Cost-effectiveness of second-line axicabtagene ciloleucel in relapsed refractory diffuse large B-cell lymphoma

Diffuse large B-cell lymphoma (DLBCL) patients with primary refractory disease or early relapse within 12 months of initial chemoimmunotherapy have a poor prognosis.¹ Based on the durable responses seen with chimeric antigen receptor T-cell therapy (CAR-T) in the third-line setting,^2-4 the ZUMA-7 (Efficacy of Axicabtagene Ciloleucel Compared to Standard of Care Therapy in Subjects With Relapsed/Refractory Diffuse Large B Cell Lymphoma) study evaluated the safety and efficacy of CAR-T in the second-line setting for patients with primary refractory/early relapse DLBCL compared with standard of care (SOC) salvage chemoimmunotherapy followed by autologous stem cell transplant (auto-SCT).⁵ It also reported a benefit in event-free survival (EFS) (24-month EFS 41% with axi-cel vs 16% with SOC; P < .001), leading to its recent US Food and Drug Administration approval.

CAR-T is associated with various toxicities such as cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS), as well as financial toxicities.⁶ Although CAR-T has significantly changed the landscape of DLBCL, its high costs pose a significant barrier to access for patients and a financial strain on hospitals. Drug acquisition with prices ranging from $373 000 to $475 000 is the largest component of CAR-T costs. Other costs include procedures, supportive care, and hospitalizations.⁷ A prior study evaluating the cost-effectiveness of CAR-T in the third-line setting for DLBCL found it may be cost-effective at a willingness-to pay (WTP) threshold of $150 000 per quality-adjusted life-year (QALY) (incremental cost-effectiveness ratio [ICER], $129 000 per QALY).⁸ 

The current study assessed the cost-effectiveness of second-line CAR-T in high-risk DLBCL modeling data from ZUMA-7.⁵ We also performed a threshold analysis to identify scenarios in which CAR-T would be cost-effective in the second-line setting for all patients with relapsed/refractory (RR)-DLBCL.

We developed a state-transition Markov model using TreeAge Pro 2021 release 1.2 (TreeAge, Williamstown, MA), simulating a cohort of US adults (mean age, 65 years) with RR-DLBCL who either had primary refractory disease or early relapse (<12 months from initial therapy).⁹ We used this model to examine the impact of the 2 treatment regimens studied in the ZUMA-7 trial: (1) SOC, defined as 2 cycles of salvage chemoimmunotherapy followed by high-dose chemotherapy and auto-SCT (if the patient responds to salvage chemoimmunotherapy) and CAR-T in the third-line setting for nonresponders and patients who relapse (2) second-line CAR-T, which consisted of conditioning chemotherapy of cyclophosphamide and fludarabine before a single infusion of axi-cel. For each treatment strategy, persons in the model transition between the following health states after completing initial therapy for RR-DLBCL: nonresponse/relapse, remission after auto-SCT, remission after CAR-T, and death (Figure 1A). One-month cycles and a lifetime horizon were used. Age-specific probability of all-cause mortality was estimated from the 2018 Centers for Disease Control and Prevention US Life Tables.¹⁰ We estimated the effects of grade 3 to 4 adverse events associated with therapy, including febrile neutropenia, anemia requiring blood transfusion, peripheral neuropathy, ICANS, and CRS.

In the SOC arm, patients with RR-DLBCL first received salvage chemoimmunotherapy. Patients who achieved a complete or partial response after salvage chemoimmunotherapy proceeded to auto-SCT in the next month. For patients without a response, they proceeded with CAR-T. CAR-T was modeled by using a similar model as that of a previously published cost-effectiveness analysis by Lin et al.⁸ We modeled third-line CAR-T as axi-cel using the ZUMA-1 (A Phase 1/2 Multicenter Study Evaluating the Safety and Efficacy of KTE-C19 in Adults With Refractory Aggressive Non-Hodgkin Lymphoma) trial and used the 4-year outcomes from the ZUMA-1 study.^2,11,12 Patients who relapsed or were refractory to third-line CAR-T were modeled as having poor survival with low quality of life and high monthly costs, as they have poor clinical outcomes with widely disparate treatment strategies.¹³ 

In the second-line CAR-T arm, patients who relapsed or had nonresponse proceeded to salvage chemoimmunotherapy and potentially transplant, as was allowed in the ZUMA-7 protocol. Given that there are limited real-world data in this setting, we modeled different long-term responses and survival scenarios. Patients who failed 3 lines of therapy in this arm were modeled as having poor survival and outcomes¹⁴ and high costs as described earlier.

We also modeled patients who achieved remission for >5 years as having a very low risk of subsequent progression with a slightly higher risk of all-cause mortality.^15,16 We also modeled that even after patients were assigned to a certain therapy, they may be unable to receive it or discontinue it due to worsening performance status or other reasons reflecting real-world practice.

With respect to modeling long-term outcomes of the SOC arm, prior studies, including CORAL (Collaborative Trial in Relapsed Aggressive Lymphoma) and SCHOLAR-1 (Retrospective Non-Hodgkin Lymphoma Research), have shown low rates of relapse in patients who responded to salvage chemotherapy + auto-SCT and were disease-free at 24 months.^1,17 Thus, we modeled a 5-year EFS in the SOC arm to be 10%, which is slightly lower than the 2-year EFS of 16% reported in the SOC arm in ZUMA-7.⁵ In our study, to inform the long-term outcomes of CAR-T, we used 4-year data from axi-cel showing excellent survival (92%) at 4 years in patients with remission at 2 years, and another study reporting 5-year outcomes of patients treated with tisagenlecleucel (tisa-cel) in the JULIET (A Phase II, Single Arm, Multicenter Trial to Determine the Efficacy and Safety of CTL019 in Adult Patients With Relapsed or Refractory Diffuse Large B-cell Lymphoma [DLBCL]) trial showing a low rate of relapse after 24 months.^11,18 However, in the ZUMA-7 trial, the median follow-up was 24.9 months, and thus we had only 2-year overall survival (OS) and EFS for second-line CAR-T. We addressed the uncertainty in long-term outcomes with second-line CAR-T by modeling several different scenarios. For the base case scenario, we modeled that 6% of patients who received second-line CAR-T progressed or needed additional lymphoma therapy between 2 and 5 years, which is a similar rate of relapse reported for CAR-T in the third-line setting.^11,12,18,19 Based on expert opinion, we also varied the 5-year EFS from 20% to 40% with second-line CAR-T. We also modeled a scenario in which fewer patients progressed with second-line CAR-T between 2 and 5 years (corresponding to a 5-year EFS of 40%) and another scenario in which the patients who avoided progression with second-line CAR-T in the first 2 years progress after 2 years (5-year EFS of 20%).

Transition probabilities, utilities, and costs were identified from published literature and public data sources, as detailed in Table 1 ^2,5,17,20-26 and supplemental Table 1 (available on the Blood Web site). Rates of remission, survival, and relapse for SOC and second-line CAR-T were obtained from the ZUMA-7 trial.⁵ We included costs related to drugs, administration costs, costs related to adverse events, transplant and cellular therapy costs, and follow-up costs. Costs were inflation-adjusted to 2021 US dollars by using the medical care component of the Consumer Price Index. Base case point estimates of cost were varied by at least ±50% for the sensitivity analysis.

The first assumption was that the effectiveness of third-line salvage chemoimmunotherapy in patients with nonresponse or relapse after second-line CAR-T is the same as when used in the third-line setting. Historically, patients with primary refractory or early relapse DLBCL who undergo salvage chemoimmunotherapy and auto-SCT have exhibited poor outcomes,^1,17 and we assumed that these poor outcomes remain when they are used after second-line CAR-T. For base case analysis, we modeled response rates for therapies post–second-line CAR-T to be 30%,²⁰ but in a sensitivity analysis, we varied it to range from 10% to 50%.^20,27 We also assumed that patients receiving second-line CAR-T would receive only steroids and no other bridging therapy, similar to the eligible population in ZUMA-7. We also assumed that patients who fail salvage chemotherapy and/or auto-SCT in the SOC arm proceed to third-line CAR-T. Finally, we assumed that patients who achieved at least 5 years of remission after auto-SCT had increased mortality compared with the general population (standardized mortality ratio [SMR], 2.2)^15,28 as did patients who received CAR-T and achieved at least 5 years of remission (SMR, 1.4).²⁹ For auto-SCT, primary nonrelapse reasons for mortality after 5 years were pulmonary toxicity (associated with carmustine), cardiac toxicity (related to high-dose chemotherapy), and infections, as well as secondary cancers.^15,28 CAR-T is also associated with some late effects, including hypogammaglobulinemia, prolonged cytopenias, late infections, immune-related late effects, cardiac toxicities, and subsequent malignancies.¹⁰ It is not necessarily related to the pulmonary or cardiac toxicity that drove 46% of the long-term nondisease-related deaths after auto-SCT.²⁹ This is why we modeled a higher SMR for auto-SCT than for CAR-T.

We calibrated our model using a previously validated, limited-memory BFGS-B optimization algorithm, defining goodness of fit as the sum of the squared distance between the published survival curves and simulated ones from our model.³⁰ The OS, progression-free survival, and EFS were calibrated at different time points for both the SOC and second-line CAR-T arms as well as for different treatment strategies after nonresponse/relapse, such as third-line CAR-T in the SOC arm (Figure 1C-D; supplemental Sections II and III).

As described earlier, the cost-effectiveness of each of the treatment strategies was reported from a health care perspective. Outcome measures were reported as ICERs (2021 US dollars per QALY) with a WTP threshold of $150 000 per QALY and the OS and EFS at 2 and 5 years.³¹ Costs and utilities were discounted by 3% annually.

Extensive one-way sensitivity analyses were performed to evaluate the effect of all defined variables according to each of the strategies, including age, life expectancy after long-term remission, and costs and outcomes with cellular therapy for RR DLBCL. For the Monte Carlo probabilistic sensitivity analysis, 10 000 iterations were performed using γ distributions for cost and β distributions for transition probabilities.

To assess the threshold efficacy for CAR-T in all patients with RR-DLBCL, an alternative scenario was modeled in which second-line CAR-T is not only used for primary refractory/early relapse patients (as we modeled earlier) but also for late relapse patients (>12 months). We compared this alternative model in which second-line CAR-T is used for both late relapse and primary refractory/early relapse patients vs a model in which second-line CAR-T is used only for primary refractory/early relapse patients and salvage chemoimmunotherapy followed by auto-SCT is used for late relapse patients (Figure 1B). The outcomes of salvage chemoimmunotherapy and auto-SCT in late relapse patients were modeled by using the CORAL study¹ (Table 2). We modeled that 65% of the RR-DLBCL cases were primary refractory/early relapse, and 35% were late relapse.^32,33 As part of a pragmatic approach, late relapse patients who failed salvage chemoimmunotherapy or relapsed after SCT could proceed to third-line CAR-T, which was modeled by using the ZUMA-1 study.² The primary purpose of this comparison was to determine the threshold efficacy for second-line CAR-T in late relapse patients for it to be cost-effective (as CAR-T has substantial costs), and these late relapse patients have excellent outcomes with the SOC salvage chemoimmunotherapy. Similar to the primary analysis, we modeled durable responses with CAR-T in those patients who remain in remission at 2 years.

In the base case analysis, we assumed that 6% of patients progressed or required additional lymphoma therapy after 24 months in the second-line CAR-T arm. Assuming a 5-year EFS of 35% with second-line CAR-T and a 5-year EFS of 10% with SOC, second-line CAR-T resulted in an additional 3.29 life-years and an incremental 2.82 QALYs (Figure 2). However, second-line CAR-T cost $771 838 compared with $508 034 for SOC (incremental cost of $263 804). Based on this, second-line CAR-T was cost-effective ($93 547 per QALY) at WTPs of $100 000 per QALY and $150 000 per QALY (Table 3).

One-way sensitivity analyses were performed on all model inputs for this model. At a WTP of $150 000 per QALY, the model was sensitive to the cost of CAR-T and the 5-year EFS for second-line CAR-T (supplemental Section V). At a WTP of $100 000 per QALY, the model was also sensitive to the cost of salvage chemoimmunotherapy and the probability of progression after third-line CAR-T (supplemental Section VI).

In this scenario, second-line CAR-T has a low rate of relapse after 2 years, which corresponds to a 5-year EFS of 40%. In this scenario, second-line CAR-T was cost-effective at WTPs of $100 000 per QALY and $150 000 per QALY (ICER $73 968 per QALY) (Tables 3 and 4).

In this scenario, 21% of patients who did not progress or require additional therapy with second-line CAR-T at 2 years progress or require additional therapy between 2 and 5 years (potentially due to T-cell exhaustion or CD19-antigen escape). This leads to an EFS of 20% with second-line CAR-T at 5 years. In this scenario, second-line CAR-T was not cost-effective at a WTP of $150 000 per QALY (ICER $233 967 per QALY) (Table 3).

A one-way sensitivity threshold analysis identified that if CAR-T therapy costs less than $972 061, it is cost-effective at a WTP of $150 000 per QALY (Figure 3A). Two-way sensitivity analyses of the costs of CAR-T and the 5-year EFS with second-line CAR-T are shown in Figure 3B. They show that if, for example, the cost of CAR-T is $500 000, second-line CAR-T has to have an EFS of at least 26.2% at 5 years to be cost-effective at a WTP of $150 000 per QALY.

As shown in Figure 3C, we performed a one-way sensitivity analysis varying the 5-year EFS with second-line CAR-T from 20% to 40%. We identified that to be cost-effective at WTPs of $150 000 per QALY and $100 000 per QALY, second-line CAR-T has to at least have an EFS of 26.4% and 33.8%, respectively, at 5 years. Next, in a two-way sensitivity analysis varying the 5-year EFS with second-line CAR-T (from 20% to 40%) and the 5-year EFS with SOC (from 10% to 14%), we showed that as long as second-line CAR-T had a 5-year EFS of at least 14% higher than that of SOC, it was cost-effective at $150 000 per QALY (Figure 3D).

Probabilistic sensitivity analysis was derived from performing 10 000 Monte Carlo model iterations for each strategy. Cost-effectiveness acceptability curves were generated for the model (Figure 4). Second-line CAR-T was the cost-effective strategy in 57% and 73% of iterations at a WTP of $100 000 and $150 000, respectively.

Because the use of second-line CAR-T may broaden in the future to include patients with late relapse, we performed a sensitivity analysis to identify the threshold effectiveness for CAR-T in late relapsed DLBCL to be cost-effective in all patients with RR-DLBCL. Two arms were modeled: in the first, both primary refractory/early relapse and late relapse patients receive second-line CAR-T, whereas in the second arm, only primary refractory/early relapse patients receive second-line CAR-T while late relapse patients receive salvage chemoimmunotherapy as a bridge to auto-SCT. We informed the inputs for salvage chemoimmunotherapy/auto-SCT in late relapse patients from the CORAL study¹ (Table 2). Second-line CAR-T was cost-effective in all patients if it had a 2-year EFS of at least 58.6% and 64.4% in late relapse patients at WTPs of $150 000 and $100 000, respectively (Figure 5A). To maintain the cost-effectiveness of second-line CAR-T in all patients, 5-year EFS had to be at least 47.2% in late relapse patients at a WTP of $150 000 (Figure 5B).

This study is the first, to our knowledge, to evaluate the cost-effectiveness of second-line CAR-T in patients with RR-DLBCL. This analysis is particularly timely and relevant given the recent US Food and Drug Administration approval of axi-cel in the second-line setting for primary refractory or early relapsed DLBCL. In the primary analysis, assuming a 5-year EFS of 35% for second-line CAR-T in primary refractory/early relapse DLBCL patients, we found that second-line CAR-T was highly cost-effective at a WTP of $100 000 per QALY (ICER $93 547). Its cost-effectiveness was dependent on a modeled long-term mortality benefit with second-line CAR-T compared with SOC, which has not been seen clinically due to the short follow-up of ZUMA-7. CAR-T was cost-effective up to a cost of $972 061 in the second-line setting in these patients and was the cost-effective strategy in 73% of iterations at a WTP of $150 000.

It is important to note, however, that the cost-effectiveness of second-line CAR-T is highly sensitive to long-term outcomes. For the primary analysis, we assumed a durable response with CAR-T in the second-line setting, with only 6% of patients progressing or requiring additional lymphoma therapy between 2 and 5 years. There are reasons to suggest that this assumption is likely to be valid. Prior studies have shown that the majority of relapses post–CAR-T are within 6 months of infusion, and the long-term results from both ZUMA-1 and JULIET suggest that being event-free at 24 months was associated with a durable response.^11,18 Hence, we expect the 5-year EFS of the ZUMA-7 trial to be largely similar to the 24-month outcomes reported. We did model both best-case and worst-case scenarios to provide the range of cost-effectiveness scenarios for second-line CAR-T as it can vary widely in different settings. In the best-case scenario, second-line CAR-T has a low rate of relapse after 2 years and remains cost-effective at a WTP of $100 000. In the worst-case scenario, second-line CAR-T has a high rate of relapse after 2 years and is no longer cost-effective (ICER $233 967). As long as the 5-year EFS with second-line CAR-T is at least 26.4%, it will remain cost-effective at a WTP $150 000, highlighting the importance of long-term outcomes with cellular therapy in determining its cost-effectiveness.

Second-line CAR-T is cost-effective in primary refractory/early relapse DLBCL despite its substantial costs, given the poor efficacy of standard care salvage chemoimmunotherapy and auto-SCT in this setting, with a 24-month EFS of 16% in ZUMA-7.⁵ Standard care was associated with poor survival, as 3.29 life-years and 2.82 QALYs were gained with CAR-T in the second-line setting compared with standard care. This was the main driver behind the cost-effectiveness of CAR-T. Indeed, we showed that even if CAR-T therapy cost $517 941 and $972 061 at WTPs of $100 000 and $150 000, it would remain cost-effective. This suggests that the high costs of progression and death after standard care offset the substantial initial cost of CAR-T therapy.

The ZUMA-7 trial⁵ chose to focus on patients with primary refractory/early relapse DLBCL as these patients have poor outcomes with standard care and comprise a high-risk subpopulation.¹ CAR-T has not yet been studied as second-line treatment in patients with late relapse (>12 months) after initial chemoimmunotherapy. Compared with the primary refractory/early relapse population, these patients have a substantially improved outcome with salvage chemoimmunotherapy followed by auto-SCT (3-year EFS of 45% vs 20%).¹ However in the future, there may be interest in using CAR-T (particularly tisa-cel and lisocabtagene maraleucel [liso-cel], associated with less CRS and ICANS than axi-cel) as second-line therapy for all patients with RR-DLBCL, both primary refractory/early relapse and late relapse patients, and particularly in older and more frail patients with comorbidities who may not be eligible for auto-SCT.^34,35 Although this patient population was not included in ZUMA-7, recent data suggest efficacy and safety of CAR-T in these patients with geriatric vulnerabilities, comorbid conditions, and functional limitations.³⁴ We modeled second-line CAR-T for all patients with RR-DLBCL and found that only if CAR-T was highly effective in late relapse patients with a 2-year EFS of at least 58.6% would it be cost-effective for all patients with RR-DLBCL. This EFS is significantly higher than the 2-year EFS of 38% reported with third-line CAR-T in the ZUMA-1 study¹² and the 2-year EFS of 41% observed with second-line CAR-T in primary refractory/early relapse patients in the ZUMA-7 study.⁵ CAR-T theoretically may have superior outcomes in the second-line setting in late relapse patients compared with the third-line setting given fewer pretreated patients and better T-cell function; however, it is unclear if this will be the case given the similar 2-year outcomes seen with axi-cel in ZUMA-7 and ZUMA-1.

There are ∼29 108 new cases of DLBCL per year in the United States, with ∼8732 patients (30%) having refractory/relapsed disease.³⁶ If CAR-T is adopted in the second-line setting for 65% of patients with RR-DLBCL with primary refractory/early relapse,^32,37 this would then lead to $1.5 billion of additional health care expenditures. Thus, despite being cost-effective, CAR-T in the second-line setting in high-risk patients adds a substantial amount to the direct health care costs. Different payors and different health care settings may not be able to afford this cost. This scenario highlights the need to reduce the costs of CAR-T using different reimbursement arrangements, outpatient CAR-T, and the potential use of off-the-shelf, allogeneic CAR-T products and dual-targeting CAR-T products, which may increase competition in the market and drive down the cost.

Prior cost-effective analyses have examined the impact of CAR-T in the third-line setting. Lin et al⁸ found that, assuming a 5-year progression-free survival of 40%, CAR-T may be cost-effective in the third-line setting (ICER $129 000 per QALY gained) compared with salvage chemoimmunotherapy and auto-SCT. We found that CAR-T was more cost-effective in the second-line setting with an ICER of $93 547 per QALY, indicating that moving CAR-T to second-line in high-risk patients improves its cost-effectiveness. We also had 4- and 5-year outcomes of axi-cel and tisa-cel in the third-line setting, respectively, which informed us that the 2-year outcomes of CAR-T in the third-line setting remained durable long term.^11,18 Unlike the study by Lin et al,⁸ we were unable to model other CAR-T products given the short-term follow-up with liso-cel in the second-line setting³⁸ and a lack of EFS improvement seen with tisa-cel in the second-line setting.³⁹ 

Although cost-effectiveness analyses can help inform the cost-to-benefit ratio of different treatment strategies, they have certain limitations. As with any oncologic modeling study, the rates of response, relapse, progression, and toxicity influence the model’s outcomes. One of the main limitations was that we lacked long-term follow-up data from ZUMA-7, as we only had the 2-year follow-up data to inform our model. In our primary analysis, we modeled a durable response with only 6% of patients progressing or needing additional therapy after 2 years. Recent long-term data on CAR-T in the third-line setting support this assumption.^11,18 We also modeled various scenarios to possibly reflect increased progression after 2 years, especially as CAR-T becomes used second-line in real-world practice. Another limitation is that we could only model one CAR-T product, as described earlier. An additional limitation is that we primarily modeled using the ZUMA-7 study, which had strict inclusion criteria, including not allowing any bridging therapy except for glucocorticoids and not allowing patients with impending organ-compromising disease.⁵ This may mean that the results from ZUMA-7 are superior to those that will be seen in real-world practice when patients with a higher burden of disease are treated. However, the TRANSFORM (Lisocabtagene maraleucel vs standard of care with salvage chemotherapy followed by autologous stem cell transplantation as second-line treatment in patients with relapsed or refractory large B-cell lymphoma) study evaluating liso-cel in the second-line setting, which allowed for one cycle of bridging therapy, had a similar median EFS to ZUMA-7.^5,38 We modeled scenarios with lower EFS to account for the possibility of patients with a higher burden of disease than that included in ZUMA-7. In addition to the strict inclusion criteria in ZUMA-7, patients who achieved a complete response or partial response to salvage chemoimmunotherapy proceeded to auto-SCT. In real-world practice, patients who achieve partial response after the first salvage therapy but still have significant residual disease often are treated with CAR-T. Thus, an additional limitation is that ZUMA-7, which informed our model inputs, may not reflect the patient population or patient management in real-world practice. We also looked at the cost only from the health care perspective, and the cost of CAR-T can vary in different settings and may be transmitted to the patient. The final limitation is that we lacked data on the feasibility of stem cell collection and the efficacy of therapies, including auto-SCT after second-line CAR-T.

In conclusion, although CAR-T in the second-line setting for primary refractory/early relapse DLBCL showed a substantial improvement in EFS without a statistically significant improvement in OS compared with SOC, this modeling analysis found that it is cost-effective at a WTP of $100 000 and $150 000 if its response is durable. These analyses can help provide support for payor coverage and clinicians’ utilization of CAR-T in the second-line setting. Widespread adoption of CAR-T in the second-line setting will lead to substantially increased costs even in a high-risk subgroup, highlighting the importance of alternative reimbursement models and other efforts to reduce its cost.

Contribution: S.K. and N.R.T. designed the study and acquired the data; S.K., N.R.T., and M.S. analyzed the data; S.K., N.R.T., M.S., A.F.H., and Y.S. interpreted the data and performed the statistical analysis; S.K., N.R.T., and A.F.H. wrote the manuscript; and S.S., P.S., L.E.B., A.V.D., M.G.M., T.S., L.L.P., J.Z., S.J.F., L.W.K., and S.T.R. helped revise the manuscript. All authors discussed the data and the analysis methods and contributed to the manuscript.

Conflict-of-interest disclosure: L.E.B. reports research funding from Merck, Inc., Amgen, AstraZeneca, and Mustang Bio; and consultancy for Novartis, Gilead, F. Hoffmann–La Roche Ltd., BeiGene, and Genentech, Inc. A.V.D. reports consultancy, honoraria, and research funding for Genentech; research funding from Takeda Oncology; consultancy and research funding from TG Therapeutics; consultancy and honoraria from AbbVie, Pharmacyclics, and BeiGene; research funding from Gilead Sciences and Secura Bio; honoraria and research funding from Bristol Myers Squibb; honoraria from Rigel Pharm; and consultancy, honoraria, and research funding from Bayer Oncology and AstraZeneca. M.G.M. reports honoraria from Janssen and EUSA; and research funding from TG Therapeutics, Epizyme, BMS, MorphoSys, and BeiGene. L.L.P. reports other (food support) from Hoffmann–La Roche; and other (travel support) from Novartis and Pfizer. T.S. reports research funding from Kite, a Gilead Company, TG Therapeutics, Celgene, and BeiGene; speakers bureau for Janssen and Seattle Genetics; consultancy, research funding, and speakers bureau for AstraZeneca and PCYC; and consultancy and other/travel support, and research funding from Juno Therapeutics. J.Z. reports research funding from Secura Bio, Daiichi Sankyo, and AbbVie; honoraria from Kyowa Kirin, Secura Bio, Seattle Genetics; and consultancy for Secura Bio, Ono, Legend, Kyowa Kirin, Myeloid Therapeutics, Verastem, and Daiichi Sankyo. S.J.F. reports consultancy and current holder of individual stocks in a privately held company for Mustang Bio and Lixte Biotechnology; and consultancy for Allogene. L.W.K. reports consultancy and current equity holder in publicly traded company for PeproMene Bio, Inc. A.F.H. reports consultancy and research funding from AstraZeneca, ADC Therapeutics, Merck, and Genentech, Bristol Myers Squibb, and Seagen; consultancy for Tubulis, Takeda, and Karyopharm; and research funding from Kite, a Gilead Company, and Gilead Sciences. The remaining authors have no conflicts to disclose.

Correspondence: Nikhil Thiruvengadam, 11234 Anderson St, MC 1516, Loma Linda, CA 92354; e-mail: nthiruvengadam@llu.edu.