How CAR-T/TCR-T Therapy in Solid Tumors Is Shaping Up: Abstracts to Watch from ASCO 2026

CAR-T cell therapy has transformed the treatment of certain hematologic malignancies, but its extension to solid tumors remains one of the most difficult problems in oncology. The barriers are well-documented: antigen heterogeneity, an immunosuppressive tumor microenvironment, poor T-cell trafficking, and the persistent risk of on-target off-tumor toxicity when targeting antigens that are expressed, even at low levels, on normal tissues. Despite these challenges, continued interest in solving these remains as evidenced by solid tumor CAR-T and TCR-T abstracts at ASCO 2026.

What follows is a thematic overview of the solid tumor CAR-T landscape as presented at this year's meeting. Rather than ranking abstracts, I have organized them by the problems they are attempting to solve, because in this field, the engineering approach matters as much as the target.

Logic-Gating for Safety and Specificity: The Tmod Platform

One of the most persistent obstacles in solid tumor CAR-T development is on-target off-tumor toxicity. Most tumor-associated antigens are also expressed at some level on normal tissues, creating a therapeutic window problem that does not exist for targets like CD19. A2 Biotherapeutics' Tmod platform attempts to address this through a logic-gated dual-receptor design: an activator receptor targeting a tumor-associated antigen (in this case, mesothelin) and a blocker receptor that recognizes HLA-A*02 on normal cells. In patients whose tumors have lost HLA-A*02 expression through loss of heterozygosity (LOH), the blocker is absent on tumor cells, allowing selective CAR-T activation.

Two abstracts from the EVEREST-2 study are relevant here. A2B694 presents initial safety and efficacy data from the Tmod construct targeting mesothelin in patients with HLA-A*02 LOH across multiple solid tumors including colorectal, pancreatic, and ovarian cancers. A2B543 adds a membrane-tethered IL-12 booster to the same construct, aiming to enhance potency while preserving the selectivity provided by the logic gate. The approach is conceptually compelling: rather than engineering greater target specificity into a single receptor, the Tmod platform uses a second receptor to create a Boolean NOT gate that protects normal tissues.

The critical question is whether the HLA-LOH patient selection requirement limits the addressable population to a degree that undermines commercial viability. LOH at HLA-A*02 occurs in a meaningful but not universal fraction of solid tumors, and the requirement for both mesothelin expression and HLA-A*02 LOH narrows eligibility further. The logic-gating concept solves a real biological problem, but its clinical impact will depend on how broadly it can be applied across target antigens and LOH frequencies.

Gastrointestinal Cancers: The Most Active Battleground

GI cancers have emerged as the most active testing ground for solid tumor CAR-T therapy, driven by the identification of several relatively tumor-selective antigens. Four ASCO 2026 abstracts illustrate the range of approaches being pursued.

Satricabtagene autoleucel (satri-cel) is the most clinically advanced program on this list. The CLDN18.2-directed autologous CAR-T has already demonstrated a statistically significant PFS improvement over physician's choice of therapy in the phase 2 CT041-ST-01 trial in previously treated advanced gastric/GEJ cancer, a notable result given how few cell therapies have achieved a randomized efficacy signal in any solid tumor. The ASCO 2026 abstract presents long-term follow-up from the sequential therapy cohort, in which satri-cel was administered after first-line treatment rather than in the heavily pretreated setting. This positioning is strategically important: if CAR-T therapy can be deployed before patients are immunologically depleted by multiple lines of chemotherapy, the persistence and efficacy of the engineered cells may improve substantially. The long-term data will be critical in assessing whether early deployment translates to durable disease control.

CHM-2101 targets Cadherin 17 (CDH17), a cell-membrane protein associated with poor prognosis in colorectal cancer, gastric cancer, and neuroendocrine tumors. Chimeric Therapeutics' third-generation CAR-T received FDA fast track designation for gastroenteropancreatic NETs and is presenting clinical and translational results from the phase 1 portion of its phase 1/2 trial. CDH17 is an interesting target because its expression pattern on GI cancers is broad, but its normal tissue distribution is restricted to the apical surface of intestinal epithelial cells, a location that is largely inaccessible to circulating CAR-T cells due to the tight junction barrier. Whether this anatomical protection is sufficient to prevent clinically significant on-target off-tumor toxicity is the key translational question this trial needs to answer.

KD-496 takes a dual-targeting approach, combining CLDN18.2 and NKG2D ligand recognition in a single CAR-T construct for advanced GI and pancreatic cancers. Dual-targeting strategies aim to mitigate antigen escape, a well-documented failure mode in solid tumors where heterogeneous antigen expression allows resistant clones to survive. However, dual-targeting also increases manufacturing complexity and introduces the challenge of balancing activation signals from two different receptors. Early safety and feasibility data will be the primary focus at this stage.

Intraperitoneal CEACAM5 CAR-T represents a different strategic approach entirely: rather than engineering around the problems of systemic CAR-T delivery to solid tumors, this program delivers IL-9-secreting CEACAM5-targeted CAR-T cells directly into the peritoneal cavity for advanced colorectal cancer with peritoneal metastases. The fast-manufactured, locally delivered, cytokine-armored construct addresses multiple barriers simultaneously, including trafficking, persistence, and the immunosuppressive TME. Intraperitoneal delivery limits the applicability to peritoneal disease, but for the subset of CRC patients with peritoneal carcinomatosis, who have particularly poor outcomes with systemic therapy, this is a rational approach. The IL-9 armoring is also noteworthy: most cytokine-armored CAR-T programs have used IL-12, IL-15, or IL-21, and IL-9's role in promoting tissue-resident memory T-cell formation is a differentiated biological rationale.

Glioblastoma: Encouraging but Early

Recurrent glioblastoma remains one of the most treatment-resistant cancers, with median overall survival historically ranging from six to nine months and no approved CAR-T therapy. Two abstracts address GBM from complementary angles.

The bivalent CAR-T update presents overall survival, safety, and neurologic function data from the ongoing phase 1 trial. Bivalent constructs targeting two GBM-associated antigens aim to reduce the risk of antigen escape, which has limited prior single-target approaches in this indication. The inclusion of neurologic function outcomes as an endpoint is important context: in GBM, even modest tumor control can have meaningful quality-of-life implications if it preserves neurological function.

A second abstract applies single-cell multi-omics and TCR sequencing to characterize the behavior of CAR-T cells after intracavitary delivery in GBM. This translational work addresses one of the fundamental unknowns in CNS-directed cell therapy: whether CAR-T cells home to the tumor, persist within the cavity, and maintain functional phenotypes over time. The identification of early tumor cavity homing and phenotypic evolution of CAR-T cells could inform rational design of next-generation constructs for GBM.

Sarcoma: Expanding the Target Landscape

Adaptimmune's afami-cel (Tecelra) remains the only FDA-approved engineered cell therapy for solid tumors, with accelerated approval in HLA-A*02-positive, MAGE-A4-expressing synovial sarcoma. However, as I have written previously, the eligible patient population is small, approximately 400 US patients with the right HLA types and antigen expression, and commercial uptake is constrained by the conditioning regimen, administration requirements, and a $727,000 list price.

The PRAME-directed TCR-T abstract in synovial sarcoma is therefore notable because it introduces a different target for the same indication. PRAME is a cancer-testis antigen with broader expression across sarcoma subtypes and other solid tumors than MAGE-A4, which could, in principle, expand the addressable patient population for engineered T-cell therapy in sarcoma. Phase 1 results will establish whether PRAME-directed TCR-T can replicate the response rates seen with afami-cel while potentially broadening eligibility. The competitive dynamics between MAGE-A4 and PRAME-directed approaches in sarcoma will be worth tracking.

Hepatocellular Carcinoma: GPC3 Advances

Ori-C101 is an armored GPC3-directed CAR-T being developed by Oricell Therapeutics for advanced hepatocellular carcinoma. GPC3 has emerged as the most actively pursued CAR-T target in HCC due to its overexpression on tumor cells and limited expression on healthy hepatocytes. Prior data from Oricell showed a 60% partial response rate by RECIST 1.1 in 10 GPC3-positive HCC patients, with two patients achieving progression-free survival of one and two years respectively, and one complete responder with no relapse at nine months. The ASCO 2026 abstract presents updated results from the phase 1b BEACON study using an optimized manufacturing process.

These are arguably the most compelling efficacy data generated by any solid tumor CAR-T program in HCC. However, several factors temper enthusiasm at this stage: the data are from a single-arm trial in China without a control group, GPC3 expression is heterogeneous across HCC patients, and the HCC treatment landscape is rapidly evolving with the emergence of combination checkpoint inhibitor and anti-VEGF regimens. Oricell has raised $110 million in pre-IPO financing and is preparing for registrational trials, so pivotal data should follow, but the path from encouraging phase 1 results to a registered therapy in HCC is not straightforward.

Small Cell Lung Cancer: CAR-T Enters the DLL3 Race

LB2102 is Legend Biotech's DLL3-targeted autologous CAR-T armored with dominant-negative TGFβR2, being evaluated in relapsed/refractory SCLC and LCNEC. DLL3 has been validated as a therapeutic target in SCLC by Amgen's bispecific T-cell engager Imdelltra (tarlatamab), which demonstrated a survival benefit over chemotherapy in the phase 3 DeLLphi-304 trial and received accelerated FDA approval.

At ASCO 2025, LB2102 reported no dose-limiting toxicities through four dose levels with an overall response rate of 16.7% and disease control rate of 66.7% across 12 patients. These numbers lag Imdelltra's approximately 40% ORR and the 50%+ ORRs reported by DLL3-directed ADC programs from Ideaya and Zai Lab. The ASCO 2026 update with continued dose escalation will be important for determining whether higher doses can close this efficacy gap. The dnTGFβR2 armoring, designed to block TGF-β-mediated immunosuppression and reduce T-cell exhaustion, is the same technology being employed in several AstraZeneca cell therapy programs and represents a rational engineering approach for the SCLC TME.

The fundamental question for LB2102 is whether a one-time autologous CAR-T infusion can offer a durable benefit that justifies the manufacturing complexity and conditioning requirements relative to off-the-shelf bispecific or ADC approaches targeting the same antigen. In SCLC, where patients typically progress rapidly and may not have time for CAR-T manufacturing, this is a particularly acute commercial and clinical challenge.

Emerging Engineering Concepts

Several additional abstracts introduce engineering approaches that are earlier in development but address fundamental limitations of current CAR-T platforms.

MICA/B-targeted CAR-T is designed to overcome two obstacles simultaneously: tumor immune escape through MICA/B shedding, and the requirement for myelosuppressive conditioning chemotherapy. By targeting the conserved α3 domain of stress ligands MICA and MICB, the approach aims for broad tumor recognition across both hematologic and solid tumors. If the conditioning-free administration proves feasible with acceptable efficacy, it would represent a meaningful step toward making cell therapy more accessible and reducing treatment-related morbidity.

NT-175 targets the TP53 R175H hotspot mutation, the most common recurrent TP53 mutation across solid tumors, found in approximately 4-7% of patients with breast, colorectal, and pancreatic cancers. Developed by AstraZeneca's Neogene Therapeutics, NT-175 is an autologous TCR-T engineered with CRISPR-mediated knockout of endogenous TCR and TGFβR2 genes. This is a conceptually important program because it targets a shared neoantigen rather than a tumor-associated antigen, meaning the target is truly tumor-specific and does not exist in normal tissue. The trade-off is a narrow eligible population (TP53 R175H positive, HLA-A*02:01 positive), but if the approach demonstrates clinical activity, the same platform logic could be extended to other shared neoantigens.

CAR001, an mRNA-engineered nanobody-CAR.BiTE γδ T-cell product, combines multiple non-standard engineering elements: mRNA-based transient expression rather than viral integration, nanobody-based antigen recognition, a secreted BiTE component, and the use of γδ T cells rather than conventional αβ T cells as the cellular chassis. The transient expression approach offers a safety advantage, as CAR expression is self-limiting, which reduces the risk of sustained off-target toxicity, but requires redosing and may limit the durability of responses.

An armored mesothelin-targeted CAR-T adds to the growing body of mesothelin-directed cell therapy data. Mesothelin is one of the most extensively studied CAR-T targets in solid tumors, with multiple prior programs yielding modest single-agent activity. The armoring strategy (not fully disclosed in the abstract title) represents an attempt to improve on earlier mesothelin CAR-T results, but the target's track record in the clinic warrants cautious interpretation of early data.

What the Landscape Tells Us

Several patterns emerge from this year's ASCO solid tumor CAR-T data.

First, the field has largely moved beyond unmodified, single-target constructs. Nearly every program presented at ASCO 2026 incorporates at least one engineering enhancement, whether armoring with cytokines (IL-9, IL-12) or dominant-negative receptors (dnTGFβR2), logic-gating, dual-targeting, CRISPR-mediated gene editing, or alternative cellular platforms (γδ T cells, iPSC-derived). This reflects a broad recognition that the first generation of solid tumor CAR-T approaches was insufficiently equipped for the solid tumor microenvironment.

Second, GI cancers are consolidating their position as the lead indication for solid tumor CAR-T, driven by the availability of relatively tumor-selective targets (CLDN18.2, CDH17, CEACAM5) and the maturation of satri-cel as the most clinically advanced program. Whether satri-cel achieves regulatory approval will be a defining moment for the entire field.

Third, the competitive landscape within each target is intensifying. DLL3 in SCLC and GPC3 in HCC are each being pursued by multiple modalities, including CAR-T, bispecifics, and ADCs, and the question is increasingly whether cell therapy can demonstrate a durability advantage that justifies its complexity. If CAR-T cannot offer meaningfully longer remissions than off-the-shelf approaches, the manufacturing burden and conditioning requirements will limit adoption regardless of response rates.

Finally, the most conceptually novel programs, including NT-175 targeting a shared TP53 neoantigen, the MICA/B CAR-T eliminating conditioning chemotherapy, and A2 Bio's logic-gating approach, remain early in clinical development. These represent the engineering ideas most likely to reshape the field if the early data support them, but solid clinical validation is still one to three years away. In the interim, the solid tumor CAR-T landscape will continue to be defined by the tension between biological ambition and clinical pragmatism.

By Sakis Paliouras, PhD