Ide-cel is a second-generation CAR which uses a lentiviral vector to transduce a BCMA targeting scFv fused to a 4-1BB co-stimulatory and CD3 signaling domains (32, 33). The pivotal phase II KarMMa study evaluated ide-cel at various doses in 128 RRMM patients who had previously received ≥ 3 prior lines of therapy including an immunomodulatory drug (IMiD), proteasome inhibitor (PI), and an anti-CD38 monoclonal antibody. Patients were infused with 150×10⁶ to 450×10⁶ CAR T cells. Overall response rate (ORR) was 73%, with 42 (33%) patients achieving a complete response (CR) or better. Measurable residual disease (MRD) negativity (at 10^-5) was achieved in 26% of all patients, but 79% of patients who achieved a ≥ CR. Notably, these response rates were relatively preserved across patients with high-risk features including penta-refractory disease, extramedullary disease, and high-risk cytogenetics. Median PFS was 8.8 months, but was 20.2 months in patients who achieved ≥ CR (11). OS for the KarMMa study was 24.8 months, but interestingly, was lower in the cohort of patients who had previously been treated with 3 prior lines (median OS 22 months) versus those who had been treated with 4 or more prior lines (median OS 25.2 months) (34). Based on these findings, ide-cel was approved for the treatment of adults with RRMM after four or more prior lines of therapy including an IMiD, PI, and anti-CD38 monoclonal antibody by the US Food and Drug Administration (FDA) in March 2021. In a post-hoc analysis of the KarMMa trial, lower levels of serum soluble BCMA, more robust blood and bone marrow CAR T expansion, and an increased ratio of naive and early memory CD4 T cells compared to senescent CD3 and CD8 T cells in the apheresis product prior to manufacturing were associated with improved response to ide-cel (35). Subsequently, real-world data for patients treated with commercial ide-cel outside the context of a clinical trial showed very similar efficacy. ORR in this population was 84% with 42% achieving ≥ CR. Median PFS (8.5 months) at 6.1 months of follow-up was similar to the KarMMa data. Notably, 75% of the patients included in this cohort would have been deemed ineligible to enroll on the KarMMa trial (36). Predictors of poor response to ide-cel in the real-world cohort included prior BCMA therapy, high-risk cytogenetics, elevated baseline ferritin level, and younger age (36, 37).

More recently ide-cel was evaluated against investigator’s choice of one of 5 SOC regimens: daratumumab, pomalidomide, and dexamethasone; daratumumab, bortezomib, and dexamethasone; ixazomib, lenalidomide, and dexamethasone; carfilzomib and dexamethasone; or elotuzumab, pomalidomide, and dexamethasone in RRMM patients who had received 2-4 prior lines of therapy including an IMiD, PI, and an anti-CD38 monoclonal antibody in the phase III KarMMa-3 trial. This trial enrolled a substantial population (43%) of patients with high-risk cytogenetics (defined as presence of del17p, t[4;14], or t[14;16]) which were evenly distributed across both arms. The ORR was 71% for ide-cel and 42% for SOC. Median PFS in the intention-to-treat population was substantially higher in the ide-cel arm (13.3 months vs 4.4 months). The hazard ratio for disease progression or death for ide-cel vs SOC was 0.49 (95% CI 0.38 - 0.65). Ide-cel showed similar improved hazard ratios for disease progression or death in patients with high-risk cytogenetics (HR 0.61; 95% CI 0.41-0.90), extramedullary disease (HR 0.40; 95% CI 0.25-0.65), and disease refractory to at least one IMiD, PI, and anti-CD38 monoclonal antibody (HR 0.46; 95% CI 0.34-0.62) (14, 38). The FDA is currently reviewing a supplemental biologics license for the approval of ide-cel in this less heavily pretreated population based on the data from the KarMMa-3 trial. Additionally, cohort 2c of the phase II KarMMa-2 trial is evaluating the efficacy of ide-cel in patients with newly diagnosed MM (NDMM) who had an inadequate response to frontline AHCT. In this population, 77% of patients achieved ≥ CR, and median PFS has not be reached at a median follow-up of 39.4 months (39).

Similar to ide-cel, cilta-cel uses a lentiviral vector to create a construct with a CD3ζ activation domain, and 4-1BB costimulatory domain. Cilta-cel’s antigen binding domain contains bispecifc scFvs targeting two distinct BCMA epitopes, VHH1 and VHH2 (40). This bi-epitope binding confers higher avidity and specificity to BCMA. Cilta-cel was evaluated in the phase Ib/II CARTITUDE-1 trial RRMM patients who had previously been treated with ≥ 3 or were double refractory to an IMiD and a PI, and had previously received an anti-CD38 antibody. Ninety-seven patients were treated with cilta-cel (29 in the phase Ib portion, 68 in the phase II portion). The population was heavily pretreated (median of 6 prior lines; 84% penta-exposed) and included 42% who were penta-refractory (12). ORR was 98%, with 95% of patients achieving a VGPR or better and 82.5% of patients achieving a sCR. MRD negativity was evaluated in 61 patients at 10^-5 and 52 patients at 10^-6. MRD negativity rates were 92% and 75% at 10^-5 and 10^-6 respectively (41). Median PFS was 34.9 months; median OS has not been reached at 27.7 months of follow-up (42). Based on these data cilta-cel was approved for the treatment RRMM patients following 4 or more prior lines of therapy, including an IMiD, PI, and an anti-CD38 monoclonal antibody in Feb 2022. Real-world data for cilta-cel is not as mature as that for ide-cel. However, in an early analysis a multi-institutional cohort of 143 patients infused with commercial cilta-cel-of whom 57% would have been ineligible for participation in the CARTITUDE-1 trial-ORR was 84% with 53% ≥ CR. Notably, 22% of patients included in this dataset were infused with an out of specification (OOS) cilta-cel product. The presence of high-risk cytogenetics (defined as the presence of del17p, t[4;14)], or t[14;16]) was associated with poorer ORR, PFS, and OS in multivariate analysis (43).

Cilta-cel was also evaluated in less heavily pretreated patients in the phase III CARTITUDE-4 trial. This trial randomized 419 RRMM patients previously treated with 1-3 prior lines to either cilta-cel or investigator’s choice of 2 SOC regimens: pomalidomide, bortezomib, dexamethasone, or daratumumab, pomalidomide, dexamethasone. The trial strongly favored the cilta-cel arm with ORR of 84.6% vs 67.3% for SOC. Responses were deeper in the cilta-cel arm (73.1% vs 21.8% ≥ CR) which translated into improved PFS (median PFS not reached for cilta-cel, 11.8 months for SOC; HR 0.26; 95% CI, 0.18 -0.38). Subgroup analysis also favored cilta-cel for all parameters tested including high-risk cytogenetics, prior lines of therapy, degree of refractoriness, and presence of extramedullary disease (13, 44). Cilta-cel is currently under evaluation for patient with suboptimal response to frontline transplant, in treatment naïve high-risk and standard risk NDMM in cohorts D, E, F of the mutli-cohort CARTITUDE-2 trial.

While these data illustrate the efficacy of both ide-cel and cilta-cel, both of these agents have important toxicities that need to be factored into any suitability discussion. Cytokine release syndrome (CRS) is a systemic inflammatory response thought to be secondary to activation of bystander immune and nonimmune cells resulting in significant cytokine release-especially IL-1, IL-2, IL-6, GM-CSF, and IFN-g (45–48). This inflammatory storm typically manifests as fever, fatigue, headache, arthralgias, and myalgias, but higher-grade manifestations including hypotension, shock, disseminated intravascular coagulation, and multiorgan system failure can occur (49, 50). Treatment of CRS ranges from supportive care with antipyretics, intravenous fluids, and supplemental oxygen for lower grade symptoms to vasopressors, the anti-IL-6 antibody tocilizumab, and high-dose corticosteroids for higher grade symptoms (50, 51). While low grade CRS is very common with both ide-cel and cilta-cel the timing of CRS onset varies for each CAR-T product (see Table 1 ). Ide-cel was associated with 84% CRS in the KarMMa trial with the majority being low grade (5% grade 3 or 4). Median onset of CRS with ide-cel in the KarMMa trial, the KarMMa-3 trial, and the real-world dataset was reliably 1 day (range 0-14) (11, 14, 36). In the CARTITUDE-1 trial 95% of patients experienced CRS with 4% grade 3 or 4. Median onset of CRS with cilta-cel in the CARTITUDE-1 trial was 7 days, while it was 8 days in the CARTITUDE-4 trial, and 9 days in the real-world cohort (12, 13, 43). These data are summarized in Table 1 .

Immune effector cell-associated neurotoxicity syndrome (ICANS) is another adverse effect common to CAR-T therapy. ICANS is a toxic encephalopathy thought to be related to endothelial cell activation and disruption of the blood brain barrier mediated by inflammatory cytokines and chemokines, which results in direct neuronal cell injury. Mild symptoms include headache, confusion, focal neurologic deficits, and impaired fine motor skills. Higher grade ICANS can manifest with aphasia, seizure, cerebral edema, and coma (49). The mainstay of ICANS management is high-dose corticosteroids; additional supportive measures including mechanical ventilation (if evidence of airway compromise) may be needed. Tocilizumab is generally only used if patients have coexisting CRS (50, 51). ICANS typically manifests later than CRS (around day 2-3 with ide-cel and days 8-10 with cilta-cel) (11–14, 36, 43). An additional neurologic toxicity seemingly unique to cilta-cel is less well understood. Termed movement and neurocognitive treatment-emergent adverse events (MNTs), they compromise a cluster of movement (e.g., micrographia, tremors), cognitive (e.g., memory loss, disturbance in attention), and personality changes (e.g. reduced facial expression, flat affect) which typically manifest after symptoms of CRS and ICANS have resolved, and unlike ICANS, are generally not responsive to steroids (52, 53). While the unique AEs associated with CAR-Ts are certainly cause for concern, with increasing ubiquity of this class of therapy and earlier recognition of, and intervention for, CRS and ICANS the rates of higher-grade AEs are decreasing in more recent trials (summarized in Table 1 ). Ideally, less high-grade CRS and ICANS will lead to less delayed neurotoxicity and MNT events.

Finally, the risk of secondary primary malignancies, long known to be an adverse effect of several myeloma therapies is largely unknown post-CAR T. However, recent data does raise concern for a possible increased risk. Specifically, rare reports of T-cell lymphomas derived from CAR T cells have been reported in several CAR T recipients. To date, only 12 cases have been reported out of the 7946 patients infused with CAR Ts, indicated predominately for B-cell lymphomas (54). Only 1 case has been reported in a myeloma patient who was treated with cilta-cel, but analysis of the patient’s apheresis product (prior to CAR T manufacture) suggested that they had several genetic mutations present at baseline and the role of the CAR is unclear (55). Similarly, myeloid malignancies have been reported with post CAR T. In the long-term follow-up of the CARTITUDE-1 trial 9 patients (9%) have been diagnosed with MDS or AML (42). However, it is unclear whether this a result of the CAR T or lymphodepleting chemotherapy. To that effect, a recent analysis of 4 patients who developed MDS after treatment with an investigational anti-BCMA CAR T showed that while none of the patients had morphologic changes consistent with MDS prior to CAR T infusion, all four patients exhibited molecular alterations associated with MDS in their pre-CAR T as well as post-CAR T therapy bone marrow with no new mutations observed after CAR T (56). Clearly further follow-up is warranted.

Cytogenetics, and more recently MRD status, have been shown to correlate with survival. In a 2022 evaluation of the impact of AHCT with quad-therapy induction in NDMM, MRD was assessed by NGS pre and post AHCT, and the group with the greatest reduction in MRD burden had high-risk cytogenetics (HRCG) demonstrating a ‘dose effect’ with stepwise greater reduction in those with 0, 1 or 2+ HRCG abnormalities (72). Those with more than 2 HRCG abnormalities – so called ultra-high risk – have worse outcomes as demonstrated in subgroup analysis of MASTER and GRIFFIN trials (73). Though among ultra-high risk patients, those who achieve MRD negativity prior to or after AHCT have improved outcomes (74). In IFM 2009 long term follow up subgroup analysis, PFS (HR 0.28, p<0001) and OS (HR 0.35, p<0.001) was longer in patients who became MRD negative (64), and in DETERMINATION, there was no PFS difference between AHCT and non-AHCT therapy in patients who achieved MRD negativity (59). (59). In the CARDAMON trial, of the 22.8% of patients who achieved MRD negativity following induction, analysis suggested there was no benefit from AHCT gained in this group (66). A large retrospective study of NDMM patients who achieved a VGPR or greater after induction therapy assessed the MRD status by next-generation flow cytometry and found pre-AHCT MRD positivity was associated with a shorter PFS (48.2 months vs 80.1 months, p<0.001) (75). Finally, the single-arm MASTER trial attempted to use MRD negativity to guide decision making in patients receiving Dara-KRd induction followed by upfront AHCT and Dara-KRd consolidation, ceasing treatment when a patient achieved two consecutive MRD-negative readings. This strategy showed promising PFS and rates of MRD negativity (76). HRCG patients in the MASTER trial had far poorer PFS, especially when their therapy was stopped-and achievement of MRD negativity.

There have been several retrospective studies evaluating responses to a second or third AHCT in the setting of RRMM, demonstrating this a feasible and safe approach, which may provide PFS benefit (77–82). Further retrospective subgroup analysis studies have demonstrated the benefit of salvage AHCT is greater in those who had a longer duration of response with their first AHCT (83, 84). This was recently called into question when long term follow-up of the GMMG ReLApsE trial did not show a difference in PFS or OS, but patients were not allowed onto the study if lenalidomide refractory, and therefore likely not generalizable to the current RRMM population (85). The interim analysis of the prospective single arm Second Chance trial shows deep responses with median PFS not reached when using Dara-KRD with salvage AHCT in the early relapse setting (86).

Melphalan retains its potent disease control even in the post CAR T setting. In a recent assessment of salvage therapies after relapse following BCMA-directed CAR T cell treatment, there appeared to be a reasonable response with 71.4% ORR, and an OS of 23.2 months in those who underwent AHCT or allogeneic HCT. Many of these patients were refractory to multiple lines of therapy (median 5 lines prior to CAR T) and the vast majority (94.9%) had had a prior AHCT (87). Salvage AHCT holds theoretical appeal in augmenting the biology of relapsed myeloma to wipe the slate clean of a heavily exposed patient. The rationale here is twofold: to gain clonal control and reset the immune milieu (88). There is a pattern of immune dysregulation and microenvironment abnormalities seen in myeloma patients with reduced NK and T cells and increased immunosuppressive cells, particularly T regulatory cells (89). This dysfunction worsens with exposure to anti-myeloma agents (90, 91). With an infusion of relatively chemotherapy naïve autologous stem cells, there opens up an opportunity for myeloma-specific immunity to be regained. In particular, the pattern of dynamics of T cell reconstitution after AHCT with a favorable ratio of T regulatory to T effector cells (92–94), may be able to be harnessed to leverage the sensitivity to immune therapies including CAR T (95). This is being tested prospectively prior to CAR T cells in NCT05393804 with the hypothesis that “fresh” non-exhausted T cells will lead to better expansion and persistence of the CAR T cell made from these cells. Furthermore, the early recovery of NK cells after AHCT may provide an opportunity to maximize potency of NK cell-therapies in this window (96).