Долгосрочная безопасность и эффективность лечения ЦАР-клеток на базе БКМА в нескольких миэломах: сравнительные данные из Cilta-cel, Ide-Sel, Anito-Sel и FCARH143.

Introduction

Over the past decade, chimeric antigen receptor (CAR) T-cell therapies have transformed the treatment landscape for several hematologic malignancies, including multiple myeloma (MM). Prior to CAR T availability, heavily pre-treated patients with relapsed/refractory MM (RRMM) had a poor prognosis, with a median progression-free survival (mPFS) of below 6 months and median overall survival (mOS) of about a year ( 1 , 2 ). However, B-cell maturation antigen (BCMA)-directed CAR T-cell therapies have transformed these inferior outcomes, underscoring their potential as durable treatment options for patients with heavily pretreated, multi-class resistant RRMM.

While early CAR T trials focused on response rates and acute toxicity, long-term follow up studies target durability, relapse kinetics, and long-term safety. This shift is particularly relevant as CAR T-cell therapy moves earlier in the treatment course, where lasting remission and reduced toxicity define clinical value. Idecabtagene vicleucel (ide-cel) and ciltacabtagene autoleucel (cilta-cel) were the first BCMA-directed T-cell therapies to be approved by the Food and Drug Administration (FDA) and thus have the longest follow up data ( 3 , 4 ). Since their approval, newer constructs such as FCARH143 and anitocabtagene autoleucel (anito-cel), have been structurally refined to improve CAR T-cell persistence in vivo and minimize toxicity ( 5 ). In this context, Tuazon et al. recently reported the five-year follow-up of a first-in-human phase I study evaluating FCARH143, a fully human BCMA-directed CAR T-cell therapy incorporating a 4-1BB costimulatory domain and a defined CD4:CD8 ratio ( 6 ). Although direct comparisons are limited by cross-trial heterogeneity, existing data allow for some discussion of different CAR T products and their respective significance on translational and clinical levels.

Comparative efficacy: survival and duration of response (DoR)

Cilta-cel and ide-cel

Cilta-cel has recently demonstrated a field reshaping efficacy in heavily pretreated RRMM with a long-term follow up report ( 7 , 8 ). A 5-year follow-up of the phase Ib/II CARTITUDE-1 trial showed that 32/97 (33%) patients treated with a single cilta-cel infusion remained progression-free with a mOS of 60.7 months, with 31/32 (97%) patients achieving a complete response (CR) or stringent CR (sCR). Importantly, patients had a median of 6 prior lines of therapy (PLT) and none received maintenance treatment post cilta-cel. More recently, in a retrospective, multicenter study of 236 patients with RRMM who were treated with cilta-cel post FDA-approval, at a median follow up of 13 months, overall response rate (ORR) and complete response rate (CRR) were 89% and 70%, respectively ( 9 ). Importantly, more than half of these patients did not meet the CARTITUDE-1 eligibility criteria, mainly due to comorbidities or treatment history.

In contrast, in the phase II KarMMa trial, where 128 patients who had at least 3 PLT for MM received ide-cel, at a median follow-up of 13.3 months, ORR and CRR were 73% and 33% respectively with a median PFS and mOS being 8.8 and 19.4 months respectively ( 10 ). In a 5-year follow up of the KarMMa trial presented at the International Myeloma Society meeting in 2024, the ORR and CRR were 76.6% and 34.3% respectively while the updated median PFS and mOS were 10.8 and 28.3 months respectively ( 11 ).

In a real-world retrospective study by Hansen et al. of 586 RRMM patients treated with cilta-cel (n=236) and ide-cel (n=350), at a median follow up of 13 and 12.6 months, respectively, cilta-cel was associated with a 2.4-fold higher chance of achieving a CR [95% confidence interval (CI): 1.63–3.60], and significantly increased PFS [hazard ratio (HR) 0.48; 95% CI: 0.36–0.63], and OS (HR 0.67; 95% CI: 0.46–0.97) compared to ide-cel ( 12 ). Together, these data underscore the efficacy and durability of cilta-cel compared to ide-cel in treating RRMM.

Anito-cel

Anito-cel is a next-generation, fully human BCMA-directed CAR T construct designed to enhance persistence and reduce immunogenicity with promising early efficacy signals from the phase II iMMagine-1 trial ( 5 ). This study evaluated 117 patients treated with anito-cel following a median of 3 PLT. At a median follow-up of 12.6 months, ORR was 97% and CR/sCR was 68%. At 12 and 18 months, PFS rates were 79% and 66%, respectively and OS rates were 95% and 90% respectively.

FCARH143

The phase I study on FCARH143, a BCMA CAR T cell with a 4-1BB costimulatory domain delivered in a predefined CD4 + :CD8 +  ratio, provides a mature 5-year follow up ( 6 ). In a small cohort of 25 patients with a median of 8 PLT who were treated with FCARH143, ORR was 100% and CR/sCR 68% (64% sCR). At 67.3 months follow-up, the mPFS and mOS were 15.5 and 32.1 months, respectively. Median DoR was 14.6 months (95% CI, 7.8–26.2) and median time to best response was 60 days. This study also reported zero treatment-related deaths, only 4% grade 3 cytokine release syndrome (CRS) events and zero grade 3 immune effector cell-associated neurotoxicity syndrome (ICANS) events by American Society for Transplantation and Cellular Therapy (ASTCT) criteria, highlighting the presumed long-term safety of FCARH143. Despite the challenge of cross-trial comparison, FCARH143 appeared to signal greater efficacy than ide-cel. However, while the sCR presents promising evidence of depth of response, the short mPFS calls into question the long-term durability of FCARH143.

CAR T constructs and relapse kinetics

Relapse kinetics, defined by metrics like time to relapse and rate of disease progression, differ significantly across BCMA-directed CAR T therapies. Unfortunately, despite high early response rates, relapse is common. According to a meta-analysis of 29 CAR T studies in patients with RRMM (n=1,395), including those treated with cilta-cel and ide-cel, the pooled ORR was 86% with a mPFS of 9.88 months and median DoR of 12.17 months ( 13 ). Across all four CAR T products outlined above, within the limitation of cross-trial comparisons, Ide-cel and FCARH143 appear to have shorter durability of disease control, reflected by shorter median PFS (10.8 vs. 15.5 months, respectively) ( 6 , 10 , 11 ). It is important to consider that the FCARH143 study enrolled a more heavily pre-treated population compared to the CARTITUDE and KarMMa studies (median 8 vs. 6 PLTs, Table 1 ) ( 6 , 8 , 10 , 11 ).

Table 1 Comparative characteristics and clinical outcomes of selected BCMA-directed CAR T therapies in relapsed/refractory multiple myeloma Variable FCARH143 Ciltacabtagene autoleucel (cilta-cel) Idecabtagene vicleucel (ide-cel) Anito-cel (CT103A) Construct/design Fully human scFv; 4-1BB costimulatory domain; CD3ζ signaling; defined 1:1 CD4:CD8 ratio; truncated EGFR for tracking Llama-derived; dual epitope BCMA-binding domains (bivalent); 4-1BB co-stim.; CD3ζ signaling Murine scFv; 4-1BB costim.; CD3ζ signaling Fully human scFv; 4-1BB co-stim.; CD3ζ signaling Trial phase/population Phase I, heavily pretreated RRMM (median 8 PLT; 80% TCR); EMD 44% Phase Ib/II (CARTITUDE-1); RRMM ≥3 PLT (median 6 PLT; 87.6% TCR); EMD 13.4% Phase III (CARTITUDE 4); RRMM (1–3 PLT; 15% TCR); EMD 21% Phase II (KarMMa); RRMM ≥3 PLT (median 6 PLT; 84% TCR); EMD 39% Phase III (KarMMa-3); RRMM (2–4 PLT; 66% TCR); EMD not specified Phase II (iMMagine-1, ASH 2025); RRMM (median 3 PLT; 86% TCR); 15% EMD N (treated patients) 25 97 176 137 254 117 Overall response rate (ORR) 100% 98% 85% (ITT) 76.6% 71% 96% Complete response rate (sCR/CR) 68% CR + sCR 83% sCR 77% ≥ CR (ITT) 34.3% CR + sCR 39% CR 74% CR + sCR Median PFS (months) 15.5 34.9 NR vs. 11.8 in cilta-cel vs. SOC (30-mo PFS rate 59.4% for cita-cel vs. 25.7% for SOC) 10.8 13.3 vs. 4 in ide-cel vs. SOC NR (12-month PFS: 82.1%, 18-month 67.4%) Median OS (months) 32.1 60.7 NR in either arm (30-mo OS rate 76.4% for cita-cel vs. 63.8% for SOC) 28.3 (12-month OS 78%) NR NR (24-month OS: 83%) Median follow-up (months) 67.3 61.3 33.6 63.6 18.6 15.9 (data cut-off: Oct 2025) MRD status Not specified Sustained MRD-negative in 26/49 evaluable patients at 10 –5 threshold ≥12 month. Sustained MRD negativity in long-term responders: 12 from a single center all MRD-negative at year 5 MRD negative for cilta-cel vs. SOC was 62% vs. 18% at 10 –5 threshold and 57% vs. 9% at 10 –6 threshold, assessed at any time before initiation of subsequent therapy. Sustained MRD negative CR: 52% in cilta-cel vs. 10% in SOC MRD-negative at 10 –5 threshold in 26% of treated and 79% of CR or sCR patients MRD-negative in 20% of patients (51/254) in ide-cel arm vs. 1% in SOC within 3 months at 10 –5 threshold MRD-negative in 95% of evaluable patients at 10 –5 threshold and 78% at 10 -6 threshold. 83% had ≥ 6 months of MRD negativity. CRS incidence (all grades) 84% (mostly G1–2); G ≥3: 4% 95%; G ≥3: 5% 75% G1–2 and 1% G3 in cita-cel group 84% any grade; 5.4% G ≥3 Not reported 85% any grade; G ≥3: 1% ICANS incidence (all grades) 24%; no grade ≥3 (ASTCT criteria) 21%; grade ≥3: 10% (includes delayed movement disorders) 18% G1–2 and 3% G3 in cilta-cel group 18% all grades; 3.6% grade ≥3 Not reported 8% any grade; G ≥3: 1% Cytopenias (neutropenia, thrombocytopenia, anemia) N =100% any grade, 25% G3–4; T =72% any grade, 25% G3–4; A =72% any grade, 25% G3–4 N = not specified; T = not specified; A = not specified N =90% any grade, all G 3–4; T =55% any grade; 42% G3–4; A =55% any grade, 36% G 3–4 *N =91% any grade, 89% G3–4; *T =63% any grade, 52% G3–4; *A =70% any grade, 60% G3–4 N =78% any grade, 76% G3–4; T =54% any grade, 42% G3–4; A =66% any grade, 51% G3–4 N =70% G3–4; T =26% G3–4; A =29% G3–4 Infections 48% any grade, 100% G4 4 infections total (% of patients not specified), 100% G3 63% any grade, 28% G3–5 *69% any grade, 22% G3–4 58% any grade; 24% G3–4; 4% G5 9% G3–4 Unique features/translational aspects Fully human scFv reduces immunogenicity; defined CD4:CD8 ratio enhances reproducibility; favorable persistence and outpatient feasibility Dual BCMA epitope binding enhances avidity; robust persistence; late-onset neurotoxicity risk Comparative study against SOC First FDA-approved BCMA CAR T; standardized manufacturing; limited persistence Comparative study against SOC Fully human scFv; favorable persistence; low toxicity enabling 8% outpatient use in Phase II trial Primary reference(s) Tuazon SA et al. , Blood 2025 ( 6 ) Jagannath S et al. , JCO 2025 ( 8 ); Lin Y et al. , ASCO Meeting 2023 ( 14 ) Einsele H et al. , Lancet Oncology , 2026 ( 15 ) Anderson L et al. , IMS 2024 Meeting ( 11 ); *Munshi NC et al. , N Engl J Med 2021 ( 10 ) Rodriguez-Otero et al. , NEJM 2023 ( 16 ); Ailawadhi S et al. , Blood 2024 ( 17 ) Patel K et al. , Blood 2025 ( 5 ) A, anemia; ASCO, American Society of Clinical Oncology; ASH, American Society of Hematology; 4-1BB, CD137 costimulatory domain; BCMA, B-cell maturation antigen; CAR, chimeric antigen receptor; CD, cluster of differentiation; CD3ζ, CD3 zeta signaling domain; CR, complete response; CRS, cytokine release syndrome; EGFR, epidermal growth factor receptor; EMD, extramedullary disease; FDA, Food and Drug Administration; G, grade; ICANS, immune effector cell-associated neurotoxicity syndrome; IMS, International Myeloma Society; ITT, intention-to-treat population; MRD, minimal residual disease; N, neutropenia, NR, not reached; OS, overall survival; PFS, progression-free survival; PLT, prior line of therapy; RRMM, relapsed/refractory multiple myeloma; scFv, single-chain variable fragment; sCR, stringent complete response; SOC, standard of care; T, thrombocytopenia; TCR, triple class refractory.

Comparative characteristics and clinical outcomes of selected BCMA-directed CAR T therapies in relapsed/refractory multiple myeloma

A, anemia; ASCO, American Society of Clinical Oncology; ASH, American Society of Hematology; 4-1BB, CD137 costimulatory domain; BCMA, B-cell maturation antigen; CAR, chimeric antigen receptor; CD, cluster of differentiation; CD3ζ, CD3 zeta signaling domain; CR, complete response; CRS, cytokine release syndrome; EGFR, epidermal growth factor receptor; EMD, extramedullary disease; FDA, Food and Drug Administration; G, grade; ICANS, immune effector cell-associated neurotoxicity syndrome; IMS, International Myeloma Society; ITT, intention-to-treat population; MRD, minimal residual disease; N, neutropenia, NR, not reached; OS, overall survival; PFS, progression-free survival; PLT, prior line of therapy; RRMM, relapsed/refractory multiple myeloma; scFv, single-chain variable fragment; sCR, stringent complete response; SOC, standard of care; T, thrombocytopenia; TCR, triple class refractory.

Additional exposure to cytotoxic therapies can induce a shift in T-cell phenotype and upregulation of exhaustion markers, diminishing the cytotoxicity of CAR T products and potentially durability of response ( 18 , 19 ). Early data from the iMMagine-1 anito-cel study revealed that 66% of patients remained progression free at 18 months ( 5 ) while no median PFS or OS for anito-cel have been reached at the time of this writing. In contrast, cilta-cel exhibits a clear long-term “tail” with one third of patients in sustained remissions off any maintenance therapy beyond five years ( 8 ).

Even though no formative head-to-head trial has compared cilta-cel and ide-cel to date, retrospective data provide strong evidence that cilta-cel may be superior in terms of PFS and OS ( 9 , 12 ). Moreover, retrospective data from the 236 patients with RRMM treated with cilta-cel post FDA-approval reveal an overall response of 89% comparable to 85% in the phase 3 trial comparing cilta-cel to standard of care (SOC) ( 9 , 15 ).

Minimal residual disease (MRD)

MRD assessment has become a critical measure of response to treatment and a surrogate marker of long-term outcomes in MM ( 20 ). However, many studies fail to consistently track MRD across all patients and clinical trial timepoints. Most studies assess MRD using bone marrow aspirates and next-generation sequencing (NGS) to determine the presence of disease on a molecular level. Results are typically reported at a sensitivity of 10 –5 , with some newer datasets reporting thresholds of 10 –6 ( 21 ).

In the CARTITUDE-1 trial, approximately one-third of patients were alive and progression-free for at least 5 years and these patients had similar baseline characteristics (i.e., high-risk cytogenetics, extramedullary disease) as those who progressed within 5 years ( 7 , 8 , 14 ). The 30-month PFS rate was 54% in the overall population compared with 75% in those with sustained MRD negativity ( 14 ). Of 49 evaluable patients with 12-month sustained MRD negativity, 26 (53%) were MRD negative at the 10 −5 threshold. A subset of 12 patients in sCR underwent MRD monitoring and 100% of them were MRD-negative (at 10 −5 threshold) at 5 years, highlighting the depth and DoR related to cilta-cel ( 8 ). Sustained MRD negativity was also associated with prolonged PFS, suggesting that it may predict long-term response.

In the phase II KarMMa trial, 26% of all 128 patients treated with ide-cel and 79% of the 42 patients in CR achieved MRD negativity at <10 –5 ( 10 ). These results differ from CARTITUDE-1, which also enrolled heavily pretreated patients ( Table 1 ) ( 15 ). In contrast, in the iMMagine-1 trial, 93% (70/75) of evaluable patients achieved MRD negativity at the 10 –5 threshold and 78% (53/68) at 10 −6 threshold, representing one of the deepest molecular responses reported to date ( 5 ).

Importantly, MRD assessments have limitations as they do not always align with other trial timepoints. As a result, correlative data (i.e. serum samples, imaging) are not always collected at the same timepoint for comparison. Additionally, cross trial variability makes it difficult to interpret MRD results across CAR T products.

Long-term safety

Acute toxicities such as CRS and ICANS are well described for BCMA-directed CAR T-cell therapies. Delayed toxicities are not as well defined in the literature but more data are emerging with longer follow up. For example, immune effector cell associated parkinsonism, enterocolitis, and neuropathy are increasingly being recognized as delayed toxicities of cilta-cel ( 22 - 24 ). Real-world follow-up data from long term registrational trials can help differentiate the safety and toxicity of individual CAR T-cell therapies, keeping in mind the impact of construct-specific binding domains, kinetics and receptor level density of engagement.

Toxicity is an important factor when considering the long-term potential of CAR T-cell therapy. Better toxicity profiles are often associated with increased safety, patient adherence, and accessibility. For patients treated with FCARH143, Tuazon et al. reported CRS and neurotoxicity in 84% (8% grade 3–4 and no grade 5) and 24% (12% grade 3 and no grade 4–5) of patients, respectively ( 6 ). In contrast, in the iMMagine-1 trial, CRS occurred in 85% and neurotoxicity in 7% of patients treated with anito-cel ( 5 ). Finally, in a multi-center real-world study comparing ide-cel to cilta-cel, Merz et al. reported similar rates of CRS and ICANS in both groups (81% and 19% for cilta-cel versus 85% and 19% for ide-cel) ( 25 ). Taken together, newer constructs like FCARH143 appear to have similar toxicity profiles to other BCMA-directed CARs.

Delayed neurotoxicity (NT) and immune effector cell-associated enterocolitis (IEC-EC) have become important therapy-related adverse events, notably more common in patients treated with cilta-cel. A retrospective study using the International Myeloma Working Group database reported parkinsonism (IEC-PKS) and cranial nerve palsies (IEC-NP) as the most common delayed NT in RRMM patients treated with CAR T (52 patients; 92% cilta-cel, 4% ide-cel, 4% investigational product) ( 26 ). In another retrospective analysis of 235 patients treated with cilta-cel, IEC-NP and IEC-PKS occurred in 6.4% and 3.8% of patients, respectively. High peak absolute lymphocyte count ≥3×10 9 /L was a significant predictor of delayed NT (negative predictive value of 98%), and early cyclophosphamide administration (1–13 days post-NT symptom onset) was associated with improvement of IEC-PKS in four patients ( 20 ).

Separately, Fortuna et al. report 14 cases of IEC-EC (2.2% cilta-cel, 0.2% ide-cel) out of 1287 patients treated with CAR T across 11 centers. Median symptom onset was 92.5 days post CAR T, and 36% of patients with IEC-EC died due to bowel perforation or sepsis ( 22 ). In a larger study of 345 patients treated with cilta-cel, total IEC-EC incidence was 1.4% with no associated deaths ( 27 ). The discrepant rates of delayed NT in patients treated with cilta-cel vs. ide-cel may be related to CAR construct design, as cilta-cel has been shown to produce high-avidity CAR T-cells which could explain the on-target, off-tumor toxicities ( 28 ).

Limitations

Comparing outcomes from various BCMA-directed CAR T-cell products is limited by distinct study populations and varying trial designs. Factors such as sample size, PLT, bridging therapy, prior therapies exposure, and disease burden can contribute to major variation in response rates. Similarly, median follow up period of studies can vary significantly, as well as MRD sampling type, depth (10 –5 vs. 10 –6 ), time, and frequency, making cross comparisons challenging. To date, none of the CAR T products have been reported in head-to-head trials with other forms of T cell redirecting therapies such as CAR T or bispecific antibodies.

CAR design and translational insights

Variations in CAR constructs and BCMA target binding play key roles in determining the safety and efficacy of BCMA-directed CAR T-cell products ( 29 ). Ide-cel and cilta-cel both incorporate a 4-1BB costimulatory domain which facilitates the formation of memory T-cells and contributes to CAR T persistence ( 7 , 8 , 10 ). However, cilta-cel features a dual BCMA-binding domain, which allows the construct to bind to two epitopes instead of one ( 7 , 30 , 31 ). This increases cilta-cel’s functional avidity and likely contributes to deeper response rates and sustained MRD negativity. This may also explain the instances of delayed neurotoxicity ( 30 ). In comparison, the anito-cel construct features a synthetic D-domain designed to engage BCMA with greater precision to achieve a fast “off-rate” enabling quick elimination of cancer cells while avoiding prolonged T-cell activation and, in turn, exhaustion ( 5 ). In addition to CAR construct design, immune microenvironment and T-cell fitness play a major role in CAR T-cell efficacy.

The 5-year follow-up from CARTITUDE-1 revealed that higher ratios of naive to mature T-cells, T-cells to neutrophils, and effector to target as well as higher baseline hemoglobin and platelets were associated with PFS at five years ( 8 ). Of note, CAR-HEMATOTOX (HT) score may not only identify high-risk candidates for severe toxicity prior to lymphodepleting chemotherapy, but also predict inferior response rates at day 90 for patients receiving CAR T ( 32 ). Additionally, in a small retrospective study, Tabbara et al. found that patients with severe or prolonged cytopenias after CAR T were at higher risk of relapse ( 33 ). Taken together, these data suggest baseline tumor burden and bone marrow reserve may be key predictors of CAR T safety and efficacy in patients with RRMM. Baseline inflammation may also disrupt optimal T-cell collection and expansion ex vivo , leading to suboptimal T-cell subset ratios and thus efficacy in vivo . Moving CAR T-cell therapy into earlier lines may be beneficial as patients are less likely to have baseline cytopenias and increased inflammation from PLT.

Conclusions

As BCMA-directed CAR T-cell therapy shifts to earlier lines of treatment, beyond the long-term efficacy signal, toxicity/safety profiles play an increasingly important role in the measure of both therapy adoption and its overall success. So far, cilta-cel has set the efficacy benchmark for sustained remission and opened the door to consideration of a “functional cure” in at least a subset of MM patients for the first time in the field. Next-generation constructs such as anito-cel and FCARH143 remain to reliably demonstrate a favorable balance of efficacy and safety in larger cohorts with longer follow up. Future studies should emphasize standardized MRD testing, extended follow-up, and comparative effectiveness in real-world populations. Understanding relapse mechanisms, including antigen escape and T-cell exhaustion, is critical to refining next-generation therapies and treatment sequencing in search of an optimal balance between long-term therapeutic efficacy and safety in all stages of MM disease.

Supplementary

The article’s supplementary files as