Development and Use of the Anti-CD19 Chimeric Antigen Receptor T-Cell Therapy Axicabtagene Ciloleucel in Large B-Cell Lymphoma

Abstract

IMPORTANCE

Axicabtagene ciloleucel, an anti-CD19-CD28-CD3ζ chimeric antigen receptor T-cell therapy, was the first US Food and Drug Administration–approved, genetically engineered T-cell therapy for adults with relapsed or refractory large B-cell lymphoma (LBCL) after 2 or more lines of systemic therapy. There has not been a US Food and Drug Administration–approved product for these cancers in more than 4 decades.

OBSERVATIONS

Unlike traditional anticancer therapies, axicabtagene ciloleucel is a patient-specific, live-cell product that has unique requirements for manufacturing, shipping, and storage, as well as for its administration and management of its adverse events. In addition, axicabtagene ciloleucel has demonstrated efficacy in patients with refractory LBCL. This review presents a timeline of the rapid clinical development of axicabtagene ciloleucel from bench to bedside, highlights how axicabtagene ciloleucel satisfies an unmet medical need for treatment of refractory LBCL, outlines the logistics of the production process and administration of axicabtagene ciloleucel, describes its mechanism of action, and summarizes the results of the pivotal study. This review also provides a survey of adverse events, with attention to the kinetics of their clinical presentation; discusses the management of adverse events; and offers suggestions for appropriate patient selection for safe administration of axicabtagene ciloleucel.

CONCLUSIONS AND RELEVANCE

The integration of axicabtagene ciloleucel therapy into standard-of-care practice for relapsed/refractory LBCL is the beginning of a paradigm shift in the treatment of patients with LBCL and is likely to lead to improvements in their survival and curability. Timely referral to centers offering the therapy is necessary for optimal patient outcomes.

Axicabtagene ciloleucel is an autologous, anti-CD19 chimeric antigen receptor (CAR) T-cell therapy with significantly improved efficacy in patients with refractory large B-cell lymphoma (LBCL) compared with other currently available treatment options.¹ Although standard-of-care therapies can cure approximately 60% of patients with newly diagnosed diffuse large B-cell lymphoma (DLBCL),^2,3 outcomes for patients with relapsed or refractory LBCL are poor.^4–8 A large, international, retrospective patient-level analysis of outcomes in refractory LBCL prior to the availability of CAR T-cell therapy, SCHOLAR-1, reported an objective response rate (ORR) of 26%, a complete response (CR) rate of 7%, and a short median overall survival of 6.3 months,¹ highlighting the need for innovative therapeutic approaches.

The antigen CD19 is an optimal target for CAR T-cell therapy given its expression across a wide range of B-cell cancers and its restricted expression in healthy tissues.^9,10 However, targeting CD19 can lead to B-cell aplasia and hypogammaglobulinemia, similar to routinely used CD20-targeting monoclonal antibody therapies, and immunoglobulin supplementation can mitigate the increased risk of infections attributable to B-cell aplasia.¹¹ B-cell recovery has been observed 2 years after receiving anti-CD19 CAR T-cell therapy in patients with ongoing CR.¹²

The cell product that became axicabtagene ciloleucel was collaboratively developed with Kite, a Gilead Company, and the National Cancer Institute (Figure 1A).^13–17 Early work revealed that, although first-generation CAR T cells (ie, CARs with only a single T-cell signaling domain) demonstrated cytotoxic activity, they lacked functional persistence.^18–20 For optimal T-cell activity, antigen recognition must occur in conjunction with appropriate costimulation, such as through CD28.²¹ In 2009, Kochenderfer and colleagues²² described the development of the CAR construct used in axicabtagene ciloleucel, which consists of an extracellular CD19-targeting, single-chain variable fragment, an intracellular CD28 costimulatory domain, and a CD3ζ signaling domain (Figure 1B). This development was quickly followed in 2010 by, to our knowledge, the first reported successful use of this CAR in a patient with B-cell lymphoma.¹³

Figure 1. Axicabtagene Ciloleucel Developmental Timeline and Structure.

Developmental timeline (A) with structure of axicabtagene ciloleucel and antigen response (B). AACR indicates American Association of Cancer Research; CAR, chimeric antigen receptor; CRADA, cooperative research and development agreement; LTR, long terminal repeat; NCI, National Cancer Institute; scFv, single-chain variable fragment; ZUMA, long-term safety and activity of axicabtagene ciloleucel in refractory large B-cell lymphoma.

^aFirst clinical data reported by Kochenderfer et al.¹³

^bAnalysis reported by Locke et al.¹⁴

^cAnalysis reported by Locke et al.¹⁵

^dFollow-up reported by Neelapu et al.¹⁶

^eFollow-up reported by Locke et al.¹⁷

Additional studies from the National Cancer Institute using the CD19-CD28-CD3ζ CAR demonstrated the importance of conditioning therapy with cyclophosphamide and fludarabine before CAR T-cell infusion. The conditioning regimen was systematically investigated and optimized in 14 different conditioning dose cohorts in early phase trials.^23,24 Conditioning therapy likely induces depletion of healthy lymphocytes, eliminating cytokine sinks and making homeostatic cytokines available for the survival and proliferation of adoptively transferred T cells,²⁵ as well as possibly eradicating immunosuppressive cells, that is, regulatory T cells and myeloid-derived suppressor cells, leading to improved CAR T-cell function.²⁵ Subsequent studies have confirmed the importance of the homeostatic cytokine environment after conditioning, notably that increased serum IL-15 promotes response to CAR T-cell therapy.^12,26,27

The high response rates observed in these early studies led to further development of this CAR construct as axicabtagene ciloleucel, which uses a closed manufacturing process. ZUMA-1, the pivotal, multicenter, phase 1/2 trial, investigated axicabtagene ciloleucel in refractory LBCL.^16,17 Axicabtagene ciloleucel demonstrated unprecedented efficacy in ZUMA-1 and was approved by the US Food and Drug Administration (FDA) in October 2017 for the treatment of adult patients with relapsed or refractory LBCL after 2 or more lines of systemic therapy.²⁸ Axicabtagene ciloleucel was the first product approved by the FDA for relapsed/refractory LBCL (Figure 2A).

Figure 2. Axicabtagene Ciloleucel Prescribing Information and Delivery.

A, Indication, limitations of use, and warning from the prescribing information.²⁸ B, Axicabtagene ciloleucel (axi-cel) process from patient to manufacturer, and back to the patient at 1 °C to 10 °C. CRS indicates cytokine release syndrome; G1, grade 1; G2, grade 2; G3, grade 3.

Manufacturing and Administration

The ZUMA-1 study demonstrated that the centralized manufacturing of axicabtagene ciloleucel was rapid and reliable with coordination of leukapheresis and shipping from multiple centers across the United States (Figure 2B). Axicabtagene ciloleucel manufacturing was initiated by harvesting patient T cells using outpatient leukapheresis of peripheral blood mononuclear cells without mobilization. Following leukapheresis, no bridging chemotherapy was allowed. Fresh apheresis material was transported at 1 °C to 10 °C to the central cell-processing facility where mononuclear cells were enriched with a closed Ficoll gradient (Sepax 2; Biosafe) system and activated with recombinant anti-CD3 monoclonal antibody and IL-2.²⁹ Activated T cells were transduced with a γ-retroviral vector containing the CAR gene and expanded in culture. Once the target dose was achieved (2 × 10⁶ CAR T cells/kg; maximum dose, 2 × 10⁸ CAR T cells), the product was washed and cryopreserved, assessed by quality control testing, and shipped to the treatment site. Before administration, the patient received conditioning chemotherapy: 3 days of cyclophosphamide, 500 mg/m²/d, and fludarabine, 30 mg/m²/d.

In ZUMA-1, this process successfully manufactured axicabtagene ciloleucel for 99% (110/111) of patients enrolled, with 91% (101/111) of patients receiving treatment despite interpatient variability in apheresis material composition and baseline lymphocyte counts, including patients with severe lymphopenia (<300 lymphocytes/μL in the peripheral blood).^16,30,31 The median turnaround time from apheresis to delivery of the final axicabtagene ciloleucel product to the treating center was 17 days. Short turnaround time was necessary for patients with rapidly growing disease and high tumor burden. The manufacturing process for commercial axicabtagene ciloleucel product is the same Good Manufacturing Practice–adherent process used for ZUMA-1.

Efficacy

In ZUMA-1, axicabtagene ciloleucel demonstrated durable efficacy in patients with refractory LBCL (Table).^{16,17,24,32–34} Eligible patients must have had progressive or stable disease as the best response to their last line of systemic therapy or have relapsed within 12 months from high-dose chemotherapy and autologous stem-cell transplantation. Phase 1 results in 7 patients with DLBCL (ORR, 71% with 4 of 7 [57%] having a CR) led to the pivotal, phase 2 portion of the trial.^14,15 The expanded phase 2 trial assessed patients in 2 cohorts: 77 patients with DLBCL (cohort 1) and 24 patients with transformed follicular lymphoma or primary mediastinal B-cell lymphoma (cohort 2).¹⁵ Cancer in these patients was highly refractory, with 26% (26/101) having a history of primary refractory disease and 53% (54/101) having a history of resistance to 2 or more consecutive lines of previous therapy.

Table.

Anti–CD-19 CAR T Cells in Clinical Trials

Study	NCI24	ZUMA-116,17	JULIET32	TRANSCEND33
ClinicalTrials.gov identifier	NCT00924326	NCT02348216	NCT02445248	NCT02631044
CAR T investigated	Anti-CD19 CAR T	Axicabtagene ciloleucel	Tisagenlecleucel	Lisocabtagene maraleucel
Patient population	15 Enrolled; 15 treated; DLBCL (n = 9); CLL (n = 4); indolent lymphoma (n = 2)	111 Enrolled; 108 treated; DLBCL (n = 84); PMBCL/TFL (n = 24)	165 Enrolled; 111 treated; 79% DLBCL; 19% TFL	134 Leukapheresed; 114 treateda; 102 treated and evaluable; DLBCL (n = 63); TFL (n = 23); TMCL/TCLL (n = 12); FL/PMBCL (n = 4)
Bridging therapy allowed	No	No	Yes	Yes
Length of follow-up	Not reported	Median, 27.1 mo	Median, 14 mo	Median, 12 mo (8 mo for duration of response)
Overall population ORR/CR, %	80/53	83/58	52/40	75/55b
Relapsed/refractory ORR	Chemorefractoryc: 63%; relapse after ASCT:100%	Refractory to ≥ second-line: 83%; relapse <12 mo after ASCT: 76%	Refractory to last line: 55%; relapsed to last line: 45%; prior transplant: 49%	Chemorefractoryc: 41%; relapse <12 mo after ASCT: 53%
Median duration of response, mo	Not reported	11.1	NR	NR
Range	1–23 mo	Not reported	Not reported	Not reported
95% CI	Not reported	4.2-NE	10.0-NR	5-NR
Median PFS, mo (95% CI)	Not reported	5.9(3.3–15.0)	2.9 (2.2–4.2)34	Not reported
Median OS	Not reported	NR (12.8-NE)	12 (7-NR)	NR (10.4-NR)
Cytokine release syndromed	Not reported	11% Grade ≥3	22% Grade 3–4	1% Grade 3–4
Neurologic events	40% Grade ≥3	32% Grade ≥3	12% Grade 3–4	13% Grade 3–4
Manufacturing success rate; median turnaround time	100% (15/15); not reported	99% (117/118); 17 de	Not reported; 54 df	99% (including 12 nonconforming); not reported

Abbreviations: ASCT, autologous stem-cell transplant; CAR, chimeric antigen receptor; CLL, chronic lymphocytic leukemia; CR, complete response; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; NCI, National Cancer Institute; NE, not estimable; NR, not reached; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; PMBCL, primary mediastinal B-cell lymphoma; TCLL, transformed chronic lymphocytic leukemia; TFL, transformed follicular lymphoma; TMCL, transformed mantle cell lymphoma.

^aTwelve patients received nonconforming product.

^bResponses in the pivotal, core population, that is, patients with not otherwise specified DLBCL or high-grade B-cell lymphoma.

^cNo achievement of partial response or CR after most recent chemotherapy in patient with DLBCL.

^dCytokine release syndrome (CRS) was graded using the Penn CRS grading scale³⁵ in JULIET and modified Lee criteria³⁶ in ZUMA-1 and TRANSCEND.

^eTime from leukapheresis to delivery at infusion center.

^fTime from enrollment to infusion.

Among the patients in the phase 2 trial who received axicabtagene ciloleucel, the best ORR was 83% and the highest CR rate was 58%.¹⁷ With a median follow-up of 27.1 months, 39% (39/101) of the patients remained in response, and 37% (37/101) remained in CR (Figure 3A). The median progression-free survival was 5.9 months, and the median overall survival was not reached (Figure 3B). Responses were dynamic, particularly in the first 3 months following infusion, with the emergence of a plateau in the progression-free survival curve at approximately 6 months. The median time to response was 1 month, and over 50% of progression events occurred by 3 months. The 1-month restaging was valuable to identify early progression necessitating alternative therapy, but 11 of 33 patients (33%) with partial response and 11 of 24 patients (46%) with stable disease at 1 month converted to CR up to 12 months after axicabtagene ciloleucel treatment, suggesting that these patients should be observed closely for deepening responses. Patients in CR or partial response at 3 months after axicabtagene ciloleucel infusion had an approximately 75% likelihood of remaining in response at 2 years without any subsequent intervention.

Figure 3. Efficacy Outcomes for Patients Treated With Axicabtagene Ciloleucel (Axi-Cel).

A, Median duration of response, 11.2 months (95% CI, 4.2-not reached [NR]). B, Median overall survival (OS), NR (95% CI, 12.8-NR); progression-free survival (PFS), 5.9 (95% CI, 3.3-15.0).

In addition, the National Cancer Institute has reported ongoing responses beyond 4 years with the same CD19-CD28-CD3ζ CAR construct as axicabtagene ciloleucel.¹² Thus,these combined results suggest that a single axicabtagene ciloleucel treatment, without consolidation or maintenance, may lead to long-term control of refractory LBCL for patients without curative options.^17,37 Although not vali-dated in the relapsed/refractory setting, analysis of first-line studies suggests that 2-year progression-free survival may be able to estimate the probability of durable remissions beyond 5 years.³⁸

Data from other anti-CD19 CAR T-cell therapies show comparable efficacy in related populations (Table). Tisagenlecleucel (previously called CTL019), targets CD19 using the same single-chain variable fragment as axicabtagene ciloleucel, a similar CD3ζ-signaling domain, but instead contains a 4–1BB costimulatory domain.³⁹ JULIET, the phase 2 study of tisagenlecleucel in relapsed/refractory DLBCL, demonstrated a 52% ORR (40% CR) with a median follow-up of 14 months (range, 0.1–26).³² Lisocabtagene maraleucel (formerly called JCAR017) targets CD19, also using the same single-chain variable fragment as axicabtagene ciloleucel, and contains a 4–1BB costimulatory domain and CD3ζ signaling domain and is manufactured with a defined 1:1 CD4:CD8 T-cell ratio. TRANSCEND NHL 001, the phase 1 study of lisocabtagene maraleucel in patients with relapsed/refractory, aggressive B-cell non-Hodgkin lymphoma, demonstrated a 75% ORR (55% CR) with a median follow-up of 12 months in its pivotal population.³³ Direct comparisons of efficacy results between the pivotal trials are impractical owing to differences in patient populations, trial de-signs, and follow-up periods. Studies coordinated by a multicenter outcome registry, such as the database managed by the Center for International Blood and Marrow Transplant Research,⁴⁰ may provide insights on the relative clinical activity of different CAR T-cell therapies.

Safety

Acute Adverse Events

Chimeric antigen receptor T-cell therapy is associated with serious and potentially life-threatening class-specific acute adverse events (AEs), namely cytokine release syndrome (CRS) and neurologic events.^36,41 ZUMA-1 demonstrated that, with appropriate AE management training of clinicians, axicabtagene ciloleucel can be safely and effectively administered at multiple sites, including many sites without prior CAR T-cell experience. Vigilant monitoring for early AE detection and rapid intervention using preestablished management guidelines can decrease the likelihood of more severe AEs. In ZUMA-1, although nearly all patients developed any grade CRS (94/101 [93%]) most commonly manifesting as pyrexia (77/101 [76%)], severe, grade 3 or higher CRS was observed in only 12 of 108 patients (11%).^16,17 The incidence of grade 3 or higher of CRS decreased from 17% in the first 69 patients to 0% in the final 39 patients treated in ZUMA-1. This improvement was likely owing to revised management guidelines with earlier use of tocilizumab for grade 2 CRS implemented during the latter part of the trial. Of 108 patients, 72 patients (67%) experienced any-grade neurologic events, with 32% having grade 3 or higher events.¹⁷ This symptom complex is also known as CAR-related encephalopathy syndrome or, most recently, immune effector-cell-associated neurotoxicity syndrome.^42,43 Similar to CRS, there was a modest decrease in grade 3 or higher neurologic events over the course of ZUMA-1, from 38% of the first 69 patients to 23% of the final 39 patients.¹⁷ There were no new axicabtagene ciloleucel–related cases of CRS or neurologic events after the primary (6-month) analysis.^16,17 With the exception of those ongoing at the time of death (n = 6), all CRS and neurologic events resolved.¹⁶

For the practicing clinician, understanding the timing of CRS and neurologic events is important to patient management. In ZUMA-1, following axicabtagene ciloleucel infusion on day 0, the median onset of CRS was on day 2 (range, 1–12), and resolution occurred at a median of day 8.¹⁶ The onset of neurologic events was slightly later, occurring at a median of day 5 (range, 1–17) with a median resolution by day 17.¹⁶ In addition, understanding the severity of CRS and neurologic events over time is helpful for understanding the clinical course of these events (Figure 4A, B). Between days 30 and 90, 3 patients were classified as having CRS: 2 with grade 1 events ongoing from before day 30 (fatigue [n = 1] and decreased appetite [n = 1], both of which resolved) and 1 death due to hemophagocytic lymphohistiocytosis. The most common neurologic events between days 30 and 90 were grade 1: encephalopathy, tremor, and memory impairment. The onset of nearly all late neurologic events began in the first 30 days post axicabtagene ciloleucel infusion, and all events resolved except for those ongoing at the time of death due to other causes.

Figure 4. Timeline of Adverse Events.

Timeline of cytokine release syndrome (A) and neurologic events (B) after axicabtagene ciloleucel therapy in ZUMA-1, with exact day of the event shown, and a visual case study illustrating management of concurrent cytokine release syndrome and neurologic events (C). Time-to-onset is calculated as (day of infusion, ie, day 0) + 1. Five patients experienced grade 1 neurologic events that were ongoing at day 90 (1 patient each with encephalopathy, hypoesthesia, tremor, and memory impairment; 1 patient with encephalopathy and leukoencephalopathy), and all events subsequently resolved. Axi-Cel indicates axicabtagene ciloleucel; CAR, chimeric antigen receptor; CRS, cytokine release syndrome; ECG, electrocardiogram; G1, grade 1; G2, grade 2; G3, grade 3; ICU, intensive care unit; IV, intravenous; and MRI, magnetic resonance imaging.

Cytokine release syndrome (64/111 [58%]) and neurologic events (23/111 [21%]) were also observed with tisagenlecleucel in the JULIET study with grade 3 to 4 in 24 patients (22%) and 13 patients (12%) patients, respectively.³² In the TRANSCEND evaluation of lisocabtagene maraleucel, CRS (27/73 [37%]) and neurologic events (18/73 [25%]) occurred, with 1 patient (1%) and 11 patients (15%) experiencing grade 3 or higher events, respectively.³³ The Penn CRS grading scale³⁵ was used in JULIET compared with ZUMA-1 and TRANSCEND, which used modified Lee criteria for CRS, and the trials had different guidelines for AE management.^16,36 A consensus grading system for CRS and neurologic toxic effects was recently published, and there are ongoing efforts to harmonize management of toxic effects.^42,43 As with efficacy data, final results from these trials and larger data sets are needed before adequate comparisons are possible.

Late AEs

Late AEs, occurring beyond 12 months and up to 2 years in ZUMA-1, were similar to those in previous reports.^12,16,17 In ZUMA-1, 52% of all patients had B-cell aplasia and hypogammaglobulinemia at study entry due to prior anti-CD20 antibody therapy, and cumulatively, 31% of patients received intravenous immunoglobulin therapy to mitigate the risk of infections. Twenty-four of 32 patients (75%) with ongoing responses had detectable B cells by 2 years post axicabtagene ciloleucel infusion. The cumulative incidence of severe grade 3 or higher infectious complications after axicabtagene ciloleucel administration was 28%.¹⁷ Between the 12- and 24-month analyses, infections were reported in only 4 patients experiencing 6 infections,and overall, 4 patients developed new serious AEs, but none were related to axicabtagene ciloleucel (grade 3 mental status changes, grade 3 lung infection, grade 3 bacteremia, and grade 4 myelodysplastic syndrome).^17,34 Follow-up for the JULIET and TRANSCEND trials is still limited, but similar long-term AE profiles are anticipated. There were no events of insertional mutagenesis reported with the use of the retroviral or lentiviral constructs in any of the 3 trials^.16,33,44

AE Management

Adverse events associated with axicabtagene ciloleucel therapy, specifically CRS and neurologic events, are generally manageable and reversible with appropriate treatment strategies. Education for multi disciplinary care teams, patients, and caregivers on CRS and neurologic events symptoms is critical to early intervention and positive outcomes.⁴² Thorough baseline assessments enable accurate detection of new-onset symptoms, and practitioners should be familiar with available grading scales and recommended treatments.^28,36,42 A comprehensive review on the assessment and management of CRS and CAR-related encephalopathy syndrome based on multiple clinical trial experiences across different anti-CD19 CAR T-cell therapies provides detailed, practical clinical guidance for practitioners managing patients treated with CAR T-cell therapy, including axicabtagene ciloleucel.⁴² There is currently a general consensus that low-grade CRS and CAR-related encephalopathy syndrome may be managed mostly with supportive care and severe events will additionally need tocilizumab, an anti–IL-6 receptor antagonist, and/or corticosteroids.^28,42 Tocilizumab was recently FDA approved for the treatment of CAR T-cell–induced CRS.⁴⁵ In ZUMA-1, 43 of 101 patients (43%) received tocilizumab and 27 of 101 patients (27%) received corticosteroids for AE management. However, the use of tocilizumab or corticosteroids did not affect the ORR, CR rate, or durability of the responses.¹⁶ The symptoms of CRS and neurologic events can often wax and wane, which can complicate their management. The case study shown in Figure 4C illustrates this concept as well as timely and appropriate escalation of care. Stage 4 refractory DLBCL in a 42-year-old man with an Eastern Cooperative Oncology Group performance score of 0 and an International Prognostic Index of 3 had progressed after 4 previous treatment regimens. This patient experienced concurrent CRS and neurologic events that, with appropriate management, fully resolved by day 12 after axicabtagene ciloleucel infusion.

Correlative Biomarkers and Mechanisms of Resistance

Correlative studies on ZUMA-1 have provided mechanistic insights into response, resistance, and AEs after axicabtagene ciloleucel therapy. Axicabtagene ciloleucel expansion peaks approximately 7 days after infusion followed by a partial contraction within the first month and a more gradual contraction thereafter.⁴⁶ Greater CAR T-cell expansion correlates with durable responses and severe neurologic events, but not grade 3 or higher CRS.^46,47 However, the lack of correlation between CAR T-cell expansion and grade 3 or higher CRS may be confounded by evolving recommendations to intervene at grade 2 CRS as opposed to previous recommendations to intervene at grade 3. In addition, because only 12 of 108 patients (11%) in ZUMA-1 had grade 3 or higher CRS,¹⁷ small patient numbers may further confound correlative analysis. Elevated IL-15 levels following conditioning chemotherapy are associated with higher CART-cell expansion and remission as well as grade 3 or higher neurologic events.^12,46 Early upregulation (day 0 and day 1 post infusion) of some myeloid-associated cytokines, such as granulocyte-macrophage colony-stimulating factor and IL-6, has also been associated with severe neurologic events.⁴⁸

Analyses of tumor samples pretreatment and at progression suggested CD19 loss as one potential mechanism of resistance.¹⁶ Expression of immune checkpoint molecules, such as CTLA-4, LAG-3, and programmed death 1 on CAR T cells and programmed death ligand 1 on tumor cells may also be involved in mediating resistance.^16,49 Additional biomarker studies may elucidate targets for AE prophylaxis while preserving CAR T-cell function and strategies for further enhancing efficacy.

Patient Selection

Axicabtagene ciloleucel represents an important treatment option for patients with relapsed or refractory LBCL. Understanding patient eligibility, and critically, identifying those who may not be appropriate candidates based on baseline factors, is important for selecting patients. In ZUMA-1, patients with previous T-cell therapy; allogeneic stem-cell transplantation; central nervous system disorders; significant cardiac disease; active autoimmune disease; history of HIV, hepatitis B, or hepatitis C infection; or active infection requiring intravenous antimicrobials were excluded. Simple urinary tract infections and uncomplicated bacterial pharyngitis that were responding to treatment did not preclude eligibility. Other than the presence of active infection and inflammatory disorders, the FDA-approved label does not list any of the above conditions as contraindications and practitioners must carefully evaluate the risks of treating patients outside the ZUMA-1 eligibility criteria. Commercially, at H Lee Moffitt Cancer Center & Research Institute and MD Anderson Cancer Center, it has been the practice to select patients for axicabtagene ciloleucel treatment based on ZUMA-1 eligibility criteria with some caveats. For example, thrombocytopenia may be the result of recent chemotherapy, so a platelet count less than 75 × 10³/μL (to convert to ×10⁹ per liter, multiply by 1) but increasing at the time of cell collection might be considered. Similarly, patients with mild cardiac, hepatic, and renal dysfunction; previous allogeneic stem-cell transplantation without active graft-vs-host-disease; and previous T-cell therapy may be considered by experienced practitioners after careful evaluation. In a small study at the National Cancer Institute, treatment of patients with prior allogeneic stem-cell transplantation with CAR T cells generated from the matched donor was found to be safe and did not increase the risk of graft-vs-host-disease.⁵⁰ Patients with active central nervous system lymphoma could benefit from CAR T-cell therapy as evidenced by case reports⁵¹; however, without stronger safety data, we recommend not to treat patients with active central nervous system lymphoma, and these patients should be an important focus of ongoing investigation.

It is necessary to reassess patients for infection and inflammation throughout the treatment journey, specifically before apheresis, conditioning chemotherapy, and axicabtagene ciloleucel infusion. Patients with preexisting active infection and inflammatory conditions are at risk of developing severe CRS and therefore should not be treated with axicabtagene ciloleucel, as cautioned in the prescribing information.^14,28,42 Axicabtagene ciloleucel administration can and should be delayed in patients with new fever or symptoms of active infection to allow diagnostic workup and antimicrobial treatment (Figure 4). In cases in which fever or other significant constitutional symptoms are believed to be due to rapidly progressive disease rather than infections, therapies to reduce the tumor burden and tumor-induced inflammatory state should be considered before axicabtagene ciloleucel infusion, even if the product has already been manufactured. In ZUMA-1, patients with the lowest quartile of tumor burden demonstrated a high ongoing response rate (67%) with minimal grade 3 or higher CRS (4%) and neurologic events (7%); as such, debulking bridging therapy may improve the efficacy and safety profiles in patients with high disease burden.⁵² Considering the risks of the therapy and given current safety data, we do not believe that axicabtagene ciloleucel should be used in patients with major comorbidities or very poor performance status.

Cost of Care and Economic Viability

Chimeric antigen receptor T-cell therapy for DLCBL is a single-use therapy with the potential for long-term disease control. While the product price of CAR T-cell therapy exceeds the 1-time cost of most other oncologic agents, almost all of the other agents require repeated use with diminishing returns for prolonged remission. Additional costs are incurred from CAR T-cell administration, including those associated with patient safety testing, apheresis, product shipping and receiving, conditioning chemotherapy, outpatient visits, inpatient hospitalization, and supportive care. Further hidden costs include the maintenance of the FDA Risk Evaluation and Mitigation Strategies program for commercial therapy, infrastructure necessary for long-term safety reporting, and the maintenance of professional accreditation.

However, CAR T-cell therapy offers value to patients with refractory LBCL who otherwise face a lifespan measured in months.⁵³ Two different post hoc analyses of cost-effectiveness of ZUMA-1 indicated a cost-effective gain in survival compared with salvage chemotherapy,^54,55 although longer follow-up is needed to confirm the cost-effectiveness of this therapy.

There is concern among medical experts and hospital administrators that Medicare & Medicaid rules do not currently allow for full cost recovery following inpatient CAR T-cell therapy administration. This restriction places hospitals at risk as the product cost must be payed upfront and carried over until the payer provides reimbursement. The amount effectively reimbursed for all required services by payers and the economic burden that the therapy might place on patients should be explored in future studies.

Future Directions

CAR T-cell therapy represents the beginning of a paradigm shift in the management of relapsed or refractory LBCL. Axicabtagene ciloleucel, with its ability to provide long-term disease control, is a much-needed therapy in this setting as patients lack curative treatment options. In an ongoing, international randomized ZUMA-7 study,⁵⁶ the efficacy of axicabtagene ciloleucel is being directly compared with the existing standard of care of salvage chemotherapy followed by autologous stem-cell transplantation in patients with DLBCL at first relapse, essentially second-line therapy.⁵⁷ Other registrational studies are evaluating this anti-CD19 CAR construct in indolent B-cell lymphoma,⁵⁸ mantle cell lymphoma,⁵⁹ and acute lymphoblastic leukemias.^60,61 In addition, combination therapy with axicabtagene ciloleucel and immune checkpoint blockade with anti-programmed death ligand 1 antibody is currently being investigated in ZUMA-6 to overcome a potential mechanism of tumor resistance.⁶² In the future, switch technologies may give practitioners the ability to precisely control CAR T-cell levels and activity to improve patient safety.^63,64 In addition, bispecific and multispecific CAR T-cell products targeting multiple tumor antigens are in development.^65,66 It is hoped that these advancements will overcome mechanisms of resistance, improve the safety profile, and continue to enhance the efficacy of CAR T-cell therapy.