Download the Report
T-cell engagers win on clean targets, and autoimmune disease is the next test
T-cell engagers (TCEs) succeed against clean, lineage-restricted targets the body can spare, such as the B cells and plasma cells behind lymphoma, leukemia and myeloma, and struggle against the shared antigens of most solid tumors, which also sit on healthy organs. Sleuth’s September 2026 analysis counts about 770 active TCE programs from about 450 developers, and 10 of the 13 approved TCEs treat blood cancers. The three solid-tumor approvals prove the rule: each hits a restricted target or is delivered locally.
The excitement has moved to autoimmune disease, where B cells are the same kind of clean target. The winner there will not simply be the deepest depleter, because patients with non-fatal disease accept less cytokine release syndrome (CRS) and infection risk, and refractory lupus is a different market from rheumatoid arthritis. About half the pipeline now engineers beyond the classic CD3 bispecific to widen the window, and pharma is building portfolios of related assets. TCEs could own the messy middle: refractory patients who need deeper depletion than biologics provide, but for whom CAR-T’s risk and logistics are too much.
About 770 programs, 13 approvals and a pipeline led by China
Most programs are preclinical, and China leads program count
Of the roughly 770 active programs, about 550 are preclinical, 15 at IND, 110 in Phase 1, 55 in Phase 1/2, 15 in Phase 2, 20 in Phase 3, one filed and 13 approved. About 500 are bispecifics and about 160 trispecifics. Chinese developers run about 360 programs, US developers about 280 and the rest of the world about 130.
The approved solid-tumor TCEs target gp100 in uveal melanoma (tebentafusp), DLL3 in small-cell lung cancer (tarlatamab) and EpCAM in malignant ascites (catumaxomab, delivered into the abdomen). Solid-tumor proof so far rests on narrow, niche targets.
Programs crowd a few targets, and next-generation formats are spread thin
CD19 and BCMA each account for about 7% of the field, followed by CD20 (about 5%), GPRC5D and PSMA (about 4% each), and EGFR, DLL3 and HER2 (about 3% each). CD3 is the T-cell arm in 79% of programs, and about half add engineering beyond a plain CD3 bispecific. Multi-antigen designs, alternative T-cell arms and peptide-HLA targeting each make up about 8% of the field, costimulated and masked designs about 6% each, and immune-tuned CD3 arms about 2%.
No next-generation format exceeds about 8% of the active base, so there is no consolidating winner yet. As surface antigens crowd, peptide-HLA engagers offer an escape by reading intracellular targets presented on the cell surface.
Deal capital has shifted to B-cell autoimmunity
TCE dealmaking peaked in 2024, at $12.8 billion in total disclosed value across 19 deals, up from $3.6 billion across four in 2023. Through September 2026, the year had reached $10.3 billion across eight deals. Autoimmune disease went from no deals before 2024 to the field’s busiest new lane, with 16 deals since then, more than twice the seven in blood cancers, and 13 of the 16 sourced from China. Capital concentrated on B-lineage targets, with BCMA in 11 deals, CD19 in 8 and CD20 in 5, against two to four each for solid-tumor targets such as PSMA, EGFR, DLL3 and HER2.
The signature transaction was UCB’s acquisition of Candid Therapeutics for about $2.2 billion in 2026, after Candid had assembled five autoimmune TCEs, mostly from China. In oncology, GSK agreed in September 2026 to pay up to $750 million for Chimagen’s myeloma trispecific, and Astellas signed an up-to-$1.7 billion collaboration with Vir for its masked PSMA engager VIR-5500 in February 2026.
M&A recurred throughout, with 12 deals in total and 5 in 2024, rather than shifting from early acquisitions to later licensing. Autoimmune buyers are paying for China-sourced B-cell depleters before durability data exists.
The clinical bar differs sharply by disease area
Blood cancers set a high response bar
In representative cross-trial comparisons, mosunetuzumab reached an 80% overall response rate with 60% complete responses, epcoritamab 63% and 40%, teclistamab 63% and 39%, elranatamab 61% and 37%, and glofitamab 56% and 43%, with talquetamab at 74% response. Durability reaches up to about two years, and tolerability, including CRS and infections, is the main barrier. Multiple myeloma is the most validated lane, with clean, lineage-restricted targets in BCMA and GPRC5D and five global approvals.
With response rates this high, a new heme TCE has to win on how its construct delivers tolerable response, not on response alone.
Solid tumors reward clean targets, and masking is the shared-antigen bet
The two systemic solid-tumor approvals win on survival rather than response. Tarlatamab (DLL3) cut the risk of death by 40% (hazard ratio 0.60) with a 35% response rate, and tebentafusp (gp100) cut it by 49% (hazard ratio 0.51) with a 9.1% response rate and 21.7-month median overall survival. Of about 80 clinical programs against shared antigens, none has passed validation. Masking, which keeps the engager inactive until tumor enzymes cleave it, is the rescue bet: in prostate cancer, masked PSMA engagers produced PSA responses in 82% of patients in a VIR-5500 dose subgroup against 30% for unmasked acapatamab, cross-trial. On the harder RECIST endpoint, masked PSMA is modest, with VIR-5500 at 45% and Janux’s JANX007 at 30% in small Phase 1 groups, and no masked program has left Phase 1.
Solid tumors are a construct question. The fourth quarter of 2026 brings early reads, including Vir’s masked HER2 engager VIR-5818, but the verdict waits for pivotal trials of VIR-5500 and Xencor’s XmAb819.
Autoimmune data show deep B-cell depletion, with durability unproven
Early human data span five engagers. A-319 (CD19) produced DORIS remission in 6 of 10 lupus patients and low disease activity in 8 of 10 at 12 months. CC312 (CD19) depleted CD19-positive plasma cells completely at its 20 µg dose, with SRI-4 responses in 5 of 6 lupus patients at week 36. Cizutamig (BCMA) produced near-complete depletion of marrow plasma cells in systemic sclerosis, velinotamig (BCMA) produced renal responses in both of two lupus nephritis patients, and CLN-978 (CD19) cleared B cells in lymph node and synovial tissue. CRS has been mostly grade 1 so far.
Attenuated, subcutaneous and masked designs fit the safety bar of a non-fatal disease, but durability remains the open question.
Next-generation levers widen the window, but the high-upside ones are unproven
Each engineering lever solves a different problem
Sleuth ranked five levers by clinical proof. Attenuated CD3, which dials down T-cell activation to cut cytokine release while preserving tumor killing, has modest upside but clean human proof: AstraZeneca’s AZD0486 reached a 96% response rate with no serious CRS. Masking has high upside and an early human signal. Two-plus-one avidity designs, with two tumor-binding arms, have modest upside and unproven decoupling from toxicity. 4-1BB costimulation carries high upside and a history of liver toxicity, and CD28 costimulation the highest upside with fatal-event precedent: Regeneron’s nezastomig recorded two grade 5 immune events and was pulled back for redesign.
The right lever depends on the indication. Masking is the solid-tumor rescue, while attenuated CD3’s gentler, global dial-down fits non-fatal autoimmune disease, where safety rather than depth is the bar.
Only dual-target myeloma trispecifics show efficacy so far
“Trispecific” covers four different bets for the third arm: a second tumor antigen, costimulation, longer half-life or a second B-cell target. Only dual-target myeloma engagers show real efficacy, because two antigens close the escape route that single-target bispecifics leave open. In Phase 1, Johnson & Johnson’s JNJ-5322 reached a 100% response rate, with 70% complete responses, in a 27-patient subgroup naive to T-cell redirection, Innovent’s IBI3003 83.3% in heavily pretreated patients and ISB2001 79%, against a 60% to 74% bar for teclistamab, talquetamab and etentamig. No trispecific has a Phase 2 or 3 readout yet, and IBI3003 is moving to Phase 3 with FDA Fast Track designation. After this snapshot, at the International Myeloma Society meeting in September 2026, J&J reported a 93.6% response rate for JNJ-5322 (ramantamig) in patients naive to T-cell-redirecting therapy.
Autoimmune TCEs compete between biologics and CAR-T
Each class reaches a different depth of the B-cell lineage. Rituximab clears circulating B cells but leaves tissue reservoirs, CD19 TCEs reach tissue B cells, BCMA TCEs reach the marrow plasma cells that sustain autoantibodies, and CAR-T reaches deepest but requires autologous manufacturing and lymphodepletion. TCEs are off the shelf, with a moderate to high safety burden. About 4.2% of TCE programs pursue lupus and about 2.7% rheumatoid arthritis, with systemic sclerosis, Sjögren’s disease, lupus nephritis, myositis and myasthenia gravis behind them, and none is approved.
Repurposed oncology TCEs such as blinatumomab and teclistamab already show grade 3 CRS in autoimmune use, which gives purpose-built autoimmune candidates room to win. The prize is the patient who needs more than a biologic but less than CAR-T, a trade-off Sleuth also tracks in the in-vivo CD19 competitive map.
What the TCE findings mean for developers and buyers
These implications are Sleuth’s interpretation of the data above.
For developers, match the construct to the disease. In blood cancers, differentiation now comes from tolerable response and convenience. In solid tumors, a shared antigen needs a construct that confines activity to the tumor, and a clean target is still the surest path. In autoimmune disease, the target profile is a safety profile first, and a program built for refractory lupus will not automatically fit rheumatoid arthritis.
For buyers, price the lever’s proof, not its upside. The levers with clean human data offer modest gains, and the ones with the biggest upside carry toxicity history or no durability data. Autoimmune TCEs from China are available early, but diligence should weigh depth of depletion against CRS and infection risk and ask how durable the reset will be.
What Sleuth expects in T-cell engagers
These predictions are Sleuth’s analytical interpretation, looking beyond what the data shows today.
- No shared-antigen engager will clear the bar before 2027. The swing read is VIR-5500’s masked-PSMA pivotal cohort, where durable tumor shrinkage would validate masking as a format.
- The trispecific premium is priced ahead of the clinical payoff. Phase 1 response rates top the bispecific bar, but IBI3003’s Phase 3 is unlikely to show a durability edge worth its added safety risk.
- CD28-costimulated engagers will remain an open question through 2027. CD28’s potency and toxicity rise together, so the strongest second signal stays stuck in Phase 1.
How Sleuth built this analysis
Pipeline scope
The pipeline covers about 770 active T-cell engager programs from about 450 developers as of September 2026, including bispecific and trispecific formats.
Targets, formats and indications
Target, format and indication labels are multi-assigned, so a program can count in several categories and shares sum to more than 100%. The field pursues about 190 indications, and about 2% of programs target infectious disease.
Deals
Deals cover licensing and M&A events from January 2021 to September 2026, using disclosed upfronts and total consideration where reported. 2026 is year to date.
Clinical comparisons
Response and safety figures are representative, single-arm and cross-trial unless a hazard ratio is shown, and are not head to head. PSA and RECIST responses come from the named dose subgroups, and CAR-T serves as a cross-modality benchmark from the literature.
Interpretation
The lever ranking is a qualitative read of mechanism and human evidence, not a measured score. The implications and predictions are Sleuth’s analytical interpretation rather than direct findings from the data.
Pipeline, deal and clinical figures are as of September 2026 unless otherwise noted.
Frequently asked questions about T-cell engagers
How many T-cell engagers are in development?
Sleuth tracked about 770 active T-cell engager programs from about 450 developers as of September 2026, including about 500 bispecifics and 160 trispecifics. Thirteen are approved and about 20 are in Phase 3.
How many T-cell engagers are approved for solid tumors?
Three: tebentafusp (gp100) in uveal melanoma, tarlatamab (DLL3) in small-cell lung cancer and catumaxomab (EpCAM), delivered into the abdomen for malignant ascites. The other 10 approvals are in blood cancers.
Why do T-cell engagers struggle in solid tumors?
Most solid-tumor antigens are shared with healthy organs, so engaging T cells against them causes off-tumor toxicity. Of about 80 clinical programs against shared antigens, none has passed validation. Masked engagers that activate only inside the tumor are the leading attempt to fix this.
Can T-cell engagers treat autoimmune disease?
Early data suggest they can deplete B cells deeply. A-319 produced DORIS remission in 6 of 10 lupus patients at 12 months, and BCMA engagers reach the marrow plasma cells that rituximab misses. No autoimmune TCE is approved, and durability is unproven.
What are trispecific T-cell engagers?
Trispecifics add a third binding arm, for a second tumor antigen, costimulation, longer half-life or a second B-cell target. Only dual-target myeloma trispecifics show strong early efficacy, with JNJ-5322 at 100% (in a naive subgroup) and IBI3003 at 83.3% response in Phase 1.
What is a masked T-cell engager?
A masked engager stays inactive until enzymes in the tumor cleave the mask. In prostate cancer, masked PSMA engagers produced PSA responses in up to 82% of patients in a dose subgroup, against 30% for an unmasked engager, though no masked program has left Phase 1.
Where is T-cell engager dealmaking going?
Capital has shifted to B-cell autoimmunity, with 16 autoimmune deals since 2024, 13 of them sourced from China. UCB acquired Candid Therapeutics for about $2.2 billion in 2026, and GSK agreed in September 2026 to pay up to $750 million for Chimagen’s myeloma trispecific.
