KYTX Miv-cel CAR-T: One-Year Durability in SPS

Kyverna Therapeutics (KYTX) said Sept 24 that a single dose of its CD19 CAR-T therapy miv-cel (KYV-101) is still working a year after infusion in stiff person syndrome (SPS), a rare autoimmune disease with no approved treatment. One-year follow-up from the registrational KYSA-8 trial showed the mobility benefit improved with time: median improvement on the timed 25-foot walk (T25FW) was 46% at week 16 and 49% at month 12 (p<0.0001). The results build on a topline Kyverna first reported in December 2025 and a primary analysis it presented at the American Academy of Neurology meeting in April.

The durability numbers matter because they underpin a rolling Biologics License Application Kyverna now says will be complete in the fourth quarter of 2026. Of the 26 treated patients, 95% kept a clinically meaningful gain of more than 20%, 92% stayed off chronic immunotherapies, and over a third finished the T25FW in under five seconds, a pace typical of healthy adults. There were no high-grade cases of cytokine release syndrome (CRS) or neurotoxicity.

SPS is driven by anti-GAD65 antibodies that disrupt inhibitory nerve signaling, causing progressive stiffness and painful muscle spasms. The disease affects roughly 6,000 people in the U.S. Miv-cel depletes the B cells that make those antibodies, aiming for a one-time immune reset rather than lifelong immunosuppression. The construct is fully human with CD28 co-stimulation, a design Kyverna argues explains its tolerability.

Safety is the other half of the story. Autoimmune CAR-T has been haunted by three patient deaths in Novartis and Bristol Myers trials, which raised questions about whether the approach could clear the bar outside cancer. Kyverna’s clean one-year safety record is a direct counterpoint.

Kyverna also released longer-term data from KYSA-6 in generalized myasthenia gravis: all seven patients hit clinically meaningful improvement on the MG-ADL and QMG scales at 24 weeks, and six of seven remained off immunosuppressants, with follow-up now past a year in five patients. The company is betting the same construct extends into a neuroimmunology franchise, with multiple sclerosis and rheumatoid arthritis among the next indications it is exploring.

What to watch: the full SPS data set drops at MS Toronto on Oct 21-23, and completing the BLA in Q4 2026 is the next real catalyst.

Source: Kyverna Therapeutics, one-year miv-cel SPS durability data (Sept 24, 2026) KYTX closed at $6.99, a micro-cap whose thesis still rests on whether the FDA accepts a single-arm, 26-patient trial as enough to approve the first CAR-T for an autoimmune disease.

What It Means for the Trade

The one-year data does one thing for the trade: it converts the durability question from an open risk into a supporting data point ahead of the rolling BLA completion in Q4 2026. The numbers that matter to a reviewer — 95% of patients keeping a clinically meaningful gain, 92% off chronic immunotherapies, no high-grade CRS or neurotoxicity — are exactly the profile that could make a single-arm, 26-patient dataset approvable in a disease with no approved treatment and no ethical case for a placebo arm. That is the entire Kyverna thesis in one regulatory question, and everything the company has released since December is aimed at answering it. The full dataset at MS Toronto on October 21-23 is the next checkpoint before the BLA completes.

Sizing discipline applies. At $6.99 this is a micro-cap where the binary cuts both ways: an approval in a ~6,000-patient US population with no competition and orphan-drug economics could re-rate the stock violently, but the FDA rejecting a single-arm trial — or asking for a controlled study — extends the timeline by years against a cash position that must fund a neuroimmunology expansion. Treat MS Toronto and the BLA completion as the tradeable events; everything between is noise.

Mechanism and Pipeline Context

Miv-cel is a fully human CD19 CAR-T with CD28 co-stimulation. CD19 sits on B cells, and in SPS the pathogenic B cells produce anti-GAD65 antibodies that disrupt inhibitory signaling in the central nervous system, producing the stiffness and spasms that define the disease. Wiping out the B-cell compartment — a one-time immune reset rather than lifelong immunosuppression — is the mechanism the whole autoimmune CAR-T field is chasing. The fully human scFv and CD28 design are Kyverna’s explanation for the tolerability gap versus the oncology CAR-Ts, where the three Novartis and Bristol Myers autoimmune trial deaths raised the bar.

The pipeline is a neuroimmunology franchise play: SPS is the lead, but KYSA-6 in generalized myasthenia gravis (all seven patients improving at 24 weeks, five with follow-up past a year) is the proof the construct generalizes, and multiple sclerosis and rheumatoid arthritis are the next fronts. If miv-cel is approvable in one autoimmune indication, the platform option value across the rest is what justifies more than a one-drug multiple.

Risks and What Could Go Wrong

1. The FDA refuses the single-arm dataset as a filing basis and demands a controlled trial — the highest-severity risk, given the agency’s general preference for randomized evidence.
2. Safety events emerge at longer follow-up or in the broader expansion — autoimmune CAR-T carries the field’s deaths history, and one serious event re-prices the whole program.
3. Cash runway — a micro-cap at $6.99 has no margin for a delayed or rejected BLA followed by a capital raise into weakness.
4. Big Pharma self-helps — if larger players advance their own autoimmune CAR-T assets faster, Kyverna’s first-mover niche erodes before it monetizes.