Impact of body mass index on anti-BCMA chimeric antigen receptor T-cell therapy outcomes in multiple myeloma Open Access

The obesity epidemic is a global health problem that has been linked to the development of numerous malignancies,¹ including multiple myeloma (MM).² Obesity also affects cancer immunotherapy responses:³ obese patients with solid tumors have superior responses to immune checkpoint inhibitors (ICI),⁴ and patients with B-cell lymphoma with increased visceral fat respond better to anti-CD19 chimeric antigen receptor (CAR) T-cell therapy.⁵ This poorly understood phenomenon is called the “obesity paradox,”⁶ and highlights the complex interaction between metabolism, antitumor immunity, and cancer biology.

CAR T-cell therapy targeting B-cell maturation antigen (BCMA) has transformed the treatment of relapsed/refractory MM.⁷ Idecabtagene vicleucel and ciltacabtagene autoleucel are the 2 current US Food and Drug Administration-approved CAR T-cell therapies in MM. Modifiable host factors, such as metabolic health and obesity, that impact response and relapse after CAR T-cell therapy are an area of active investigation. A report from China evaluating investigational anti-BCMA CAR T-cell constructs found that patients with body mass index (BMI) ≥25 kg/m² had prolonged progression-free survival (PFS) compared with those with BMI <25 kg/m².⁸ However, the impact of BMI on outcomes in a United States cohort has not been described.

Because BMI conflates adiposity and lean mass, its association with clinical outcomes may be complex. We hypothesized that normal weight patients have more favorable tumor biology, whereas obesity may paradoxically confer enhanced immunotherapy sensitivity and CAR T-cell therapy efficacy, as seen in other cancers.^6,9,10 Thus, the intermediary overweight state may represent a relatively immunometabolically adverse phenotype without the compensatory benefits seen at either end of the BMI spectrum (Figure 1A). To evaluate this hypothesis, we modeled the relationship between BMI and CAR T-cell therapy outcomes using a restricted cubic spline¹¹—an approach that captures nonlinear associations between continuous variables—to evaluate how BMI influences CAR T-cell therapy outcomes.

We retrospectively studied patients treated with anti-BCMA CAR T-cell therapy for relapsed/refractory MM at Massachusetts General Hospital from 2016 to 2023. The study was approved by our institutional review board, and was conducted in accordance with the Declaration of Helsinki. Baseline characteristics including BMI and inflammatory serologic markers were recorded on the start date of lymphodepletion (LD). The Kruskal-Wallis test and Fisher exact test were used to compare continuous and categorical variables, respectively. Responses were defined according to International Myeloma Working Group criteria.¹² High-risk cytogenetics were defined as del(17p), t(4;14), or t(14;16), by fluorescence in situ hybridization, per International Myeloma Working Group criteria.¹² PFS was defined as the time from CAR T-cell infusion to either disease progression or death. Overall survival (OS) was defined as the time from CAR T-cell infusion to death. Cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS) were graded according to American Society for Transplantation and Cellular Therapy (ASTCT) guidelines.¹³ A Cox proportional hazards regression model incorporating a restricted cubic spline with 3 degrees of freedom¹¹ was used to evaluate the relationship between BMI and PFS. PFS and OS probabilities were estimated using the Kaplan-Meier method, and compared between subgroups with log-rank tests. Univariable and multivariate Cox proportional hazards models were developed for PFS and OS; hazard ratios (HR) were reported with 95% confidence intervals (CI). Statistically significant predictors (P < .05) in univariable models were included in multivariable models. BMI and baseline inflammatory laboratory value associations were modeled using linear regression. Analyses were performed with R 4.5.1.

Among 134 patients with MM (56.7% male) treated with anti-BCMA CAR T-cell therapy (idecabtagene vicleucel 51.5%, ciltacabtagene autoleucel 23.9%, investigational 24.6%), the median age was 67 years (39-85 years), median number of prior therapy lines was 4 (1-16), and 53% had prior autologous hematopoietic cell transplant (AHCT). At CAR T-cell infusion, 52.2% had extramedullary disease and 38.1% had high-risk cytogenetics. Median apheresis-to-infusion time was 41 days, and 73.9% received bridging therapy. Median BMI was 26.7 kg/m² (17.9-55.2 kg/m²). The median follow-up time from CAR T-cell infusion was 19.9 months (95% CI, 16.6-29.8).

A Cox model with restricted cubic spline revealed a nonlinear BMI-PFS relationship with increased HR in the mid-range (Figure 1B). Patients were grouped as normal weight (n = 49, 36.6%), overweight (n = 44, 32.8%), and obese (n = 41, 30.6%). Baseline characteristics were generally balanced between groups (Table 1), with more males in the overweight and obese groups (P < .001). In the overweight group, more patients received prior AHCT (P = .003), and had lower pre-LD white blood cell (P = .007) and platelet counts (P = .040). Diabetes was more prevalent in the obese group (26.8% vs 6.1% and 13.6%; P = .24); hypertension/hyperlipidemia and pre-LD C-reactive protein (CRP) also positively trended with BMI.

Best overall response rates were 87.8%, 75.0%, and 87.8% (P = .126), and complete response rates were 42.9%, 36.4%, and 56.1% (P = .185) in the normal weight, overweight, and obese groups, respectively (Figure 1C). More patients in the overweight group failed to achieve at least a partial response compared with the combined normal weight and obese groups (25.0% vs 12.2%, P = .040). For the entire cohort, 12-month PFS and OS were 47.8% (95% CI, 39.6-57.8) and 75.8% (95% CI, 68.4-84.0), respectively. Stratified by BMI, outcomes were worst in the overweight group (Figure 1D-E): 12-month PFS was 28.8% vs 51.9% (normal weight) and 62.6% (obese), P < .001; 12-month OS was 61.4% vs 82.9% (normal weight) and 84.2% (obese), P = .006. In univariable Cox models (supplemental Tables 1 and 2), overweight BMI predicted inferior PFS (HR, 2.35 [95% CI, 1.5-3.67]; P < .001) and OS (HR, 2.38 [95% CI, 1.32-4.28]; P = .004). In multivariable models adjusted for significant univariable covariates (eg, sex, number of prior lines of therapy, prior AHCT, CAR T-cell construct, high-risk cytogenetics, pre-LD hemoglobin, platelets, and CRP), overweight BMI remained independently associated with worse PFS (HR, 1.69 [95% CI, 1.03-2.77]; P = .038) and worse OS (HR, 2.33 [95% CI, 1.08-5.032]; P = .031; supplemental Tables 3 and 4). Kaplan-Meier analyses of PFS stratified by extramedullary disease, prior AHCT, prior therapy-line burden, and CAR T-cell construct consistently showed inferior outcomes in the overweight group (Figure 2A-I).

In the entire cohort, any-grade CRS occurred in 89.6% of patients, with grade ≥2 in 29.8% and grade ≥3 in 3.7%; any-grade ICANS occurred in 23.1%, with grade ≥2 in 16.4% and grade ≥3 in 6.7%. The 10-day cumulative incidence of grade ≥2 CRS trended higher in normal weight (36.7%) vs overweight (18.2%) and obese (26.8%) groups (P = .15; supplemental Figure 1A); grade ≥2 ICANS did not significantly differ (supplemental Figure 1B). Given known obesity-inflammation links, we modeled the relationship of BMI and pre-LD ferritin, CRP, and lactate dehydrogenase, but found no significant linear associations (ferritin P = .575, R² = 0.002; CRP P = .567, R² = 0.003; lactate dehydrogenase P = .316, R² = 0.007; supplemental Figure 2A-C).

Collectively, our data support a U-shaped BMI-outcome relationship after anti-BCMA CAR T-cell therapy in MM, with worse responses and survival in patients who were overweight compared with those with normal weight and those with obesity. A potential unifying hypothetical framework is an immunometabolic valley resulting in lower CAR T-cell efficacy in the overweight BMI range, with favorable tumor biology on one end of the BMI spectrum (normal weight) and improved CAR T-cell functionality or sensitivity on the other (obese; Figure 1A). Adipose tissue promotes plasma cell survival/proliferation via interleukin-6, interleukin-1b, tumor necrosis factor-α, and adipokines,¹ and obesity is a risk factor for MM progression.¹ In the longitudinal CoMMpass study, patients with obesity had significantly shorter PFS and OS compared with normal weight patients.¹⁴ Increased visceral fat was also associated with worse treatment response in a German study,¹⁵ thus obesity may negatively impact therapy responsiveness.¹⁶ Conversely, patients with obesity have superior ICI responses in numerous cancers.⁶ Proposed explanations for this paradox include the obese T-cell exhaustion phenotype, distinct cytokine profiles, and increased tumor immunogenicity due to impaired immunoediting that promote ICI responsiveness.^17,18 However, it is unclear if obesity influences CAR T-cell outcomes. Prior work on BMI and anti-CD19 CAR T-cell response in B-cell lymphoma and acute lymphoblastic leukemia have been mixed: 1 study reported improved outcomes with obesity,¹⁹ whereas other studies found no association.^20,21 Given that CAR T-cell efficacy is the result of myriad factors, including tumor biology, quality of the apheresis T-cell pool, product phenotype, host responses to LD, preinfusion inflammation, postinfusion expansion, immune cell trafficking, and CAR T-cell persistence, our observation that obesity correlates with better outcomes after anti-BCMA CAR T-cell therapy in MM warrants targeted mechanistic study.

Caveats of our study include (1) the single-center design and CAR T-cell product heterogeneity, resulting in possible construct-specific effects, (2) unintentional selection bias of patient referrals that may have enriched the overweight group for unmeasured adverse risk factors, (3) combination of underweight (BMI < 18.5 kg/m²) with normal weight (BMI 18.5-25 kg/m²) groups together as BMI <25 kg/m², as the underweight cohort was too small to analyze independently, thus the potential effect of cachexia on anti-MM CAR T-cell responses remains unaddressed, and (4) assessment of BMI only at time of LD, without longitudinal trajectories, as preinfusion weight loss has been linked to worse outcomes with anti-CD19 CAR T-cell therapy.²² 

Finally, BMI is an imperfect surrogate that may not reflect adipose mass or metabolic health.²³ To better delineate adipose-specific effects, studies utilizing comprehensive body composition analysis are warranted.⁵ Nevertheless, we provide evidence that both normal weight and obese patients with MM have improved outcomes with anti-BCMA CAR T-cell therapy compared with those who were overweight. Further validation and mechanistic research are essential to elucidate the underlying biological processes. This research could uncover novel therapeutic targets or modifiable factors to be leveraged to improve CAR T-cell therapy outcomes in MM.

A.Y. receives research funding from Amgen, Bristol Myers Squibb, GSK, Johnson & Johnson, and Sanofi. N.R. receives funding from the Multiple Myeloma Research Fund Challenge Award, and the Paula and Rodger Riney Foundation.

Contributions: R.C., K.M., D.C., and N.R. conceived the project idea; R.C., K.M., A.J., C.M., S.S.H., A. Barselau, and D.C. were responsible for collecting patient data and constructing a retrospective patient database; K.M. performed the statistical analyses; R.C., K.M., and D.C. wrote the manuscript; A.Y., A. Branagan, B.P., M.F., N.R., and D.C. were the physicians who took care of the patients included in the study and helped provide data for analysis; and A.Y., A. Branagan, B.P., M.F., and N.R. provided feedback on the manuscript.

Conflict-of-interest disclosure: A.Y. is a consultant for AbbVie, Adaptive Biotechnologies, Amgen, Bristol Myers Squibb (BMS), Celgene, GSK, Johnson & Johnson, Karyopharm, Oncopeptides, Pfizer, Prothena, Regeneron, Sanofi, Sebia, and Takeda. M.F. is a consultant for BMS, Kite/Gilead, Johnson & Johnson/Legend, Cytoagents, and SOBI. The remaining authors declare no competing financial interests.

Correspondence: Diana Cirstea, Center for Multiple Myeloma, Yawkey Building Suite 9A, Massachusetts General Hospital, Boston, MA 02114; email: dcirstea@mgh.harvard.edu.