The Platform Trial Takeover: Neurology’s Last Best Chance to Stop Failing Patients

Alzheimer’s research has burned through roughly $42 billion in failed clinical trials over the past two decades. Motor neuron disease kills half its patients within three years of diagnosis. Parkinson’s disease affects more than 10 million people globally, and the last genuinely disease-modifying approval for a neurodegenerative condition remains, depending on how generously you score the evidence, a matter of active scientific debate. The traditional two-arm, single-compound, sequential trial model did not create this crisis — but it has reliably failed to end it. And now, three converging signals in 2025 and 2026 point to a structural shift that the clinical operations community has not yet named: the Neurodegenerative Platform Pivot.

The Lancet published a Comment in June 2026 calling for multi-arm multi-stage (MAMS) platform trials to become the standard architecture for neurological drug development. The authors are not making a theoretical argument. They are pointing at a field where neurodegenerative biology has resisted every sequential trial design thrown at it, while the platform model — in oncology, in COVID-19, in HIV — has demonstrated that running multiple candidate arms through a shared control group, with pre-specified interim stopping rules, changes the math of drug discovery.

The math, in neurology, is overdue for changing.

The Architecture the Field Abandoned Too Soon

Open the FDA’s 2019 guidance on adaptive design clinical trials for drugs and biologics. The document is genuinely progressive. It defines adaptive design as allowing “prospectively planned modifications based on accumulating data,” covers Bayesian adaptive frameworks, and explicitly encourages sponsors to use computer simulations to justify their designs. The FDA, on paper, endorses the intellectual architecture that MAMS trials depend on. But read further and the operational reality becomes apparent: the guidance places the entire burden of simulation, justification, and pre-specification on the sponsor, with no template, no therapeutic-area-specific precedent, and no commitment to accelerated review timelines for adaptive submissions in neurology. The gap between regulatory permission and regulatory support is where neurodegenerative MAMS trials go to die.

Meanwhile, the science has moved. A MAMS platform design for progressive multiple sclerosis, modeled by researchers at Amsterdam UMC, is projected to reduce trial duration by approximately half compared to running separate two-arm Phase II and Phase III trials sequentially. The same design is estimated to recruit around 1,900 participants — roughly two-thirds the size of what the conventional approach requires — and complete in approximately six years instead of twelve. For a disease like progressive MS, where patients are accumulating irreversible neurological damage while sponsors run sequential single-drug studies, six years is not an efficiency gain. It is a clinical outcome.

The efficiency argument becomes impossible to dismiss when you stack it against the cost data. Published modeling for MAMS trials in progressive multiple sclerosis shows the design is approximately two-thirds the size of traditional sequential trials — smaller sample, shorter duration, shared control arm. That shared control arm is the key mechanism: instead of enrolling a new placebo group for every compound tested, the platform reuses control data across arms, each arm governed by its own interim analysis and stopping rule. Drop a failing drug early. Add a promising new candidate without rebuilding the infrastructure. The platform does not care which compound wins — it is designed to find out faster than any single-drug study can.

The Signal Coming from Motor Neuron Disease

The most operationally instructive example running right now is not in a boardroom white paper — it is the MND-SMART trial, the first MAMS Phase III study for motor neuron disease to launch in Europe. The first participant was randomised on February 27, 2020. Since then, the platform has done exactly what a platform is supposed to do: it has cycled compounds through without rebuilding its infrastructure from scratch each time. In 2023, the trial launched amantadine — a drug originally developed for Parkinson’s disease — as its third experimental arm. In April 2026, tacimus entered the platform as another repurposed candidate. MND-SMART is also now planning to extend its architecture toward Alzheimer’s disease.

That last detail deserves a pause. A platform trial originally built for motor neuron disease is now architecting an expansion into Alzheimer’s. The infrastructure — the randomisation system, the safety monitoring board, the master protocol, the site relationships — transfers. The scientific and operational investment compounds rather than depreciates. Compare that to the conventional Alzheimer’s trial model, where each sponsor builds a separate Phase II, runs it for three to five years, fails, and files the data in a regulatory drawer while the next sponsor starts from zero with a structurally identical design.

The counterintuitive truth here is that platform trials are not primarily a scientific innovation. They are an institutional one. The hard part of a MAMS trial is not the Bayesian interim analysis or the adaptive randomisation algorithm — those are solved problems with established statistical frameworks. The hard part is convincing academic medical centers, patient advocacy groups, funders, and regulators to commit to an infrastructure before they know which drug it will test. Traditional trial funding flows to compounds, not platforms. That misalignment between how money moves and how platforms work is the real friction point.

What Sponsors and Sites Are Not Seeing Coming

The operational implication that most clinical ops teams are missing is this: as MAMS architectures scale in neurology, the competitive advantage will belong to sites that have already operationalized master protocol infrastructure, not to those with the most neurology-specialist investigators. A site running MND-SMART already has trained coordinators, calibrated outcome assessors, established IRB relationships with the platform’s master protocol, and familiarity with adaptive randomisation handoffs. When MND-SMART adds tacimus or expands to Alzheimer’s, that site does not re-onboard. It activates. A site that has never run a platform trial will spend six to nine months on setup that a platform-experienced site completes in weeks.

For sponsors, the implication is starker. The FDA’s adaptive design guidance explicitly advises sponsors to request a Type B meeting before submitting an adaptive design protocol, and to submit “adequate details of the adaptive design at the time of IND submission.” That is procedurally sound advice — but it assumes the sponsor is building a single adaptive trial for a single compound. MAMS neurology platforms involve master protocols with multiple INDs, staggered arm additions, and shared control data that spans multiple regulatory submissions. There is no FDA guidance template for that complexity in a neurodegeneration setting. Sponsors who treat MAMS submissions like standard adaptive trials will encounter review friction that adds twelve to eighteen months before a single patient is randomised.

Technology vendors face a different kind of disruption. The RTSM and EDC systems built for conventional two-arm trials handle one randomisation schema, one set of eligibility criteria, and one data lock. A MAMS platform with three active arms — each at a different stage of interim analysis, each potentially dropping or adding — requires systems capable of dynamically updating randomisation algorithms mid-trial without breaking audit trails or triggering protocol amendments that require fresh IRB review. Most commercial IRT vendors have not built for that use case. Those that have, including a small number of platform-specialist vendors who cut their teeth on oncology MAMS trials like the I-SPY 2 breast cancer platform, are positioned to dominate neurology deployment as the field converts.

The Lancet’s June 2026 call for MAMS adoption in neurological disease is not a prediction. It is a description of where the evidence already points. MND-SMART is live. The Amsterdam UMC modeling is published. The FDA’s adaptive design guidance exists. What remains is the institutional will to stop funding single-drug neurology trials that have a structural failure rate baked into their architecture, and to redirect that capital into platforms that can test, discard, and advance candidates at the speed the disease burden demands.

Twelve months from now, the sponsors still running standalone Phase II trials in Alzheimer’s and Parkinson’s will face a specific problem: their control arms will be statistically underpowered relative to MAMS platforms that have accumulated shared control data across multiple arms and years. Regulators comparing a 400-patient two-arm Parkinson’s trial against a MAMS platform with 1,900 participants and three completed interim analyses will not need to say which design they prefer. The data density will say it for them — and sponsors without a platform strategy will find themselves renegotiating their entire development programs from a position of structural disadvantage.

References

1. The Lancet — “Multi-arm multi-stage platform trials for neurological disease: accelerating progress” (2026)
2. FDA — “Adaptive Design Clinical Trials for Drugs and Biologics: Guidance for Industry” (2019)
3. PMC / NCI — MND-SMART: First MAMS Phase III Trial for Motor Neuron Disease in Europe
4. Amsterdam UMC — “Designing multi-arm multistage adaptive trials for neuroprotection in progressive multiple sclerosis”
5. University of East Anglia — MAMS trial design modeling for progressive multiple sclerosis (accepted manuscript)
6. MND SMART — Platform trial update including amantadine and tacimus arm launches (2026)