Genetic Epilepsy Team Australia’s 2025 Genetic Epilepsy Conference

Prof Piero Perucca: Thank you, Sarah, and it’s an absolute privilege to be here this morning and would like really to thank the organizers for the invitation. It’s also daunting for me to follow such a wonderful presentation and powerful such as the one that Lee has provided on Gemma. So I’ve been tasked today with speaking about single participant trials for novel epilepsy therapies, and I realized that some questions have already come through in with the previous presentations.

And I do hope I can tackle some of those queries like Ingrid. I do have a fair few of disclosures but I view them just like Ingrid as a reflection of how we want to engage with different stakeholders with industry. When I was studying medicine, I like to say this story often. I was always told that industry is it’s evil and it’s, we should be, we shouldn’t be working with them.

And the reality is that this is just so untrue if [00:01:00] we want to actually move the field forward. If we want to identify new treatments, we actually need to work together. And this is what at least for I can speak for myself and I’m sure for Ingrid, we we strive every day and we look forward to doing so.

I’d like to start with a slide on conventional randomized controlled trials. And this really is the cornerstone or of evidence-based medicine. They define effectiveness of one or more therapeutic interventions at the population level by providing overall average results, and this is how they work.

So the conventional randomized controlled trials involve getting a large number of participants. Some will receive treatment X and some will receive a placebo or potentially another active treatment, and some participants will respond to treatment X. There will also be [00:02:00] some participants who would, will be responding to placebo, and that’s a real issue in clinical trials.

There will also be some participants as highlighted in blue, who will display a partial response to treatment and to treatment X in this case. And that’s difficult to quantify in the scheme of a conventional randomized controlled trial. But at the end of the day one will just tally up the number of participants who respond to treatment X and those who do not respond to treatment X and what we’ll be able to conclude.

That treatment x is or will not be effective at a population at large. However, what these trials do not allow us to answer is whether a specific treatment works for a specific individual. And that’s a real issue, and you will really need a completely different design to answer that specific question.

The design that can answer that question are N oh one trials. [00:03:00] Before delving into the details of N oh one trials, I like to say that the term N oh one trials is often used quite loosely in the medical community. Also in the literature referring to open label, uncontrolled trials of a medication in an individual, essentially they’re often used to designate the prescription of a medication to a person.

But prescribing a medication to a person is something we do every time, every day in clinical practice. But that doesn’t make that trial an N oh one trial. N oh one trials are rigorous experiments and are defined as single participant crossover studies where there are repeated cycle paired cycles with an active intervention, which we would call X in this case, and a control intervention, which we’ll call y with monitoring of personalized treatment outcomes at the individual level.

And I’d really like to stress two words in that sentence, personalized and individual. So what these trials [00:04:00] allow us to do is actually to personalize the outcome by tailoring the assessment of a specific outcome in an individual. They allow us to tailor the design of the trial to the most disabling manifestation or symptom in that particular individual.

Something that cannot be done with large scale, randomized controlled trials. And essentially this is a scheme of that design. Where’s my mouse? So you’ll have an individual who’s going through these paired cycles of treatment X or in this case treatment-wise, placebo. And we’ll be continuously asking the question, are they responding to treatment?

And we record that response or lack of response. And ultimately, by adding up that information, we’ll be able to determine whether there is a relationship between treatment x in this case and their response in a specific individual. Now, NO one trials can include key components of conventional randomized controlled trials [00:05:00] such as blinding randomization of treatment sequence.

And I think this is quite underestimated. There is also the possibility of. Aggregating results of NO one trials conducted in similar conditions, in the same, in the very same disease to determine generalizability, whether that same treatment works at a at the larger population of individuals with that same condition.

And this is also acknowledged, by the way, by regulatory authorities that understand and recognize that for, especially for rare diseases it’s really a trade off between putting together limited quantity of high quality evidence, as well as large quantities of smaller quanti or smaller or lesser quality evidence.

There are certain conditions that are necessary for conducting n one trials. One pertains to the nature of the disorder. This, the disorder needs to be chronic, [00:06:00] stable, slowly progressive, frequently re recurring and relapsing. That’s the quintessential element of epilepsy. The the nature of the treatment should also be specific.

There should be significant individual differences in treatment effects. Uncertainty as to what is the best treatment in that condition rapid onset of action, and it may also add rapid offset of action of treatment brief and safe washout periods. The outcome assessment needs to be validated.

You need to be able to have repeated measures as you’re going through repeated pair cycles of treatments, and possibly use biomarkers to quantify the treatment effects. And importantly, you need stakeholders, different stakeholders, namely consumers, healthcare providers, and the healthcare system willing to engage and work together to make this work.

Now, the very first N oh one trial was actually published in 1986 in the new England Journal of Medicine by Gordon Gaas, [00:07:00] a an outstanding clinician scientist with his colleagues in Toronto was able to conduct this N oh one trial in an, in a, in an individual with poorly controlled asthma.

Just like to read actually this excerpt from the abstract, because I think it’s quite telling, this was again, written 40 years ago. So it starts by saying, although the treatment of an individual patient in routine clinical practice has been likened to an experiment, the method is so susceptible to bias that we have to come to demand multi patient, double blind, randomized, controlled trials of matter of matters of efficacy.

Unfortunately, such trials have not, or cannot be carried out for many clinical disorders. Even when they have been executed, their results may be difficult to extrapolate to individual patients. So what I’m telling you today actually was. Was already written four decades ago. And to resolve this problem, we have begun to use double-blind randomized trials in which a single patient undergoes a series of pairs of [00:08:00] treatment consisting of one active and one placebo, or alternate treatment per pair with the order determined by random allocation, appropriate treatment targets.

Signs, symptoms or laboratory tests are used as the measure of efficacy. And the trial is continued until efficacy is established or disproved. And we describe such trial which resulted in a dramatically beneficial modification of treatment in a patient with partially reversible airflow limitation, which was poorly controlled asthma.

So this is exactly the quintessential principle and beautifully written, I would say, for N oh one trials. So why are for a design that was developed 40 years ago, why is it not more commonly utilized or is it so to answer this question, I wanted to first outline potential applications of N oh one trial designs in medicine.

And they can be used to assess the efficacy of treatment in cases of heterogeneity, of treatment effects in conditions with or without established therapies to assess tolerability [00:09:00] in that individual to identify the appropriate dose for an individual to assess the treatment effects in conditions with variability in manifestations.

This is the case of the rare genetic epilepsies requiring individualized outcomes because an outcome will be will potentially be more dis to an individual as opposed to another one to investigate treatment effects in conditions which are rare to assess the effects of treatment withdrawal to explore treatment effects in relation to the natural course of the disease, and to assess treatment effects in people who do not meet criteria for conventional randomized controlled trials.

So you can see how wide is the applicability of the N oh one design. As I said, this was developed for four decades ago, 40, 40 years ago. And have we learned our lessons in the last 40 years? Have we followed that initial blueprint? So to answer that question Steinberg and colleagues published in neurology in 2022, a systematic review of N oh one [00:10:00] trials conducted across areas of neurology.

They identified 40 N oh one trials 19 were individual N oh one trials in 21 were series of N oh one trials. And this is a breakdown of the different conditions in which the trials were conducted. The majority were conducted in neuromuscular disorders. Others in neurodegenerative mo or movement disorders.

12% of all trials were conducted in paroxysmal disorders. That’s epilepsy. And when looking at the fine print, there were only two studies actually performed in epilepsy. And the remaining trials were done in sleep disorders traumatic spinal or brain injury and in other conditions.

The most common type of intervention that was tested WA was a pharmacological intervention accounting for 85% of all trials. And you can see that in the vast majority of cases, that treatment was an off-label use of a specific treatment. But the next question is, were these trials [00:11:00] actually well conducted?

And to do this, the authors utilize two measures. One’s called the SENSE score, the other one is the jut score to determine whether these trials were carried out appropriately both in terms of quality of reporting and quality of design. And for both scores, they higher the score on a scale from zero to a hundred, the better the trial.

And you can see that the on average the scores were 50 here and 55 there not great. And in fact, the authors concluded overall methodologic quality of design and reporting of N one trials are suboptimal and can easily be improved in future trials by always including key items such as a description of the methods utilized for blinding and randomization.

Our washout period identification of a primary outcome measure rules for identification of responders, and recording of whether post-trial treatment decisions align with the trial results. So that’s a bit I would say disappointing that in 40 years [00:12:00] we didn’t do better and that we didn’t follow the blueprint that was developed by Gaad and colleagues in 1986.

What about epilepsy? The first N one trial to be conducted for in, in epilepsy was that was utilized to test a a treatment or a medication called smide in the space of focal epilepsy. And this was an effort led by Michael Prier, colleagues at Cincinnati. And they apply the N oh one design successfully to demonstrate actually the, that this treatment was efficacious.

Sadly this treatment was not further developed due to financial rather than scientific reasons, but that’s another issue, I think, in drug development. But they were able to demonstrate the efficacy and safety of the zide in a first clinical trial of an investigational anti-seizure medication using the single patient design.

And they concluded that this trial establishes the suitability of the N one design to investigate promising [00:13:00] new anti-seizure medications as this blueprint in epilepsy been followed. Not really. So I conducted a systematic review before this putting together, this talk to actually look at how many trials have been done in n one trials have been done in epilepsy, excluding n one trials, looking at precision therapies for monogenic epilepsy, which I’ll touch base in a second.

But by doing a systematic review of identified only five publications in this space which is a bit sad, I must say. One is indeed a study by Michael Previ, which I mentioned earlier. And then there were other four publications, which included either one participant four participants. You can see that the type of intervention assess was different.

It ranged from a developmental anti an investigational anti-seizure medication to an established medication like valproate to hippocampal stimulation to swimming goggles as well. That was tested in a one trial design into transcranial direct [00:14:00] current stimulation. Different designs were used, but more importantly, we’re looking at the methods.

There was very ample variability and things were not done systematically or appropriately as we would’ve liked. So why would N oh one trials be a deal for monogenic or single gene epilepsies? And this is a slide that Andreas had shown earlier which really provides a chronology of landmark discoveries in epilepsy in epilepsy genetics.

It all started with CRNA four which was identified by Ingrid Sam and colleagues in Germany in a large Australian pedigree with autosomal dominant sleep related promoter epilepsy. And I’ll get back to that in a second. That opened the era termed as channelopathy era where a number of discoveries were done in large pedigrees.

Then there were the dark ages where a number of investigators claimed the association between a specific gene [00:15:00] and a specific form of epilepsy, which did not withstand the test of time. We did not, which did not withstand replication. And then there was the advent of next generation sequencing which really propelled a a new era in epilepsy, genetic with a pace of discoveries which has been tremendous whereby even the most enthusiast really had a hard time to follow the almost weekly discoveries that has slow slowed down in recent times, but then has picked up again, driven by large collaborative efforts.

So really the explosion of epilepsy gene discoveries in the past 15 years has opened opportunities to improve the treatment of epilepsy. And we’ve gone from a model which is a one size fits all approach to one where we can strive for, which is to identify a precision. Treatment. Now, there are challenges in in treatment testing in monogenic epilepsies.

And indeed for the vast majority of the [00:16:00] monogenic epilepsies, conventional large scale, randomized controlled trials are not a physical option. And that’s because of the rarity of rare genetic epilepsies and the variability in symptom manifestations. It is nearly impossible indeed to achieve this sufficient statistical power to assess treatment outcomes.

And even when conventional large scale randomized controlled trials can be performed, a large proportion of people will be excluded because these trials requires strict inclusion and exclusion criteria, unfortunately. And so not everyone will, not, will make it into a trial. And that’s a real issue.

But that’s the almost also the only way for these trials to be conducted because they are at the end of the day experiments. And when you have an experiment, you need to set your framework. N one trials however, can overcome these challenges in treatment selection, at least in rare genetic epilepsies.

And this has already been acknowledged by many colleagues. We’ve heard Andreas we’ve heard Ingrid. This is [00:17:00] Sarah Zen, who just published this year a retrospective open-label cohort study testing amitriptyline as a precision therapy for case in Q2 and three related neurodevelopmental disorders.

13 individuals were enrolled. Eight demonstrated at least minimal improvement in a, in two or more domains with improvements in alertness and communication be the most frequently reported. But here’s where I started scratching my head, because they claim in those with reported improvements, amitriptyline was discontinued for individuals, but continued improvements we’re seen to the same or greater extent compared to treatment.

So was treatment actually working for them or was this just a natural course of disease? And indeed and I completely agree with with them, they concluded overall the effect of amitriptyline remains unclear and formal N oh one trials are needed to investigate a precise effects of amitriptyline in [00:18:00] casein q ga function related neurodevelopmental disorders.

What’s the evidence? What are the publications around N oh one trials in monogenic epilepsies? So I’ll get back to that same gene, CRNA four, which was the very first gene described in epilepsy by Ingrid Sam and colleagues in Germany. They reported that a missense variant in in this gene was indeed associated with autosomal dominant nocturnal frontal lobe epilepsy, which has now been renamed as autosomal dominant sleep related IMO epilepsy.

This was a large Australian family, as you can see, as large as it gets with many individuals affected. Now one of those individuals had uncontrolled epilepsy and was the subject of an N one trial performed here in Australia by J by John Willowby in in south Australia, and in colleagues. And I’ll just take an an excerpt from that publication.

The woman, which at the time was age 33 years at seizures since age 12 years, she was initially responsive to Valproate [00:19:00] and Dan Carbamazepine and had infrequent seizures approximately two sequentially one day every six months until the age of 29. At this age, the patient ceased smoking, having started at age 13 years and have been smoked 16 cigarettes per day for many years.

And shortly after, during a distressed distressing personal experience, seizures became frequent and disruptive through her life. Although they remained briefed, they proved refractory to various treatment including carbamazepine, vpr, gabapentin, and clonazepam. So there was a signal there that potentially nicotine could be efficacious in, in, in this person.

And this led really to test a precise approach of utilizing the N one trial design of nicotine patches as an add on treatment for uncontrolled seizures in in, in, in this person with autosomal dominant sleep related right promoter epilepsy due to a pathogenic variant in CRNA four. And this was such a beautiful NO one trial started, which started with an initial baseline [00:20:00] just recording seizures without any treatment, 97 days, and you can see if you can see my miles.

So that’s the baseline period. So it is in light blue and each square, each box is a day. And within each box, you’ll see the number of seizures in that given day. So there were seizures where this person had 10 seizures. Others, other days where there were no seizures at all, but it gave you a sense of the actual frequency or seizure control.

Then the person started nicotine patches and that’s this blue, darker blue box. And that went on for 294 days. And you can see without making any strange statistical calculations that there were only four days with seizures over those 297 days, which is amazing. And after that open label phase, then a double blind randomized phase where everyone was blinded, that included the person, their clinicians whoever [00:21:00] prescribed the patches they actually had no idea what the person was taking.

And you can see here that there were two periods in, in, in dash lines, in dark dash lines, their periods of active treatment. And instead, we where placebo was administered was was in light blue. And again, you don’t need any statistics to see that there was not a single day with seizures during periods of active treatment.

With the nicotine patch and the seizures only occurred while placebo was administered. We always liked statistics, so statistics was done. And you can see that the p value was extremely small to just validate the difference in between treatment with nicotine patches and placebo. Si. Now this trial was just beautifully done and it ticked all the boxes when you look at measures of internal validity, such as defining the study design, selecting the appropriate control, randomization, sampling of behavior, blinding of the of the individual and the healthcare professional [00:22:00] blinding of the assessors.

And also it ticked many boxes in terms of external validity. Obviously the two boxes that were not ticked were replication and generalization because this was an N oh one trial done in one single individual. Many of you will say beautiful example. How many publications are out there on on N oh one trial testing precision approaches for monogenic epilepsies?

And this is my next slide. What if I tell you that there’s not a single N oh one trial in monogenic epilepsies and, the late Matthew Perry or better I would say the friend’s character will absolutely be quite confused, isn’t it? And so we are, what’s the issue with doing NO one trials in rare conditions?

And this was beautifully tackled by this publication that assessed the quality of the data in the space of rare genetic neurodevelopmental disorders. And they concluded that in that at the time, this was in 2021, only 12 studies complied with the fundamental [00:23:00] criteria of a controlled multiple crossover trial showing limited use and reporting of N one trials for rare genetic neurodevelopmental disorders.

The issue fundamentally is that the literature is full of publications stating that n one trials were conducted when this were simply the prescription of a medication individual that’s not an N oh one trial. And when you actually dissect those publications, you’ll end up with just a handful of of individuals and of studies, which meet the standards of an N oh one trial.

And the authors here went on to identify the challenges and provide recommendations to conduct appropriate NO one trials in rare genetic developmental disorders. And this. Relate to making sure that you clarify the disorder beforehand. You acquire baseline characteristics. You target the symptoms and the symptom that is most disabling for that individual.

You make clear what the eligibility criteria are. You define how, what, and how many [00:24:00] concurrent medications are allowed. You define the comorbid conditions. Obviously there are also recommendations around the inappropriate appropriate intervention, the design of the study, which includes also devising ap priori from the very beginning, the appropriate analysis that then will be carried out when the trial is completed.

Defining the outcome measures, and I’d like to emphasize one will define the outcome measures which are appropriate for that person. So if a person has a rare genetic epilepsy where seizures are fully controlled by the most disabling symptom is ataxia or lack of coordination, then that will be the primary outcome of that NO one trial.

And aligned with picking the most appropriate outcome, one, we’ll have to pick the most appropriate analysis. So we’ve listened. We’ve actually listened and I’m absolutely pleased to share that we have, after many years of work, we have developed a master protocol for testing of candidate precision therapies [00:25:00] in monogenic epilepsies.

This is the prime master protocol, which includes N one randomized controlled trials, which utilize double-blind design use are placebo controlled. And this is I’d really like to acknowledge as I further go into the details of it, the tremendous work of Drsi Wang, who is a young neurologist, who doing a PhD.

And this is at the very core of our PhD. She’s been a tremendous engine behind this. And it’s been a lot of work to set this up. I must say many years of work. And it involves many of you or here that includes Ingrid, includes Sarah, includes Catherine. This is currently set up at the Austin, but we are really hoping to go Australia wide, so everyone has a chance to potentially access an NO one trial.

So how does this work? You get signed up into the master protocol. The master protocol is a framework to conduct trials in parallel. This is widely used in other conditions such as oncology, such as infectious diseases, but it’s not, it has not been utilized in epilepsy as of yet. And so we’re [00:26:00] very proud to introduce this.

The patient, the family signs up into the prime master protocol, and then they can access available precision therapies, which are then tested with a one randomized controlled trial. Where the person goes through period cycles of treatment or placebo. There may be an initial pre randomized phase period where it’s an open label phase where the aim is to identify the most appropriate dose for that person.

And the dose is then tested in the randomized of blind phase. And if that treatment works that treatment will be provided in, in, in an open label extension, and then will be incorporated into the person’s routine management. So this is incredibly powerful because it allows us really to test promising candidate precision therapies for rare monogenic epilepsies.

And here’s an example, a short list of potential candidate epilepsy precision therapies, which are suitable to be tested [00:27:00] using the N oh one trial design for Pyridine Amin, for kcna KC and A one and kcna two related epilepsy, at least gain function variants. And I’m, very pleased to say that we do have an arm, an N one randomized controlled trial assessing it for purine.

This is already active and approved by our ethics committee at the Austin. And we have already a participant in that arm. But you can see that there’s a long list of potential suitable candidates, and this was just a slide that needed to put together. But they could have several more slides and they’re all promising.

But the evidence for their promise is based on single case reports or small uncontrolled observational studies, which really do not provide sufficient evidence to in terms of rigor and of their effectiveness. I’d also like to emphasize that one important aspect of N one trials is that it could also disprove whether a particular treatment is actually working.

And that’s important because a treatment may not be [00:28:00] safe for that person and there is no need to continue on that treatment. Some including some of us have advocated that N oh one trials should also be incorporated in routine clinical care, even to test available existing medications. And and suggestions for this is for instance, to utilize this approach to improve the clinical management of an individual to test, for instance, a treatment approved for the very same indication in that population.

And instead utilizing the research n one trial design to test treatments which are non-licensed which may be off label which may be applied to answer question that goes beyond that individual. Because as I mentioned earlier, one of the advantages of N oh one trials is the possibility of aggregating results when those trials are performed in the same disease to determine whether the results actually are generalizable to the same population.

In conclusion, I hope I convinced you that N oh one trials are [00:29:00] powerful approaches to, to assess the effectiveness of therapeutic interventions at the individual level. They’re particularly suitable to assess interventions in rare conditions, such as rare monogenic epilepsies and genetic neurodevelopmental disorders, but they’re greatly underutilized.

Most previous NO one trials across neurology have been fraught with several limitations ranging from design issues to issues with the statistical analysis, which really hamper the in interpretation of the findings. And well-designed and conducted NO one trials are are warranted to investigate off labor repurposed treatments as well as novel therapies.

For rare monogenic epilepsys. So before I conclude, I just wanted to acknowledge the wonderful team that we have at the Austin, the fantastic collaborators students, postdoc research assistants that have ab they’re absolutely critical for our research and for some of the results shared with you, but importantly, you [00:30:00] patients families, consumers who are at the very center of our world, of our research and we and the drive for for all of our work and of course all the funding that has made the work that presented possible.

Thank you. Thank you.