The Board Brief

Monday, September 28, 2026 · 7 stories worth your attention

Kyverna cell therapy shows sustained benefit in autoimmune disease

Why it matters: Reports that many treated patients remained off other therapies suggest cell therapy could change how some autoimmune conditions are managed. Boards should distinguish that potential from evidence of durable benefit, safety and value across broader patient populations.

Relevant to: CMO, Board Quality Committee

Patient impact: If sustained in larger studies, treatment-free periods could reduce the burden of ongoing therapy for some patients; long-term safety and durability remain uncertain.

The strategic question is whether an intensive, potentially one-time intervention can deliver a sufficiently durable improvement to justify its clinical risks and care requirements. Reports of patients remaining off other therapies are encouraging, but being off treatment is not, by itself, a complete measure of disease control. Boards should look for the number treated, follow-up duration, disease activity, relapse rates, adverse events and reasons patients did or did not resume medication before treating this as a practice-changing result. Transferability is also uncertain. Autoimmune diseases differ in their biology and treatment options, so a result in one condition should not be assumed to apply to others. Cell therapy requires patient selection, preparation, acute monitoring and long-term follow-up; those demands may limit which organisations can provide it safely even if efficacy is confirmed. For providers, the near-term task is not broad deployment but readiness assessment: identify where appropriate patients would be referred, who would manage complications, and how outcomes would be tracked after treatment. Boards should ask management to set evidence thresholds before committing resources or changing pathways. A credible decision would compare sustained disease control and quality of life against treatment-related harms and the outcomes achievable with existing therapies, rather than relying on the appeal of a period without medication.

Question for Boards: What minimum evidence of sustained disease control, safety and follow-up would we require before referring patients for this therapy or building a delivery pathway?

Clinical AI screening reaches more than one million patients across three countries

Why it matters: Deployment in India, Thailand and Australia shows that a clinical AI tool can operate at substantial scale across distinct settings. For boards, the key question is whether that reach is matched by reliable performance, effective follow-up and appropriate oversight in each setting.

Relevant to: CMO, COO, Board Quality Committee

Patient impact: More than one million patients have been screened. The available claims do not establish whether screening improved diagnosis, treatment or health outcomes.

Screening more than a million patients across India, Thailand and Australia is a meaningful deployment milestone, but scale is not the same as clinical benefit. The board-level distinction is between an algorithm that can be installed in multiple settings and a screening service that consistently connects patients to appropriate care. Differences in patient populations, equipment, staffing and referral capacity can all change what a successful deployment looks like. For organisations considering similar expansion, the transferable lesson is to govern the whole pathway rather than the model alone. Before rollout, leaders should specify which patients are eligible, what happens after a positive or inconclusive result, who handles exceptions, and how performance will be checked at each site. Reporting should distinguish the number screened from diagnostic accuracy, completed referrals and downstream outcomes. Aggregated results may conceal a site where follow-up is weak or performance differs for a particular patient group. The research was funded by Alphabet Inc. or a subsidiary, which makes independent scrutiny of methods and outcomes especially important; funding alone does not determine the validity of the work. Boards should treat the reported reach as evidence that cross-setting implementation is possible, while requiring site-level clinical and operational evidence before assuming the approach is effective or transferable to their own system.

Question for Boards: What site-level evidence would we require to show that an AI screening rollout leads to accurate results and completed follow-up—not just a high number of patients screened?

Embryo Editing and Selection Call for Coordinated Governance

Why it matters: Early embryo-editing research is advancing while a professional society says polygenic embryo selection is not ready for clinical use. Boards overseeing fertility or reproductive-genetics services need to distinguish research progress from evidence sufficient to offer patients a service.

Relevant to: Board Quality Committee, CMO

Patient impact: Clear oversight can help prevent patients from mistaking experimental editing results or uncertain polygenic predictions for clinically validated options.

The governance challenge is not that editing and selection are the same procedure. It is that both can be presented to prospective parents as ways to influence a future child's health, despite very different evidence and risk profiles. Reported adenine base-editing results in early embryos suggest technical progress, including editing at two targets without the major chromosomal abnormalities observed in some earlier experiments. They do not establish clinical safety, long-term outcomes or a basis for offering embryo editing as care. Meanwhile, the American Society for Reproductive Medicine says preimplantation genetic testing for polygenic disorders is not ready for clinical use because its predictive value remains uncertain and ethical concerns persist. For boards, a single oversight framework should create a consistent boundary between permitted research, validated clinical testing and services that should not be offered. That framework should still assess each technique on its own terms: editing raises questions about unintended and potentially heritable changes; polygenic selection raises questions about prediction, embryo ranking and how uncertainty is communicated. Review should cover proposed indications, evidence thresholds, consent materials, laboratory oversight and any claims made by commercial partners. This is especially transferable to organisations that both deliver fertility care and collaborate with genetics researchers: governance gaps can emerge where a research result is translated into patient-facing language before clinical utility has been established. The immediate decision is not whether to adopt a new intervention, but who can authorise such a transition and on what evidence.

Question for Boards: What evidence and independent review would our organisation require before any embryo-editing or polygenic-selection method moves from research into patient-facing services or marketing?

Single-cell AI maps Alzheimer’s-associated cellular patterns at population scale

Why it matters: A framework that links single-cell data to Alzheimer’s-related phenotypes could help research organizations investigate disease heterogeneity. Boards should distinguish this research capability from a validated clinical test or treatment-selection tool.

Relevant to: Chief Research Officer, Board Research Committee

The strategic development is the combination of scale and phenotype linkage. Using data from more than 6 million prefrontal-cortex nuclei across more than 1,000 brains, the PsychAD researchers developed PASCode to assign aggregated phenotype labels to cells. They identified approximately 1.5 million phenotype-associated cells among 584 donors with Alzheimer’s-related phenotypes, and report better label-assignment performance than individual differential-abundance methods. That gives research teams a way to examine which cellular patterns track with measured disease features, rather than treating Alzheimer’s as a single undifferentiated condition. For boards overseeing neuroscience research, the immediate decision is about evidence generation, not clinical deployment. A useful next test would be whether the associations reproduce in independent cohorts and remain informative across differences in sample collection, brain region and phenotype definition. Boards should also ask whether the method identifies biological signals that existing analyses miss, rather than merely assigning labels more efficiently. Transferability will depend on access to sufficiently large, well-characterized tissue datasets and the ability to link cellular measurements to reliable donor phenotypes. The work may help prioritize hypotheses for mechanistic studies or drug-target research, but performance on post-mortem single-cell data does not establish that a living patient can be diagnosed, stratified or treated differently. Governance should keep those research and clinical claims separate.

Question for Boards: What independent-cohort validation and evidence of added biological insight would we require before using PASCode-derived associations to prioritize Alzheimer’s research investments?

Phase 2 trial reports higher response rates with evorpacept combination in HER2-positive gastric cancer

Why it matters: A randomized phase 2 trial found higher investigator-assessed response rates when evorpacept was added to a three-drug regimen for advanced HER2-positive gastric or gastroesophageal junction cancer. The larger difference in patients with HER2 confirmed by a fresh biopsy makes testing practices important to interpreting and potentially applying the finding.

Relevant to: CMO, Board Quality Committee

Patient impact: The evorpacept combination produced a higher investigator-assessed objective response rate than the comparison regimen in this phase 2 trial. Whether that translates into longer survival or better quality of life is not established by the supplied findings.

The board-level significance is not that a new four-drug regimen is ready for routine adoption. It is that a randomized phase 2 signal could change how oncology services prepare for subsequent evidence and, potentially, for selecting patients. Among all 127 randomized patients, investigator-assessed objective response rates were 40.3% with evorpacept added to trastuzumab, ramucirumab and paclitaxel, versus 26.6% with the three-drug regimen alone. In the subgroup with HER2 positivity established by a fresh biopsy, the corresponding rates were 54.8% and 23.1%. That larger subgroup difference is promising, but it should not be treated as proof that repeating a biopsy improves outcomes or that the regimen benefits every HER2-positive patient equally. For health systems, the transferable issue is the interface between trial eligibility and real-world diagnostic pathways. If later studies confirm a clinically meaningful benefit in a biomarker-defined population, services may need reliable access to tissue sampling, timely pathology and coordinated treatment decisions. Those capabilities matter most where obtaining a new biopsy is difficult or would delay care. Response rate is an early measure, and investigator assessment leaves important questions about durability, survival, toxicity and patient experience unresolved. Boards should ask oncology leaders to track confirmatory evidence before changing formularies or pathways, while identifying whether current HER2 testing practices could support equitable access if the evidence matures.

Question for Boards: If a confirmatory trial establishes benefit, could our oncology pathway obtain and assess a fresh HER2 biopsy promptly without delaying treatment or excluding patients who cannot undergo repeat sampling?

Study examines enterocolitis after BCMA CAR-T therapy

Why it matters: Enterocolitis associated with ciltacabtagene autoleucel raises a care-delivery question for CAR-T programs: how to recognize, investigate and manage gastrointestinal symptoms after treatment. A small study reports improvement with upadacitinib in two affected patients, but does not establish a standard treatment.

Relevant to: CMO, Board Quality Committee

Patient impact: Two patients with CAR-T enterocolitis had clinical, endoscopic and histologic improvement after upadacitinib. The finding is promising but too limited to establish how reliably the treatment works or which patients might benefit.

For boards overseeing CAR-T services, the strategic issue is not whether to adopt upadacitinib on the strength of two responses. It is whether the service can detect and assess a potentially consequential complication after patients have moved beyond the immediate treatment window. Gastrointestinal symptoms following CAR-T therapy may involve multiple possible causes, so an escalation pathway needs timely specialist assessment rather than an assumption that every case has the same mechanism or treatment. The study’s comparison groups—10 patients with CAR-T enterocolitis, seven CAR-T-treated controls without it and 26 healthy volunteers—provide a basis for investigating the condition. They do not establish its incidence among all recipients or prove that upadacitinib will be effective more broadly. Clinical, endoscopic and histologic improvement in two people is an important signal for further evaluation, not a protocol-ready estimate of benefit or risk. Boards can act on that distinction now. They can ask whether the CAR-T program has clear responsibility for follow-up, access to gastroenterology and pathology review, and a process for recording suspected cases and outcomes. This approach is transferable to other advanced-therapy complications: strengthen recognition and evidence collection while keeping experimental treatment decisions under careful clinical oversight.

Question for Boards: How does our CAR-T program identify and investigate post-treatment enterocolitis, and who reviews outcomes before an emerging treatment such as upadacitinib becomes routine practice?

Phase 3 Evidence Points to a New Second-Line Option for Small-Cell Lung Cancer

Why it matters: Phase 3 data support tambotatug pelitecan as a potential alternative to topotecan for patients with relapsed small-cell lung cancer. Boards should prepare for a possible change in treatment pathways while distinguishing trial support from regulatory approval and routine availability.

Relevant to: CMO, Board Quality Committee

Patient impact: A new second-line option could expand treatment choices for patients whose cancer has relapsed, but the supplied evidence does not establish the size of any survival, quality-of-life or toxicity benefit.

Relapsed small-cell lung cancer is a setting in which the established second-line option, topotecan, offers modest benefit and considerable toxicity. Phase 3 support for tambotatug pelitecan therefore warrants attention: the relevant question is not simply whether another drug can enter the pathway, but whether it changes the balance of benefit and burden for patients who may have limited tolerance for further treatment. The board-level task is to require a disciplined comparison before changing practice. Clinical leaders should examine the trial’s absolute outcomes, adverse events, eligibility criteria and comparator arm, then assess how closely the enrolled population resembles patients treated locally. An antibody–drug conjugate may also require different preparation, administration and toxicity-management processes from existing therapy. Those operational requirements should be tested against expected clinical value rather than assumed to be routine. This finding is transferable as an evaluation framework across oncology services, not yet as a blanket treatment recommendation. Phase 3 evidence supports consideration of a new option; it does not, on the information supplied, establish regulatory status, reimbursement, price or the magnitude of patient benefit. Boards can ask for a concise adoption plan that links any formulary decision to approvals, full trial results and monitoring of outcomes in practice.

Question for Boards: What absolute benefit and toxicity differences versus topotecan would justify adding tambotatug pelitecan to our relapsed small-cell lung cancer pathway, and what approvals and operational safeguards would be required first?

Health Board Intel — Evidence-led intelligence for healthcare leaders