Growing evidence for CAR-T and its expanding use across hematologic malignancies are fueling comparisons with stem cell transplantation. Their relative roles differ sharply by disease and clinical setting.
In large B-cell lymphoma, CAR-T has moved ahead of the traditional salvage chemotherapy and autologous hematopoietic cell transplantation (auto-HCT) pathway for many patients with primary refractory or early-relapsed disease. In B-cell acute lymphoblastic leukemia, CAR-T can provide durable control or bridge to allogeneic HCT (allo-HCT). In multiple myeloma, CAR-T has reached first relapse and is now being compared directly with frontline auto-HCT. In acute myeloid leukemia, allo-HCT remains central and CAR-T is still investigational.
The latest EBMT activity survey, covering 2024, recorded 47,204 HCT procedures. Allo-HCT rose 2.6% from 2023. CAR-T use rose 24.5% to 6,082 patients, almost all receiving autologous products, concentrated in B-cell malignancies and myeloma (Passweg et al., 2026).
Three Different Sources of Antitumor Effect
Auto-HCT does not replace malignant hematopoiesis with a healthy donor system. It makes high-dose therapy possible, because that therapy would otherwise cause prolonged or irreversible marrow failure. The patient’s own stem cells are collected beforehand and reinfused after conditioning for hematopoietic rescue. The antitumor effect comes from the high-dose therapy itself.
Allo-HCT adds donor-derived immune activity. The graft-versus-leukemia or graft-versus-lymphoma effect recognizes malignant cells across multiple antigens and histocompatibility differences. The immune pressure is broad, and the costs are graft-versus-host disease (GVHD), infection, organ toxicity and non-relapse mortality.
CAR-T cells are engineered to recognize specific surface antigens, with CD19 and B-cell maturation antigen (BCMA) among the best established targets. They have to expand, stay functionally fit and persist against an often immunosuppressive microenvironment. CAR-T also carries distinct acute and longer-term toxicities. Relapse can follow antigen loss or downregulation, T-cell exhaustion or insufficient persistence (Rafei et al., 2026).
You can also read:
LBCL: CAR-T Has Changed the Salvage-to-Transplant Paradigm
The traditional second-line strategy for transplant-eligible LBCL was salvage chemoimmunotherapy, then high-dose chemotherapy and auto-HCT in chemosensitive disease. It performed poorly in primary refractory and early-relapsed lymphoma, where many patients never responded well enough to reach transplant.
ZUMA-7 compared axicabtagene ciloleucel (axi-cel) with standard salvage therapy followed by auto-HCT in responding patients with primary refractory LBCL or relapse within 12 months. At a median follow-up of 47.2 months, estimated 4-year overall survival was 54.6% with axi-cel versus 46.0% with standard care (HR for death, 0.73) (Westin et al., 2023).
TRANSFORM independently showed superior event-free survival with liso-cel against the same transplant-based strategy: median event-free survival was not reached versus 2.4 months with standard care (Abramson et al., 2023). Chemosensitivity was no longer a mandatory gateway, and patients could go directly to CAR-T as potentially curative therapy.
That shouldn’t be extended to every LBCL relapse, though. Current EBMT recommendations still treat auto-HCT as a valid option for early but chemosensitive relapse. Patients with later relapse can do reasonably well with auto-HCT and haven’t been studied in the same randomized second-line comparisons (Snowden et al., 2025).
The Balance Varies by Lymphoma
Follicular lymphoma uses CAR-T for later-line relapsed or refractory disease, but auto-HCT still fits selected patients, including those with POD24 (progression of disease within 24 months of first-line therapy) who respond to salvage therapy. In mantle cell lymphoma, CAR-T now has an established role after treatment failure, and frontline auto-HCT is being reconsidered as BTK inhibitor-based therapy moves earlier.
Allo-HCT is largely reserved for selected patients after CAR-T failure or when CAR-T is unavailable. In Hodgkin and T-cell lymphomas, transplantation retains a much larger role and CAR-T remains investigational (Greco et al., 2025).
In some cases, bispecific antibodies can provide disease control without either CAR-T or transplantation, which affects where each fits.
You can also read:
B-ALL: CAR-T Remission Does Not Settle the Transplant Decision
The relationship with allo-HCT in B-ALL is more complex, since both CAR-T and allo-HCT can produce deep remissions with curative potential. In ZUMA-3, brexucabtagene autoleucel produced high remission rates in adults with relapsed or refractory B-ALL. Among responders, median duration of remission was 44.2 months in the 14 patients who subsequently underwent allo-HCT and 18.6 months in the 43 who did not. The comparison was exploratory and non-randomized (Shah et al., 2023).
Patients who went to transplant had received fewer prior therapies, had lower baseline marrow blast burden, and were far less likely to have had a previous allo-HCT, so the longer remission can’t be credited to transplantation alone. Durable remissions also occurred without any subsequent transplant.
MRD negativity after CAR-T doesn’t fully settle relapse risk. CD19-positive relapse can occur as CAR-T activity wanes, and CD19-negative relapse can occur through antigen escape. Blinatumomab, inotuzumab ozogamicin, and increasingly potent BCR::ABL1 inhibitors can all produce deep responses before CAR-T or allo-HCT enters the sequence (Kantarjian and Jabbour, 2025).
T-ALL: Engineering Around Fratricide
No CAR-T product is currently approved for T-ALL, and allo-HCT remains the main curative option for relapsed or refractory disease. Target biology is the obstacle: CD7 and similar candidates sit on malignant and normal T cells alike, including the T cells used to manufacture the CAR product. The cells kill each other during production, a phenomenon called fratricide.
Fratricide-resistant autologous CD7 CAR-T cells have shown real activity. In a 17-patient study, 16 reached MRD-negative complete remission, with durable responses both with and without subsequent allo-HCT. Two problems remain: GVHD risk from the T-cell receptor, and lymphodepletion that has to clear the patient’s existing T cells without destroying the CAR-T product (Oh et al., 2024).
Gene editing now addresses all three together. Base-edited, donor-derived CAR7 cells disrupt CD7 to prevent fratricide, TCRαβ to limit GVHD, and CD52 to allow alemtuzumab-based lymphodepletion. The approach has moved from first-in-human proof of concept to a phase 1 study of universal BE-CAR7 cells in relapsed or refractory T-ALL.
CRISPR/Cas9-edited allogeneic CD7 products are advancing in parallel, including constructs that delete both CD7 and TRAC. Off-the-shelf CAR-T in T-ALL is increasingly plausible, but the evidence is still early-phase (Chiesa et al., 2026).
AML: CAR-T Confronts Shared Myeloid Antigens
The most accessible antigens on AML cells are also present on normal hematopoietic stem/progenitor cells or mature myeloid cells, which risks severe, prolonged myelotoxicity. Intraclonal heterogeneity adds to the problem: antigen-low or antigen-negative populations survive highly specific immune pressure and cause relapse.
A 2026 systematic assessment of 63 reported AML CAR targets found early clinical studies concentrated mainly on CD123, CD33, and CLL-1. Durable complete remissions remained limited across most approaches. CLL-1 studies produced encouraging responses in some small cohorts, but prolonged neutropenia and other myelotoxicity remain limitations (Blood Cancer Journal, 2026). Current EBMT recommendations keep allo-HCT central for eligible adverse-risk AML and for many intermediate-risk patients.
Making Myeloid CAR-T Controllable
The allogeneic Allo-RevCAR01-T cells express a universal receptor that doesn’t recognize CD123 directly. A separately administered bispecific adapter, R-TM123, bridges CD123 on the leukemic cell to the RevCAR on the T cell. Activity depends entirely on the adapter, so it can be titrated, interrupted, and restarted through its administration. Earlier work with the platform showed recovery of leukocytes and platelets after the adapter was withdrawn.
The phase Ia RevSTAR-123 study, presented at EHA 2026, enrolled 17 patients with relapsed/refractory or MRD-positive CD123-positive AML. Repeated R-TM123 cycles produced renewed CAR-T activation and expansion, with clearer responses at higher doses. CRS occurred in 76%, mostly low-grade, with no GVHD or ICANS. It is still early-phase, but it shows that activity against a shared myeloid antigen doesn’t have to stay switched on continuously.
You can also read:
Editing the Graft Instead
Gene-edited transplantation approaches the same constraint from the opposite direction. CRISPR-Cas9 editing can remove CD33 from donor hematopoietic stem and progenitor cells, producing a hematopoietic system protected from CD33-directed therapy, and residual AML stays CD33-positive.
A 2026 phase 1/2a study evaluated the CD33-deleted allogeneic graft tremtelectogene empogeditemcel (trem-cel), followed by gemtuzumab ozogamicin maintenance, in high-risk AML or MDS. CD33 editing persisted across donor-derived myeloid, B-cell, and NK-cell lineages (DiPersio et al., 2026).
The same principle could eventually open up antigens that are currently too dangerous to target. In AML, transplantation may become part of the engineering that makes targeted cellular therapy possible, and CAR-T may not replace it.
Multiple Myeloma: CAR-T Has Reached First Relapse
Myeloma permits the most direct comparison with auto-HCT. The transplant provides no donor-derived antimyeloma immunity: high-dose melphalan supplies the antitumor effect, and the autologous graft restores hematopoiesis. Modern quadruplet induction and maintenance have strengthened the transplant-based strategy, so historical comparisons are uninformative.
Cilta-cel has already moved into early relapse. CARTITUDE-4 enrolled lenalidomide-refractory patients after one to three prior lines. At extended follow-up, 30-month progression-free survival (PFS) was 59.4% with cilta-cel versus 25.7% with standard therapy, and 30-month overall survival (OS) was 76.4% versus 63.8%.
Five-year data from CARTITUDE-2 cohort A, presented at the 2026 International Myeloma Society meeting, add a different kind of evidence. Twenty patients with relapsed or refractory MM after one to three prior lines received a single cilta-cel infusion with no maintenance. At a median follow-up of 60.7 months, 10 of 20 were alive and progression-free at five years. Median PFS was 60.5 months and 5-year OS was 69.2%. Sustained control without continuous maintenance is notable in a disease usually managed with ongoing therapy (IMS 2026, Abstract PA-288).
CARTITUDE-6 compares cilta-cel directly with auto-HCT after D-VRd induction in 759 transplant-eligible, newly diagnosed patients. Both arms receive the same contemporary induction, so the trial isolates the contribution of high-dose melphalan and stem-cell rescue against cilta-cel at the same point in frontline therapy. Results have not been reported yet (NCT05257083).
You can also read:
Bridging the CAR-T Manufacturing Interval
Disease has to stay controlled while autologous CAR-T is manufactured. Bridging therapy is common in aggressive lymphoma and multiple myeloma, and its goal is disease control, not necessarily the deepest possible response. The choice of bridging agent can affect what follows: prolonged cytopenias, infection, organ toxicity and lymphocyte depletion can compromise later treatment, and prior antigen-directed therapy may limit target availability and CAR-T activity.
This creates a selection point that trial efficacy comparisons don’t show: some patients referred for CAR-T never reach infusion because of progression, clinical deterioration, manufacturing problems, or logistical delay.
CAR-T Is Being Redesigned Around Its Current Limitations
In-vivo CAR-T could remove several constraints. The first-in-human IASO206 study generated BCMA-directed CAR-T cells directly in patients with relapsed or refractory MM after a single IV infusion, without apheresis, ex-vivo manufacturing, or lymphodepleting chemotherapy. Among the first 10 efficacy-evaluable patients, both the objective response rate and MRD-negativity rate were 90% (IMS 2026, LBA-11).
The first-in-human LB2501 study, presented at EHA 2026, extended in-vivo CAR-T to relapsed or refractory B-cell non-Hodgkin lymphoma, with dual CD19/CD20 targeting to address antigen escape. A single infusion of the T-cell-targeted lentiviral vector produced CAR-T expansion without lymphodepleting chemotherapy. All six patients treated at the higher dose level responded and five achieved a complete response, though median follow-up was only 2.2 months (EHA 2026, LB5006).
Other dual- and multi-antigen constructs are in development to reduce antigen escape. Allogeneic CAR-T could remove patient-specific manufacturing but still has to overcome alloreactivity, host rejection, and limited persistence (Rafei et al., 2026).
Watch more on OncoDaily TV
Access Can Determine Which Therapy Is Actually Available
Regulatory approval doesn’t make CAR-T equally available across health systems. Access depends on reimbursement, accredited treatment centers, apheresis and cell-processing infrastructure, manufacturing capacity, and referral pathways. Differences remain substantial even within Europe: eight countries still had no reimbursed EU-approved CAR-T product in April 2025, with particularly marked gaps across Central and Eastern Europe.
Transplantation has different access constraints. Auto-HCT requires successful stem-cell mobilization and collection, and allo-HCT adds HLA typing, donor identification and work-up, and graft procurement. Both depend on specialized transplant centers and inpatient capacity. Matched unrelated and haploidentical transplantation have broadened donor availability, but access is uneven internationally.
Commercial CAR-T carries a high upfront product cost on top of leukapheresis, hospitalization, toxicity management, and follow-up. Transplant costs are distributed across conditioning, graft collection or procurement, hospitalization, complications, and longer-term care. The relative financial burden varies substantially by health system and reimbursement model.
In the 2024 EBMT survey, rates across participating countries ranged from 2.7 to 422.8 per 10 million population for allo-HCT and from 0.3 to 210.5 for CAR-T. In a UK real-world mantle-cell lymphoma cohort, 119 patients were approved for brexu-cel but only 83 were ultimately infused.
CAR-T vs Stem Cell Transplantation in 2026
CAR-T has changed established transplant pathways in some settings and is being compared directly against transplantation in others. Allo-HCT still provides therapeutic functions that current CAR-T approaches cannot consistently reproduce. The technology is converging faster than the disease-by-disease picture suggests, and its clinical reach also depends on healthcare systems.
You can also read:
FAQ
Can CAR-T cure blood cancer without a stem cell transplant?
Yes, in some settings. CAR-T can produce durable remissions without subsequent transplantation, particularly in large B-cell lymphoma, where CAR-T itself is used as potentially curative therapy. In B-ALL, the situation is less settled: some patients remain in long-term remission after CAR-T alone, while others proceed to allo-HCT because of relapse risk.
Can CAR-T be used after a stem cell transplant?
Yes. CAR-T can be given to selected patients whose disease relapses after transplantation. Previous allo-HCT creates additional considerations, including the origin of the T cells, prior graft-versus-host disease, immunosuppression, cytopenias, and infection risk. Prior transplantation therefore changes how CAR-T is planned but does not automatically exclude it.
Why is CAR-T harder to develop for AML and T-ALL?
Finding a safe target is much harder. Many AML targets, including CD33 and CD123, are also expressed on normal hematopoietic cells, so sustained targeting can damage normal myelopoiesis. In T-ALL, targets such as CD7 may be present on both malignant T cells and the T cells used to manufacture CAR-T, causing fratricide. Gene editing and switchable CAR platforms are being developed to overcome these problems.
Could CAR-T eventually replace frontline stem cell transplantation in multiple myeloma?
This is being tested prospectively. CARTITUDE-6 directly compares cilta-cel with auto-HCT after the same D-VRd induction in transplant-eligible newly diagnosed MM. The trial therefore provides a direct comparison between high-dose melphalan with stem-cell rescue and BCMA-directed CAR-T at the same point in frontline treatment. Results are not yet available.
