How the Field Built the First Consensus Clinical Trial Framework for Charcot-Marie-Tooth Disease

Earlier this month, a landmark set of consensus recommendations for designing clinical trials in Charcot-Marie-Tooth (CMT) disease was published in the Journal of the Peripheral Nervous System, marking the first time patient advocacy organizations, clinicians, and pharmaceutical companies have aligned on a unified framework for CMT trial design. The publication came from ToPIC: CMT, a collaborative group involving five patient organizations including the Muscular Dystrophy Association, alongside academic clinicians and eight pharmaceutical companies. With no approved treatments currently available for CMT, the framework is designed to give drug developers a clearer, more consistent roadmap as they design studies and work toward regulatory submission.

Key recommendations included flexible trial designs to limit placebo exposure, enrollment based on clinical phenotype rather than strict genotype stratification where appropriate, function-based endpoints such as the CMT Health Index, and a defined role for biomarkers including neurofilament light chain, PMP22 expression, and MRI-based muscle fat fraction. The framework also supports early intervention strategies and outlines practical considerations for pediatric enrollment, safety databases, and gene therapy products.

In a conversation with NeurologyLive, Brian Lin, PhD, Research Director at the Muscular Dystrophy Association and one of the contributors to the framework, addresses how phenotype-based enrollment can be structured to avoid obscuring subtype-specific signals, which of the candidate biomarkers is closest to regulatory-grade validation, how the FDA has approached flexible trial designs in rare neurological diseases, and how the group guarded against commercial timelines shaping the scientific recommendations.

NeurologyLive: The framework recommends phenotype-based enrollment over strict genotype stratification in certain contexts. Scientifically, how do you ensure that approach doesn't obscure subtype-specific treatment signals in trial data?

Brian Lin, PhD: Yes, I think one of the biggest scientific concerns with phenotype-based enrollment is that you could potentially dilute a strong efficacy signal that exists in a particular genetic subtype. So, I think the key is to plan for that upfront, both in how you design the trial and how you analyze the data. In CMT, if there’s a biological reason to think a treatment might work differently in different genetic subtypes, you want to prespecify how those groups are going to be evaluated. You also have to be careful about differences in natural history. Different forms of CMT can progress at different rates, and if you don’t account for that, you could potentially confuse differences in disease progression with a treatment effect.

So, where it makes sense, the statistical analysis plan should prespecify important genetic or mechanistic groupings, such as demyelinating versus axonal, as well as X-linked disease. At the same time, having a common primary endpoint is important so you can look at the overall treatment effect and then use secondary or exploratory endpoints to really understand what’s happening within particular subgroups. Biomarkers can also be particularly valuable like molecular markers, electrophysiology, imaging, protein expression, transcriptomics, or other measures of target engagement and biological response.

Neurofilament light chain, PMP22 expression, and MRI-based muscle fat fraction are all named as candidate biomarkers. Which of those is closest to regulatory-grade validation, and what does that path still require?

Of those three, I’d probably put MRI-based muscle fat fraction closest to being a mature clinical-development biomarker for CMT. There’s fairly strong evidence that quantitative MRI can reliably detect muscle degeneration and progression over relatively short periods, across multiple CMT subtypes, and it has already been used as a pharmacodynamic measure in CMT trials. That said, I’d distinguish between being a strong research or pharmacodynamic biomarker and being a regulatory-grade surrogate endpoint.

To get to that level, we still need broader longitudinal validation across subtypes, better standardization across sites and scanners, and most importantly, interventional trial data showing that changes in fat fraction with treatment translate into preservation of meaningful clinical function.

NfL is promising, but we still need to understand whether it consistently tracks progression in CMT and whether it changes with effective treatment. It has been shown to be elevated in multiple CMT subtypes and has been very useful in diseases like ALS, but the CMT data are less mature. PMP22 is biologically very interesting, particularly for CMT1A, because it is so closely tied to the disease mechanism. But of the three, I’d say it is probably the furthest from regulatory validation because of challenges around measurement, variability, and establishing a clear relationship between changes in PMP22 and meaningful clinical outcomes.

The guidance supports single-participant designs and externally controlled trials to limit placebo exposure. What level of evidence does FDA typically require before accepting those designs in a rare neurological disease, and has there been any informal alignment with the agency on that?

FDA has shown increasing flexibility around trial design in rare diseases, but the standard for substantial evidence of effectiveness hasn’t changed. Importantly, FDA has also recognized that in very small populations, individuals can sometimes serve as their own controls, using a well-characterized baseline or natural history. That can be particularly relevant when there simply aren’t enough patients to conduct a conventional randomized controlled trial.

For single-participant or externally controlled designs you still need to provide a very convincing case that the treatment effect is real and not being driven by natural history, bias, or other confounding factors. For an external-control design, that means having a strong and well-characterized natural history, carefully considering prognostic factors, using objective endpoints where possible, and demonstrating a treatment effect that is compelling enough to distinguish itself from expected disease progression.

For very small populations, including different CMT subtypes, FDA’s newer Rare Disease Evidence Principles also provide a framework where a well-controlled study can potentially be supported by robust confirmatory evidence, including mechanistic, pharmacodynamic, natural-history, and external-control data.

I think the key message from FDA that has been echoed throughout many discussions and public-facing meetings has been that early alignment with the agency is critical around the specific trial design and overall evidentiary package. Acceptability will ultimately depend on the specific disease, therapy, endpoints, natural-history data, and totality of the evidence.

Early intervention is emphasized, but enrolling pediatric CMT patients introduces its own complexity. How did the group define what "scientifically and ethically justified" looks like in practice?

I think the key is balancing the potential benefit of treating earlier against the additional considerations of enrolling children. Scientifically, there needs to be a strong rationale that the disease is already causing meaningful functional changes and that intervening earlier could actually alter the trajectory of disease.

Ethically, the potential benefit of preventing irreversible disease progression has to be weighed against the known and unknown risks of the intervention, the availability of other treatments, and the burden of participating in a clinical trial. That includes being thoughtful about placebo exposure and the long-term consequences of treating a developing child. So, for our group, scientifically and ethically justified meant that we have evidence of a meaningful therapeutic window, we can reasonably measure whether we’re changing the disease trajectory, and the potential benefit of intervening early is sufficient to justify the risks.

With eight pharmaceutical companies involved, how do you prevent this framework from being shaped too heavily by commercial timelines rather than what the science actually supports?

I think having eight pharmaceutical companies involved was a real strength because it brought a range of perspectives and real-world experience in clinical development. From the start, the group was very conscious of keeping the recommendations grounded in science, and they were developed through a broad consensus process involving academic researchers, clinicians, patients, and industry, independent of individual development timelines.

We also wanted the recommendations to distinguish between what the field knows today and where evidence is still emerging. And importantly, the framework isn’t meant to be static. It needs to evolve as our understanding of natural history, biomarkers, endpoints, and treatment response improves. Having industry involved can actually help with that, because these companies are generating much of the clinical and natural-history data that will ultimately help refine the framework.

REFERENCE
1. Abrams CK, Bruhn S, Burns J, et al. A unified clinical trial framework for Charcot-Marie-Tooth disease. J Peripher Nerv Syst. 2026;31(3):e70146. doi:10.1111/jns.70146