Acquired mutations in patients with relapsed/refractory CLL who progressed in the ALPINE study

Bruton tyrosine kinase (BTK) is a cytoplasmic, nonreceptor tyrosine kinase primarily expressed in hematopoietic cells, particularly in B lymphocytes. BTK mediates the B-cell receptor (BCR) signaling cascade, resulting in cell proliferation, differentiation, and survival.¹ Because aberrant BTK-mediated BCR signaling contributes to pathogenicity, BTK is an important target for the treatment of patients with chronic lymphocytic leukemia (CLL).^1,2 Irreversible covalent BTK inhibitors (cBTKis), including ibrutinib, zanubrutinib, and acalabrutinib, have transformed the therapy landscape of CLL and are currently recommended for disease treatment.^3-5 

cBTKis abate BTK kinase activity and subsequent BCR signaling by binding to the BTK C481 amino acid residue and blocking the adenosine triphosphate binding pocket. Despite the effectiveness of cBTKis in treating CLL, a small number of patients acquire cysteine (C)-to-serine (S) substitutions at position 481 (C481S) that disrupt cBTKi binding, restore BTK catalytic activity, and lead to drug resistance.⁶ Other variants at BTK position 481 (eg, kinase-activating C481T, kinase-impaired C481G, and kinase-dead C481F, C481R, and C481Y) are observed in some patients with CLL, though at a much lower frequency than C481S.^7-9 

Although less common, variant cBTKi resistance mutations at other loci (non-C481) within the kinase domain of BTK, which often co-occur with C481 mutations, have been observed in patients with CLL who progress on cBTKi treatment. Interestingly, clinical observations suggest that certain non-C481 mutations seem to be specific to the cBTKi administered. In some patients who developed drug resistance to zanubrutinib or ibrutinib, kinase-impaired BTK mutations at L528, in addition to mutations at C481, have been observed.^10-12 Some patients treated with acalabrutinib or ibrutinib acquire BTK T474 gatekeeper mutations in addition to C481 mutations.^11,13-15 Although evidence is sparse on their prevalence, these non-C481 mutations may confer cross-resistance to noncovalent BTKis (ncBTKis), including pirtobrutinib, which is not dependent on the C481 binding site and is a promising therapy recently approved in the United States for patients resistant to cBTKis.^16-18 Many patients who develop resistance to the ncBTKi pirtobrutinib acquire predominantly non-C481 BTK T474I and L528W mutations, with V416L, A428D, M437R, and other mutations also observed at progression.^10,11,19,20 

Although there is evidence that certain mutations in BTK confer resistance to BTKis, the rate and timing of their occurrence in patients with BTKi-resistant relapsed/refractory (R/R) CLL are unclear. Early studies suggested a high rate (up to ∼70%-80%) of BTK mutations in patients with R/R CLL (and without Richter transformation [RT]) who were treated with ibrutinib^7,9,11,21-27 and acalabrutinib.^{12-15,25,28-31} However, more recent studies that include data from real-world cohorts suggest lower rates (∼50%) of BTK mutations in patients with CLL at progressive disease (PD).^9,26,32 The high variability in reported BTK mutation rates across studies is likely due to small patient populations who may differ as to whether disease progression occurred or patients came off for adverse events, and who differ in demographics, baseline characteristics, and the type, number, and duration of previous therapies. Variability in the sensitivity of mutation detection platforms may also have an impact.

Although BTK mutations are the most frequently reported mechanism of acquired cBTKi resistance, some patients develop cBTKi resistance in their absence. Little is known about these alternate cBTKi resistance mechanisms.^22,33,34 Often, but not always, mutations in the BTK domain co-occur with gain-of-function mutations in phospholipase C gamma 2 (PLCG2), an enzymatic substrate of BTK that, when mutated, reconstitutes BCR signaling.^6,13,35,36 However, BTK/PLCG2 comutations are infrequent, and solo PLCG2 mutations associated with cBTKi resistance are even more rare.²⁴ Whether or not mutations in other genes, chromosomal abnormalities, epigenetic profiles, or other mechanisms play a role in cBTKi resistance is currently poorly understood or undiscovered.³⁷ 

Although multiple smaller or retrospective studies have evaluated cBTKi resistance mutations in patients with R/R CLL, to date, data from large, randomized trials are limited. Here, we describe mutational data from the large, randomized, multicenter phase 3 ALPINE trial in patients with R/R CLL. We describe cBTKi resistance mutation data from 57 patients who were sampled at a median follow-up time of 25.7 months; accordingly, these patients represent a relatively high rate of early progression after treatment with either zanubrutinib or ibrutinib in the head-to-head ALPINE study (www.ClinicalTrials.gov identifier: NCT03734016).

Full details of the ALPINE trial, including patient selection criteria and study design, have been reported previously.³² In the ALPINE study, PD was determined by independent review committee (IRC; n = 139; zanubrutinib, n = 54; ibrutinib, n = 85) and/or by investigator (INV; n = 132; zanubrutinib, n = 54; ibrutinib, n = 78) using International Workshop on CLL criteria.^3,32 

The trial was approved by the institutional review board or independent ethics committee at each trial site and conducted in accordance with the principles of the Declaration of Helsinki, the International Council for Harmonisation Good Clinical Practice guidelines, and all applicable regulatory requirements. All the patients provided written informed consent. All authors had full access to the data and vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol and statistical analysis plan.

Peripheral blood mononuclear cell samples were collected at baseline (median CLL fraction, 67.1% [range, 0.05-90.87]) and at PD or after PD and before subsequent CLL therapy from 57 patients who progressed on the ALPINE study³² (zanubrutinib, n = 26/57; ibrutinib, n = 31/57), representing 40.2% of the total PD patients based on INV assessment (n = 53/132; Figure 1A; supplemental Table 1); 91.2% (n = 52/57) of these patients had matched baseline samples (supplemental Table 2).

Gene mutation analysis was performed using a high-sensitivity PredicineHEME-targeted gene next-generation sequencing (NGS) panel that covers all the exons of 106 genes, including BTK and PLCG2, with an average sequencing depth of 4749 reads in these 2 genes; 27 putative CLL driver genes with prognostic value in CLL identified by Knisbacher et al³⁸ were also represented on this panel (supplemental Table 3). The validated limit of detection for all mutations represented on this panel is 0.1% for hot spot and 0.25% for non–hot spot mutations.³⁹ 

Data reported include all BTK and PLCG2 mutations present at variant allele frequency (VAF) ≥0.25%. For all other genes, pathogenic mutations with VAF ≥1% are reported. Pathogenicity was assessed using VarSome. For samples collected at or after PD, the cancer cell fraction (CCF) was approximated using the following equation: VAF/(absolute lymphocyte count/white blood cell count). For 4 samples, absolute lymphocyte count/white blood cell count data were not collected at the same time as the PD sample collection. In these cases, CCF was estimated using the maximum VAF of mutations without copy number gains from NGS data. For all samples in this study, sequencing reads were manually inspected to confirm the absence of BTK/PLCG2 mutations within the validated limit of detection and/or assay performance criteria.

For BTK/PLCG2 or CLL driver gene mutation rate analysis, only patients with paired baseline and PD samples and without RT at PD (n = 52; zanubrutinib, n = 24; ibrutinib, n = 28) were included. Del(17p), del(11q), trisomy 12, del(13q), immunoglobulin heavy-chain variable region (IGHV) mutation status, and complex karyotype (CK) status were assessed in baseline samples as described in Brown et al.³² 

To explore the association between acquired BTK and/or PLCG2 mutations and other key genomic markers, including driver genes and chromosome alterations, 2 × 2 contingency tables and Fisher exact test were conducted.

Of the 57 patients with PD included in this analysis (zanubrutinib, n = 26/57; ibrutinib, n = 31/57), 79% were male. The median age was 65 years, and median number of previous treatments was 1 (zanubrutinib: range, 1-3; ibrutinib: range, 1-7). Baseline characteristics, including sex, age, region, disease stage (Eastern Cooperative Oncology Group performance status, bulky disease), and del(17p)/TP53 status, of the 57 patients with PD samples were representative of the overall study population (intent to treat; supplemental Table 4). In patients with baseline samples and without RT (n = 52), 9.6% (5/52) had del(17p), 17.3% (9/52) had del(11q), 21.2% (11/52) had trisomy 12, 55.8% (29/52) had del(13q), 84.3% (43/51) had unmutated IGHV, and 35.7% (10/28) had CK ≥3 (Figure 2). The 3 patients without RT and without baseline samples were ibrutinib treated and did not have BTK mutations at PD.

The overall study median follow-up time was 29.6 months, and for the 57 patients with PD, the median follow-up time was 25.7 months from the final progression-free survival analysis at a data cutoff of 8 August 2022.³² A breakdown of patient PD assessments (INV vs IRC) is given in supplemental Table 1. The types of PD as assessed by INV are found in supplemental Table 5. Of patients with PD assessed by INV only, 15 of 26 patients treated with zanubrutinib had PD during treatment and 8 of 26 patients had PD after treatment. In ibrutinib-treated patients, 26 of 31 patients had PD during treatment, and 4 of 31 patients had PD after treatment discontinuation (Figure 1A; supplemental Table 1). The median treatment duration for zanubrutinib-treated patients was 20.2 months (n = 26/57 [range, 4.4-39.9]) and for ibrutinib-treated patients was 16.8 months (n = 31/57 [range, 3.5-36.2]). For the 57 patients with PD, median treatment duration in patients with PD without BTK mutations was 16.8 months (n = 19/52 [range, 5.0-33.3]) for zanubrutinib-treated patients and 15.9 months (n = 25/52 [range, 5.9-29.4]) for ibrutinib-treated patients. Of patients with BTK mutations at PD (n = 8), median total treatment duration was 29.7 months (n = 5/8 [range, 18.4-34.6]) for zanubrutinib-treated patients and 30.8 months (n = 3/8 [range, 11.8-34.5]) for ibrutinib-treated patients. Furthermore, 2 zanubrutinib-treated patients and 2 ibrutinib-treated patients, none with BTK mutations, developed RT during the treatment (supplemental Tables 2 and 5).

No BTK mutations were detected at baseline (of 52 patients with evaluable samples). At PD, 8 patients (zanubrutinib, n = 5/24 [20.8%]; ibrutinib, n = 3/28 [10.7%]) acquired mutations in BTK (Figure 2; Table 1), and among INV-assessed patients who experienced PD during treatment without a drug hold (n = 33; Figure 1A), 7 patients (zanubrutinib, n = 4/14 [28.6%]; ibrutinib, n = 3/19 [15.8%]) acquired BTK mutations (Figure 1A; Table 1). All patients (n = 9) with acquired BTK and/or PLCG2 mutations had PD assessed by both IRC and INV during treatment, except 1 ibrutinib-treated patient who had PD assessed only by INV after treatment. Seven ibrutinib-treated patients with INV-assessed PD, none with BTK/PLCG2 mutations, were among the patients who experienced a drug hold (defined as no treatment for ≥7 days within 6 weeks before PD; Figure 1A). One patient treated with zanubrutinib experienced a drug hold within 6 weeks before INV-assessed PD (Figure 1A).^32,40 

Of the patients who acquired BTK mutations, the median number of acquired BTK mutations was 3 (range, 1-6) for patients treated with zanubrutinib and 1 (range, 1-2) for patients treated with ibrutinib. The median CCF of acquired BTK mutations was 5.31% (range, 0.47%-77.72%) for patients treated with zanubrutinib and 1.84% (range, 1.09%-13.71%) for patients treated with ibrutinib (see Table 1 for individual patient data). Of 18 single-nucleotide variants in BTK, 77.8% (n = 14/18; zanubrutinib, n = 11/14; ibrutinib, n = 3/4) were at C481 (Figure 1B; Figure 3). Non-C481 mutations were detected in 12.5% of patients (n = 3/24) who progressed on treatment with zanubrutinib (2 with L528W [CCF = 9.58% and 17.6%] and 1 with A428D [CCF = 37.03%]); these 3 patients had progression while on the study treatment (n = 3/16 [18.8%]; Table 1; Figures 1B and 3). With exception of 1 variant of uncertain significance (BTK D43H mutation, CCF = 1.09), no non-C481 mutations were detected in patients who progressed on treatment with ibrutinib. One ibrutinib-treated patient who progressed after treatment discontinuation had a C481 mutation (Figure 1A; Table 1).

Two ibrutinib-treated patients (patients 3 and 4) acquired PLCG2 mutations at PD (Table 1). Patient 3 had both BTK and PLCG2 mutations (Table 1). No PLCG2 mutations were detected in patients treated with zanubrutinib.

Most patients in this study received subsequent lines of therapy after study treatment discontinuation due to disease progression (zanubrutinib arm, 18/26 [69.2%]; ibrutinib arm, 21/31 [67.7%]), including all patients with acquired BTK and/or PLCG2 mutations (according to September 2023 data cutoff). The median time from study drug discontinuation to start of next line of therapy was 3.5 weeks (range, 0.1-103.0) for patients treated with zanubrutinib and 5.3 weeks (range, 0.1-77.0) for patients treated with ibrutinib. Additional details on the next line of therapy after the study treatment are provided in supplemental Table 6.

At baseline, 92.3% of patients (n = 48/52) had mutations (single-nucleotide variant or indel) in 1 of the 27 CLL driver genes described by Knisbacher et al,³⁸ which were represented on the gene panel used in this study; 32 of 52 patients (61.5%) had at least 2 driver gene mutations, and the median number of mutations in these patients was 3 (range, 1-5; Figure 2). The median number of baseline driver mutations did not differ between patients who later developed BTK mutations (median, 2 [range, 1-4]) vs patients who retained wild-type BTK at PD (median, 2 [range, 0-5]; P = .33; supplemental Table 7). However, having ≥2 baseline driver gene mutations was associated with the later acquisition of BTK and/or PLCG2 mutations compared with patients with <2 mutations (P = .025; supplemental Table 8). At disease progression, the median number of driver gene mutations was 2 (range, 1-4) in patients with BTK mutations and 1 (range, 0-4) for patients with wild-type BTK (P = .08). One ibrutinib-treated patient (patient 3) had coacquired BTK, PLCG2, and TP53 mutations at PD.

Of the 27 driver genes included in our targeted panel, at baseline, the most frequently detected driver gene mutations were in NOTCH1 (n = 21; zanubrutinib, n = 11; ibrutinib, n = 10), TP53 (n = 19; zanubrutinib, n = 9; ibrutinib, n = 10), BRAF (n = 10, 5 patients in each arm), SF3B1 (n = 8; zanubrutinib, n = 3; ibrutinib, n = 5), and ATM (n = 8; zanubrutinib, n = 2; ibrutinib, n = 6). Copy number aberrations in 9 of 27 genes were observed in 23 patients and most frequently in CCND2 (chromosome 12; n = 10, amplification [amp]; zanubrutinib, n = 9; ibrutinib, n = 1), ATM (chromosome 11; n = 8, deletion [del]; 4 patients in each arm), TP53 (chromosome 17; n = 6, del; zanubrutinib: n = 2; ibrutinib: n = 4), and KMT2D (chromosome 12; n = 6, amp; zanubrutinib, n = 5; ibrutinib, n = 1; supplemental Table 9). At PD, acquired driver gene mutations were observed in 1 patient in the zanubrutinib arm (who also had TP53 and XPO1 mutations) and 5 patients in the ibrutinib arm (1 with TP53, 1 with SETD2, 1 with SF3B1, and 2 with ASXL1 mutations). Acquired copy number aberrations in driver genes were observed in 10 patients (KRAS [chromosome 12] amp, 3 zanubrutinib; NRAS [chromosome 1] amp, 2 ibrutinib; CDKN1B [chromosome 12] amp, 2 zanubrutinib, 1 ibrutinib; BIRC3 [chromosome 11] del, 2 ibrutinib).

To understand whether driver gene mutations were more or less likely to occur in patients with BTK mutations, associations between the 2 were evaluated. No clear associations between baseline driver gene mutations and BTK mutational status were detected based on this data set. Driver gene mutations at either baseline or PD were not associated with poor disease prognostic factors, such as del(17p), IGHV mutation, or CK status (Figure 2).

In this report, we describe cBTKi resistance mutation data from 57 patients with R/R CLL from the ALPINE study, representing those with a relatively high rate of early progression at a median follow-up time of 25.7 months. Most patients (n = 33/41 [80.5%]) did not acquire BTK or PLCG2 mutations at disease progression as of this data cutoff. Among zanubrutinib-treated patients in this study, 5 were found to have acquired BTK mutations, 3 of whom had a non-C481 mutation either with or without a C481 mutation. Among the ibrutinib-treated patients, all 3 patients with BTK mutations had C481 mutations, including 1 with a non-C481 mutation. Overall, patient numbers from this study are too limited to draw definitive conclusions on the frequency of C481 vs non-C481 mutations among patients with CLL who were treated with either zanubrutinib or ibrutinib. In this study, patients from both study arms with BTK mutations were treated for a significantly longer period of time than patients without BTK mutations, suggesting, as previously reported,⁷ that the likelihood of developing a BTK mutation increases with treatment duration. Hence, the overall low incidence of patients with BTK mutations in this study is likely because most patients sampled here were the earliest to relapse in the overall study population, and spent a shorter amount of time on cBTKi treatment (median treatment duration, 17 months⁴¹). Our results suggest that the mechanisms of early progression on cBTKi may be largely independent of BTK mutations, and our reported mutation frequencies are therefore not likely to be representative of all patients with CLL with disease progression on these BTKis.

Much speculation has focused on the incidence of the kinase-impaired BTK L528W mutation in patients with progression on zanubrutinib following the report by Blombery et al.¹⁰ Blombery et al¹⁰ analyzed data from patients with already confirmed BTK mutations from an undefined total number of patients with CLL who were tested at the Peter MacCallum Cancer Centre between 2017 and February 2022.¹⁰ This contrasts with the current study, in which all PD patients with available samples were evaluated and only 8.3% (2/24) of zanubrutinib progressors were found to have L528W mutations. In this study, no L528W mutations were found among ibrutinib progressors to date, although they have been previously reported to occur at a low frequency among ibrutinib-treated patients.^11,42 In addition, because no patients treated with acalabrutinib were included in this analysis, direct comparisons of the incidence of L528W in zanubrutinib- vs acalabrutinib-treated patients cannot be extrapolated from this report. Further research with larger cohorts will be necessary to clearly define the frequency of L528W and other BTK mutations among patients with disease progression on zanubrutinib.

Recently, data on mutations among patients progressing on the ELEVATE-RR study revealed that at a median follow-up time of 41 months, 66% of acalabrutinib-treated and 37% of ibrutinib-treated patients had BTK mutations,¹² rates significantly higher than those observed in this study. This finding likely reflects the longer follow-up time on the study drug in ELEVATE-RR compared with the patients reported here from the ALPINE study (the patients analyzed in this manuscript tended to have relatively early disease progression). Furthermore, potential differences in the sensitivity of mutation detection methods and platforms could lead to variation in reported mutation frequencies. In the ELEVATE-RR study, the BTK T474 mutation, which, to date, has not been observed in zanubrutinib-treated patients, was more prevalent in acalabrutinib-treated patients than ibrutinib-treated patients. The observations that certain non-C481 mutations may be specific to the type of cBTKi used may have implications for choosing subsequent lines of therapy. For example, acquired resistance mutations seem to be driven by the structure of each inhibitor and how it binds to BTK.^43,44 As data from larger patient cohorts become available, a more detailed understanding of BTK mutation patterns and their implications for the use of different BTKis will emerge.

The low incidence (17%) of patients with BTK/PLCG2 mutations in this study highlights that at this short median follow-up time of 25.7 months, this data set includes more patients who relapsed early in the ALPINE study. Thus, in these patients, other mechanisms within or outside of the BCR pathway, potentially including mutations in other genes, may be driving disease progression. The mutational analysis of a subset of the CLL driver genes described by Knisbacher et al³⁸ revealed that, at baseline, most patients analyzed in this study (92.3%) had mutations in at least 1 driver gene, with a median of 3 mutations per patient. This reveals that most patients had poor genetic prognostic factors at the time of enrollment, which may be one of the contributing factors for the development of clinical resistance. However, the functional roles of these gene mutations in the patients described here are unknown. Interestingly, patients with baseline mutations in ≥2 driver genes were more likely to acquire BTK mutations at PD in this study. Some patients (9.6%) acquired mutations in additional driver genes at disease progression, suggesting that clonal evolution happened before or at disease progression in these patients. Although patient numbers were low to fully illustrate clinical implications, baseline genetic aberrations including del(11q), del(17p), del13q, trisomy 12, CK, or IGHV did not appear in this study to be correlated with the occurrence of cBTKi resistance mutations or mutations in driver genes.

In addition to mutations in BTK, PLCG2, known driver genes, and genomic characteristics, treatment-related factors may have contributed to disease progression in the patients from the ALPINE study reported here and in other CLL patient populations. For example, 7 ibrutinib-treated patients who progressed during treatment were also subject to a ≥7-day drug hold within 6 weeks before PD, which could also potentially lead to disease progression.

Caution is needed when comparing mutation frequencies across studies as patient populations, sampling methods, duration and type of previous treatment, and timing of sampling differ among studies and may contribute to erroneous interpretation of results. For example, in studies in which retrospective samples are requested, there may be bias in the types of clinical cases that INVs are likely to submit. As another example, studies that select for CD19⁺ cells before mutation analysis may have increased sensitivity in detecting low-frequency variants. Reported mutation frequencies also depend on the type of mutation detection platform used and the source of sample material. For example, in this study, the mutational analysis was done using peripheral blood mononuclear cell samples. Mutations that may be present in the lymph node and/or bone marrow compartments therefore remain unknown. Finally, it is important to note that a major limitation of this study is the short follow-up time, such that most patients with progression in this analysis were early progressors. As reported mutation rates are likely dependent on these factors and the number, type, and duration of previous and current treatments, caution should be given when interpreting the data and comparing across studies.

In conclusion, we report data on mutational analysis from patients with disease progression on an unbiased, randomized, multicenter, phase 3 trial in patients with R/R CLL. With a median follow-up time of 25.7 months, the short cBTKi treatment duration and early progression were associated with a low rate of BTK and PLCG2 mutations. The short follow-up time in these patients represents a limitation of this study but at the same time highlights that alternative mechanisms of resistance, such as those discussed by Lampson and Brown,⁴⁵ may be driving disease progression in these patients. Follow-up analyses of the resistance mechanisms in these patients, along with mutational analysis from patients with later progression, will be needed to confirm this. These data suggest that salvage therapy with ncBTKis, including pirtobrutinib,¹¹ or BTK-targeted protein degraders that have in vitro activity against several non-C481 mutations, should remain viable treatment options after zanubrutinib or ibrutinib treatment, although clinical studies will be required to demonstrate the efficacy of these agents in this population.⁴³ 

Contribution: J.L., A.I., T.S., and A.C. designed the research, analyzed the data, generated the figures, and drafted the manuscript; R.H. and Y.S. provided computational and statistical data analyses and generated the figures; J.R.B., B.E., N.L., S.O., C.S.T., L.Q., and M.S. enrolled the patients and collected the clinical data; and all authors contributed to data interpretation and reviewed the manuscript.

Conflict-of-interest disclosure: J.R.B. served as a consultant for AbbVie, Acerta/AstraZeneca, Alloplex Biotherapeutics, BeiGene, Galapagos NV, Genentech/Roche, Grifols Worldwide Operations, InnoCare Pharma Inc, iOnctura, Kite, Loxo@Lilly, Merck, Numab Therapeutics, Pfizer, and Pharmacyclics; and received research funding from BeiGene, Gilead, iOnctura, Loxo@Lilly, MEI Pharma, and TG Therapeutics. J.L. is employed and may own stock in BeiGene. B.E. served on advisory boards of Janssen, AbbVie, BeiGene, AstraZeneca, MSD, and Lilly; served on speakers bureaus of Roche, AbbVie, BeiGene, AstraZeneca, and MSD; received honoraria from Roche, AbbVie, BeiGene, AstraZeneca, and MSD; received research funding/grants from Janssen, Gilead, Roche, AbbVie, BeiGene, and AstraZeneca; and received travel expenses from BeiGene. N.L. served as a consultant for AbbVie, AstraZeneca, BeiGene, Lilly, Genentech, Janssen, and Pharmacyclics; and received research funding from AbbVie, AstraZeneca, BeiGene, Lilly, Genentech, Octapharma, Oncternal, MingSight, and TG Therapeutics. S.O. served as a consultant for AbbVie, AstraZeneca, BeiGene, Lilly, Janssen, Johnson & Johnson, Pfizer, and Pharmacyclics; received research funding from BeiGene, Lilly, Pfizer, Pharmacyclics, and Regeneron; and reports membership in CLL Society (unpaid). C.S.T. received research funding from Janssen, AbbVie, and BeiGene; and received honoraria from Janssen, AbbVie, BeiGene, Lilly, and AstraZeneca. L.Q. reports consulting or advisory roles at BeiGene, Johnson & Johnson, Sanofi, and MSD; received research funding from BeiGene, Johnson & Johnson, and Pfizer; and served on the speakers bureaus of BeiGene, Johnson & Johnson, Pfizer, AstraZeneca, and Roche. R.H., Y.S., A.I., T.S., and A.C. are employed and may own stock in BeiGene. M.S. served as a consultant for AbbVie, Genentech, AstraZeneca, Genmab, Janssen, BeiGene, Bristol Myers Squibb (BMS), MorphoSys/Incyte, Kite Pharma, Lilly, Fate Therapeutics, Nurix, and Merck; received research funding from Mustang Bio, Genentech, AbbVie, BeiGene, AstraZeneca, Genmab, MorphoSys/Incyte, and Vincerx; holds stock in Koi Biotherapeutics; and reports employment at BMS (spouse).

Correspondence: Jennifer R. Brown, Dana-Farber Cancer Institute, 450 Brookline Ave, Boston, MA 02215; email: jennifer_brown@dfci.harvard.edu.