Kyverna: Can One Miv-cel Infusion Sustain Function Without Chronic Immunotherapy in Refractory Stiff-Person Syndrome?

Decision question

If FDA approves miv-cel for SPS, will treated US patients keep the walking and disability gains seen in KYSA-8, without restarting chronic SPS immunotherapy, in enough numbers and centers to change how refractory SPS is treated?

Prospective initiation

Kyverna began a rolling BLA, submitted the chemistry, manufacturing and controls module by August 11, 2026, plans to request priority review under its RMAT designation and targets completion of the submission in the fourth quarter of 2026. A 2027 launch is a company aim conditional on approval. No BLA acceptance, action date or approval exists. [S10, S17]

Catalyst-scope firewall

KYSA-6 in myasthenia gravis, the progressive multiple sclerosis program, the terminated lupus nephritis and systemic sclerosis studies, KYV-102 (a separate whole-blood manufacturing product) and Kyverna's financing are context, and none is credited to this catalyst. Events in other companies' CAR T programs are context, not miv-cel observations. [S11, S21]

Catalyst definition and scope

What the record establishes

A registrational, single-arm, open-label Phase 2 trial treated 26 adults and reported statistically significant within-person improvement on its primary walking endpoint at week 16, with follow-up to a median of 12.0 months. FDA has granted miv-cel RMAT and Orphan Drug designations in SPS, and a commercial supply agreement with ElevateBio BaseCamp was signed on July 23, 2026. [S2, S9, S14]

What the record does not establish

No randomized or blinded comparison of miv-cel, peer-reviewed full KYSA-8 paper, public protocol or statistical analysis plan, or participant-level dataset was located. The registry lists 27 enrolled against 26 treated and analyzed, and no source opened explains the difference. No BLA acceptance, label, commercial treatment count, or publicly announced price or coverage policy was located. [S18, S2]

Dated record

Mechanism

Case reports2024-25

▶

Single-arm trial, n=26Week 16 primary

▶

One-year follow-upCutoff Jul. 10, 2026

▶

Rolling BLA begunCMC module, Aug. 2026

▶

BLA submission completeCompany target Q4 2026

▶

FDA acceptance, then actionDates unknown

▶

Durable function in routine careNot established

Established in primary recordsNot yet established

Biology boundary

SPS is associated with antibodies against GAD65 or the glycine receptor, and the trial enrolled only antibody-positive patients. Whether these antibodies cause symptoms, and whether lowering them explains benefit, is a hypothesis rather than an established mechanism. CD19 B-cell depletion is the intended effect, and B cells returned in 17 of 26 patients by one year, so whether benefit outlasts B-cell return is the central durability question. Kyverna reports efficacy was maintained with reconstitution, without per-patient data. [S2, S5]

Class considerations

FDA required boxed-warning and class labeling changes for marketed CD19- and BCMA-directed autologous CAR T products after investigating T-cell malignancies, and noted that patients should be monitored lifelong for secondary malignancy. Kyverna's long-term follow-up study is registered for up to 15 years. These are class and study commitments, not observed miv-cel events. [S11, S19]

Why the decision is the right anchor

FDA action decides whether miv-cel may be marketed, and it can be an approval or a complete response letter. Availability then needs both approval and launch, and an approval fixes the labeled population every later measure depends on.

Trial design

How KYSA-8 was run

Who was treated

Walking-aid counts are read from the poster figure percentages (12%, 35%) and the stated 12 aid users. [S2]

Reading the population

All 26 had received IVIg or SCIg and had an inadequate response to at least one immunotherapy, so the result speaks to refractory, antibody-positive SPS in centers able to deliver CAR T, not to all SPS. SPS immunotherapies were stopped before conditioning, although the posters do not state when baseline tests were done relative to withdrawal. [S2]

Walking outcomes

Primary Endpoint and One-Year Walking Time

[Chart C1: Median within-person reduction in T25FW time from baseline. Unit: median % reduction in T25FW time from each patient's baseline. Window: Week 16 (primary) and month 12; cutoff Jul 10 2026. Provenance: Reported fact; within-person, no control (S2; S3).]

Median walking time over 25 feet fell 46% from baseline at week 16 and 49% at one year.

These are medians of each patient's own percentage change; p = 0.0003 at week 16 and p < 0.0001 at one year test against no change, not against a control. The one-year median time is not derived here from the baseline median, because a median of changes cannot be applied to a median baseline. [S2, S3]

Clinically Meaningful Walking Improvement

[Chart C2: Patients with at least 20% T25FW improvement. Unit: patients. Window: Week 16 and month 12. Provenance: Reported fact; threshold derived from MS literature (S2).]

21 of 26 patients improved by at least 20% at week 16 and 20 of 26 at one year.

The 20% threshold comes from multiple sclerosis research rather than SPS-specific validation. Kyverna reports that 20 of the 21 week-16 responders kept the response, and about a third walked 25 feet in under 5 seconds at one year. [S2, S7]

Walking Aids Over Time

[Chart C3: Walking-aid use for T25FW, all treated patients. Unit: % of 26 (as printed; may not sum to 100 due to rounding). Window: Baseline, week 16, month 12. Provenance: Reported fact (poster figure) (S2).]

Patients walking without an aid rose from 54% to 85%; 8 of the 12 baseline aid users no longer needed one at week 16 and one year.

The 8 of 12 figure applies to baseline aid users only. At one year 15% of all patients, about 4, still used bilateral aids. [S2]

Function and disability

Six-Minute Walk Distance

[Chart C4: Median change in 6-minute walk distance. Unit: metres (median within-person change). Window: Week 16 and week 52. Provenance: Reported fact; within-person (S2; S6).]

Median walking distance rose 89 m by week 16 and 119 m by one year, from a baseline median of 240.5 m.

The one-year value comes from 25 patients. Kyverna compares the gain with a 20 m minimally important change taken from multiple sclerosis literature. [S2, S6]

Disability, Ambulation, Stiffness and Sensitivity

[Chart C5: Mean improvement on secondary scales, week 16 and one year. Unit: points (signed mean change; negative = improvement). Window: Week 16 (cutoff Nov 26 2025) and one year (cutoff Jul 10 2026). Provenance: Reported fact; within-person (S2; S8).]

Mean scores improved on all four scales at both visits; three changed little between week 16 and one year, and the sensitivity score improved less at one year.

Signed mean changes, week 16 then one year: Rankin −0.8 to −1.0; ambulation −1.6 to −1.7; stiffness −1.5 to −1.6; sensitivity −3.2 to −2.8 (one-year p ≤ 0.0003, within-person). [S2, S8]

Disability Level Before and After

[Chart C6: Modified Rankin scale distribution, baseline and month 12. Unit: % of 26. Window: Baseline and month 12. Provenance: Read from poster shift-plot figure (colour-coded segments) (S2).]

At one year 13 of 26 patients (50%) were at Rankin 0-2, against 1 of 26 at baseline, while 11 of 26 (42%) remained at Rankin 4.

Bar labels are the poster's rounded percentages; counts are the only whole numbers of 26 consistent with them (12%, 8% and 31% give 3, 2 and 8). Rankin 4 means moderately severe disability, unable to walk or attend to bodily needs without help. [S2]

Durability

Freedom From Chronic SPS Immunotherapy

[Chart C7: Chronic SPS immunotherapy status at one year. Unit: patients. Window: One-year cutoff Jul 10 2026 (median follow-up 12.0 months). Provenance: Reported fact (S2).]

24 of 26 patients remained off chronic SPS immunotherapy at one year; two restarted IVIg after month 8, one despite continued benefit.

A separate patient received replacement IVIg for low antibody levels; whether that patient overlaps with these counts is not stated, so the numbers are not combined. Use of symptomatic drugs such as benzodiazepines and baclofen was not reported. [S2]

Durability evidence at one year

Earlier evidence for persistence

The poster cites initial SPS and myasthenia gravis patients treated outside the trial with benefit beyond 24 months without chronic immunotherapy, and published compassionate-use cases describe walking gains within three to six months. These are single cases or small series with different follow-up, so they support plausibility rather than a durability rate. [S2, S24]

What would show durability

Two things matter: whether patients whose B cells have returned keep their walking and disability gains, and how many restart IVIg or other chronic immunotherapy over the following years. Both need per-cohort reporting with retained numbers and elapsed time since infusion.

Safety

Treatment-Related Adverse Events

[Chart C8: Treatment-related adverse events, one-year analysis. Unit: patients. Window: Infusion to Jul 10 2026 cutoff. Provenance: Reported fact (S2).]

Cytokine release syndrome occurred in 24 of 26 patients, all grade 1 or 2, and three patients had serious treatment-related events.

Grade 3/4 neutropenia occurred in 5 patients; 4 resolved after a median of 121.5 days and one had residual grade 1 neutropenia at study end. No immune-effector hemophagocytic syndrome was seen. [S2]

Severe events behind the grades

Primary analysis versus one year

Reading the safety record

No high-grade CRS or neurotoxicity does not mean no severe toxicity: the regimen requires chemotherapy conditioning, an initial inpatient stay, and produced a grade 4 airway event and prolonged neutropenia. Twenty-six patients followed for about a year cannot exclude rarer or later harms, including the class malignancy risk that FDA labels for marketed CAR T products. [S2, S3, S11]

Comparator problem

What the natural-history study adds

In routine care at Johns Hopkins and the University of Colorado, walking-aid use rose 73% over time and Rankin scores did not improve, and Kyverna reports that most patients showed minimal or no walking improvement. That supports the view that SPS rarely improves on its own. However, patients, timing and measurement differ from KYSA-8, so it frames the unmet need rather than providing a treatment comparison. [S4, S8]

Why the signal still matters

The within-patient changes are consistent across walking, disability, stiffness and sensitivity scales, sustained at one year in most patients, and accompanied by B-cell depletion and antibody falls. That concordant persistence is encouraging. However, without a control group or rater-blinding detail, the treatment-specific share of the change cannot be estimated, and differences in design rule out comparing its size with earlier IVIg or rituximab trials. [S2]

What a randomized test could look like

Kedrion's Phase 3 IVIg trial (randomized, quadruple-masked, n=38, T25FW at week 24; not yet recruiting as of October 23, with a November 2026 estimated start) and Jefferson's efgartigimod crossover trial (n=26, recruiting) will measure placebo-condition walking changes in SPS, which will help frame open-label results. [S25]

Ecosystem and roles

Supply chain unknowns

Turnaround from collection to infusion, release failures and capacity were not disclosed in any filing opened, and the ElevateBio agreement redacts quantities and prices. Kyverna's October investor deck, inspected by the packet reviewer but blocked to me (HTTP 403), states manufacturing success above 98%; its denominator, cutoff, definition of success and reproducibility are unknown, so it is a company claim, not an established release rate. For a product made individually from each patient's cells, every failed batch delays treatment. [S14]

Licences and obligations

Kyverna's licence with Kite was terminated in October 2025, with $3.8M still owed by December 31, 2026 in cash or stock. Its Intellia agreement covers a separate allogeneic product (KYV-201), not miv-cel supply, and KYV-102 is a separate whole-blood product with its own IND. [S11]

Commercial build

Kyverna had 134 full-time employees at December 31, 2025, and by August 2026 reported hiring commercial leadership, a Chief Commercial Officer and Chief Financial Officer, commercial site activation and payer engagement. [S12, S10]

Delivery and access

The delivery funnel to track

Population estimates, labeled as sponsor figures

Coverage

Medicare's national coverage determination for CAR T covers autologous treatment for cancer and is silent on autoimmune use, which leaves the Medicare pathway for an SPS approval undefined rather than excluded. Commercial payers' current SPS policies concern IVIg and similar off-label therapy, and no miv-cel coverage policy has been located. [S26]

What cannot be inferred

Diagnosed and immunotherapy-treated patient estimates are not eligible or paying patients, the prevalent pool is finite rather than recurring, and no price, net revenue, uptake rate or center count is projected here.

Competition and current care

What competition means here

No efficacy ranking is drawn because no trial compares these products directly and their populations, designs and endpoints differ. Miv-cel's proposed difference is a one-time treatment that stops chronic immunotherapy, traded against conditioning chemotherapy, inpatient monitoring and CAR T risks. IVIg already has randomized SPS evidence from a small crossover trial, and a positive larger randomized IVIg or efgartigimod trial would add to the controlled evidence behind a lower-intensity option. [S2, S25]

Class context

FDA has required boxed-warning and class labeling changes for marketed CD19- and BCMA-directed autologous CAR T products over T-cell malignancy, with lifelong monitoring. That is class labeling context, not a miv-cel observation, although it shapes how regulators, centers and payers view autoimmune CAR T. [S11, S29]

Resources

Kyverna Resources, Kept Separate by Status

[Chart C9: Kyverna resources by status. Unit: US$ millions. Window: Balances Dec 31 2025 and Jun 30 2026; facility terms Oct 31 2025, amended Jul 8 2026. Provenance: Reported fact; not additive (S13; S11; S15).]

Cash and securities fell from $279.3M to $199.4M in the first half of 2026; $125M of the loan facility is undrawn and conditional.

The undrawn amount includes $15M of Term A available to December 31, 2026, conditional Term B and C loans and a discretionary Term D, so it is not added to cash. [S13, S11, S15]

Why resources matter to the catalyst

Completing the BLA, preparing centers and building supply all precede revenue, and the loan's minimum-revenue covenant begins with the quarter ending June 30, 2027, subject to cash and market-capitalization exceptions. The $300M shelf registration is capacity to issue securities, not cash. [S15, S11]

Counterevidence

How I weigh the alternatives

The first two decide whether the thesis question can be answered in the patient's favor, because they concern whether the signal is real and lasting. Regulatory and toxicity risks decide whether the question gets tested at all. Delivery limits decide how many patients the answer reaches.

Current challenges at October 23, 2026

Expected breadth

Expected breadth

★★★★☆

Four stars, forecast against one ecosystem: US specialist care for immunotherapy-refractory, antibody-positive stiff-person syndrome, covering the neuroimmunology and CAR T centers, infusion providers and payers that treat these patients. If the catalyst is fully realized, an approved one-time therapy would become the first FDA-approved SPS treatment and would give eligible patients an alternative to indefinite off-label IVIg, after a trial in which measured walking time improved, and 8 of 12 baseline walking-aid users no longer needed an aid for the timed test. I stop short of five because the regimen's intensity, center access and fitness requirements will keep many patients on current care, because 42% of trial patients remained at Rankin 4 at one year, and because lower-intensity therapies are seeking randomized evidence. [S2, S12]

Why the change would be broad within this ecosystem

Kyverna estimates about 6,000 diagnosed US patients, of whom 2,000-2,500 receive off-label immunotherapy; how many would meet a label and be fit for conditioning is unknown, so the estimate is a judgment about the potential change in care rather than a count of eligible patients. For those who are treated, the change would be one of kind rather than dose: a single course in place of recurring IVIg, after which 24 of 26 trial patients were off chronic SPS immunotherapy at one year. [S12, S2]

Why it would not be complete

Symptomatic drugs, physical therapy and specialist follow-up would continue, and the trial excluded patients bedridden for more than three months and those over 75. IVIg already has small randomized SPS evidence, Kedrion's Phase 3 IVIg trial is due to start, an efgartigimod crossover trial is recruiting and allogeneic CAR T programs are entering the disease, so the reference treatment could be contested even after an approval. [S18, S22, S25]

Historical and forward context

Kyverna says miv-cel could become the first approved CAR T therapy for an autoimmune disease, and I located no approved autoimmune CAR T. The BLA submission is targeted for completion in the fourth quarter of 2026, with FDA acceptance and action dates unknown, so the estimate is set now and does not move with those milestones. [S6, S10]

What four stars does not mean

Stars estimate how much of this one ecosystem a fully realized catalyst would reach. They make no statement about approval odds, durability or uptake, and they do not move with disclosure quality.

Realized Depth

Realized Depth is reserved for outcomes that actually occur after the FDA action, such as approval, commercial infusions and durable function in treated patients. No post-approval catalyst outcome has occurred, so no Realized Depth is recorded and no outcome grade is issued.

Admissibility and horizon

What counts as durable function off chronic SPS immunotherapy

Treated patients keep measured walking and disability gains without restarting chronic SPS immunotherapy. A pass needs approval with a label covering the studied population, disclosed commercial treatment counts, and follow-up of named cohorts that reports retained numbers, walking measures and chronic SPS immunotherapy status together, rescue treatment, attrition and elapsed time since infusion, through a horizon measured from the actual FDA action. Replacement IVIg for low antibody levels is tracked as a separate category. The evidence required is qualitative and joint: walking and disability measures together with chronic SPS immunotherapy status, by named cohort, with the treated denominator, retention, rescue treatment and follow-up time since infusion, plus commercial infusion counts, activated centers and payer access where disclosed. Missing items are recorded as missing; no numerical threshold is preset, low disclosure is a tracking gap rather than failure, and revenue alone cannot pass the test.

Sponsor targets versus observations

Admissibility verdict

The catalyst is admissible, although the action date is not yet set and durability measurement depends on Kyverna publishing follow-up with denominators.

Tracking standard

Period integrity and absence of evidence

Within-person trial changes are never compared with real-world counts or other trials' arms, durability figures carry their denominator and cutoff, and Kyverna results in other diseases neither confirm nor contradict this catalyst. A missing figure is recorded as not located, and an inaccessible document as inaccessible.

Falsifiers

Next measurement

Baseline values to measure against

Data still needed for an independent reading

Oct. 22-23 additions captured

A corrected file of the October 21 poster, created October 22 (clinical cutoff unchanged at July 10), changed one figure label, the week-4 six-minute-walk count from 26 to 25. Registry records for Kedrion's IVIg trial and CRISPR's CTX112 trial show last updates on October 22. A Bristol Myers Squibb release dated October 23 on zola-cel in multiple sclerosis was supplied in the owner packet but could not be opened (HTTP 403) and concerns a different product. [S1, S2, S25, S28]

Standards and disclosures

Comparator and breadth boundaries

No cross-trial ranking is drawn. Expected breadth on page 15 estimates reach within US refractory SPS care; admissibility on page 16 decides whether the outcome can be graded.

Citations and access

Bracketed codes such as [S2] point to the source register on pages 20 and 21. Sources supplied by the owner that I could not open are marked as such, and no figure from them is presented as verified. This document is not a recommendation, rating, price target or medical advice.

Primary sources

Primary sources

Disclosures

Bottom line

In 26 patients with refractory, antibody-positive stiff-person syndrome, one miv-cel infusion was followed by a median 49% within-patient reduction in 25-foot walk time at one year, with 24 of 26 off chronic SPS immunotherapy. The study had no control arm, the regimen carries real toxicity and the BLA is not yet complete. The open question is whether an approval turns that signal into durable function without chronic SPS immunotherapy for patients in routine US care. Four stars records how much of refractory SPS care a fully realized result would reach, and nothing about which way it will go. [S2, S10]

Next measurement

Kyverna's third-quarter report, then completion and FDA acceptance of the BLA. I will check the FDA decision as soon as it is public, then label scope, commercial infusions and the first long-term follow-up disclosures, recording each against this baseline.

What would move my reading

Source hierarchy and completeness

Facts about miv-cel come from Kyverna's SEC filings, trial registry records, conference posters and releases; competitor facts come from registries and published trials. The register supports the reported summaries used here; it is not exhaustive, because no protocol, statistical plan, peer-reviewed full paper or participant-level dataset was available, and two owner-supplied documents could not be opened (HTTP 403).

Grading and terminal rule

No outcome grade is issued and no Realized Depth is recorded until outcomes occur. Expected breadth is four stars against US specialist care for immunotherapy-refractory, antibody-positive SPS. Terminal review falls five years after the actual FDA action date.

Investment-risk disclosure. Independent research for informational and educational purposes. Not personalized investment advice, an offer, a solicitation, a recommendation, a rating, a price target, a stock prediction, an expected-security-return analysis or medical advice. Investing involves risk, including possible loss of principal. Figures are drawn from the primary sources listed on pages 20 and 21 and are current as of the evidence date printed on page 1.