Delayed neutrophil recovery following BCMA CAR-T therapy in Duffy-null myeloma does not impact severe infections or survival

TO THE EDITOR:

Chimeric antigen receptor T cells (CAR-Ts) targeting B-cell maturation antigen (BCMA) have shown remarkable efficacy in the treatment of multiple myeloma (MM), with high response rates and some patients achieving durable remissions.^1,2 Despite improved disease control, a subset of patients develop prolonged cytopenias,^3-7 which represent a significant source of morbidity and unmet need. Prolonged or profound neutropenia after CAR-T in particular is associated with disease progression, severe infections, and mortality, making it a critical area of ongoing investigation.^8-11 Prior retrospective analyses have identified risk factors for delayed absolute neutrophil count (ANC) recovery, including elevated baseline inflammatory markers, prior lines of therapy, advanced age, and self-reported Black race.^4,6,7,12-16 Given the expanding use of CAR-T earlier in the disease course and the disproportionate incidence of MM in Black patients, further work is needed to elucidate drivers of post–CAR-T hematologic toxicity in this population.

In recent years, there has been growing acknowledgment that self-reported Black individuals are disproportionately deemed ineligible for clinical trials or chemotherapy due to lower ANC.^17,18 It is now recognized that ANC differences are largely driven by the Duffy-null genotype, a genetic variant in ACKR1 highly prevalent in people of African descent, including approximately two-thirds of African-Americans.^19-22 This variant is associated with decreased circulating ANC without an apparent increase in infections,^21,22 likely due to diminished ability to mobilize marginated granulocytes.^23,24 Although guidelines defining a reference neutrophil range in the Duffy-null population are in development,²⁵ there are currently no clear standards for treatment and supportive care for Duffy-null patients requiring myelotoxic therapies. We hypothesized that delayed post–CAR-T neutrophil recovery in self-reported Black patients with MM may be driven by the high prevalence of Duffy-null phenotype in this population.

Expanded blood typing, including Duffy testing, is routinely performed at our institution before CD38 antibody exposure and is therefore widely available in our patients with MM. To assess the role of Duffy antigen status in ANC trajectory after BCMA-directed CAR-T, we reviewed medical records for patients with MM at our center between 2017 and 2024 under an Institutional Review Board-approved protocol. Laboratory values were collected from lymphodepletion until 100 days, the time of clinical restaging after cell therapy, and censored at disease progression. Given the association of Duffy-null phenotype with mild baseline neutropenia, sustained ANC recovery was defined as the first of 30 consecutive days with ANC ≥1.5 × 10³ cells per μL.

A total of 189 BCMA CAR-T patients were identified, and 152 patients (80%) with at least 1 month of relapse-free follow-up and Duffy testing were included. Self-reported race was documented in 140 patients and was Black in 33 patients (24%). Duffy-null phenotype was present in 28 patients (18%), including 25 of 33 Black patients (76%), 1 of 19 Hispanic patients, and 2 others without documented ethnicity; all White and Asian patients were Duffy non-null. Duffy-null (n = 28) and non-null patients (n = 124) otherwise had similar demographic and disease characteristics at CAR-T infusion, including median age (60 vs 62 years), prior lines of therapy (5 vs 5), International Staging System stage II to III disease (36% vs 39%), and drug refractoriness (triple-refractory, 68% vs 78%; penta-refractory, 25% vs 32%). Duffy-null patients had a lower rate of prior autologous stem cell transplant (64% vs 79%) or at least 1 high-risk cytogenetic abnormality (39% vs 53%), although these differences did not reach statistical significance. The 2 groups had similar incidence of overall (89% vs 86%) and grade 3+ cytokine release syndrome (0% vs 3%; supplemental Table 1).

Duffy-null and non-null patients had similar baseline ANC values before lymphodepletion (2.2 vs 2.6; P = .14). When comparing ANC trajectory after CAR-T, Duffy-null patients had significantly prolonged recovery to ANC ≥1.5 × 10³ cells per μL (median, 68 vs 40 days; P = .04; Figure 1A). Duffy-null patients showed a trend toward lower ANC at day 30 (0.9 vs 1.4; P = .07), with significantly lower values at 60 days (1.7 vs 2.6; P = .01) and 90 days (1.6 vs 2.3; P = .002; Figure 1B) after CAR-T. Conversely, Duffy-null patients had no difference in platelet recovery ≥50 × 10³/μL (P = .18) and faster recovery to platelets ≥100 × 10³/μL (median, 27 vs 38 days; P = .03). The 2 groups had similar use of granulocyte colony-stimulating factor in the first 100 days; 89% of Duffy-null patients and 84% of non-null patients required at least 1 dose (median 4 vs 5), which may be related to delayed ANC recovery in Duffy-null individuals without the prolonged severe neutropenia triggering additional growth factor. Notably, Duffy-null patients had increased incidence of infections within 100 days post-CAR-T (79% vs 40%; P = .0003), but this was driven primarily by increased grade 1 viral upper respiratory infections not requiring treatment (64% vs 31%; P = .002). Excluding grade 1 events, Duffy-null patients developed a similar rate of overall (25% vs 19%; P = .60) and severe grade 3+ infections (3.6% vs 7.3%; P = .69; supplemental Table 3). One Duffy-null patient died from grade 5 bacteremia, which occurred after ANC recovery.

The CAR-HEMATOTOX (HT) score has previously been validated to predict prolonged neutropenia, severe infections, and mortality after BCMA-directed CAR-T.^9,13 To adjust for the impact of factors included in this score, such as baseline inflammatory markers, we performed multivariable analysis examining the contribution of Duffy status and HT score to ANC recovery in our cohort. The Duffy-null and non-null groups showed similar proportions of HT-high patients (46% vs 44%; supplemental Table 1). After adjusting for Duffy status, age, prior lines of therapy, and HT score using a Cox proportional hazards model, Duffy-null phenotype remained significant with a hazard ratio of 0.47 (95% confidence interval, 0.27-0.83; P = .009) for ANC recovery (Figure 2A). We next stratified neutrophil recovery by both Duffy status and HT score (Figure 2B). Patients who were Duffy-null and HT-low (n = 15) closely approximated the recovery trajectory of those who were Duffy non-null and HT-high (n = 54; median, 53 vs 65 days). However, despite similarly delayed ANC recovery to HT-high patients, Duffy-null phenotype did not drive a corresponding risk for poor outcomes, as the Duffy-null group showed similar median progression-free survival (31 vs 22 months; P = .33) and overall survival (not reached vs not reached; P = .57) compared with non-null patients (supplemental Figure 1). Furthermore, HT score stratified patients by progression-free survival independently of Duffy status (supplemental Figure 2).

Limitations include the small sample of Duffy-null patients and increased rate of low-grade viral infections, which may contribute to neutropenia and partially confound recovery analysis. Nonetheless, in a subgroup analysis excluding patients with grade 1 infections, Duffy-null individuals maintained a similarly delayed time to ANC recovery compared with non-null patients (63 vs 42 days); although not statistically significant due to small subgroups, this suggests that the contribution of Duffy status to prolonged ANC recovery was likely independent of the observed difference in low-grade infections. Additionally, 61% of CAR-T patients in our cohort were treated on clinical trials, which may have strict exclusion criteria for cytopenias. These data may not represent the full spectrum of Duffy-null patients, as individuals who did not qualify for CAR-T trials due to baseline neutropenia would be excluded; it is possible that differences in ANC recovery may be even more pronounced if Duffy-null patients with lower baseline ANC values could be included. Further research is needed to define potential differences in response to therapy and more tailored ANC safety thresholds and inclusion criteria for Duffy-null persons to ensure equitable clinical trial access.

In summary, in this large single-center analysis of BCMA-directed CAR-T for MM, Duffy-null patients showed delayed neutrophil recovery after CAR-T independent of other validated risk factors, such as HT score, without the associated deleterious effects on survival or severe infections. These data indicate that differential neutrophil trajectory in Duffy-null individuals does not affect outcomes and highlights the need to establish consensus guidelines to direct posttreatment monitoring and supportive care in Duffy-null patients with MM and other hematologic diseases.

This retrospective research study consisting of patient medical record review and analysis was approved by the Institutional Review Board of the Mount Sinai School of Medicine (study ID 24-00505).

Acknowledgments: The authors thank Irene Protasio-Deslandes and the Mount Sinai Hospital Blood Bank for assistance in confirming Duffy antigen status for patients in our cohort. The authors acknowledge the philanthropic support and clinical and research staff at the Center of Excellence for Multiple Myeloma at Mount Sinai.

This study was supported by research funding from the National Cancer Institute grants R01 CA244899, CA252222, K12CA270375, and P30CA196521 (S.P.); as well as the National Institutes of Health KL2 Scholar Award 1KL2TR004421-01, the Multiple Myeloma Research Foundation Research Scholar Award, and the Robert A. Winn Diversity in Clinical Trials Career Development Award sponsored by the Bristol-Myers Squibb Foundation (S.T.).

Contribution: Z.M.A., S.P., and S.T. conceptualized the study; Z.M.A., S.B., D.P., J.C., and S.T. collected data and performed chart review; Z.M.A. and S.T. analyzed the data; and Z.M.A., S.P., and S.T. wrote and edited the manuscript with approval from all authors.

Conflict-of-interest disclosure: D.P. reports honoraria from Sanofi. L.J.S. reports consulting/advisory board fees from Janssen Pharmaceuticals. C.R. reports advisory board fees from Janssen Pharmaceuticals, Takeda Pharmaceuticals, Bristol Myers Squibb, Amgen, and Karyopharm Therapeutics. A.C.R. reports consulting fees from Johnson & Johnson, Adaptive, Bristol Myers Squibb, and Sanofi. S.J. reports consulting fees from Janssen Pharmaceuticals, Bristol Myers Squibb, Caribou Biosciences, Legend Biotech, Regeneron Pharmaceuticals, Takeda Pharmaceuticals, Sanofi, and Poseida Therapeutics; and advisory board fees from Janssen Pharmaceuticals, Bristol Myers Squibb, and GlaxoSmithKline Pharmaceuticals; and data safety monitoring board fees from Janssen Pharmaceuticals, Sanofi, and Genmab. H.J.C. reports employment at Multiple Myeloma Research Foundation; and research support from Genentech Roche, BMS, and Takeda Pharmaceuticals. S.P. reports advisory board for Grail; and research support from Celgene/BMS Corporation, Grail, and Caribou. The remaining authors declare no competing financial interests.

Correspondence: Santiago Thibaud, Hematology and Medical Oncology, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, Box 1185, New York, NY 10029; email: santiago.thibaud@mountsinai.org.