Before the two firsts — a quick personal note.

June got away from me, in the best possible way. Rather than at my desk, I spent the month on the ground across Saxony — in the rooms where cell and gene therapy actually meets the real world: explaining immune cells to curious kids with origami DNA at Dresden’s Long Night of Science; trading ideas with Leipzig’s city officials on why life sciences belong at the centre of public policy; and sitting down with payers and innovators at a SaxoCell × DIANA forum on the question that quietly decides everything — reimbursement.

Which is exactly why this fortnight’s news lands the way it does. China just booked two different “firsts” in two weeks — but as these photos remind me, approval is only the first mile. Here’s what changed, and what it buys.

TL;DR — what this issue covers:

• The commercial first — CARsgen’s satri-cel wins the world’s first CAR-T approval in a solid tumor, on a statistically significant progression-free survival win.

• The credentialing first — Qihan’s QT-019B becomes the first China-developed cell therapy to hold the FDA’s “triple crown”: Fast Track, RMAT, and Breakthrough — earned in autoimmune disease, on U.S. soil.

• The pattern — China is now competitive at both ends of the value chain in the same fortnight: speed-to-market at home and FDA-grade quality signals abroad.

• The thesis — a de-risked, commercial asset and an FDA-credentialed platform asset are exactly what ex-China partners in-license, against a deal market where upfronts are up ~230%.

• The honest caveats — modest benefit, n=6 signals, designations that aren’t approvals, and geopolitics that Western buyers now price in.

For a decade the question was when China would catch up. This fortnight, the more useful question is what the West now has to buy from it.

The spine: two different kinds of “first”

It’s easy to read these as two more China biotech headlines. They’re not the same headline twice. They sit on the two different axes that actually decide whether a country leads a modality.

The first axis is velocity — can you get a genuinely novel therapy through your own regulator and into patients faster than anyone else? The second is credentialing — can you clear the highest evidentiary bar in the world, the FDA‘s, on quality and conviction?

Historically China was charitably granted the first (fast, cheap, domestic) and denied the second (not FDA-grade, fast-follower). This fortnight it took both, with two different assets, two different companies, two different disease areas.

That combination — not either headline alone — is the signal.

Pillar 1 — The commercial first: satri-cel cracks the solid-tumor wall

On June 22, 2026, China’s NMPA approved CARsgen‘s satricabtagene autoleucel — satri-cel, CT041, branded Kaileimei® — making it the world’s first CAR-T cell therapy approved in a solid tumor.

For a field that has spent a decade running into the wall of the solid-tumor microenvironment, that sentence is the whole story.

The label is narrow and earned: Claudin18.2-positive, HER2-negative advanced gastric or gastro-oesophageal junction (G/GEJ) adenocarcinoma, in patients who have failed at least two prior lines. The pivotal trial, CT041-ST-01 (NCT04581473), published in The Lancet, is the first randomized controlled trial of CAR-T in solid tumors anywhere. It randomized 156 heavily pretreated patients 2:1 to satri-cel or treatment of physician’s choice (TPC).

The numbers, stated honestly:

• Progression-free survival (the primary endpoint, and a clean win): median 3.25 vs 1.77 months, hazard ratio 0.37 (95% CI 0.24–0.56), p<0.0001.

• Objective response rate: 22% vs 4% in the intention-to-treat population (30% among patients with measurable disease) — meaningful in a setting where late-line response is usually near zero.

• Overall survival: median 7.92 vs 5.49 months, HR 0.69 (95% CI 0.46–1.05). A favorable trend, but the confidence interval crosses 1 — so call it directional, not proven.

This is real, and it is modest. The right framing isn’t “cure”; it’s “the wall has a door.” Median PFS measured in single-digit months tells you the magnitude; the HR of 0.37 and a near-total ORR gap against control tells you the biology is doing something no prior solid-tumor CAR-T has reproducibly done.

Safety is consistent with the mechanism: cytokine release syndrome in the vast majority of treated patients (mostly low-grade) and deep, expected cytopenias.

The part worth a callout for anyone thinking about platforms rather than products: CARsgen pairs satri-cel with a proprietary lymphodepletion regimen that adds low-dose nab-paclitaxel to standard cyclophosphamide/fludarabine, intended to loosen the immunosuppressive stroma and drive CAR-T infiltration into the tumor.

That’s a portable idea — a conditioning strategy for solid-tumor cell therapy generally — not just one product’s footnote.

Access is already moving from label to patient. Jiahui International Cancer Center in Shanghai is positioning to treat both domestic and international patients, with a dedicated international-patient service.

And the program is expanding at the front of the disease: a Phase I in pancreatic adjuvant therapy, plus earlier-line and perioperative G/GEJ studies — the direction that matters, because cell therapy works better the less beaten-down the patient.

Pillar 2 — The credentialing first: Qihan’s QT-019B earns the FDA “triple crown”

The second first is quieter and, for the long-term thesis, arguably bigger.

Qihan Biotech‘s QT-019B has become the first China-developed cell therapy to hold all three of the FDA’s top expedited designations at once — Fast Track (November 2025), then Regenerative Medicine Advanced Therapy (RMAT) and Breakthrough Therapy.

Call it the triple crown: the agency’s highest-conviction stack of “we want to move fast with you” signals, earned on Western soil rather than inferred from a domestic filing.

What it is matters as much as what it’s won. QT-019B is the hard version of allogeneic CAR-T: an off-the-shelf, gene-edited, dual-target CD19/BCMA cell therapy. To survive in a patient who didn’t donate the cells, it’s engineered on both sides of the rejection problem — the T-cell receptor is knocked out to limit graft-versus-host disease, and multiplex edits confer hypoimmunity so the patient’s own T and NK cells are less likely to recognize and clear it.

Persistence in the face of host rejection is the thing most allogeneic programs fail at; that’s the bar Qihan is being credentialed against.

And the indication is the tell. This isn’t oncology — it’s refractory systemic lupus erythematosus (SLE), the spearhead of the autoimmune CAR-T wave, where the addressable population dwarfs any single cancer.

Early investigator-initiated data are striking if still small: a 100% response signal in SLE-associated immune thrombocytopenia (six evaluable patients — five complete responses, one partial), with complete responses in four of five antiphospholipid-syndrome ITP patients across a broader 20-patient autoimmune cohort. The U.S. IND is cleared, dosing in the Phase I/IIa has begun, and there’s a parallel China CDE IND.

The credentialing point stands on its own: a Chinese company built a genuinely difficult allogeneic platform and got the FDA to fast-track it — in the disease area with the largest commercial prize. That’s not fast-follower behavior.

Pillar 3 — The pattern: velocity and credentialing

Here’s why the two belong in one issue. Satri-cel proves China can take a first-in-class asset through its own regulator to commercial reality faster than anyone else managed in solid tumors. QT-019B proves China can clear the FDA’s quality-and-conviction bar abroad. Domestic speed and foreign credibility — in the same two weeks.

The old framing — China = fast, cheap, derivative — survives only if you can argue China is good at exactly one of those axes. This fortnight breaks the argument on both.

When a country can simultaneously be first-to-market at home and FDA-credentialed abroad, the “fast-follower” label stops describing reality and starts costing you money.

Pillar 4 — The investor thesis: this is the China-to-West licensing setup

Why does a portfolio care? Because these two assets are, almost to spec, what ex-China partners in-license.

Western pharma in-licensing wants one of two things: a de-risked, now-commercial asset with human efficacy and a regulatory stamp, or an FDA-credentialed platform it can build a franchise on.

The fortnight produced one of each — satri-cel (commercial, randomized data, world-first label) and QT-019B (triple-crowned, platform, autoimmune TAM). That’s not a coincidence; it’s the maturation of the supply side of the licensing market.

And the price of that supply has re-rated. The era of treating Chinese assets as the bargain bin is over: average licensing upfronts have risen roughly 230%, from about $52M to $172M between 2022 and early 2026, prompting analysts to declare China no longer the “bargain basement” of biopharma.

China out-licensing hit a record ~$137.7B in 2025, with advanced modalities — CGT among them — making up roughly half of deal volume. Two world-firsts land directly into that bidding environment.

So the watch isn’t the approval. It’s the ex-China path. For satri-cel, that’s the North American ELIMYN18.2 study (CT041-ST-02) in CLDN18.2-positive gastric and pancreatic cancer — the bridge to a Western label or a partner.

For Qihan, it’s the U.S. Phase I/IIa execution and the first Western partnership or financing event. Those are the catalysts that convert “China led this fortnight” into “someone paid for it.”

The honest caveats (because credibility is the product)

A confident thesis that hides its weak points isn’t worth much. Here are mine.

Satri-cel’s benefit is real but modest, and the approval is China-only. Single-digit-month PFS, an OS signal that didn’t clear significance, and severe-but-expected toxicity. “International access” today means medical travel to one Shanghai center — not a foreign label. The early-line and perioperative expansion is genuinely promising but still early, partly investigator-initiated.

Qihan is early and small. A 100% signal in six patients is a reason to watch, not a reason to extrapolate. Designations accelerate review; they are not approvals, and the FDA grants — and occasionally rescinds — them on evolving data.

Geopolitics is now a line item. China-origin assets carry BIOSECURE-style counterparty and political risk that Western acquirers actively price in. That can compress multiples even on excellent science — and it cuts both ways on timing.

None of this negates the thesis. It sizes it. The shift is in direction and credibility, not in any single readout going China’s way.

What I’m watching from here

The catalysts that turn this from narrative into P&L:

• Any ex-China licensing move on satri-cel, and tangible ELIMYN18.2 progress in North America.

• Qihan’s U.S. trial execution and a first Western partner or financing event.

• Satri-cel real-world uptake, plus the pancreatic-adjuvant and perioperative/earlier-line reads — the data that decides whether this is a niche third-line drug or a platform.

• Deal-size discipline — whether upfronts for credentialed China CGT keep climbing, which is the market voting on exactly this thesis.

Names to watch

Not investment advice — a map of where the operating leverage sits. Do your own work.

• CARsgen Therapeutics ($2171.HK) — the cleanest public expression. Now holds the world’s only solid-tumor CAR-T approval, a portable conditioning platform, and an active Western bridge (ELIMYN18.2). The whole bull case is execution on label expansion and an ex-China deal.

• Qihan Biotech (private) — no ticker yet; watch for a Western partnership, a financing round, or an eventual IPO. The triple crown is the kind of credential that tends to precede one of those.

• The China CGT basket / acquirers — the broader read is that ex-China pharma needs to buy from this pipeline. Watch which large caps move first on credentialed China cell therapy; the deal, not the approval, is the tell.

For a decade the question was when China would catch up. This fortnight, it’s worth asking what the West now has to buy from them.

This post is for informational purposes only and is not investment advice. I am not a financial advisor. Early-stage clinical signals and regulatory designations are not approvals. Do your own research and consult a professional before making any investment decision.

Sources

• CARsgen — approval of satri-cel, world’s first solid-tumor CAR-T (PR Newswire)

• FiercePharma — CARsgen’s gastric CAR-T nabs world-first China approval

• The Lancet — CT041-ST-01 phase 2 pivotal trial (satri-cel)

• OncLive — satri-cel PFS/ORR detail in CLDN18.2+ G/GEJ cancer

• CGTLive — ELIMYN18.2 and the nab-paclitaxel lymphodepletion regimen

• PR Newswire — Jiahui International Cancer Center patient access

• BioSpace — Qihan QT-019B receives FDA RMAT and Breakthrough

• AllSci — QT-019B achieves the FDA “triple crown”

• BioSpace — FDA clearance of QT-019B IND for refractory SLE

• Business Wire — Qihan initiates dosing in Phase I/IIa

• FierceBiotech — China no longer the “bargain basement” (230% upfront jump)

• PharmaSource — China out-licensing hits record $137.7B in 2025

By Anthony Ao · THE NO VC Playbook · novc.substack.com

If we want to predict the future, a good starting point is to look at what we currently have on our plate.

To look ahead to 2027 and consider how the Cell and Gene Therapy landscape could change, let’s look at the full picture right now. I have picked five leading drug candidates (plus seven runner-ups) that could be pillars of the CGT industry next year.

Before we dive in, I am not saying that these are the “top five blockbuster drugs” or endorsing their chance of clinical success. Rather, I take these five shots as five tests of their respective questions and ask, “If the drug succeeds, what would shift in the field?”

If several of these programs progress, 2027 could strengthen the case that advanced therapy is becoming broader, more chronic-disease-oriented, and more pharma-owned.

If several stall, the lesson will not be that cell and gene therapy “failed.” It will be that the next bottleneck is no longer only biology. It is manufacturing, CMC, release testing, payer design, site-of-care logistics, long-term monitoring, and operational repeatability.

You can think of it as a flowchart, or a decision tree.

How I Chose these Five Programs

From the sea of hundreds of clinical trial assets, I applied the following five lenses to select those worth our attention.

That last filter is the most important. The point is not to predict five winners but to identify five assets that can tell us what kind of CGT market the big pharmas are actually building.

A. Five Programs, Five Market Questions

1. Anito-cel: the autologous CAR-T scale question

Company / status: Gilead / Kite, via the Arcellx lineage. Gilead completed its acquisition of Arcellx in April 2026, so the asset should not be described as an unchanged equal co-development structure.

Asset: anitocabtagene autoleucel, or anito-cel. Earlier materials refer to the product as CART-ddBCMA / ddBCMA.

Modality: investigational autologous BCMA-directed CAR-T cell therapy. The construct uses Arcellx’s D-Domain binder and includes 4-1BB and CD3-zeta signaling domains.

Indication / 2027 marker: Gilead announced that the FDA accepted the BLA for anito-cel as a fourth-line treatment for relapsed / refractory multiple myeloma, with a PDUFA target action date of December 23, 2026. The BLA is supported by the Phase 1 study NCT04155749 and the Phase 2 iMMagine-1 study NCT05396885.

Anito-cel is still investigational. 2027 becomes a launch, access, and positioning year only if the FDA approves it near the end of 2026.

What it tests: whether a differentiated autologous BCMA CAR-T can still become a scalable pharma franchise in a crowded myeloma market.

The strategic questions:

• If anito-cel is approved and launches well, does autologous CAR-T remain a durable commercial model?

• Or does the field still need allogeneic and in vivo approaches to solve access, cost, turnaround time, and treatment-center capacity?

2. Sura-vec: the chronic retina gene therapy question

Company: AbbVie / REGENXBIO.

Asset: surabgene lomparvovec, or sura-vec / ABBV-RGX-314.

Modality: investigational NAV AAV8 gene therapy encoding an anti-VEGF antibody fragment designed to inhibit VEGF activity in the eye.

Indication / 2027 marker: sura-vec is being developed in wet age-related macular degeneration and diabetic retinopathy. REGENXBIO’s May 2026 update said it expected topline data with AbbVie from the ATMOSPHERE and ASCENT pivotal wet AMD trials in Q4 2026, with global regulatory submissions expected in 2027 if the data support filing.

This should not be framed as “AAV will enter chronic disease” in general. The cleaner claim is narrower: sura-vec could test whether an ocular AAV therapy can produce a scalable product-market fit in a high-volume chronic retina market.

What it tests: whether a one-time or low-frequency ocular gene therapy can compete against established repeat anti-VEGF injection workflows.

The strategic question:

• If sura-vec data are strong enough to support filing, does gene therapy become a credible chronic-retina business model?

• Or do procedure burden, bilateral disease, durability expectations, payer treatment of upfront cost, and retina-specialist workflow limit adoption?

3. Arlo-cel: the multi-target myeloma question

Company: Bristol Myers Squibb, through the Juno Therapeutics lineage.

Asset: arlocabtagene autoleucel, or arlo-cel / BMS-986393. It is also identified as CC-95266.

Modality: investigational autologous CAR-T cell therapy targeting GPRC5D, a myeloma target distinct from BCMA. BMS describes QUINTESSENTIAL as a Phase 2 study of arlo-cel in relapsed / refractory multiple myeloma.

Indication / 2027 marker: the QUINTESSENTIAL study, NCT06297226, is an open-label, single-arm, multicenter Phase 2 study. BMS’s trial-in-progress poster lists about 175 planned patients across roughly 47 centers and gives an estimated primary-analysis completion in Q1 2027. ClinicalTrials.gov lists estimated primary completion as June 30, 2027.

This is not a fixed 2027 readout; the safer framing is a 2027 primary-analysis / primary-completion window. Arlo-cel remains investigational.

What it tests: whether myeloma becomes a multi-target cell therapy market after BCMA.

The strategic question:

• If arlo-cel produces competitive data, does pharma need multiple CAR-T targets inside the same tumor type?

• Does myeloma become the template for serial cell-therapy sequencing: BCMA first, GPRC5D next, and perhaps other targets or modalities around them?

4. Zimislecel: the metabolic cell therapy question

Company: Vertex.

Asset: zimislecel, formerly VX-880.

Modality: investigational allogeneic, stem-cell-derived, fully differentiated islet-cell therapy. Vertex describes zimislecel as an insulin-producing islet-cell therapy for type 1 diabetes.

Indication / patient population: type 1 diabetes patients with impaired awareness of hypoglycemia and severe hypoglycemic events. The current approach uses standard immunosuppression.

2027 marker: this is not a clean 2027 approval-cycle anchor. In its Q1 2026 update, Vertex said it had completed an internal manufacturing analysis, resumed dosing, treated multiple patients after resumption, and expected to provide updated study-completion timelines later in 2026.

This is not a broad diabetes cure. The current evidence base is a narrow, high-risk type 1 diabetes population requiring immunosuppression. The strategic read-through is broader than the current addressable population, but the factual claim must remain narrow.

What it tests: whether cell replacement can become credible metabolic medicine.

The strategic question:

• If dosing, CMC, durability, and benefit-risk continue to improve, does cell therapy become a serious metabolic-disease category?

• If timelines keep slipping, does the field learn that manufacturing reliability, release testing, and regulatory package quality now matter as much as proof-of-concept biology?

5. VERVE-102: the in vivo editing confidence question

Company / status: Lilly, after its completed acquisition of Verve Therapeutics in July 2025.

Asset: VERVE-102.

Modality: investigational in vivo base-editing medicine targeting PCSK9 in the liver after a single intravenous infusion. Lilly describes the product as using an mRNA encoding an adenine base editor plus a guide RNA targeting PCSK9, encapsulated in a lipid nanoparticle using GalNAc-LNP delivery technology.

Indication / patient population: the Phase 1b Heart-2 study is enrolling adults with heterozygous familial hypercholesterolemia or premature coronary artery disease who require additional LDL-C lowering despite maximally tolerated oral therapies.

2027 marker: Lilly reported interim Phase 1b Heart-2 data in May 2026, including dose-dependent reductions in PCSK9 and LDL-C, with durability observed up to 18 months. Lilly also said it planned to begin Phase 2 enrollment by the end of 2026.

This is not a 2027 approval story. It is a platform-confidence story. The phrase “cardiometabolic prevention” should be used carefully: the current enrolled population is high-risk LDL-C lowering, not broad primary prevention.

What it tests: whether one-time in vivo editing can earn enough safety, durability, and clinical-development confidence to become more than a rare-disease platform.

The strategic question:

• Can one-time genetic medicine compete in chronic cardiovascular risk management, where the standards for safety, off-target confidence, redosing, ethics, price, and follow-up are much stricter than in severe rare disease?

B. Additional Names Worth Tracking

Here are seven runner-ups that I think are worth tracking, without going as deep as the mentioned five shots.

1. Kelonia / proposed Lilly acquisition — KLN-1010

   1. Why it matters: KLN-1010 is the in vivo CAR-T wildcard. Kelonia describes KLN-1010 as an investigational in vivo gene therapy designed to generate anti-BCMA CAR-T cells in the body after a single dose.

   2. Status detail: Lilly announced an agreement to acquire Kelonia in April 2026, but Kelonia’s May 31, 2026 update still described the proposed acquisition as pending transaction close. The accurate framing is “Kelonia, under proposed Lilly acquisition,” not “Lilly-owned Kelonia.” This is a Phase 1 / early clinical story, not a near-term 2027 approval story.

2. J&J / Legend — CARVYKTI, cilta-cel

   1. Why it matters: CARVYKTI is the benchmark BCMA CAR-T product for the myeloma discussion.

   2. Status detail: CARTITUDE-6 compares DVRd followed by ciltacabtagene autoleucel versus DVRd followed by autologous stem-cell transplant in newly diagnosed multiple myeloma. CARVYKTI should not be described as approved in newly diagnosed myeloma. The FDA label is for adult relapsed / refractory multiple myeloma after at least one prior line of therapy including a proteasome inhibitor and an immunomodulatory agent, and in patients refractory to lenalidomide.

3. Bayer / BlueRock — bemdaneprocel

   1. Why it matters: this is the neurodegeneration horizon signal.

   2. Status detail: Bayer announced in September 2025 that the first Parkinson’s patient had been treated in BlueRock’s pivotal Phase 3 exPDite-2 trial. Bayer described the study as randomized, sham-surgery-controlled, and double-blind, with approximately 102 participants and a week-78 primary endpoint. Treat it as a horizon signal for whether cell therapy can move into neurodegeneration, not a clean 2027 readout.

4. Roche / Poseida — P-BCMA-ALLO1

   1. Why it matters: this is the allogeneic myeloma watchpoint.

   2. Status detail: Roche’s Poseida acquisition announcement highlighted donor-derived, off-the-shelf cell therapies and identified P-BCMA-ALLO1 as a lead allogeneic CAR-T targeting BCMA. Roche also stated that P-BCMA-ALLO1 had RMAT designation for relapsed / refractory multiple myeloma after three or more prior lines of therapy. Use it as a strategic comparator to autologous CAR-T, not as proof that off-the-shelf CAR-T has solved efficacy, persistence, safety, or commercial adoption.

5. Intellia / Regeneron — nex-z

   1. Why it matters: this is the in vivo CRISPR risk case.

   2. Status detail: Intellia announced in March 2026 that the FDA removed the clinical hold on the MAGNITUDE Phase 3 trial of nex-z in ATTR-CM. The company said prior holds followed liver-safety findings, and that mitigation measures included enhanced liver monitoring, steroid guidance, and additional eligibility exclusions. Use nex-z to illustrate how safety, monitoring, and regulatory confidence shape in vivo editing platform value, not as a simple verdict on CRISPR.

6. Sarepta / Roche — ELEVIDYS

   1. Why it matters: this is the AAV safety and regulatory cautionary example.

   2. Status detail: in November 2025, the FDA added a boxed warning for acute serious liver injury and acute liver failure and limited the ELEVIDYS indication to ambulatory Duchenne muscular dystrophy patients age four years and older with a confirmed DMD mutation. The FDA said the action followed reports of fatal acute liver failure in non-ambulatory patients. Use ELEVIDYS as a cautionary example for post-approval AAV safety and scrutiny, not as evidence that AAV as a class is broadly nonviable.

7. AstraZeneca / EsoBiotec

   1. Why it matters: this is the platform-deal example.

   2. Status detail: AstraZeneca announced in March 2025 that it would acquire EsoBiotec, including its in vivo delivery platform. AstraZeneca described EsoBiotec’s ENaBL platform as using targeted lentiviruses to deliver genetic instructions directly to immune cells through an intravenous injection, with the potential to reduce barriers associated with ex vivo manufacturing. This is not a product catalyst; it is a sign that pharma is buying delivery capability and manufacturing-model optionality.

C. If x is approved, then…

If anito-cel is approved and arlo-cel data support continued development

• Myeloma starts to look less like a single-target BCMA market and more like a multi-target cell therapy market.

• BCMA may be the first layer rather than the end state.

The next questions become sequencing, slot allocation, treatment-center throughput, toxicity management, and competition with bispecific antibodies.

The strategic implication: pharma may need more than one cell therapy asset inside the same tumor type. But that only becomes durable if the clinical benefit is strong enough to justify the operational burden.

If sura-vec pivotal data are positive enough to support filing

• Ocular gene therapy moves closer to high-volume chronic disease.

• The question shifts from “can AAV biology work?” to “can the system scale it?”

That system includes retina-specialist workflow, surgical delivery, durability expectations, payer treatment of upfront cost, bilateral disease, safety monitoring, and patient willingness to choose a procedure over repeated injections.

The strategic implication: gene therapy would no longer be only a rare-disease business model. But sura-vec would prove that point only for a specific ocular AAV setting, not for all chronic disease.

If zimislecel dosing and CMC stabilize

• Cell therapy becomes more credible in metabolic disease.

But the first addressable use case remains much narrower than broad diabetes care: high-risk type 1 diabetes patients, standard immunosuppression, long-term monitoring, and a complex benefit-risk calculation.

The strategic implication: cell therapy can move outside oncology, but only if manufacturing, patient selection, durability, and follow-up are strong enough.

If VERVE-102 continues to show durable LDL-C lowering and acceptable safety

• In vivo editing becomes a broader cardiometabolic platform conversation.

But prevention-oriented markets are unforgiving. A product like this must answer safety, off-target risk, redosing limitations, long-term monitoring, ethics, pricing, and patient-selection questions before anyone should extrapolate too aggressively.

The strategic implication: the market may begin valuing in vivo editing platforms beyond rare disease, but 2027 would still be about confidence-building, not approval.

If safety, CMC, or site-of-care limits slow the field

This may be the most important scenario.

If several programs stall, the lesson will be that the value of the field has shifted from scientific possibility to operational competence.

The strategic implication: the next CGT winner may be the company that solves manufacturing, delivery, safety management, reimbursement, and site-of-care design before it tries to sell the miracle.

Closing

The 2027 CGT market will not be decided by whether the science can produce miracles. That question has already been answered in quite a handful of settings.

The harder question is whether companies can turn those miracles into scalable systems that apply to a wider context.

1. Anito-cel and arlo-cel test whether autologous CAR-T can remain a pharma-scale business in myeloma.

2. Sura-vec tests whether ocular gene therapy can enter a high-volume chronic market.

3. Zimislecel tests whether cell replacement can become metabolic medicine, initially in a narrow high-risk T1D setting.

4. VERVE-102 tests whether in vivo editing can move from high-risk LDL-C lowering toward a broader cardiometabolic platform conversation.

These are the hypotheses that the big pharmas are asking with these assets, and a year from now, we will have the data and a good judgment and answer to each of these questions.

• If the programs progress, 2027 could make advanced therapy feel broader, more chronic-disease-oriented, and more pharma-owned.

• If they stall, 2027 could still be clarifying: the next CGT winner may be the company that solves manufacturing, delivery, safety management, and reimbursement before it tries to sell the miracle.

Sources

1. Lilly to acquire Kelonia Therapeutics to advance in vivo CAR-T cell therapies (Kelonia)

2. A single dose of Lilly’s PCSK9 base editor, VERVE-102, reduced PCSK9 by up to 88% and LDL-C by up to 62% (Eli Lilly)

3. Gilead Sciences Completes Acquisition of Arcellx Ahead of Potential Commercial Launch of Anito-cel (Gilead)

4. Gilead Sciences to Acquire Arcellx to Maximize Long-Term Potential of Anito-cel (Gilead)

5. Kite and Arcellx Continue Momentum with Advances in Anito-cel Multiple Myeloma Program (Gilead)

6. ABBV-RGX-314 for Retinal Diseases (REGENXBIO)

7. REGENXBIO Reports First Quarter 2026 Financial Results (REGENXBIO)

8. Study of Arlocabtagene Autoleucel (BMS-986393), NCT06297226 (ClinicalTrials.gov)

9. Vertex R&D Pipeline: Type 1 Diabetes (Vertex)

10. A Safety, Tolerability, and Efficacy Study of VX-880, NCT04786262 (ClinicalTrials.gov)

11. Vertex Reports First Quarter 2026 Financial Results (Vertex)

12. Lilly completes acquisition of Verve Therapeutics (Eli Lilly)

13. Kelonia Therapeutics Doses First Patient in Phase 1 INMMYCAR Study (Kelonia)

14. A Study to Evaluate a Novel Gene Therapy, NCT07075185 (ClinicalTrials.gov)

15. CARTITUDE-6: Daratumumab, Bortezomib, Lenalidomide and Dexamethasone, NCT05257083 (ClinicalTrials.gov)

16. CARVYKTI Package Insert and Medication Guide (FDA)

17. First Parkinson’s disease patient treated in BlueRock’s pivotal Phase III trial of bemdaneprocel (Bayer)

18. Roche to acquire Poseida Therapeutics (Roche)

19. Intellia Announces FDA Lift of Clinical Hold on MAGNITUDE Phase 3 Trial in ATTR-CM (Intellia)

20. FDA Approves New Safety Warning and Revised Indication Limiting Use for Elevidys (FDA)

21. AstraZeneca to acquire EsoBiotec to advance cell therapy ambition (AstraZeneca)

The better way to read it is a broader replacement-cycle question:
What comes after first-generation autologous CD19 CAR-T?

I do not mean that Kite or Gilead has formally declared KITE-753 a “Yescarta replacement.” KITE-753 is the most obvious public candidate for Kite’s next-generation autologous lymphoma architecture because it is designed around the three pressure points that define the first wave of CD19 CAR-T: antigen escape, inflammatory toxicity, and manufacturing time.

Kite’s EHA 2026 package describes KITE-753 as an investigational bicistronic autologous CAR-T that combines anti-CD19 and anti-CD20 targeting, dual costimulation through CD28 and 4-1BB, and a novel manufacturing process intended to preserve T-cell fitness. In plain language: two targets, two costimulatory domains, and a process designed to deliver a fitter product faster. (Kite/Gilead EHA 2026 release)

Kite is publicly advancing the kind of architecture one would expect if the company was preparing for a post-Yescarta product cycle.

The larger market structure is more interesting than the single asset, and it resolves along one axis: how the cell gets made. Autologous, then allogeneic, then in vivo.

That progression — not any single readout — is the spine of the next five years, and KITE-753 is one move along it.

Read against that spine, the honest first answer to “what replaces Yescarta” is that the opening moves are not replacements at all. The incumbent modality defends itself two ways.

That is the real issue. And it deserves the same scrutiny I just gave KITE-753 — which is exactly where most CGT commentary, including the bullish in vivo case, quietly stops.

KITE-753 turns “next generation” into a commercial question

KITE-753 matters because it is a next-generation construct from a sponsor that already owns commercial CAR-T infrastructure.

The product design is explicit. KITE-753 is a bicistronic autologous CAR-T directed against CD19 and CD20. The KITE DuoCore construct combines CD28 and 4-1BB costimulation. The manufacturing process is intended to preserve T-cell fitness. (Kite/Gilead EHA 2026 release)

Earlier Tandem and ASH-related materials explained why Kite is pushing this construct forward.

In CAR-T-naive patients at dose level 3, KITE-753 reported an 86% overall response rate and a 79% complete response rate after the post-cutoff update, with a median follow-up of 4.0 months. The same presentation described a rapid leukapheresis-to-delivery process. (Kite TCT 2026 deck)

Those numbers should not be over-read. The peer-reviewed Tandem abstract was appropriately cautious: as of March 18, 2025, only three KITE-753 patients had received dose level 3, all were in complete response at 4.4 months, and the expansion phase was ongoing. (Transplantation and Cellular Therapy abstract)

So the point is not that the asset is de-risked. It is not.

If KITE-753 moves quickly into pivotal or Phase 3 development in large B-cell lymphoma, it becomes the first obvious internal successor candidate capable of pressuring the economics and positioning of axi-cel/Yescarta. Because the product is built around the limitations that would force a successor strategy in the first place.

The relocation play: CAR-T leaves oncology

The first answer to “what’s next” is not a replacement. It is a relocation. The first defensive move for ex vivo CAR-T is not a better lymphoma cell — it is the same immune-reset logic of CD19 CAR-T, moved into autoimmune disease.

The academic signal

The premise is no longer theoretical. The academic signal came through clearly in the NEJM case series from Müller, Mackensen, Schett, and colleagues: CD19 CAR-T was evaluated in 15 patients with severe systemic lupus erythematosus, idiopathic inflammatory myositis, or systemic sclerosis, and the authors concluded that the approach appeared feasible, safe, and efficacious in that case series. (NEJM; PubMed)

The commercial question is different: which sponsor can convert immune reset from academic proof-of-concept into a reproducible, reimbursable, label-ready product?

Cabaletta (rese-cel)

Cabaletta is the purest public autologous autoimmune test. Rese-cel, formerly CABA-201, is a fully human CD19 CAR-T with 4-1BB costimulation. Cabaletta’s EULAR 2026 update ties the program to registrational timing, including a planned systemic sclerosis registrational program in 4Q26, mid-2027 topline data for the dermatomyositis/antisynthetase syndrome registrational cohort, and a 2H27 BLA strategy for adult and juvenile dermatomyositis. (Cabaletta EULAR 2026 update)

That is not a Yescarta-replacement story inside lymphoma. It is a category-expansion story. It says the same personalized cell-therapy infrastructure can earn a new durability premium in diseases where chronic biologics and immunosuppression dominate.

Novartis (rap-cel)

Novartis is testing whether rapid-manufactured rapcabtagene autoleucel can become the incumbent-grade autoimmune entrant. Its EULAR 2026 agenda included data for rap-cel/YTB323 in severe refractory SLE, idiopathic inflammatory myopathies, and diffuse cutaneous systemic sclerosis. (Novartis EULAR 2026 release) Its Phase 2 AUTOGRAPH-SLE/LN study describes one-time rap-cel after lymphodepletion in active refractory SLE or lupus nephritis, anchoring the program in a prospective development path rather than a one-off academic series. (Novartis AUTOGRAPH-SLE/LN trial page)

Kyverna (miv-cel)

Kyverna is trying to turn neuroimmunology and rheumatology into a franchise. Mivocabtagene autoleucel, or miv-cel/KYV-101, is an autologous fully human CD19 CAR-T. At EULAR 2026, Kyverna highlighted updated miv-cel data in ACPA-positive, treatment-refractory rheumatoid arthritis and noted that the Phase 2 portion of the COMPARE trial had been initiated and fully enrolled. (Kyverna EULAR 2026 update)

Fate (FT819 / FT839)

Fate complicates the lane because FT819 is not autologous. It is an off-the-shelf, iPSC-derived, CD19-targeting CAR-T product candidate. Fate’s EULAR 2026 update positions FT819 in SLE with less-intensive conditioning and outpatient/community-hospital administration, while FT839 adds dual CD19/CD38 targeting and “Sword and Shield” engineering. (Fate EULAR 2026 update)

Why relocation holds — or breaks

The common thread is that autoimmune CAR-T is trying to pull the CAR-T value proposition into diseases where the comparator is chronic disease control rather than another one-time oncology cell therapy.

That can be commercially defensible if remissions are durable, logistics become outpatient-friendly, and payers accept a one-time price in exchange for avoided chronic biologic spend.

It is fragile if lymphodepletion, infection risk, relapse after B-cell reconstitution, or uneven durability erodes the immune-reset narrative.

The re-engineering play: better cells for blood cancers

The re-engineering play is the more direct attempt at replacement: a better CAR-T for hematology, built by changing the architecture itself.

KITE-753 is the autologous exemplar because it combines dual antigen targeting, dual costimulation, and manufacturing optimization inside a sponsor that already knows how to commercialize CAR-T.

But the same architectural impulse is visible across allogeneic programs.

Caribou (CB-011)

Caribou‘s CB-011 is an allogeneic anti-BCMA CAR-T in multiple myeloma with immune cloaking. Caribou’s RMAT release reported 92% ORR, 75% or greater CR, and 91% MRD negativity in a 12-patient BCMA-naive recommended-dose cohort, while emphasizing no observed GVHD at any dose level. (Caribou RMAT release)

Cellectis (eti-cel)

Cellectis‘ eti-cel is a cleaner dual-target lymphoma analogue. The company describes eti-cel as an allogeneic dual CAR-T targeting CD20 and CD22 in relapsed/refractory non-Hodgkin lymphoma. Cellectis reported preliminary Phase 1 data with an 88% ORR and a 63% complete response rate at the current dose level, and its 2026 strategy update framed further development around improving response depth and durability. (Cellectis ASH 2025 update; Cellectis 2026 strategy)

The approval reality check

This is where discipline matters.

The FDA’s approved cellular and gene therapy product list includes multiple autologous CAR-T products, including Kymriah, Yescarta, Tecartus, Breyanzi, Abecma, Carvykti, and Aucatzyl. As of the FDA’s April 23, 2026 update, I do not see an approved allogeneic CAR-T product on that list. (FDA approved cellular and gene therapy products)

Therefore, allogeneic CAR-T should be treated as architecture under clinical validation, not as a de-risked commercial substitute.

Stated plainly, the spine is this:

Autologous owns the approved base.
Allogeneic is the most active architectural challenger.
In vivo is the potential category breaker.

The 24-month counter-scenario: in vivo CAR-T eats the stack

The counter-scenario is simple: while ex vivo sponsors compete on better cells, in vivo sponsors try to delete the cell factory.

The literal claim that manufacturing cost goes to zero is too strong. Vectors, LNPs, quality systems, release testing, cold chain, clinical monitoring, and pharmacovigilance do not disappear.

But the primary sources do support a directionally radical idea: eliminate patient-specific apheresis, ex vivo engineering, expansion, reinfusion logistics, and often lymphodepleting chemotherapy.

If that works clinically, the cost stack does not just shrink. The basis of competition changes.

Capstan / AbbVie (CPTX2309)

Capstan‘s CPTX2309 is the most important clinical autoimmune signal in this counter-scenario. AbbVie acquired Capstan, including CPTX2309, a Phase 1 in vivo targeted lipid nanoparticle anti-CD19 CAR-T candidate for B-cell-mediated autoimmune diseases. AbbVie described the asset as a tLNP platform designed to deliver RNA payloads capable of engineering specific cell types in vivo. (AbbVie-Capstan acquisition agreement; AbbVie completion release)

Kelonia / Lilly (KLN-1010)

Kelonia is the oncology counterweight. Lilly agreed to acquire Kelonia for up to $7.0 billion, including $3.25 billion upfront. Kelonia’s lead program, KLN-1010, is a Phase 1 lentiviral in vivo CAR-T for relapsed/refractory multiple myeloma. Lilly framed the acquisition around in vivo gene delivery and integration technology with potentially broad applicability. (Lilly-Kelonia acquisition release)

That is the capital signal that should discipline every ex vivo bull case. A major pharma was willing to pay an autologous-CAR-T-scale price for a platform whose core promise is to make autologous logistics unnecessary.

The broader deal cluster

The broader deal cluster says the same thing.

Kite agreed to acquire Interius BioTherapeutics for $350 million to bring in an integrating in vivo platform designed to generate CAR-T cells directly inside the patient’s body. (Kite-Interius acquisition release)

AstraZeneca acquired EsoBiotec, whose in vivo cell-therapy platform is designed to engineer immune cells inside the body and potentially deliver cell therapy in minutes rather than weeks. (AstraZeneca-EsoBiotec acquisition release; completion release)

Umoja‘s UB-VV111 has FDA Fast Track designation in relapsed/refractory large B-cell lymphoma and chronic lymphocytic leukemia. Umoja says UB-VV111 generates CD19-directed CAR-T cells in vivo and is designed to address high manufacturing costs, long wait times, burdensome treatment processes, and limited product availability. (Umoja UB-VV111 Fast Track release)

Tessera remains preclinical, but its ASGCT 2026 update keeps the platform in the strategic conversation: an all-RNA Gene Writing approach delivered in vivo, with CAR-T applications included in the company’s preclinical disclosure package. (Tessera ASGCT 2026 update)

Legend’s human signal (LB2501)

Legend adds the newest oncology signal — and it earns the same discipline I applied to KITE-753’s three patients. Reuters reported on June 2, 2026 that Legend’s in vivo candidate LB2501 produced responses in all six patients in its higher-dose group, with five complete responses. Six patients, one dose level, no durability yet: by the standard I just used on Kite, that is a hypothesis, not a result. What makes it matter is not the response rate. It is that an in vivo construct generated a human oncology response at all — the category cleared a bar it had never cleared before. (Reuters)

Integrating vs. transient: the variable that sets the clock

The bull case treats “in vivo” as one category.

It is two, and the split is the part of the story that actually decides the 2027-2028 clock.

1. Most of the named programs deliver an integrating lentiviral vector inside the patient.
   Kelonia’s KLN-1010 is lentiviral; Interius, Umoja’s UB-VV111, and EsoBiotec’s platform are all lentiviral; so is Legend’s LB2501 — the one generating the human oncology data the bulls are pointing at. Integration is what makes the CAR durable. It is also what carries insertional-mutagenesis risk — and that risk is now being run inside the body, stripped of the ex vivo manufacturing, integration-site analysis, and release testing that currently sit between an engineered cell and a patient. The single most scrutinized step in cell therapy is being moved out of a GMP suite and into a vein. That, not efficacy, is the most plausible reason a regulator slows the clock — and it lands hardest on exactly the program with the best early data.

2. The other camp goes non-viral.
   Capstan’s CPTX2309 is an mRNA-LNP: no integration, no insertional mutagenesis — and, because the CAR is expressed transiently, an open question about whether a single pass depletes B cells deeply and durably enough to matter. Tessera’s RNA gene writing is non-viral by design, trading the genotoxicity problem for the burden of editing the genome accurately without a vector to carry the payload. Either way, the choice is a trade, not a free lunch.

No in vivo program gets safety and durability for free. Each architecture still has to prove the exact thing the other one is handed.

What the capital does and does not prove

The deal cluster is the strongest argument for in vivo, and the one most likely to be over-read. Pharma buys conviction; it also buys optionality and defensive hedges, and from the outside the two look identical.

A $350 million Interius deal is a cheap option on a platform, not a verdict on it. The AbbVie–Capstan and Lilly–Kelonia sums are harder to wave away — those are closer to the price of belief than the price of insurance — but even a multibillion-dollar bet can be a large company buying the right not to be late.

So the honest read is narrower than “pharma paid, therefore in vivo wins.”

It is this: the people holding the most CAR-T data are no longer willing to bet only on a better cell. That is a real signal. It is not yet a result.

The question is whether one or two programs can show enough pharmacodynamic B-cell depletion, early clinical response, manageable CRS/ICANS, and repeatable vector or LNP manufacturing to convince physicians and payers that ex vivo improvements are incremental while in vivo changes the treatment model.

If Capstan, Kelonia, Interius, Umoja, Tessera, Legend, or another in vivo program produces credible human data by 2027-2028, KITE-753’s benchmark will no longer be Yescarta alone.

It will be a vial or infusion that asks why the patient ever needed an ex vivo cell factory.

What would confirm the replacement-cycle thesis?

There are five signals I am watching. Hit three of these by mid-2027, and the replacement cycle is real, datable, and was callable early. Hit zero or one and KITE-753 stays an elegant dataset in search of a market.

1. KITE-753 pivotal initiation, on the clock.
A registrational or Phase 3 start in large B-cell lymphoma within twelve months of the EHA 2026 update — by mid-2027.

Slower than that, and “lifecycle management” was only an aspiration.

2. A named architectural answer from BMS or Novartis.
A dual-target, dual-costimulation, or rapid-manufactured successor to Breyanzi or Kymriah in the clinic by the end of 2027.

Silence through 2027 says the incumbents see no replacement cycle worth chasing.

3. An economic comparator a CFO would recognize.
A sponsor deck, label, or health-economics dataset that prices faster manufacturing or better durability as share defense — reduced bridging, outpatient eligibility, a premium justified against incumbent CAR-T.

4. Autoimmune durability that survives the payer.
Drug-free remission past twelve months in a registrational cohort, plus a first coverage decision that treats one-time reset as cheaper than chronic biologic spend.

Response headlines do not count; durability after B-cell return, and a payer “yes,” do.

5. In vivo data that clear the operational bar.
The first in vivo CAR-T with n>20, durable responses past six months, CRS/ICANS a community center can manage, and — for the integrating lentiviral programs — a clean enough safety readout to satisfy a regulator.

The KITE-753 thesis

KITE-753 is the right anchor for the ex vivo view: a Big-3 sponsor putting a genuinely next-generation autologous architecture — dual antigen, dual costimulation, faster manufacturing — into the public record.

• The incumbent modality is buying time two ways, by relocation and by re-engineering, and both presume the cell factory still has room to run.

• Allogeneic is the most credible of those bets and still has not produced a single approved product.

• The capital, meanwhile, has stopped flowing only toward better cells and started flowing toward no cells — not uniformly, not all conviction, but unmistakably in one direction.

KITE-753 is the best version of the old question: which cell wins. The bet worth making is that the question itself is expiring — that what comes after Yescarta is not a better cell, but no cell factory at all.

Capital is the symptom. Displacement is the cause.

A company survives only if it has a program that clears all three — which is why a diversified pharma can lose a single program and continue, while a pure-play cannot.

A cell therapy can produce excellent clinical responses and still become strategically unattractive if another modality becomes “good enough,” easier to prescribe, faster to deliver, and cheaper to scale.

Conversely, a smaller company can remain valuable if its biology is compelling and its indication has not yet been colonized by a lower-friction competitor.

This is the CAR-T triangle:

Efficacy proves the therapy can matter.
Safety proves it can be used.
Competitive capital proves it can survive.

Two legs can create a promising program. All three are required to create a durable company.

The triangle

Efficacy asks whether the therapy changes the disease at a clinically meaningful magnitude.

In lymphoma, that means complete response, durability, and survival-related read-throughs.

In autoimmunity, it means deep B-cell depletion, clinical remission, steroid-sparing benefit, durability, and a credible retreatment logic.

Safety asks whether the biology can be delivered to real patients.

That means cytokine-release syndrome, ICANS, infections, cytopenias, hypogammaglobulinemia, organ-specific risk, and the tolerability of lymphodepletion or any preconditioning regimen.

Competitive capital asks whether the program remains fundable after the market has seen the alternatives.

The money is the visible output. The hidden input is displacement risk.

A board, pharma portfolio committee, or crossover investor will not only ask, “Does this CAR-T work?”

It will ask:

“Will enough physicians and patients choose this product over something available off the shelf, with less infrastructure and faster time to treatment?”

That is why the third leg is not simply a cash balance. Capital tends to disappear after the competitive position deteriorates. When that happens, the proximate cause is a financing decision. The root cause is a change in the market’s read-through of the asset.

Case study 1: JNJ-4496 won biology, then lost the market

On April 30, 2026, Johnson & Johnson announced that it would discontinue development of JNJ-90014496, also called JNJ-4496, a dual-targeting autologous anti-CD19/CD20 CAR-T for relapsed or refractory large B-cell lymphoma. J&J described the discontinuation as a strategic business decision based on portfolio priorities and its assessment of the evolving LBCL treatment landscape. [1]

That decision was striking because the early clinical signal was strong.

In June 2025, J&J reported first Phase 1b data in CAR-T-naive relapsed or refractory LBCL. At the recommended Phase 2 dose of 75 × 10⁶ CAR-positive T cells, the complete response rate was 80% in one manufacturing cohort and 75% in another. Safety was not blank-check benign, as the immune-effector toxicity profile looked clinically manageable: among 25 treated patients, J&J reported no grade 3 or 4 CRS, two ICANS events including one grade 3 event, grade 3 or 4 treatment-emergent adverse events in 84%, grade 3 or 4 neutropenia in 72%, serious treatment-emergent adverse events in 28%, and one grade 3 infection. [2]

J&J had also paid meaningfully for the asset.

In May 2023, Janssen entered a worldwide collaboration and license agreement with Cellular Biomedicine Group for C-CAR039 and C-CAR066, with CBMG disclosing a $245 million upfront payment. [3]

So the scientific readout did not look like the obvious reason to stop.

J&J did not name a single competitor modality in the discontinuation statement. My interpretation is that the market readout did.

• EPKINLY’s pivotal LBCL data supporting accelerated approval showed an overall response rate of 61%, complete response rate of 38%, and median duration of response of 15.6 months. [4]

• COLUMVI’s FDA approval was supported by a 56% overall response rate and 43% complete response rate in a single-arm study, and Roche described it as a fixed-duration bispecific antibody regimen. [5]

These complete response rates are numerically below the early JNJ-4496 signal, but the operating model is different.

That is the third leg.

JNJ-4496 did not need to lose to a better CAR-T. It needed only to face a modality that was sufficiently active and much easier to deliver.

For a pharma portfolio committee, the registration question then changes from:

“Can we get this product approved?”

to:

“After substantial additional registration and launch investment, will this product earn a durable place in an LBCL market already served by off-the-shelf bispecifics and incumbent CAR-Ts?”

The lesson is not that CAR-T is over.

The lesson is that, in crowded oncology indications, efficacy and safety can be necessary but still insufficient. Competitive displacement can convert a scientifically attractive asset into a strategically unattractive one.

Case study 2: Fate Therapeutics shows another third-leg failure mode

Fate Therapeutics is a useful counterexample because its failure mode was not the same as JNJ-4496’s.

Fate’s problem was not a named bispecific antibody displacing its iPSC-derived cell-therapy platform. Its problem was that a major partner withdrew, forcing the company to reprioritize around assets and indications where the platform had a clearer reason to exist.

In April 2020, Fate and Janssen announced a collaboration to develop iPSC-derived cell therapies for cancer. The transaction included $50 million upfront, a $50 million equity investment, up to $1.8 billion in development and regulatory milestones, and up to $1.2 billion in commercial milestones, making the total headline opportunity roughly $3 billion. [6]

In January 2023, Fate disclosed that Janssen had elected to terminate the collaboration. Fate later discontinued several oncology programs, including FT516, FT596, FT538, and FT536, and concentrated resources on FT819 and other prioritized candidates. [7]

This is still a third-leg event, but the chain is different.

• JNJ-4496 shows competitive displacement.

• Fate shows partner withdrawal and internal portfolio reset.

From inside a biotech, however, the practical result is similar. The asset may retain biological promise, but the funding path collapses unless the company can reposition into a more defensible clinical setting.

Fate’s repositioning is important.

Fate survived strategically by moving the platform toward a terrain where cell therapy may have a differentiated mechanistic rationale and where the competitive map is less saturated than LBCL.

Case study 3: Cabaletta is the live test

Cabaletta’s lead program, rese-cel, formerly CABA-201, is an autologous CD19-directed CAR-T using a 4-1BB costimulatory domain for autoimmune diseases.

The strategic question is whether Cabaletta can clear all three legs before a non-cell competitor does to autoimmune disease what bispecific antibodies did to LBCL.

On efficacy, the early autoimmune data are no longer merely theoretical.

At ACR Convergence 2025, Cabaletta reported that seven of eight lupus patients with sufficient follow-up achieved either DORIS remission or renal response: three of four systemic lupus erythematosus patients with at least three months of follow-up achieved DORIS remission, the fourth had pure class V lupus nephritis and achieved complete renal response, and three of four lupus nephritis patients with sufficient follow-up achieved renal response. In myositis, all four patients with dermatomyositis or antisynthetase syndrome who reached the week-16 landmark met the total improvement score threshold. [11]

On safety and dose delivery, the story is still evolving.

Cabaletta’s autoimmune strategy originally relied on lymphodepletion, but the company is now explicitly testing a no-preconditioning approach, meaning rese-cel dosing without lymphodepleting chemotherapy.

In May 2026, Cabaletta presented no-preconditioning RESET-PV data: among four patients dosed at the lowest rese-cel dose, three had CD19-positive B-cell depletion below 0.1 cells/µL, two with at least six months of follow-up achieved clinically meaningful complete remission, and the company reported no dose-limiting toxicities, no ICANS, and one grade 1 CRS event. [12]

On regulatory path, Cabaletta has reduced some clinical-development uncertainty but not eliminated it.

On balance-sheet capital, the surface picture changed materially in May 2026.

The unresolved question is no longer about whether Cabaletta can raise money in the abstract.

It just did.

The sharper question is whether a lower-friction modality can enter lupus, myositis, or systemic sclerosis quickly enough to change the strategic underwriting of rese-cel.

No bispecific is approved in any autoimmune indication today — but the first candidates have already entered the clinic. Roche’s mosunetuzumab (CD20×CD3) is in Phase 1 in SLE, and Xencor’s XmAb657 (CD19×CD3), purpose-built for autoimmunity, has begun first-in-human testing in idiopathic inflammatory myopathy — i.e., in myositis, rese-cel’s own lead indication. rese-cel leads these programs by years, so I still treat autoimmune as a less-displaced setting than LBCL. But less-displaced is not undisplaced: the displacement clock has started, and the thesis now rests on rese-cel converting its head start into approvals and durability data before a bispecific proves CAR-T-like outcomes with off-the-shelf convenience.

• JNJ-4496 had an impressive response rate in an indication where bispecific antibodies had already reset physician expectations around convenience.

• Cabaletta has early clinical activity in indications where deep B-cell depletion may matter and where the modality battle is not yet decided.

My prediction is conditional but clear:

The key phrase is “if all three.” Two legs are not enough.

Why two legs are not enough

The triangle is useful because it explains why apparently contradictory outcomes can all be rational.

Efficacy plus safety, without competitive capital, gives JNJ-4496.

In my interpretation of J&J’s stated portfolio decision, a high complete-response signal and manageable CRS/ICANS profile were not enough to preserve an LBCL program in a market already moving toward off-the-shelf bispecifics and away from marginally differentiated autologous cell therapies. [1, 2, 4, 5]

Efficacy plus capital, without enough safety, is especially fragile in autoimmunity.

Oncology accepts more toxicity because the treated population is often refractory, relapsed, and facing near-term mortality.

Autoimmune disease is different.

A therapy for lupus, myositis, pemphigus, or systemic sclerosis must justify CRS risk, infection risk, cytopenias, hypogammaglobulinemia, fertility-relevant chemotherapy exposure, and hospitalization burden against a lower mortality baseline.

This is why data without conditioning or preconditioning matter. Not because they prove the entire platform, but because they directly attack the tolerability problem that will determine adoption outside oncology. [12, 9]

Safety plus capital, without differentiated efficacy, gives a platform story rather than a product.

Fate’s Janssen collaboration had substantial headline economics, but when the partnership ended Fate still had to discontinue multiple programs and concentrate on a narrower set of candidates. [6, 7]

Capital can rent time. It cannot substitute for a clinical effect that changes the sponsor’s strategic calculus.

All three legs create the ideal company.

The ideal CAR-T company produces clinically decisive efficacy, delivers it with a safety profile appropriate to the target population, and operates in an indication where the competitive landscape still rewards cell therapy’s complexity.

If any leg is missing, the company faces a trade-off:

Better data but worse access.
Safer delivery but weaker efficacy.
Enough cash today but no strategic reason for the next dollar tomorrow.

What would make this view wrong

First: competitive displacement in autoimmunity.

The main risk to treating autoimmune as a less-displaced setting is a non-cell modality that achieves comparable functional B-cell depletion, durable remission, and steroid-sparing benefit with easier logistics.

XmAb657 (CD19×CD3) is in Phase 1 in myositis, rese-cel’s lead indication; mosunetuzumab (CD20×CD3) is in Phase 1 in SLE. If either delivers CAR-T-like depletion with bispecific logistics, it compresses rese-cel’s window the way bispecifics compressed JNJ-4496’s in LBCL.

Third: timing risk.

Cabaletta’s May 2026 financing reduces immediate balance-sheet pressure, but it does not remove execution risk.

The company still has to generate enough clinical and regulatory evidence before the next capital window closes or before competitors change the perceived value of the asset. [15, 16, 14]

Most investors begin with safety because it is the familiar biotech risk. I begin with competitive displacement because it is the third leg’s leading indicator.

By the time a program is described as “capital constrained,” the market has often already decided that something else is more fundable.

Conclusion

JNJ-4496’s discontinuation is an inflection point, not an obituary for CAR-T.

It shows that clinical potency is no longer enough in indications where easier modalities have become good enough.

Fate shows that even a well-funded platform can be forced to reposition when the partner and portfolio logic break.

Cabaletta is the live experiment on the other side of the map: an autoimmune CAR-T company with early efficacy, an evolving safety strategy, a refreshed balance sheet, and a competitive landscape that, unlike LBCL, is not yet defined by an already approved off-the-shelf substitute.

The companies most likely to survive the next five years will not be the ones with the most elegant constructs alone.

They will be the ones that choose indications where CAR-T’s biological advantage is decisive, where toxicity is proportionate to disease severity, and where the next dollar of capital sees strategic leverage rather than imminent displacement.

Disclosure

I work with Phacilitate and SaxoCell. This analysis is based only on public sources and does not use confidential information, employer-owned information, or private relationship capital. Phacilitate’s convening role and my SaxoCell work did not direct this analysis. As of publication, I hold a position in Cabaletta Bio (CABA) and no direct public equity positions in Johnson & Johnson, Fate Therapeutics, Roche, or Xencor. This article is not investment advice.

References

1. Johnson & Johnson. “Johnson & Johnson statement on investigational programs in large B-cell lymphoma.” April 30, 2026.

2. Johnson & Johnson. “Johnson & Johnson’s dual-targeting CAR-T cell therapy shows encouraging first results in large B-cell lymphoma.” June 13, 2025.

3. Cellular Biomedicine Group. “Cellular Biomedicine Group Announces Exclusive Collaboration and License Agreement with Janssen to Develop and Commercialize anti-CD19 & CD20 Bi-Specific and anti-CD20 CAR-Ts for Non-Hodgkin Lymphoma.” May 2, 2023.

4. FDA. “FDA grants accelerated approval to epcoritamab-bysp for relapsed or refractory diffuse large B-cell lymphoma and high-grade B-cell lymphoma.” May 19, 2023.

5. Genentech. “FDA approves Genentech’s Columvi, the first and only bispecific antibody with a fixed-duration treatment for people with relapsed or refractory diffuse large B-cell lymphoma.” June 15, 2023.

6. Fate Therapeutics. “Fate Therapeutics Announces Worldwide Collaboration with Janssen for Novel iPSC-derived Cell-based Cancer Immunotherapies.” April 2, 2020.

7. Fate Therapeutics. “Fate Therapeutics Announces Termination of Collaboration Agreement with Janssen.” January 5, 2023.

8. Fate Therapeutics. “Fate Therapeutics Receives Regenerative Medicine Advanced Therapy Designation from FDA for FT819 to Treat Moderate to Severe Systemic Lupus Erythematosus.” April 14, 2025.

9. Fate Therapeutics. “Fate Therapeutics Showcases FT819 Clinical Activity in SLE without the use of Conditioning Chemotherapy at the 2026 ASGCT Annual Meeting.” May 11, 2026.

10. Fate Therapeutics. “Fate Therapeutics Announces Presentations at 2026 ASCO and EULAR Annual Meetings Highlighting Off-the-Shelf CAR T-cell Therapy Pipeline for Cancer and Autoimmune Diseases.” May 21, 2026.

11. Cabaletta Bio. “Cabaletta Bio Presents Positive Clinical Data and Development Updates for Rese-cel at ACR Convergence 2025.” October 27, 2025.

12. Cabaletta Bio. “Cabaletta Bio Presents Preconditioning-free Clinical Data and Automated Manufacturing Translational Data for Rese-cel at ASGCT 2026 Annual Meeting.” May 14, 2026.

13. Cabaletta Bio. “Cabaletta Bio Announces 2027 Rese-cel BLA Submission Anticipated in Myositis Following Recent FDA Alignment on Registrational Cohorts.” May 15, 2025.

14. Cabaletta Bio. “Cabaletta Bio Announces 2026 Strategic Priorities.” January 12, 2026.

15. Cabaletta Bio. “Cabaletta Bio Reports First Quarter 2026 Financial Results and Provides Business Update.” May 14, 2026.

16. Cabaletta Bio. “Cabaletta Bio Announces Pricing of $150 Million Underwritten Offering.” May 4, 2026.

1. Autologous CAR-T retains its durability advantage in refractory hematologic cancers.

2. Allogeneic continues to chase scale.

3. In vivo is becoming the third path — and over the last fourteen months it has attracted more capital, faster, than either of the other two did at the same stage.

In the last fourteen months, AbbVie has paid $2.1 billion for Capstan Therapeutics. Eli Lilly has paid up to $2.4 billion for Orna Therapeutics and up to another $7 billion for Kelonia. Bristol Myers Squibb has paid $1.5 billion for Orbital Therapeutics. AstraZeneca has picked up EsoBiotec for up to $1 billion. Gilead’s Kite has acquired Interius BioTherapeutics for $350 million. Combined deal value: well north of $14 billion, on top of separate licensing transactions like Kite’s deal with Pregene Biopharma (up to $1.64 billion) and a string of late-stage venture rounds for the independents.

The new paper, led by Angela R. Corrigan with E. John Wherry, Drew Weissman, and the Michael Betts lab, is the latest academic data point in that thesis. It is also, in a way, the academic-side mirror of the commercial bet.

Weissman is a co-founder of Capstan — the company AbbVie just bought.

The paper is what that platform’s underlying biology looks like while still wearing its university lab coat.

What the paper actually shows

The researchers used a lipid nanoparticle decorated with fractalkine — the chemokine also known as CX3CL1 — to deliver mRNA selectively to T cells expressing its receptor, CX3CR1. CX3CR1 is not the antigen-specific T-cell receptor; it is a chemokine receptor associated with cytotoxic effector and antigen-experienced CD8 T-cell states. Using fractalkine as a targeting ligand lets the nanoparticle home in on a defined functional state rather than a broad cell type.

The numbers are striking. In mice, the platform reached up to ~90% of effector CD8 T cells in blood and spleen. In rhesus macaques, the figure was close to 100% of peripheral effector T cells. Expression remained transient, which is the design point — the goal is to hand the cells a short-lived instruction, not to rewrite them.

The instructions tested in the paper were illustrative. In mice, the cargo was IL-2 mRNA, which prompted treated T cells to secrete the cytokine. In macaques, the cargo was mRNA encoding CD62L — a homing molecule that effector T cells normally lack. That is more interesting than it sounds. CD62L is the badge that lets T cells enter lymph nodes; effector T cells, having shed it, are confined largely to the vasculature. Restoring CD62L expression, even briefly, expands where these cells can go.

The authors describe the platform as enabling rapid, efficient, and transient in vivo modification of effector T cells. The intervention is not gene editing. It is closer to a temporary software update, delivered to a defined cell state, that fades as the mRNA decays.

The commercial landscape

The Penn paper sits inside a field that has, very quickly, gone from elegant preprints to billion-dollar M&A. It is worth knowing who is in it.

Capstan Therapeutics (AbbVie, $2.1 billion, June 2025) is the most direct commercial analogue to the Corrigan paper. The platform uses targeted lipid nanoparticles to deliver mRNA to specific T-cell subsets. Lead asset CPTX2309 is a CD8-targeted anti-CD19 in vivo CAR-T in Phase 1 for B-cell-driven autoimmune disease. The deal produced roughly a 6x return on the ~$340 million of venture capital Capstan had raised, including a $165 million Series A in 2022 backed by the VC arms of Pfizer, Bayer, Novartis, Lilly, and BMS.

Orna Therapeutics (Eli Lilly, up to $2.4 billion, February 2026) brings a circular-RNA-plus-LNP platform, with a lead anti-CD19 in vivo CAR for B-cell autoimmune disease. Lilly’s strategic logic became clearer two months later when it agreed to acquire Kelonia.

Kelonia Therapeutics (Eli Lilly, up to $7 billion, April 2026) is the largest deal in the category. Kelonia’s platform uses a lentiviral iGPS delivery system rather than LNPs, and its lead asset KLN-1010 is a BCMA-targeting in vivo CAR-T for relapsed/refractory multiple myeloma. Kelonia surprised the field by presenting first-in-human data at ASH 2025: three patients, all reaching minimal residual disease–negative status by one month, holding through three. The dataset is small, but it is the first credible clinical signal for in vivo CAR-T in cancer.

Orbital Therapeutics (BMS, $1.5 billion, October 2025) is the other circular-RNA-plus-LNP play.

EsoBiotec (AstraZeneca, up to $1 billion, March 2025) and Interius BioTherapeutics (Gilead/Kite, $350 million, August 2025) both run on lentiviral vectors rather than LNPs. Interius’s INT2104 was the first in vivo CAR-T to enter clinical trials in Europe, in January 2025.

The independent players still in the field are now among the most-watched companies in cell therapy:

Umoja Biopharma runs a VivoVec lentivirus-LNP hybrid platform with multiple programs in or approaching the clinic. UB-VV111 (B-cell cancers) is partnered with AbbVie and has FDA Fast Track designation; UB-VV400/410 (CD22) is in Phase 1. Umoja raised $100 million in its 2025 Series C and is widely seen as the most obvious remaining acquisition target.

Renagade Therapeutics is building a multi-organ targeted LNP platform.

Sana Biotechnology is the only major public-market name in the cohort.

Myeloid Therapeutics is pursuing mRNA-LNP and presented first-in-human in vivo mRNA CAR data at ASCO 2025.

Pregene Biopharma, China-based, signed a $1.64 billion deal with Kite later in 2025 ($120 million upfront), signaling that US and European pharma majors are willing to source the technology offshore. Suzhou-based Starna Therapeutics also has in vivo CAR-T in the clinic, backed by a $44 million Series B that included Lilly Asia Ventures.

In the same window, AstraZeneca closed its $1.2 billion acquisition of Gracell (autologous manufacturing) and Roche acquired Poseida for up to $1.5 billion (allogeneic). The cell therapy stack as a whole is consolidating, but the in vivo branch is consolidating fastest.

What’s still open

A lot.

Kelonia’s ASH data covers three patients. The Capstan, Orna, and EsoBiotec programs are all in Phase 1. Most in vivo CAR-T candidates are aimed first at autoimmune disease rather than cancer — partly because the durability bar is lower, partly because reduced toxicity fits autoimmune use better. Whether the same approaches work in solid tumors, where target cells are rare, exhausted, or tissue-resident, is unresolved.

Several technical questions cut across the field.

• Can ligand-conjugated LNPs be dosed repeatedly without immunogenicity?

• How well does blood-compartment targeting translate to lymphoid tissue, bone marrow, or tumor stroma?

• What is the right durability profile for a given indication — transient for autoimmune reset, longer for oncology?

• How predictive are mouse and macaque models for human disease?

• Can the platforms be manufactured at clinical scale with the reproducibility regulators expect from a ligand-conjugated, formulation-dependent product?

The bottom line

The Corrigan paper does not, on its own, change the field. The acquisitions changed the field. The paper is the academic correlate — a clean demonstration that targeted mRNA-LNP delivery can hit a well-defined immune-cell state at high efficiency in both mice and primates, with transient expression and minimal off-target activation.

What it signals, for anyone tracking advanced therapies as a business: cell therapy is not pivoting from ex vivo to in vivo. It is bifurcating.

The question for platform companies, investors, and strategic partners is no longer whether in vivo immune-cell programming matters. The capital markets have answered that. The question is which delivery architecture — targeted LNP-mRNA, lentiviral, circular-RNA-LNP, or lentivirus-LNP hybrid — wins which indication, and which of the independent players still has the strategic optionality to define the next round of deals.

For a field that did not have a clinical proof-of-concept eighteen months ago, that is a remarkable position to be in.

Sources

• Angela R. Corrigan et al., “In vivo reprogramming of cytotoxic effector CD8 T cells via fractalkine-conjugated mRNA-LNPs,” Science Immunology 11, no. 119 (May 8, 2026): eaec3436. DOI: 10.1126/sciimmunol.aec3436.

• AbbVie / Capstan Therapeutics acquisition (June 30, 2025).

• Eli Lilly / Orna Therapeutics acquisition (February 9, 2026).

• Eli Lilly / Kelonia Therapeutics acquisition (April 21, 2026).

• BMS / Orbital Therapeutics acquisition (October 10, 2025).

• AstraZeneca / EsoBiotec acquisition (March 17, 2025).

• Gilead-Kite / Interius BioTherapeutics acquisition (August 2025).

• Kite / Pregene Biopharma licensing deal (2025).

• Kelonia first-in-human KLN-1010 data, ASH 2025 Annual Meeting late-breaking session.

• Umoja Biopharma $100M Series C (2025); AbbVie partnership for UB-VV111.

• Foundational biology: Jung et al., J. Immunol. (1988); Bazan et al., Nature (1997); Imai et al., Cell (1997); Böttcher et al., Nat. Commun. (2015); Gerlach et al., Immunity (2016); Zwijnenburg et al., Immunity (2023); Billingsley et al., Small (2024).

• Nobel Prize in Physiology or Medicine 2023, awarded jointly to Katalin Karikó and Drew Weissman.

Key readouts

• 21 Chinese CGT companies completed new financings in Q1 2026

• RMB 2.5B+ total disclosed financing

• nearly 70% of disclosed deals were RMB 100M+ rounds

• capital spread across CAR-T, TCR-T, iPSC, NK, regenerative medicine, and adjacent gene-delivery approaches

• rounds included angel, Pre-A, B, C, and Pre-IPO financings

• the largest disclosed round was Oricell’s approximately RMB 491M C1 financing

What did Q1 2026 actually show?

The biggest signal from Q1 2026 is not simply that China’s CGT market remained active. It is that the market is starting to look more structured, with financing spreading across modalities, stages, and company types.1

In Q1 2026, 21 Chinese CGT companies completed financings totaling more than RMB 2.5 billion, with nearly 70% of disclosed deals at or above the RMB 100 million level.2 Just as important, the quarter was not limited to early-stage activity. It included a mix of early-stage, mid-stage, and later-stage rounds,3 suggesting that investors were willing to finance programs beyond the earliest conceptual stage, even though financing data alone do not establish scientific validation or commercial readiness.

More broadly, this activity sits on top of a very large registration base: one industry dataset based on Qichacha reported 25,965 active or existing enterprises with “cell and gene therapy” in their business scope as of April 30, 2025.4 Surely, this should not be read as 25,965 therapeutic developers, but it does show how widely CGT-related business activity has entered China’s company-registration landscape.

Which companies pulled in the largest rounds?

Q1 2026 showed the size and seriousness of China’s CGT financing machine.

Oricell’s approximately RMB 491 million C1 round was among the largest disclosed financings of the quarter.5

Shize Biotech completed RMB 400 million in B/B+ and C1 financing for iPSC-derived allogeneic neural cell therapy.6

Regend Therapeutics raised RMB 350 million in Series C to advance regenerative medicine.7

Xinjing Zhiyuan completed an over-RMB 200 million Series B financing in solid-tumor TCR-T.8

Yimufeng completed an approximately RMB 200 million Pre-IPO financing in next-generation CAR-T.9

Other named rounds across NK, regenerative medicine, and adjacent platform categories also contributed to the quarter’s breadth.10

The point is not that one company won. The point is that capital spread across multiple modality bets while a meaningful share of larger rounds clustered in companies already advancing beyond the earliest financing stages.11

Why does modality breadth matter?

That breadth is one of the clearest signals from the quarter. The financing map spans CAR-T, TCR-T, iPSC, NK, regenerative medicine, and adjacent gene-delivery or nucleic-acid approaches.

The disease map is broad as well. It stretches from hematologic malignancies into solid tumors, CNS indications, organ regeneration, diabetes-related applications, veterinary stem-cell development, and anti-aging.12 This is what an ecosystem looks like when it starts to widen both by modality and by use case.

The quarter’s stage distribution matters even more. This was not just a seed-stage enthusiasm cycle built on concept decks and platform slogans. The quarter included angel and Pre-A financings in frontier areas, but it also included B, C, and Pre-IPO rounds.

That pattern suggests that the Chinese CGT market can support early discovery, translational build-out, and later-stage clinical execution, even though financing data alone do not establish scientific validation or commercial readiness.13

A serious ecosystem does not only fund one modality. It funds multiple competing routes to the same strategic bottleneck: scalable, controllable, accessible advanced therapies.

Is capital becoming more selective?

Just as important, the capital does not look indiscriminate.

Hot money funds categories. Selective capital funds execution.

A plausible read from Q1 is that investors are becoming more selective about what deserves real money. The companies attracting larger rounds are not just telling an exciting science story. They appear to be offering at least one of three attributes that investors may find more credible: clearer translational plans, a more manufacturable or scalable platform logic, or a more advanced path toward clinical execution. That is why Q1 2026 should not be read simply as a strong quarter for CGT financing. The better reading is that investors are starting to reward clearer translational plans, more manufacturable platform logic, and more advanced clinical execution.

You can see that in the modalities being rewarded. In CAR-T, the cited financings point to interest in groups trying to push beyond the old ex vivo, hematologic-only template and tackle harder problems around solid tumors, in vivo approaches, tumor heterogeneity, exhaustion, and the suppressive microenvironment. In iPSC, the investment case presented in the cited materials centers on off-the-shelf logic, standardization, and the possibility of industrialized production. In NK, the financing narrative is likewise tied to manufacturability, logistics, and allogeneic usability.

Across modalities, the practical question appears to be shifting from “does the biology work” to “can this biology be produced, controlled, scaled, and delivered?”14

That is the real significance of Q1 2026. The quarter suggests that China’s CGT market is moving from thematic enthusiasm toward more structured capital allocation. It is increasingly a story not only about scientific ambition, but also about platform design, manufacturing logic, regulatory progression, and financing depth lining up in the companies attracting capital.

What should global CGT readers take away?

For global CGT readers, the message is simple: China should no longer be read only as a venue for development efficiency or fast-follow platform building.

It is increasingly becoming a source of platform companies, industrial-scale ambitions, and capital-backed advanced-therapy execution.

That does not mean every one of these companies will win. It does not mean commercialization bottlenecks have disappeared. And it does not mean the category is de-risked. But it does mean that the market has become large and deep enough for capital to express real preference rather than generic excitement.

That is the bigger signal from Q1 2026: China’s CGT market is no longer just active. It is starting to look structured.

A team centered on Tongji Hospital in Wuhan, China, has now reported one of the earliest phase 1 clinical datasets for in vivo BCMA-directed CAR-T generation. It is also a reminder that China-linked translational teams are moving quickly in frontier CGT.1

Key readouts

• 5 heavily pretreated patients with relapsed/refractory multiple myeloma

• 4/5 achieved objective responses

• all evaluable responders were MRD-negative at 10⁻⁵ by Day 60

• all 5 patients had grade 3 or higher adverse events

• CRS in 4/5 patients

• one grade 1 ICANS event

• one death during follow-up, attributed in the paper to extramedullary lesion-related spinal cord compression

Leukapheresis, ex vivo manufacturing, lymphodepleting chemotherapy, cell expansion, and the long vein-to-vein timeline may one day feel a lot less inevitable if in vivo CAR-T generation proves reproducible and controllable.

What exactly is ESO-T01?

ESO-T01 is an intravenously administered in vivo CAR-T generation platform designed to induce anti-BCMA CAR-T cells after infusion. So, this is not just another conventional ex vivo-manufactured CAR-T product.

What did the first five patients actually show?

Here are the results from the five treated patients. CRS occurred in 4 out of 5. One patient developed grade 1 ICANS, and one patient died during follow-up; the paper attributes the death to extramedullary lesion-related spinal cord compression.

On the efficacy side, 4 out of 5 achieved objective responses. By Day 60, all evaluable responders were MRD-negative at 10⁻⁵.

All five treated patients had relapsed or refractory multiple myeloma and were heavily pretreated.

So my read is this: the efficacy signal is encouraging, but the safety profile remains intensive and should be interpreted cautiously in a five-patient phase 1 study.

What do Day 4, Day 14, and Day 28 actually tell us?

The most encouraging thing in a study like this may be the kinetics. CAR transgene copies became detectable after Day 4, expanded to a peak around Day 14, and then declined by Day 28.

And that expansion also came with an encouraging specificity signal in the reported assay. At peak expansion, the frequency of CAR-positive non-T cells remained below 1%.

What still has to be proven?

Many early criticisms about the feasibility of in vivo CAR-T focused on off-target transduction. This early clinical dataset suggests predominantly T-cell-restricted transduction in peripheral blood at peak expansion, with the obvious caveat that this is not the same as proving perfect T-cell specificity in every tissue or context.

If in vivo CAR-T generation proves reproducible and controllable, it could eventually reduce several manufacturing and scheduling constraints associated with conventional CAR-T treatment. As a result, the astronomical cost of CAR-T cell therapy will go down. The modality is no longer just theoretical. The next questions are less about whether the concept can work at all and more about control, reproducibility, and safety.