Cancer cachexia and weight loss before CAR T-cell therapy for lymphoma are independently associated with poor outcomes

Chimeric antigen receptor (CAR) T-cell therapy has transformed the treatment of non-Hodgkin lymphoma (NHL). Despite impressive advancements in the field, more than half of patients with large B-cell lymphoma (LBCL) are either refractory to or relapse after CAR T-cell therapy¹ and face dismal oncologic outcomes afterward.^2,3 Over the past few years, several prognostic markers of short- and long-term response to CAR T-cell therapy among patients with NHL have been identified. These markers include baseline tumor burden,⁴ prior treatment exposures,⁵ attributes of the CAR T-cell product,⁶ response to bridging chemotherapy,⁷ and scores or classifications constructed from laboratory measurements before CAR infusion, such as CAR-HEMATOTOX and INFLAMIX.^8-10 Identifying predictive biomarkers at early time points is critical to identify high-risk patients who might benefit from post-CAR treatment strategies and to design interventions that could address potentially modifiable factors and alter the trajectory of CAR T-cell recipients.

A potentially modifiable, patient- and disease-specific risk factor of interest is cancer cachexia. Several attempts have been made to accurately define this complex metabolic and immune syndrome characterized by weight loss (including muscle loss) and progressive functional decline.¹¹ Fearon et al defined cancer cachexia as weight loss >5% over past 6 months or body mass index (BMI) <20 kg/m² and weight loss >2% or sarcopenia.¹² Evans et al, on the contrary, defined cachexia as weight loss of >5% in 12 months plus 3 of the following criteria: decreased muscle strength, fatigue, anorexia, low fat-free mass index, and abnormal biochemistry (C-reactive protein [CRP] >5.0 mg/L, interleukin-6 [IL-6] >4.0 pg/mL, hemoglobin [Hb] <12 g/dL, and albumin <3.2 g/dL).¹³ In lymphoma, malnutrition, defined as an imbalance of energy and nutrients to the patient’s needs, is thought to negatively affect response to therapy, as summarized in a recent review.¹⁴ A recent study evaluated 70 patients treated with CAR T-cell therapy for NHL and showed that those with >5% weight loss between initial diagnosis and CAR T-cell infusion had worse overall survival (OS) in univariate analysis. Although intriguing, the study was limited by the small sample size, lack of multivariable adjustment, and the long interval in the measurement of weight loss because, patients may have received multiple treatments between diagnosis and CAR T-cell therapy.¹⁵ Additionally, detailed body composition studies have highlighted that the relationship between adipose and muscle tissue distribution and CAR T-cell outcomes is nuanced, with particularly unfavorable outcomes observed in patients with low abdominal visceral fat deposits and sarcopenia.^16,17 

We hypothesized that cancer cachexia could be a prognostic risk factor for CAR T-cell therapy and conducted a retrospective analysis in a large single-institution cohort of CAR T-cell recipients for lymphoma. We focused on weight loss to define cachexia rather than measurements of cytokines and body composition, because weight is universally available in all oncology practices. In our analysis, we investigated the relationship between cancer cachexia (as measured by weight loss) in the 3 months preceding CAR T-cell therapy, inflammatory status, tumor burden, response to bridging chemotherapy, disease status at the time of infusion, and treatment outcomes.

We retrospectively identified unique adult patients, aged ≥18 years, who received CAR T-cell treatment for US Food and Drug Administration–approved indications for lymphoma at Memorial Sloan Kettering Cancer Center (MSK) from 2017 to 2023. CAR T-cell products included axicabtagene ciloleucel (axi-cel), tisagenlecleucel (tisa-cel), brexucabtagene autoleucel (brexu-cel), and lisocabtagene maraleucel (liso-cel). We included patients for whom the following information was available (supplemental Figure 1): height and weight at the time of lymphodepleting chemotherapy (LDC) and 3 months (±20 days) prior, survival status, last follow-up, and assessment of cytokine release syndrome (CRS). The study was approved by the MSK Institutional Review Board.

The primary objective of the study was to determine the association between weight loss and OS among CAR T-cell recipients with lymphoma. The secondary objectives of the study were to determine the association between weight loss, response to CAR T-cell therapy, event-free survival (EFS), cumulative incidence of relapse (CIR), and nonrelapse mortality (NRM).

BMI was defined as weight in kilograms divided by height in meters squared (kg/m²). BMI categories were created per Centers for Disease Control (CDC) criteria.¹⁸ The first examined time point was the earliest recorded value between 110 and 70 days before CAR T-cell infusion (–3 months). This time point was selected as the baseline time because clinical decision-making regarding proceeding or not with CAR T-cell treatment usually occurs in the 3 months preceding the infusion, making it a relevant time point for potential intervention in the future. The second examined time point was at the time of LDC. We chose the day of LDC initiation rather than CAR T-cell infusion because patients may receive IV hydration around the time of inpatient chemotherapy, influencing their weight. We also examined a third time point, namely the earliest recorded value between 200 and 160 days before CAR infusion (–6 months). Weight loss was defined as a decrease in BMI of >5% between 2 time points.¹⁹ 

Tumor burden was approximated with serum lactate dehydrogenase (LDH) and total metabolic tumor volume (TMTV) on positron emission tomography (PET) scans before CAR T-cell infusion.⁴ Disease response was assessed per the Lugano criteria.²⁰ CAR T-cell products were grouped by costimulatory domain: CD28 products included axi-cel and brexu-cel; and 41BB products included tisa-cel and liso-cel.

OS was defined as the time from CAR T-cell infusion to the date of death. EFS was defined as the time from infusion to death, relapse, progression, or initiation of next treatment, whichever came first. CRS and immune effector cell–associated neurotoxicity syndrome (ICANS) were defined and graded as per American Society for Transplantation and Cellular Therapy (ASTCT) criteria.²¹ Immune effector cell–associated hematotoxicity (ICAHT) was defined and graded as previously described,²² using a published automated grading algorithm.²³ Early ICAHT concerns neutropenia in the first 30 days after CAR T-cell infusion, whereas late ICAHT describes neutropenia beyond day 30 (captured until day 100 in this study cohort).

Wilcoxon rank-sum, Fisher exact, and χ² tests were used to compare clinical and treatment characteristics by weight loss categories. Mean and standard deviation was calculated for laboratory values, and medians and interquartile ranges were provided for other continuous variables, as noted in the tables and figures. CAR-related toxicity and day 28 responses were similarly compared by weight loss categories. Multivariable logistic regression models further assessed these associations, adjusting for key clinical factors. Inflammatory markers were compared by weight loss categories using a Wilcoxon rank-sum test. Kaplan-Meier methods were used to estimate OS and EFS probabilities based on weight loss. Multivariable Cox regression models were used to adjust for patient and clinical characteristics. Cumulative incidence functions were used to estimate the incidence of NRM and relapse. For NRM, competing events included relapse, and for relapse, death in absence of relapse was considered a competing event. Gray test was used to compare these end points by weight loss categories.

During the study period from 2017 to 2023, a total of 259 patients with lymphoma were treated with US Food and Drug Administration–approved CAR T-cell therapy at MSK and had complete data for inclusion in our study. Supplemental Figure 1 shows the number of patients who did not have complete data to include in our analysis. There were 103 patients who did not have BMI data available at 3 months before CAR T-cell treatment, likely because they were referred to our center specifically for CAR T-cell treatment. Those patients were not statistically different from the 259 with fully available data regarding age, sex, performance status, TMTV, disease status at the time of CAR, and CAR T-cell product. They were more likely to have mantle cell lymphoma (17% vs 8.1%; P = .028; data not shown).

Table 1 shows baseline disease and patient characteristics of the patients included in our analysis. Most patients (n = 222 [85%]) were treated for large BCL, whereas 21 (8.1%) were treated for mantle cell lymphoma and 16 (6.2%) for follicular lymphoma. The cohort was predominantly male (64%) and aged >65 years (52%) at the time of infusion. Over half of patients (58%) were overweight or obese 3 months before CAR T-cell infusion. Regarding ethnicity, a higher proportion of Hispanic patients (87.5%) than non-Hispanic patients (56%) were overweight or obese at that time point (P = .006). There were no statistically significant differences in overweight/obese rates with respect to race. At the time of apheresis, 75% of patients had progressive or stable disease, and subsequently, 77% received bridging therapy. Of the 200 patients who received bridging, 53 (27%) received intensive chemotherapy, 52 (26%) received a polatuzumab-based regimen, 40 (20%) received radiation alone, 25 (13%) received lenalidomide, a Bruton tyrosine kinase inhibitor, or a BCL2 inhibitor, 7 (3.5%) received steroids alone, and the remaining 23 (11.5%) received other bridging regimens. At the time of CAR T-cell infusion, 59% had stable or progressive disease. Almost half of patients (47%) received axi-cel, whereas 26% received liso-cel, 22% received tisa-cel, and 4.2% received brexu-cel. At the –3-month time point, 26% patients had a BMI in the obese category, 32% in the overweight category, 40% in the normal category, and 2.7% in the underweight category.

About a quarter of patients (n = 68 [26%]) had weight loss >5% between –3-month and LDC time points (Figure 1; supplemental Figure 2). The median number of days between the –3-month and LDC time points was 94 (range, 70-110). Patients with weight loss were more likely to have a Karnofsky performance status <90 (P = .005) and higher TMTV as measured by PET (P < .001) at the time of CAR T-cell infusion (Table 1). Although bridging therapy was given to most patients irrespective of weight loss, those with weight loss were significantly more likely to receive bridging therapy (P = .001). Patients with weight loss also had a nonsignificant trend for higher rates of stable or progressive disease at the time of apheresis (P = .05) and at the time of infusion (P = .09). Prior treatments, including number of lines, autologous hematopoietic stem cell transplant, allogeneic hematopoietic stem cell transplant, and CAR T-cell treatment, were not significantly different between the 2 groups. A higher proportion of patients with weight loss (59%) received axi-cel than patients without weight loss (43%; P = .034). Two patients with weight loss (3%) and 1 patient without weight loss (0.5%) had manufacturing failures during the CAR T-cell manufacturing process (P = .11).

Patients with weight loss had significantly worse cytopenia and higher inflammatory markers than those without at 3 different time points (apheresis, lymphodepletion, and CAR infusion; supplemental Table 1). At apheresis, patients with weight loss had lower Hb, lower albumin, and higher LDH. At lymphodepletion, they had lower Hb, lower albumin, higher LDH, higher CRP, higher ferritin, higher IL-6, and higher tumor necrosis factor α (supplemental Figure 3). On the day of CAR T-cell infusion (day 0), they had lower Hb, lower absolute neutrophil count, lower albumin, higher CRP, higher ferritin, higher IL-6, higher tumor necrosis factor α, higher LDH, and higher IL-10. Patients with weight loss had a significantly higher mean CAR-HEMATOTOX score calculated based on laboratory values at lymphodepletion (1.98 vs 1.17; P = .008). A high-risk score ≥2 was seen in 50% patients with weight loss and 25% patients without weight loss.

Lastly, we examined the overlap between weight loss and the laboratory criteria for cachexia outlined by Evans et al, as described above.¹¹ Patients with weight loss were significantly more likely to meet each individual Evans criterion for cachexia (supplemental Table 2); 88% met at least 1 criterion, compared with 55% of patients without weight loss (P < .001). We excluded Hb <12 g/dL from our assessment of meeting any Evans criteria because of the high prevalence of anemia across all our patients.

We assessed the incidence of CAR T-cell–related adverse events among patients with and without weight loss (Table 2). Patients with weight loss did not have any difference in the rate of CRS (any and grade 3+) and ICANS (any) compared with patients who did not lose weight. However, they had higher rates of grade 3+ ICANS (P = .016) and grade 3+ early ICAHT (P = .012). However, in separate multivariable logistic regression models, which included the lymphoma histology, LDH at lymphodepletion, and CAR T-cell product costimulatory domain, neither of the associations with severe ICANS or ICAHT was maintained (data not shown). There was no difference for late ICAHT incidence between the 2 groups.

Patients with weight loss were numerically more likely to develop an infection within the first 100 days from infusion than those without (63% vs 45%; P = .11 in a cumulative incidence analysis with death as a competing risk). Supplemental Table 3 describes the infections that patients developed. Use of IV immunoglobulin did not differ between the groups: 4.7% of patients without weight loss and 4.4% of patients with weight loss received IV immunoglobulin before CAR T-cell infusion (P > .9), whereas 22% of patients without weight loss and 26% of patients with weight loss (P = .5) received IV immunoglobulin after CAR T-cell infusion.

We evaluated the best response after CAR T-cell therapy among patients in our cohorts. Patients with weight loss were less likely than those without to be in complete remission (49% vs 75%) and more likely to have progressive disease (28% vs 15%; P < .001; Table 3). In a multivariable logistic regression model adjusting for age, sex, performance status, diagnosis, LDH at LDC, disease status at infusion, CRP, and received CAR T-cell products, patients with weight loss were less likely to achieve complete remission than those without (odds ratio, 0.39; 95% confidence interval, 0.19-0.79; P = .009; Table 4). The conclusion was the same when TMTV was used in the model as a measure of tumor burden instead of LDH (data not shown).

We then assessed survival outcomes among patients with weight loss and those without. At a median follow-up of 12.4 months, calculated using the reverse Kaplan-Meier method, patients with weight loss had significantly inferior OS compared with those without (median OS, 10.5 months vs not reached; P < .0001; Figure 2A). This remained statistically significant (P = .02) in a multivariable Cox proportional hazards model incorporating age, sex, BMI category 3 months before CAR T-cell therapy, performance status, diagnosis, LDH, disease status at the time of CAR, CRP, and CAR product costimulatory domain (supplemental Table 4). Patients with weight loss had worse EFS (Figure 2B) in both unadjusted models and when adjusted for the variables described above (supplemental Table 5). The conclusion was the same for OS and EFS when TMTV was used in the model as a measure of tumor burden instead of LDH (data not shown). Patients with more profound weight loss (>10% BMI decline in 3 months before CAR) had even worse OS (4.1 months vs 43.8 months; P < .001; Figure 2C). In a spline model, OS appeared to be associated with weight loss in a continuous fashion (Figure 2D). To better understand the association with worse OS among those with weight loss, we plotted the CIR and NRM. Patients with weight loss had higher CIR with death as a competing risk (P = .02) and a trend toward higher NRM at 12 months (11% vs 6.7%) than those without (P = .059; Figure 3B). Supplemental Table 6 shows the causes of death among the patients in the 2 groups. There were 12 patients (18%) without weight loss and 7 patients (17%) with weight loss who died with infection as a contributing cause of death.

Lastly, given the use of a 6-month interval for the definition of cachexia in the Fearon criteria,¹⁰ we performed an exploratory analysis assessing the importance of weight loss between –6 months and LDC. Sufficient data were obtained for 175 patients: 122 (70%) with no weight loss and 53 (30%) with weight loss. Those with weight loss between –6 months and LDC had worse OS and EFS in both univariate and multivariable models, adjusted for the same variables described above (data not shown).

As access to CAR T-cell therapy increases, it remains imperative to identify predictive and prognostic markers of response, survival, and toxicity among patients who receive it. Identifying markers in the months-long period before CAR T-cell therapy is especially relevant clinically, because it provides an opportunity for early intervention to improve outcomes. Recent analyses have shown that the median time from consultation with a cell therapy physician until CAR infusion is 52 days (∼2 months), highlighting the importance of this period.²⁴ Weight loss is a readily available and easily obtainable data point for oncologists worldwide. In this study of patients with lymphoma receiving CD19 CAR T-cell therapy, we show that weight loss (>5% BMI reduction in the 3 months preceding CAR T-cell therapy) is associated with worse response rates, OS, and EFS independent of previously established prognostic markers such as tumor burden, performance status, and inflammation. We also show that there is a substantial overlap between weight loss and biochemical components of cachexia (as defined by Evans et al),¹³ illustrating the utility of weight loss as an easily available marker.

There is likely a complex relationship between inflammatory status, disease burden, and weight loss that drives cancer outcomes. Here, we attempt to elucidate this relationship in multivariable models. In our cohort, weight loss was an independent predictor of poor OS and EFS even when controlling for LDH and tumor status at infusion (disease burden) and CRP (inflammatory status). Thus, we postulate the model shown in Figure 4, in which weight loss, inflammation, and tumor burden interact with each other and independently contribute to poor CAR T-cell outcomes.

The exact mechanism through which weight loss might drive excess mortality in patients receiving CAR T cells is not known. Our findings suggest that the increased risk of relapse is the main driver of worse OS, although there was also a trend toward increased NRM among the patients with weight loss. The rate of infections was numerically higher among those with weight loss, but that was not statistically significant in competing risk analysis. In our cohort, patients with weight loss also experienced lower responses to CAR T-cell therapy, even after adjustment for known risk factors such as disease burden and inflammatory state. Lastly, it is possible that patients with weight loss have few (if any) options available to them if they relapse after CAR T-cell therapy as a result of their performance status, which would influence their OS probabilities.

Although the exact mechanism for higher proportions of refractory disease in the weight loss group is unknown, prior studies have reported that patients with cancer cachexia exhibit impaired immune function and exhausted T-cell phenotypes.²⁵ Notably, patients with cachexia have worse outcomes in the context of immune checkpoint treatment,²⁶ which raises the possibility that patients with weight loss have an immune-suppressive tumor microenvironment.^27,28 In the context of allogeneic stem cell transplantation, another treatment modality that at least partially relies on the antileukemic effect of T cells, weight loss is associated with inferior OS and NRM.²⁹ Because the median time between apheresis and CAR T-cell infusion in our study was 37 days, it is possible that the weight loss process had started before apheresis for many of our patients; thus, the apheresed T cells might have already been affected by the immunosuppressive effects of cachexia. Allogeneic CAR T cells could help address this concern in the future. In summary, there is evidence to suggest that cancer treatment modalities that rely on T-cell function (such as immune checkpoint inhibitors, allogeneic stem cell transplantation, and CAR T cells) do not work as well in the context of weight loss and cachexia. It remains to be investigated whether this is true in the context of bispecific antibodies, which engage endogenous T cells. Whether bispecific antibodies could constitute a reasonable choice for patients with lymphoma with significant weight loss, who might have good CAR T-cell outcomes, remains to be seen. Of note, there is evidence that obese patients might have better outcomes when treated with immune checkpoint inhibitors,^30,31 whereas obesity is thought of as an adverse risk factor in allogeneic stem cell transplant.³² In our study, we did not observe a relationship between obesity and CAR T-cell outcomes among CAR T-cell recipients, which is consistent with prior reports.³³ 

Our findings suggest that weight loss in the months preceding CAR T-cell therapy can be a helpful, universally available “alarm signal” for clinicians. Unlike the measurement of cytokines and disease burden evaluation with PET scans, weight measurements do not add any cost to the care of patients receiving CAR T-cell therapy and are available to essentially all clinicians, because oncology patients are weighed frequently for chemotherapy and other dosing purposes. This can help identify patients who might not fare as well with existing commercial CAR products and could be considered for clinical trials or other post-CAR consolidative approaches to improve long-term remission rates.

One intriguing question is whether interventions to reverse weight loss or influence the relationship between weight loss and inflammation can help improve outcomes of patients receiving CAR T-cell therapy. Although our data cannot answer this question, data from patients with cancer undergoing surgery and solid organ transplant recipients suggest that prehabilitation programs can improve outcomes,^34,35 and early data from the autologous transplantation setting demonstrate some improvement in performance status.³⁶ Thus, it seems plausible that a prehabilitation program encompassing both nutritional interventions to prevent weight loss and exercise programming to enhance performance status could improve CAR T-cell therapy outcomes among patients with lymphoma. An exercise-focused such program is currently under investigation in a clinical trial among older patients with hematologic malignancies scheduled to receive CAR T-cell therapy (NCT05763563). As shown in Figure 4, there are existing strategies to modify tumor burden (through bridging therapy), and prophylactic agents to modify inflammatory state are under active investigation.^37,38 However, there are currently no standard interventions to address weight loss, which represents an opportunity for action. We hope future studies can elucidate which components of weight loss matter more (adipose tissue vs muscle mass) and what interventions can deliver the maximal impact on outcomes.

Strengths of our study include the large population across diseases and products, our ability to adjust estimates based on many covariates, and the availability of historical BMI data at a defined time point. Our study is limited by its retrospective nature, which results in possible unmeasured covariates that could confound our findings. In addition, it is possible that the variables we used to represent tumor burden (metabolic tumor volume and LDH), disease progression (PET comparisons), and inflammation (CRP and ferritin) inadequately capture those processes, leading to inadequate controlling in multivariable modeling. Lastly, our data set does not include documentation of infections before CAR T-cell therapy, which could be associated with both weight loss and poor outcomes.

In summary, we present, to our knowledge, the first body of evidence that weight loss in the 3 months preceding CAR T-cell therapy is an independent poor prognostic marker for CAR T-cell recipients with lymphoma, may serve as a potentially modifiable factor in addition to known biomarkers of poor response such as disease burden and baseline inflammation, and warrants further investigation in CAR T-cell therapy recipients.

This research was supported in part by National Institutes of Health/National Cancer Institute (NIH/NCI) Cancer Center Support Grant (P30CA008748 [U.A.S.]). The work of M.C. was supported by an Alfonso Martin Escudero grant. A.R.-D. is supported by a grant from Fundación Española de Hematología y Hemoterapia. R.S. reports grant support from the NIH-NCI (K08-CA282987). U.A.S. is supported by the Memorial Sloan Kettering Cancer Center (MSK) Paul Calabresi Career Development Award for Clinical Oncology (K12CA184746), Paula and Rodger Riney Foundation, Allen Foundation Inc, Parker Institute for Cancer Immunotherapy at MSK, International Myeloma Society, HealthTree Foundation, Willow Foundation, and David Drelich, CFP Irrevocable Trust. U.A.S. is also supported by the American Society of Hematology Clinical Research Training Institute, Transdisciplinary Research in Energetics, and Cancer training workshop R25CA203650 (principal investigator, Melinda Irwin).

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Contribution: Y.V. designed and performed the analysis and drafted the manuscript; S.D. supervised the data analysis; K.R., M.C., A.L.D.A., A.R.-D., E.L., G.C., and I.L. extracted patient data for analysis; H.S. and A. Bedmutha performed the total metabolic tumor volume analysis; A.B., G.L.S., M.S., M.-A.P., G.S., and M.L.P. edited the manuscript; U.A.S. and J.H.P. oversaw and designed the study and edited the manuscript; and all authors approved the final manuscript.

Conflict-of-interest disclosure: Y.V. received a 1-time consultancy fee from EastRx. K.R. received research funding, consultancy fees, honoraria, and travel support from Kite/Gilead; honoraria from Novartis; consultancy fees and honoraria from Bristol Myers Squibb (BMS)/Celgene; and travel support from Pierre Fabre. A.L.D.A. received research funding from Kite/Gilead. M.-A.P. reports honoraria from Adicet Bio, Allogene, AlloVir, Caribou Biosciences, Celgene, BMS, Equillium, ExeVir, ImmPACT Bio, Incyte, Karyopharm, Kite/Gilead, Merck, Miltenyi Biotec, MorphoSys, Nektar Therapeutics, Novartis, Omeros, Orca Bio, Sanofi, Syncopation, VectivBio AG, and Vor Biopharma; serves on data safety monitoring boards (DSMBs) for Cidara Therapeutics and Sellas Life Sciences; and the scientific advisory board of NexImmune; has ownership interests in NexImmune, Omeros, and Orca Bio; and has received institutional research support for clinical trials from Allogene, Incyte, Kite/Gilead, Miltenyi Biotec, Nektar Therapeutics, and Novartis. I.L. received travel support from Kite/Gilead. U.A.S. reports support research funding support from Celgene/BMS and Janssen to the institution; research support from Sabinsa Pharmaceuticals and M&M Labs to the institution; and personal fees from Janssen Biotech, Sanofi, BMS, and i3Health outside the submitted work. G.L.S. reports research funding to the institution from Janssen, Amgen, BMS, Beyond Spring, and GPCR; and is on the DSMB for Arcellx. M.S. served as a paid consultant for McKinsey & Company, Angiocrine Bioscience, Inc, and Omeros Corporation; received research funding from Angiocrine Bioscience, Inc, Omeros Corporation, and Amgen, Inc; served on ad hoc advisory boards for Kite, a Gilead company, and Miltenyi Biotec; and received honoraria from i3Health, Medscape, and Cancer Network for continuing medical education (CME)-related activity and from IDEOlogy. A. Boardman has received compensation for participating in consulting activities with BMS. M.A.-P. reports honoraria and consulting fees from BMS, Cellectar, Ceramedx, Juno, Kite, MustangBio, Garuda Therapeutics, Novartis, Pluto Immunotherapeutics, Rheos, Seres Therapeutics, Smart Immune, Thymofox, and Synthekine; and other fees from Juno and Seres. J.H.P. received consulting fees from AffyImmune Therapeutics, Amgen, Autolus, Be Biopharma, BeiGene, Bright Pharmaceutical Services, Inc, Curocell, Kite, Medpace, Minerva Biotechnologies, Pfizer, Servier, Sobi, Synthekine, and Takeda; received honoraria from OncLive, Physician Education Resource, and MJH Life Sciences; serves on the scientific advisory boards of Allogene Therapeutics and Artiva Biotherapeutics; and received institutional research funding from Autolus, Genentech, Fate Therapeutics, Incyte, Servier, and Takeda. The remaining authors declare no competing financial interests.

Correspondence: Jae H. Park, Department of Medicine Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York NY 10065; email: [email protected].