17% receiving tocilizumab. One patient experienced low-grade ICANS. Grade 3+ infections occurred in 41% of patients. In total, 15% of patients suffered AEs leading to death, including 8 deaths due to COVID-19 infection, 4 deaths due to non–COVID-19 pneumonia, 1 death due to progressive multifocal leukoencephalopathy (PML), and 1 death due to Escherichia sepsis. As of August 2024, the European Commission has approved odronextamab to treat adult patients with R/R FL after 2 or more lines of systemic therapy, while evaluation by the FDA and data from odronextamab SC dosing cohorts are pending.

Overall, data from phase 2 studies of mosunetuzumab, epcoritamab, and odronextamab in R/R FL after 2 or more prior lines of therapy revealed not only high response rates with frequent and durable CRs but also how each agent needed optimization to account for its CRS risk profile. Mosunetuzumab, epcoritamab, and odronextamab are converging on similar ORRs and CR rates (Table 2) and relatively comparable CRS profiles (Table 3) after dosing and premedication optimization. Open questions focus on the risk/benefit balance of each schedule. The marked increased frequency of administration of epcoritamab and odronextamab compared to mosunetuzumab leads to significant differences in patient time toxicity, defined as time spent by the patient coordinating care, travel, wait, and treatment times at the health care facility, hospital, urgent/emergent facility for side effects, and follow-up visits.²¹ Other considerations include side effects from premedications (eg, up to approximately 5  ×  total prednisone-equivalent exposure with odronextamab vs mosunetuzumab), and long-term toxicities of B-cell depletion from indefinite vs time-limited treatment, potentially resulting in a significant increase in infection risk. Answers to these questions are also evolving as we learn more about BsAb retreatment, which appears reasonably effective in the small cohorts described to date.²²

CAR T for relapsed FL

To date, 3 autologous anti-CD19 CAR-T products have received FDA accelerated approval for the treatment of R/R FL in the third-line setting or beyond based on phase 2 studies. In the ZUMA-5 trial, axicabtagene-ciloleucel (axi-cel) was studied in patients with R/R FL after 2 or more prior lines of therapy including an anti-CD20 mAb and alkylator.^4,23 For the 127-patient cohort, median time from leukapheresis to axi-cel delivery was 17 days. The ORR was 94% with a CR rate of 79% among 124 patients receiving axi-cel, with no ORR difference by POD24 status but an 11% lower CR rate in patients having experienced POD24. The 36-month DOCR was 62%, and the median PFS was 40.2 months with a 36-month PFS of 54%. Grade 3-plus cytopenias were present on or after day 30 in 33% of patients, most commonly neutropenia (27%). CRS of any grade occurred in 78% of patients (grade 3+, 6%), requiring tocilizumab in 50%, corticosteroids in 18%, and vasopressors in 5%. Neurological AEs occurred in 56% (grade 3+, 15%) of patients. Grade 3-plus infections occurred in 15% of patients, while 27% required Ig replacement. Of 32 reported deaths after axi-cel infusion, 10 were due to progressive FL, 10 were due to infection, and 5 were due to other malignancies, among other causes. Eleven patients who relapsed after responding to axi-cel were retreated, and all patients responded to axi-cel retreatment.

The ELARA trial explored tisagenlecleucel (tisa-cel) in adults with R/R FL after 2 or more prior lines of therapy.^3,24 In this 98-patient cohort, the median time from manufacturing start to product release was 24 days, and the median time from enrollment to infusion was 46 days. The ORR was 86% with a CR rate of 69%. Notably, the CR rate was 59% vs 85% in patients with or without prior POD24, respectively. After a median follow-up of 29 months, 24-month PFS and DOCR were 57% and 66%, respectively. In total, 49% of patients experienced CRS (all grade 1-2), neurological events occurred in 37% (grade 3+, 3%), and ICANS occurred in 4% (grade 3+, 1%). Among 47 patients experiencing CRS, 16 received tocilizumab and 3 received steroids. Prophylactic Ig replacement was administered to 33 patients (34%). Grade 3-plus infections occurred in 9% of patients. All 13 deaths occurred more than 30 days after tisa-cel infusion, and causes included progressive FL in 7 patients, 2 other malignancies, and 1 case each of hemophagocytic lymphohistiocytosis, possible PML, postallogeneic transplantation complications, and pneumonia.

Lisocabtagene maraleucel (liso-cel) was studied in the TRANSCEND FL trial, focused on patients with R/R FL in the third line and beyond. It also enrolled a cohort of high-risk patients with biopsy-confirmed R/R FL receiving liso-cel in second-line therapy (POD24 with frontline treatment ≤6 months from original diagnosis or high tumor burden according to Groupe d'Etude des Lymphomes Folliculaires criteria with prior anti-CD20 mAb and alkylator).⁷ In the third-line setting or beyond, the ORR was 97% with a CR rate of 94%, which were similar in patients with or without prior POD24. After a median follow-up of 18.9 months, the 12-month PFS was 81%. In the 25-patient high-risk second-line population, the ORR and CR rates were both 96%. After a median follow-up of 18.1 months, the 12-month PFS was 91%. Incorporating all 130 safety-evaluable patients, CRS occurred in 58% of patients (grade 3+, 1%), neurologic AEs occurred in 15% of patients (grade 3+, 2%), and grade 3-plus infections occurred in 5% of patients. Twelve deaths occurred after liso-cel infusion, and causes included progressive lymphoma in 4 patients, acute myeloid leukemia in 2 patients, COVID-19 in 2 patients, and 1 case each of macrophage activation syndrome, PML, heart failure, and erythema multiforme.

Clinical case follow-up and our approach to sequencing therapy in R/R FL

Our patient with R/R FL, considering fifth-line therapy at 54 years of age, received single-agent SC mosunetuzumab on-study and achieved a CR after cycle 2, experiencing no CRS or hospitalizations. He continued to work throughout treatment and received the planned 8 cycles of therapy. He is now 2 years post treatment and remains in CR. CAR T, retreatment with mosunetuzumab, or lenalidomide-based therapy could be considered if his disease recurs.

This real-life example reflects our current practice for the management of R/R FL in the third-line setting or beyond, which is shaped by several factors. First, BsAbs are off-the-shelf agents immediately available to patients needing therapy. While ZUMA-5, ELARA, and TRANSCEND FL demonstrated typical leukapheresis-to-infusion durations, this does not include other requirements such as pretesting, insurance approval, and slot allocation for standard-of-care CAR T (Figure 2). Even highly efficient cellular therapy programs report referral-to-infusion times of 2 to 3 months, potentially requiring lead-in or bridging therapy for patients with symptomatic R/R FL.²⁵ Recent prior bendamustine may also influence the sequencing of BsAbs and CAR T. ZUMA-5 outcomes stratified by bendamustine exposure indicate that the potency of axi-cel (and likely other CAR-T products) can diminish drastically: the median PFS was fewer than 8 months in patients treated within 1 year of bendamustine exposure. This may be even more relevant in the future if CAR T is considered as second- line therapy for patients experiencing POD24, where time from bendamustine may be shorter. Preliminary data, albeit retrospective, have not seen this trend with BsAbs despite patients with prior bendamustine exposure having lower median CD3+ cell counts pretreatment.²⁶

Figure 2.

Steps and time to first response assessment. Key phases of care before and after BsAb or CAR T treatment from referral to first response assessment. Note that timing of CAR T could be delayed or prolonged at several points, most commonly due to insurance authorization (eg, single-payer agreement negotiations) and the need for lead-in treatment in highly symptomatic patients. PET/CT, positron emission tomography/computed tomography; PFTs, pulmonary function tests; SW, social work; TTE, transthoracic echocardiogram.

Other important considerations include therapeutic time and toxicity burdens.²¹ Among BsAbs, mosunetuzumab in particular offers a transparent treatment burden and AE profile that is reasonably predictable and likely familiar to patients. CAR T, often billed as “one-and-done,” is unlikely to impose less total time burden on patients compared to mosunetuzumab when considering all required encounters. Importantly, time toxicity for epcoritamab and odronextamab may be substantially higher, as outlined previously, with weekly or twice-weekly visits initially plus potentially indefinite treatment. Additional clinical trial and real-world follow-up are needed to more accurately assess any increased infection risk from indefinite epcoritamab or odronextamab treatment outside the COVID-19 pandemic.

It remains unknown whether CAR T offers curative potential for patients with R/R FL, thus the decision as to where to sequence CAR T must be considered differently than R/R diffuse large B-cell lymphoma (DLBCL), for which CAR T is given with curative intent. The most mature data for CAR T for R/R FL is axi-cel, with a PFS of less than 50% at 4 years.²³ The frequency of grade 3-plus neurologic AEs seen with third-line axi-cel also raises questions regarding the generalizability of this approach to real-world patients with R/R FL, a disease with a median age at initial diagnosis of 65 years. Comparisons of grade 5 toxicities between BsAbs and CAR T (Table 4), as well as among agents within each class, were heavily confounded by the COVID-19 pandemic. All, or nearly all, accruals for trials of axi-cel and tisa-cel in R/R FL were completed prior to the start of the pandemic, while mosunetuzumab accrual was completed 5 months after the start of the pandemic. In contrast, epcoritamab, odronextamab, and liso-cel studies all accrued most patients during the height of the pandemic. Grade 5 AEs from BsAbs were largely driven by COVID-19 infections in trials utilizing indefinite treatment, and CAR-T safety also appears heavily driven by context and product choice. Overall, liso-cel appears to strike the most acceptable balance between efficacy and safety among the approved CAR-T agents for R/R FL; however, a longer follow-up is needed to assess the durability of the remarkable PFS seen at 12 months.

Table 4.

Causes of death on BsAb or CAR-T trials

Agent	N	Median follow-up for assessing deaths (months)	Available descriptions of deaths
Mosunetuzumab	90	18.3	8 deaths: 6 progressive FL and 2 other causes
Epcoritamab	128	17.4	13 deaths: 6 COVID-19 infection, 7 other causes; deaths due to progressive FL not reported
Odronextamab	128	26.6	19 deaths: 8 COVID-19 infection, 4 non-COVID-19 pneumonia, 1 PML, 1 Escherichia sepsis
Axicabtagene-ciloleucel	127	41.7	32 deaths: 10 progressive FL, 10 infectious causes, 5 other malignancies, 7 other causes
Tisagenlecleucel	97	28.9	13 deaths: 7 progressive FL, 1 PML, 1 hemophagocytic lymphohistiocytosis, 4 other causes; all >30 days after infusion
Lisocabtagene-maraleucel	130	17.9	12 deaths postinfusion: 4 progressive FL, 1 macrophage-activation syndrome, 1 PML, 6 other causes

To date, we predominantly have data from BsAb trials in R/R DLBCL to support their activity post-CAR-T relapse, but emerging data exploring the reverse approach may be applicable to FL. One study analyzed patients with R/R DLBCL receiving CAR-T post-BsAb exposure; efficacy and safety were comparable to a matched control cohort of patients receiving CAR T without prior BsAb exposure.²⁷ Preclinical studies are exploring how continuous BsAb dosing impacts the T-cell exhaustion that can be deleterious to future CAR-T production,²⁸ but we lack evidence of associated clinical consequences, and even brief BsAb treatment-free intervals appear to ameliorate BsAb-induced T-cell dysfunction in preclinical models. Additionally, because current BsAbs target CD20 while CAR-T products target CD19, surface antigen loss by itself is fortunately unlikely to be a concerning mechanism of cross-resistance. We look forward to future correlative science to elucidate effector T-cell and tumor-intrinsic factors that may aid in both prognostication post immunotherapy and personalized treatment selection.

For patients with R/R FL and confirmed aggressive HT, decision-making is different given the lower response rates with BsAbs, notably mosunetuzumab, in DLBCL compared to indolent B-NHL.²⁹ While CR rates after CAR T are also lower in R/R DLBCL than R/R FL, axi-cel, tisa-cel, and liso-cel likely offer a higher chance of initial cure to patients with transformed FL, compared to BsAbs.^30,31 Considering many patients with transformed FL have prior anthracycline exposure, CAR T offers a greater upside to this population and is our preference in this setting. We apply this in practice to patients with a high suspicion for aggressive HT in whom a biopsy is infeasible but who have concerning features (eg, very high standardized uptake values on positron emission tomographic scan, elevated lactate dehydrogenase level, rapid disease growth). By extension, given the greater efficacy of epcoritamab over mosunetuzumab in R/R DLBCL the former may be preferred for “aggressive”-behaving FL when choosing between BsAbs.³² Nonetheless, every effort should be made to obtain a biopsy in this situation, including an “invasive procedure” if required.

Conclusions and future directions

The pace of advances in R/R FL management affords the luxury of debating between multiple highly active therapies. Most patients with FL do not require all approved agents during their disease course. That said, we lack head-to-head comparisons between therapies in the multiple R/R setting, and these are unlikely to be forthcoming. Exciting phase 3 studies are exploring BsAbs or CAR T in earlier lines of therapy vs standard options, while phase 2 trials are assessing BsAb monotherapy or combinations with lenalidomide or other novel agents in the frontline setting.^33-37 The incorporation of new agents into earlier lines of therapy continues to influence treatment sequencing in later lines, especially as longer follow-ups of ongoing studies shed light on any curative potential of BsAbs or CAR T for FL.

Conflict-of-interest disclosure

David A. Russler-Germain: research funding: Lymphoma Research Foundation, Institute for Follicular Lymphoma Innovation; consultancy: Regeneron; advisory board: AstraZeneca.

Nancy L. Bartlett: research funding: ADC Therapeutics, Bristol Myers Squibb, Celgene, Gilead/Kite Pharma, Merck, Millenium, Pharmacyclics, F. Hoffmann-La Roche/Genentech, Seattle Genetics; advisory committee: Foresight Diagnostics, Kite, F. Hoffmann-La Roche/Genentech, Seattle Genetics.

Off-label drug use

David A. Russler-Germain: nothing to disclose.

Nancy L. Bartlett: nothing to disclose.