Bridging intensity is associated with impaired hematopoietic recovery after BCMA CAR-T therapy for multiple myeloma Open Access

Chimeric antigen receptor (CAR) T-cell (CAR-T) therapy has led to a paradigm shift in the treatment of advanced B-cell malignancies.^1,2 Safety and efficacy of B-cell maturation antigen (BCMA)–targeted idecabtagene vicleucel (ide-cel) and ciltacabtagene autoleucel (cilta-cel) in relapsed/refractory multiple myeloma (RRMM) have been demonstrated in clinical trials^3,4 and real-world studies.^5,6 Despite increasing experience and scientific advances, guiding patients through the pre– and post–CAR-T period remains challenging. The long time span from T-cell collection to actual CAR-T infusion often requires the application of bridging therapies. Considering the diverse therapeutic landscape in MM, ranging from conventional chemotherapy to targeted therapies and immunotherapies, there is still notable heterogeneity and limited evidence regarding the optimal selection of bridging agents.⁷ 

Bridging therapies are critical for disease control or debulking. This is not only important to avoid disease-related complications and sequelae but also to mitigate the risk for post–CAR-T toxicities.^7-9 Numerous studies have demonstrated associations between baseline tumor burden and neurological toxicities or cytokine release syndrome (CRS).^10-12 At the same time, bridging therapies are considered a risk factor for post–CAR-T cytopenias.^13-18 These were recently termed “immune effector cell–associated hematotoxicity” (ICAHT) and represent the most common high-grade adverse event of CAR-T therapy.^18-21 Both, early and late cytopenias can substantially affect quality of life and morbidity.^17,20,21 Moreover, profound and protracted neutropenia can predispose for infections, which are the main determinant of nonrelapse mortality.^22-24 However, the negative impact of bridging on post–CAR-T hematopoietic reconstitution remains controversial, because there are also studies reporting no or contrary associations after CD19- and BCMA-directed CAR-T therapy.^25,26 To date, studies on the clinical implications of different bridging approaches in RRMM are limited, and effects on the longitudinal recovery of different cell lineages have, to our knowledge, not been conclusively clarified.^14,15 

Overall, it remains unclear to what extent bridging therapy influences cytopenias after CAR-T therapy in RRMM. To elucidate the impact of different intensities of cytotoxic (CTX) bridging on post–CAR-T hematopoietic recovery and cytopenia-associated complications, we performed a multicenter international retrospective study including 158 patients with RRMM who received an infusion of anti-BCMA CAR-Ts.

We analyzed 158 patients with RRMM who were consecutively treated with either cilta-cel (n = 22) or ide-cel (n = 136) at 3 university hospitals (Heidelberg University Hospital, University Hospital of Würzburg, and Dana-Farber Cancer Institute) until the end of May 2024 (Figure 1A). Data cutoff was 22 August 2024. Patients had received lymphodepletion with fludarabine and cyclophosphamide (n = 153) according to the manufacturer’s instructions,^3,4 or bendamustine (n = 5) due to the fludarabine shortage²⁷ (Table 1). Clinical data and institutional standard operating procedures have already been partially published.²⁸ Data were extracted from the electronic patient management software and the original medical records whenever available. The total observation period for toxicities ranged from day 0 (day of CAR-T infusion) to day 90. The time period until day 30 was referred to as the early post–CAR-T period, and the period between day 31 and 90 was referred to as the late post–CAR-T period.^28,29 Unless otherwise stated, laboratory values before lymphodepletion were collected with a leniency period of up to 5 days.²⁸ Missing data on complications were mainly because of a limited follow-up time, external care with limited data access, or loss to follow-up.

The term holding therapy refers to any antimyeloma treatment applied in the period before leukapheresis,^30,31 provided that the time period between last exposure and final leukapheresis was <60 days. Bridging therapy was defined as any treatment applied between leukapheresis and lymphodepletion.^7,13,30,31 Based on exposure to classical CTX chemotherapy, we differentiated between non-CTX (no CTX agents; eg, daratumumab/carfilzomib/dexamethasone), intermediate CTX (1-2 CTX agents; eg, carfilzomib/cyclophosphamide/dexamethasone, cyclophosphamide/etoposide/dexamethasone), and intensive CTX (≥3 CTX agents or high-dose therapy with autologous stem cell transplantation; eg, cisplatin, doxorubicin, cyclophosphamide, and etoposide) approaches (Figure 1A-B; supplemental Tables 1 and 2). Patients without systemic treatment were classified as non-CTX. If 2 different regimens were applied, patients were assigned according to the most CTX treatment. An independent radiotherapy expert ruled out a relevant effect on bone marrow function when patients received radiotherapy alone or concurrently, which was therefore not an exclusion criterion for the further analysis.

Cytopenias were graded according to Common Terminology Criteria for Adverse Events, version 5.0. Grade ≥3 cytopenias were defined as severe. Nadir values were based on the lowest documented cell count, regardless of the underlying cause.²⁸ Only patients with repetitive measurements (≥2 time points) during the respective observation period were included,²⁸ unless the criteria for severe cytopenia were met. For analysis of hematopoietic reconstitution over time, blood cell counts available closest in time to the following time points were recorded: day 0 (−2 days), day 7 (±3 days), day 14 (±3 days), day 21 (±3 days), day 28 (±3 days), day 35 (±3 days), day 42 (±3 days), day 50 (±5 days), day 63 (±7 days), and day 90 (±7 days). Following the approach described by Juluri et al,³² time to neutrophil recovery was defined as the time from the first absolute neutrophil count (ANC) drop of <1000/μL to 3 consecutive measurements obtained on different days with an ANC of ≥1000/μL. Time to platelet recovery was defined as the time from the first platelet drop of <50 × 10⁹/L to 3 consecutive measurements with a platelet count of ≥50 × 10⁹/L. The first measurement was counted as the day of recovery. In the few cases in which recovery occurred without available confirmatory measurements, these were counted as recovery events. The thresholds were derived from CTC grade ≥3 and selected based on case numbers and clinical relevance. The phenotypes of neutrophil recovery (quick, intermittent, and aplastic) were defined according to Rejeski et al.¹⁹ 

Treatment response was assessed by the investigators on the basis of the International Myeloma Working Group criteria,³³ with the usual adjustments and limitations due to the retrospective design.⁵ An absence of response data was classified as “not available.” Patients with no measurable disease and no evidence of disease progression were classified as “not evaluable.” Disease stage according to the Revised International Staging System,^34,35 drug refractoriness, and previous lines of therapy^36,37 were determined according to international consensus. Extramedullary disease included bone-associated (paramedullary) and bone-independent soft tissue masses³⁸; extraosseous disease referred exclusively to the latter form.³⁸ High-risk cytogenetics was defined as presence of del(17p), t(4;14), and/or t(14;16).^3,4 Chromosome 1q21 abnormalities (+1q) included 1q gain or amplification.^39,40 CRS and immune effector cell–associated neurotoxicity syndrome (ICANS) were graded according to the American Society for Transplantation and Cellular Therapy.⁴¹ Infections were defined and graded as previously described.^42,43 For the analysis of severe infections, only grade ≥3 events with an onset between CAR-T infusion and day 90 were included. An infection event required clinical, microbiological, or radiological evidence. Fever alone was not counted as an infection event.⁴² Nonrelapse mortality and causes of death were defined according to the literature.^42,44 

Quantification of immune cell subsets in the peripheral blood was performed by flow cytometry as part of routine diagnostics at each center. The immune status was determined ≤14 days before lymphodepletion.

R (version 4.4.1), GraphPad Prism (version 10.4.0), and BioRender were used for data analysis and visualization. Continuous variables were compared by nonparametric Mann-Whitney U or Kruskal-Wallis tests, and percentages by the Fisher exact or χ² tests. Longitudinal blood cell counts were analyzed using a linear mixed model with random patient effect and group, time, and group-time interaction as fixed effects.²⁸ Univariate and multivariate logistic regression analyses were performed to examine associations between severe cytopenias and selected binary or continuous variables. Associations with time to neutrophil and platelet recovery, time to severe infection, progression-free survival (PFS), and overall survival (OS) were analyzed by univariate or multivariate Cox proportional hazards regression. Multivariate models were based on a priori–selected variables reflecting differences in baseline characteristics and treatment center^28,30 and established risk factors for post–CAR-T cytopenias.^20,21,25 Kaplan-Meier curves were used to estimate PFS and OS and compared by log-rank test. Kaplan-Meier estimates were further used to obtain cumulative risk curves for time to neutrophil and platelet recovery. Competing risk analysis of recovery times was not performed because of the low incidence of concurrent events. In case of loss to follow-up, disease progression, start of a new systemic treatment, or death, patients were censored. For time to severe infection, we performed an additional competing risk analysis treating the latter 3 event types as competing risks and using the Aalen-Johansen estimator to calculate cumulative incidence functions. Groups were compared by the Gray test. Severe infection density until day 90 was defined as the mean number of events per 100 patient days considering the total days at risk.⁴² All P values were 2-sided and considered as statistically significant at P <.05.

This study was approved by institutional committees and boards (ethics committee of the Medical Faculty of Heidelberg University [S-096/2017] and the Dana-Farber Institutional Review Board [18-340]). Written informed consent was obtained from all patients. All procedures were conducted in accordance with the Declaration of Helsinki.

Baseline characteristics for the total cohort (n = 158) are provided in Table 1. The median age was 64 years (interquartile range [IQR], 57-70), and 55 patients (35%) were female. Of 143 patients with available data, 7 (5%) had an Eastern Cooperative Oncology Group performance status score of ≥2, 11 of 136 (8%) had International Staging System stage III, 65 of 155 (42%) had extramedullary disease, 59 of 149 (40%) had high-risk cytogenetics, and 20 of 96 (21%) had a bone marrow burden of ≥50%. Patients had received a median of 5 previous therapy lines (IQR, 5-7). Triple-class and penta-drug refractory disease were found in 129 of 157 (82%) and 44 of 155 (28%) evaluable patients, respectively. The median vein-to-vein time was 53 days (IQR, 46-63).

Of the total cohort, 82 patients (52%) were assigned to the non-CTX, 55 patients (35%) to the intermediate CTX, and 21 patients (13%) to the intensive CTX bridging therapy group (Figure 1A-B). Details on the administered agents and regimens are shown in supplemental Tables 1 and 2. Five patients had received radiotherapy during the bridging period, whereby the effect on the bone marrow reserve was considered negligible after assessment by an independent radiotherapy expert. Differences regarding the distribution of patient and disease characteristics are outlined in Table 1. Comparing the allocation with a treatment group between the holding and bridging periods at the individual patient level (supplemental Figure 1), we observed continuity of the previous treatment group in 79% (n = 65), 62% (n = 34), and 43% (n = 9) of patients in the non-CTX, intermediate CTX, and intensive CTX bridging groups, respectively. We found no significant differences in the distribution of treatment responses before leukapheresis or lymphodepletion according to bridging intensity (Figure 1C; supplemental Table 3).

A comparison of laboratory parameters before lymphodepletion is provided in supplemental Table 4. The intensive CTX bridging group showed a higher frequency of patients with increased LDH (57%) compared with the non-CTX (31%) and intermediate CTX (11%) groups (P < .001). Moreover, we observed higher median ferritin levels in the intensive CTX group (369 ng/mL) compared with the others (191 ng/mL and 141 ng/mL, respectively; P = .13). Significant differences were seen in CD3⁺ and CD4⁺ T-cell and CD19⁺ B-cell counts, with higher levels in the non-CTX group than in the intermediate CTX group.

ANC, platelet, and hemoglobin levels before lymphodepletion are illustrated in Figure 1E-G. CTC grades of cytopenias are given in supplemental Tables 5 and 6. Patients in the intermediate CTX and intensive CTX groups showed lower hemoglobin values (median: 10.3 g/dL and 10.2 g/dL, respectively) compared with the non-CTX group (median, 11.05 g/dL; P = .01 and P = .03, respectively), whereas no significant differences were found for neutrophils and platelets. However, the intensive CTX group exhibited a higher rate of severe cytopenia of any kind (33%) and of severe thrombocytopenia (19%) compared with the non-CTX (9.9% and 4.9%, respectively) and intermediate CTX groups (9.4% and 1.8%, respectively; P = .02 and P = .03, respectively).

Most patients developed CRS (n = 134 [85%]), including 2 grade 3 events (1.3%). Patients in the intermediate CTX bridging group showed a significantly increased rate of CRS grade ≥2 events compared with all others (51% vs 26%; odds ratio [OR], 2.92; 95% confidence interval [CI], 1.50-5.85; P = .002). ICANS affected 16 patients (10%), with 4 grade ≥3 cases (2.5%) and an even distribution of ICANS cases across groups. Further information on safety is given in supplemental Table 7.

Of the total cohort (n = 158), 124 patients (78%) achieved a partial response or better, and 55 patients (35%) had a complete response or better. No significant associations between best overall response and bridging groups were found (supplemental Table 3). With a median follow-up of 9.8 months (95% CI, 8.2-11.2), median PFS was 10.6 months (95% CI, 8.6-16.3), and median OS was not reached (supplemental Figure 2A-B). The intensive CTX group showed a significantly inferior PFS than the non-CTX group (hazard ratio [HR], 2.49; 95% CI, 1.27-4.87; P = .008; Figure 1D). Moreover, OS tended to be shorter among patients with intensive CTX bridging, albeit not significant (HR, 2.04; 95% CI, 0.77-5.40; P = .15; supplemental Figure 2C).

The analysis of post–CAR-T ANC levels over time for the 3 bridging groups showed a lymphodepletion-associated drop, followed by a recovery (Figure 2A). Intensive CTX bridging was associated with a second drop, coinciding with the late post–CAR-T period. Average ANC levels across time points were significantly different between the non-CTX and intensive CTX groups (P = .01), and a global trend toward lower ANC levels was seen for the intermediate CTX group (P = .09). Moreover, there was a significant interaction between group and time points (P < .001), indicating different longitudinal recovery patterns (supplemental Table 8).

The median time to neutrophil recovery after the first ANC drop of <1000/μL was 21 days (95% CI, 15-42) in the non-CTX group, 21 days (95% CI, 16-38) in the intermediate CTX group, and 63 days (95% CI, 29 to not estimable) in the intensive CTX group (Figure 2B). Patients in the latter group showed a significantly prolonged neutrophil recovery compared with all others (HR, 0.49; P = .04; supplemental Table 9).

Moreover, we observed significant differences between the 3 groups when comparing early and late ANC nadir values (Figure 2C-D). The univariate analysis showed significant associations between CTX bridging and severe early neutropenia (OR, 3.52; P = .005), and intensive CTX bridging and severe late neutropenia (OR, 2.89; P = .05; supplemental Tables 10 and 11).

The 3 bridging groups showed further differences in the distribution of neutrophil recovery phenotypes (P = .02) as defined by Rejeski et al¹⁹ (Figure 2E; supplemental Table 7), with an increasing proportion of intermittent recovery in parallel with cytotoxicity, and the highest proportion of aplastic phenotypes in the intensive CTX group. We observed a trend toward increasing rates of granulocyte-colony stimulating factor administrations during the early (46% vs 55% vs 70%; P = .3) and late (23% vs 34% vs 40%; P = .3) post–CAR-T periods across the 3 groups (supplemental Table 7). The recovery of leukocytes and lymphocytes is illustrated in supplemental Figure 3.

Severe infections affected 16 patients (20%) in the non-CTX group, 11 patients (20%) in the intermediate CTX group, and 8 patients (38%) in the intensive CTX group. In a univariate cause-specific Cox regression analysis, with the non-CTX group as reference, intensive CTX bridging showed a nearly significant association with the time to severe infection (HR, 2.32; 95% CI, 0.99-5.43; P = .05), whereas no association was observed for the intermediate CTX group (HR, 0.98; 95% CI, 0.45-2.11; P = .95). Adjusting for competing risks, the Aalan-Johansen estimator of the corresponding cumulative incidence curves is shown in Figure 2F. Moreover, the intensive CTX group was characterized by a higher severe infection density until day 90 (0.55) than the non-CTX (0.37) and intermediate CTX (0.31) groups (Figure 2G). Further details on the total of 47 severe infection events are provided in supplemental Figure 4.

Longitudinal platelet recovery depending on bridging intensity is shown in Figure 3A. We observed a second, delayed decline of the platelet count, which was most pronounced in the intensive CTX group, with significant differences between the non-CTX and intensive CTX groups across time points (P = .006). Moreover, there was a significant interaction between group and time points (P < .001; supplemental Table 8).

CTX bridging was significantly associated with a longer time to platelet recovery in univariate analysis (HR, 0.57; P = .05), with a median of 26 days (95% CI, 18-35) in the non-CTX group, 42 days (95% CI, 35-65) in the intermediate CTX group, and 55 days (95% CI, 21 to not estimable) in the intensive CTX group (Figure 3B; supplemental Table 12). Moreover, intensive CTX bridging was associated with lower platelet nadir values (Figure 3C-D) and an increased risk for severe thrombocytopenia during the early (OR, 4.11; P = .009) and late (OR, 9.06; P < .001) post–CAR-T periods (supplemental Table 13 and 14).

CTX bridging resulted in a significantly higher rate of platelet transfusions during the late post–CAR-T period compared with the non-CTX group (OR, 5.63; P = .008), and numerically higher rates of patients receiving thrombopoietin receptor agonists and stem cell boosts (Figure 3E-G; supplemental Table 7).

Although the 3 bridging groups showed a similar recovery pattern of hemoglobin values during the early post–CAR-T period, there was a second decline among patients with intensive CTX bridging (Figure 4A). We observed significant differences across time points between the non-CTX and intensive CTX groups (P = .02), and time-group interaction analysis confirmed distinct temporal dynamics of hemoglobin recovery, driven by the intensive CTX group (P < .001; supplemental Table 8).

Intensive CTX bridging was significantly associated with lower hemoglobin nadirs and severe anemia during the late (OR, 5.4; P = .003), but not early, post–CAR-T period (Figure 4B-C; supplemental Tables 15 and 16). Moreover, CTX bridging led to a significantly higher need for late red blood cell transfusions (OR, 4.16; P = .02; Figure 4E; supplemental Table 7).

To further understand the impact of bridging intensity on post–CAR-T cytopenias, we performed multivariate modeling accounting for potential confounders and including established risk factors (Tables 2-4). Intensive CTX bridging was confirmed as an independent risk factor for severe early and late thrombocytopenia, severe late anemia, and a prolonged time to neutrophil recovery. Nearly significant associations were observed for severe late neutropenia and time to platelet recovery. Apart from bridging intensity, we identified CRS grade ≥2, preexisting grade ≥3 cytopenias, and disease progression before lymphodepletion as relevant risk markers of post–CAR-T cytopenias.

Here, we introduce a simple classification of bridging therapies, which can be easily transferred to everyday clinical practice and future studies. We found associations between CTX bridging intensity and temporal dynamics and phenotypes of post–CAR-T hematopoietic reconstitution. Intensive CTX bridging was associated with delayed recovery and a second drop in cell counts, leading to increased rates of severe cytopenias during the late post–CAR-T period.

Late cytopenias and the biphasic nature of hematopoietic recovery cannot be explained by lymphodepletion- or CRS-associated toxicity alone^19,45 and seem to be driven by other mechanisms, such as redistribution and delayed immune effects of CAR-Ts,^16,46 inhibition of hematopoietic stem cell maturation through paracrine effects⁴⁷ or B-cell recovery related–perturbations of stromal derived factor 1 levels.⁴⁵ Other explanations include cumulative bone marrow toxicity related to previous therapy lines and myeloma disease itself.^18,21 Based on our findings, it can be hypothesized that intensive CTX bridging adds more damage to hematopoietic stem cells, inducing acute and latent long-term bone marrow injury.⁴⁸ The reduced hematopoietic reserve associated with the latter form is masked under homeostatic conditions,⁴⁸ but could undermine the later resistance to a “second hit” such as CAR-T–related inflammation and shifts in the bone marrow cytokine milieu. Of note, no differences between groups were observed for the time from initial diagnosis or the number of previous therapy lines, and a relevant proportion of patients in the intensive CTX group had received a less CTX holding therapy before. Although intensive CTX therapies were associated with a higher rate of severe cytopenias of any kind before lymphodepletion, differences regarding baseline cell counts were discrete. It is therefore noteworthy that differences regarding recovery patterns were most pronounced during the late post–CAR-T period, and that CTX bridging intensity was associated with increased levels of supportive care, including transfusions, growth factors, and stem cell boosts.

There is accumulating evidence for negative effects of bridging therapies on hematopoietic recovery after anti-BCMA-^14,15 and anti-CD19 CAR-T therapy,^13,16 and bridging therapies were taken up as a risk factor for ICAHT in the current consensus guidelines.¹⁸ Moreover, studies on bridging in RRMM have mainly demonstrated associations between alkylator-based regimens and inferior post–CAR-T outcomes,^14,15 which is consistent with our findings. However, there are also studies that question the negative effects of bridging therapies on hematopoietic recovery. Logue et al observed an association between more time from last bridging and any grade ≥3 cytopenia on day 30 after ide-cel,²⁵ and Xia et al reported an increased probability for cytopenias after CD19 CAR-T in subgroups with lower bridging rates.²⁶ Moreover, debulking appears to be critical for protection from other immune-associated toxicities and favorable outcomes.^10-13,49 Overall, there seems to be a fragile balance for the use of CTX bridging therapies between disease control, potentially leading to better outcomes and facilitating recovery of the hematopoietic niche,²⁶ and impaired reconstitution by directly damaging the hematopoietic system, resulting in transfusion dependency and other complications.

The intensive CTX group was characterized by an increased tendency to severe infections. This could be explained by the impaired post–CAR-T recovery of immune cells but also by preceding negative effects of CTX chemotherapy on the humoral and cellular immune function and regulation,^50,51 damage to mucosal barriers^50,52 and endothelial dysfunction.^53,54 

The major limitation of this study is the retrospective design. In addition, the cilta-cel cohort was characterized by limited patient numbers, a longer vein-to-vein time, and a shorter follow-up period. Another limitation arises from the fact that bridging groups showed differences regarding baseline characteristics and outcomes. Moreover, the results have to be interpreted under consideration of the multifactorial genesis of ICAHT.²¹ Aggressive disease manifestations are often countered by more intensive treatments, and differences in disease burden and dynamics before and after CAR-T infusion might be closely linked to increased toxicity. This makes it difficult to draw final conclusions regarding disease-, CAR-T–, and bridging-related effects. However, we aimed to address this point and adjust for potential confounders by including multivariate models and different end points.

In summary, our data underscore the relevance of intensive bridging as a risk factor for post–CAR-T cytopenias. In combination with clinical scores,^19,28,55 bridging intensity should be considered for risk stratification before BCMA CAR-T therapy and risk-adapted planning and implementation of post–CAR-T clinical care under consideration of current guidelines.¹⁸ High-risk patients could benefit from alternative lymphodepletion regimens^27,56,57; more frequent follow-ups; intensified prophylaxis; and supportive measures, including growth factors, transfusions, and immunoglobulin replacement. The availability of cryopreserved stem cells should be given special consideration when planning bridging and CAR-T therapy. If not available, alternatives to intensive bridging should be thoroughly evaluated, and a combined T-cell and stem cell collection approach may provide a solution for high-risk patients in the future.⁵⁸ Although caution toward high-dose alkylators is also recommended in the current International Myeloma Working Group guidelines on treatment sequences,³¹ targeted therapies, and novel immunotherapies, including short-term use of bispecifics,^59,60 appear to be promising in this context. The recent approval and increasing use of CAR-T therapies in earlier lines will open up new options for less hematotoxic bridging approaches and should be the preferred indication for cellular immunotherapies.

The authors thank all patients and the personnel at the participating centers who contributed to this work.

J.H.F. is funded through the International Myeloma Society and the Paula and Rodger Riney Foundation. A.S.S. has received funding from the National Cancer Institute (K08CA252174) and the Department of Defense (CA210827). M.S.R. has received research funding from the Dietmar Hopp Foundation.

Contribution: J.H.F., M.S.R., and S.S. conceptualized the study; J.H.F., T.H., and K.R. developed the methodology; J.H.F., X.Z., V.W., P.C., L.G., J.K., D.S., and K.H. performed the investigation; J.H.F. was responsible for analysis and visualization; M.S.R., S.S., L.R., A.S.S., O.N., E.K.M., C.S.M., M.J.F., N.W., H.G., A.S., M.H., M.S., C.M.-T., M.T., H.E., P.D., and N.C.M. were responsible for study supervision; J.H.F. wrote the original draft; M.S.R., K.R., and A.S.S. reviewed and edited the manuscript; M.S.R. was the study guarantor; and all authors read and approved the final manuscript.

Conflict-of-interest disclosure: J.H.F. reports an advisory role for Pfizer; honoraria from Bristol Myers Squibb (BMS) and Stemline Therapeutics; and travel and congress participation grants from Janssen-Cilag. K.R. reports research funding, honoraria, and travel support from Kite/Gilead; honoraria from Novartis and BMS/Celgene; travel support from Pierre-Fabre; and a consulting role for Kite/Gilead and BMS/Celgene. H.G. reports grants and/or provision of investigational medicinal products from Amgen, Array Biopharma/Pfizer, BMS/Celgene, Chugai, Dietmar Hopp Foundation, Janssen, Johns Hopkins University, Mundipharma GmbH, and Sanofi; research support from Amgen, BMS, Celgene, GlycoMimetics Inc, GlaxoSmithKline (GSK), Heidelberg Pharma, Hoffmann-La Roche, Karyopharm, Janssen, Incyte Corporation, Millenium Pharmaceuticals Inc, Molecular Partners, Merck Sharp & Dohme, MorphoSys AG, Pfizer, Sanofi, Takeda, and Novartis; has participated in advisory boards from Adaptive Biotechnology, Amgen, BMS, Janssen, and Sanofi; received honoraria from Amgen, BMS, Chugai, GSK, Janssen, Novartis, Sanofi, and Pfizer; and received support for attending meetings and/or travel support from Amgen, BMS, GSK, Janssen, Novartis, Sanofi, and Pfizer. A.S.S. reports consulting fees from Novartis. M.S.R. reports a consulting or advisory role for BMS, Amgen, GSK, Janssen, Sanofi, Pfizer, AbbVie, and Takeda; research funding from BMS, Janssen, Sanofi, and Heidelberg Pharma; travel support from BMS, Amgen, and Janssen; and honoraria from BMS, Janssen, AbbVie, and Sanofi. The remaining authors declare no competing financial interests.

Correspondence: Jan H. Frenking, Internal Medicine V, Heidelberg University Hospital, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany; email: jan.frenking@med.uni-heidelberg.de; and Marc S. Raab, Internal Medicine V, Heidelberg University Hospital, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany; email: marc.raab@med.uni-heidelberg.de.