de WEGER, Mr Anton, Executive Lead, Australian Epilepsy Project FORGASZ, Ms Rebecca, Strategy and Government Affairs Lead, Australian Epilepsy Project MALJEVIC, Associate Professor Snezana, Deputy Lead, Epilepsy and Neurodevelopment Research, Florey Institute of Neuroscience and Mental Health TAILBY, Dr Chris, Chief Scientist, Australian Epilepsy Project [10:12] CHAIR: Welcome. Thank you very much to all of you for appearing before the committee today. I understand that information on parliamentary privilege and the protection and obligations of witnesses giving evidence to Senate committees has been provided to you. I now invite each organisation to make a short opening statement. Just a reminder that we have your submissions, and folks have lots of questions, so if each organisation could keep their statement to around three minutes that would be great. Mr de Weger : Thank you, Chair and members of the committee, for the opportunity to be here today. We are honoured to contribute to this inquiry and to improve the lives of people living with epilepsy. The committee has received over 200 submissions. The picture they paint collectively is of a healthcare system that is failing people with epilepsy. Agonising delays in accessing specialist and expert care was the single most universal theme across all submissions. Our colleagues in ANZAN and the ESA, the professional associations for neurology and epileptologists, provided data showing that wait times for new adult patients at epilepsy clinics exceed 900 days in some states. These delays are the entry point to a cascade of harm, with devastating impacts on physical and mental health, family life and careers told in many of the submissions by individuals and families. Depression and anxiety are cited in over 70 per cent of submissions as almost universal comorbidities of epilepsy that are virtually unaddressed in standard epilepsy care. For those with drug resistant epilepsy, multiple submissions describe trial-and-error medication odysseys over years and even decades, with waits of 10, 15 or even 20-plus years before being assessed for potentially curative brain surgery. Adding to all this is the multiplier effect of geographic inequality, raised in dozens of submissions from individuals and families. The Australian Epilepsy Project has engineered a bold world-first solution that specifically seeks to address these systematic problems. It's the national, physical and digital infrastructure that provides advanced epilepsy testing to people regardless of where they live. The testing includes specialised MRI scans, screening for anxiety and depression, cognitive assessment and genetic testing. We provide an integrated report of findings back to the referring neurologist, which helps to identify the causes of optimal treatment pathways for patients' epilepsy. We already have proven results identifying the cause of epilepsy for up to 19 per cent of our patients whose previous MRI scans were normal and reducing the delay from diagnosis to seizure freedom to as little as two years. We can't solve all the problems in epilepsy care, but we know we can be part of the solution. In 2021 the Australian government backed our vision with a $30 million MRFF Frontiers grant. We recently received, and are extremely grateful for Minister Butler's announcement of, a second tranche of funding of $30 million over four years. We have the potential to transform the standard of care for epilepsy at a national level, but we need policy level support, including expedited regulatory pathways and reimbursement pathways through MSAC, to ensure potential is translated into action. Many submissions have recommended a national coordinated approach to clinical care pathways for epilepsy. We would add to those recommendations and specifically call for the AEP to be recognised as a key infrastructure and implementation partner in any such coordinated national plan. We recommend our submission to the committee, which provides further details about the points we have raised in this opening statement, and we welcome questions about the AEP or any other aspects of epilepsy that we might be able to assist with. CHAIR: Associate Professor Maljevic. Prof. Maljevic : Thank you for the opportunity to appear today. You have just heard about epilepsy as a common neurological condition affecting many Australians, but I'm going to talk about a smaller and particularly severe group of epilepsies—developmental and epileptic encephalopathies, or DEEs. They typically begin in infancy or early childhood and are caused by genetic variants that often arise for the first time in the child that is affected. These are not simply epilepsy disorders; the same genetic change that causes seizures is also profoundly disrupting brain development. Many children cannot walk or talk, require constant care and face a substantially elevated risk of premature death. Currently, there are no effective treatments. Twenty years ago, the biggest challenge for these families was getting a diagnosis at all. Families were sent home with the label of idiopathic epilepsy, which basically meant, 'We don't know what your child has.' That has dramatically changed with advances in genetics, driven significantly by Australian researchers. Genetic cause has been identified in about 50 to 60 per cent of DEE cases, with more than 1,000 implicated genes. But knowing the gene is only the beginning of the story; the next challenge is learning what to do with that information. This is where discovery science plays a critical role. At the Florey, we study how disease-causing genetic variants affect brain development and function, and we use that knowledge to design therapies that target the underlying cause, not just symptoms. As an example, our pre-clinical research on a specific genetic type of epilepsy, DEE epilepsy, directly contributed to gene targeted therapy that is now in advanced clinical trials. Just this week, one received the FDA Breakthrough Therapy designation, which is the highest level of recognition for a promising new treatment. That therapy did not emerge from translation alone; it emerged from years of discovery science. We are now at a point where advances in genetics technology and our understanding of brain biology have made targeted therapies genuinely feasible for the first time. The knowledge needed to develop the next generation of these therapies exists in Australian laboratories, but discovery science requires sustained long-term investment and that pipeline is under pressure. The opportunity before this committee is to support a stronger pathway from diagnosis to understanding and from understanding to better outcomes for Australian families. Without that investment, diagnoses remain answers; with it, they become treatments. Thank you. I welcome your questions. CHAIR: Senator Askew. Senator ASKEW: Thank you for your submissions and for taking the time to come along today. It's been really interesting to read through so many of the different submissions that we've looked at. I was quite interested to read the submission from the Australian Epilepsy Project, especially in regard to development in areas such as AI. Could you give us a little bit more background in what that is going to look like, how you see AI and why that is so important as part of going forward? Mr de Weger : We start off with a fixed protocol of testing. What we do is take all of the tests that you would do on the most severe and drug-resistant cases of epilepsy just before surgery, and we do them at the beginning of somebody's journey. That gives us a very fixed dataset with the latest data at high resolution using the latest capabilities. We then process all of that data. We bring it into a centralised location, and we process it to extract out biomarkers from the imaging, the cognitive testing and the genetics. That is quite a complicated process that does use some AI in some cases—more likely machine learning than the complex large language models we're seeing at the moment. We then combine all of those biomarkers into a centralised location where we keep track of the provenance of that data. We then use other correlations analysis to determine what the relevant concerns for this individual are based on how they relate to all of the other people that have gone through the process, including our volunteers. So we do use complicated and the latest capabilities to be able to process this very large amount of data per participant, and it is also related to how they relate to everybody else that has been through that gives the information that we need. We then refine that down to the most salient points, which we send through to the clinician. We've based what we send through to the clinician on years of research from our expert clinicians as to what they need to see to make the best possible diagnosis and treatment decisions. Senator ASKEW: So the AI is just a tool in there. Mr de Weger : Yes. Senator ASKEW: How often is surgery required in cases of epilepsy diagnosis? How comfortable are people with the use of AI in that? Mr de Weger : I'll go to Chris to talk about the surgery. Dr Tailby : Roughly 30 per cent of cases end up being refractory to medication; they don't respond to medication. In cases of focal epilepsy, that's the trigger to look for whether there is a surgical option. The exact percentage I'm not sure of—I'd have to take that on notice—but that step of reaching drug refractoriness should precipitate the referral for investigation of whether surgery is an option in that case. This is where something like the AEP can really help with its advanced MRIs. We've found that, to date, 10 per cent of cases have been missed on previous MRIs. It's being picked up through the AEP MRIs that there's a lesion that's likely causative of the epilepsy, and that offers a potential surgical target. The cognitive testing can also give clues as to where the focus of the epilepsy may be coming from. All this information can then be fed into a comprehensive epilepsy program that does that detailed work-up for the surgical planning itself. Mr de Weger : In relation to the use of AI, we've had 1,500 people go through our process at the moment, and we haven't had people exclude themselves from participating because of our use of AI. We're very confident that we're using it in the right kind of way and very concerned that we are using it in a way that is appropriate for medical usage. Senator ASKEW: I noticed you spoke in your opening statement about needing policy-level support. Is that to actually facilitate the use of AI, or is that just the broader picture? Ms Forgasz : I think there are a couple of things at policy level. One is the Commonwealth supported coordinated approach to developing a coordinated care pathway for epilepsy. There's a great precedent in cancer care with optimal care pathways, and there's work underway by ENDA that's been previously mentioned—we have a good relationship with ENDA—which is looking at an optimal care pathway for epilepsy. Something where something like the AEP can be built in at the appropriate touchpoint—something that's endorsed and supported from a national perspective would be one. Regulatory pathways are really important for us as we transition from being a research project to being able to be a sustainable and commercially viable service on an ongoing basis. There is a pathway for software as medical device. The landscape for AI is still evolving, but it's very well known, for a lot of medtech companies, that they do have to go overseas first to get that regulatory approval because the process in Australia can be very slow and drawn out, and we want to expedite the availability of this technology to patients in Australia, so an expedited regulatory pathway is really important. Finally, there is currently no reimbursement pathway for digital health technologies through the MBS. Whether it is a reimbursement pathway through the MBS or an alternative funding model—there have been studies into alternative funding models that might be better suited to a more national infrastructure, like ours—policy support for a funding mechanism is recommended to make this affordable, universally, as part of the standard of care. Senator ASKEW: What would the alternative look like? You're saying that they've been looking at it. What does that look like? Ms Forgasz : A kind of model that we've looked at in the past that we feel might be applicable is something like BreastScreen Australia. That's one of the national health programs that are funded at a national level for everyone. There's still a bit of a challenge with those because they get implemented state by state. Each state has a different operating model and a different infrastructure for how that's delivered, which is really complicated. A program like ours is already operating at a national scale. The technology is available everywhere. Something like a national health program might be an alternative funding model rather than a reimbursement pathway or trying to work, on a state-by-state level, through the public hospital system. Senator ASKEW: Associate Professor Maljevic, you were talking about the epilepsy treatments that are currently in development. Can you summarise whether there is a significant deficit in research? Do you feel like we've actually got the research requirements underway? What sort of support do you need in that area? Prof. Maljevic : Obviously, it wouldn't be news that research is underfunded. We are having a really good opportunity to jump on board with novel therapies. I think there is a great knowledge in Australia. I want to emphasise here that AEP was possible and that our work is possible because there was a sustained funding that Melbourne epilepsy groups had over years. For 20 years, over program grants, we were all working together and interacting and building that knowledge base and that experience. That is now not really the case anymore. Obviously, AEP is a great success. The other people staying in research are fighting with current grant cycles every couple of years. That is where those of us working on rare epilepsies especially see the opportunity to maybe ask for some coordinated funding—something like the Childhood Dementia Initiative achieved, because that is a great example of how something that was fragmented, as it is discussed here a lot, came under one umbrella and is now getting research, because the area is definitely underfunded, and we are noticing that in losing people to industry, to teaching positions and to going abroad. That's probably the biggest loss we are having. Mr de Weger : I fully agree with what Snez is saying there. We're really at the beginning. We're collecting a brand new, world-leading dataset on epilepsy that doesn't really exist anywhere else in the world, but that's just the starting point of it. We still need to analyse that, and there's a lot of research still to go from where we are, and it is really important that we do that here in Australia and then translate that into outcomes in Australia as well. Senator ASKEW: Excellent. CHAIR: Senator Ruston. Senator RUSTON: This question is probably for the Florey institute. You talk about pharmaceutical development for ultrarare conditions being commercially unattractive without specific policy mechanisms to support that. I'm really keen to understand. Have you thought through what policy mechanisms are needed in order to better support people who live with epilepsy? Prof. Maljevic : Thank you for that question. It's a really important one. We know that companies won't jump on board with ultrarare diseases because they are simply not interesting. But, when we have one or a few patients, or even 10 patients, with a specific change in a gene, there is great potential nowadays with the advancement of technologies to develop gene targeted therapy that is a particular precision medicine treatment for that patient. However, at this stage, that kind of treatment and those kinds of developments are happening around the world. In the US there is a compassionate use; there is N-of-1 or N-of-few approval methods that go through the FDA. So it is regulated. There is a bit more control of what happens. We know that families here are really interested because we, as researchers, get approached, but there are no guardrails around N-of-1 studies here. We don't have any guidelines for how that should work. So it is left to the multidisciplinary committees at clinics to maybe approve some compassionate use of something that is available. We are working on development of those therapies, but we don't know how they will actually reach a clinic or a patient because those rules are not there, and I think they are really needed. That's a conversation probably to be had between the TGA, NHMRC, Health and consumers to decide how we regulate those trials. This is happening only in one or a few patients, but we do want to keep the knowledge and the outcomes and whatever we learn from those treatments in rare patients, for the ones that are coming next. Senator RUSTON: Have you had much involvement with the health technology assessment review? If you have a reasonable understanding of it, is there anything in that review that you think would play into a better pathway for new and innovative technologies and treatments for people living with epilepsy? Prof. Maljevic : Unfortunately not, but I can take that on notice if that's of interest. Senator RUSTON: With respect to the challenges that you're talking about and you've discussed this morning in response to me and others, I'd like to understand if the recommendations of the health technology assessment review would in any way respond to those policy mechanisms that you say are not there or are not adequate in providing expedited access—or any access, for that matter—to these treatments. Chair, wind me up when you need to. You know me; I'll talk for hours. CHAIR: I'll give you one more question, Senator Ruston. Senator RUSTON: Okay. The other thing that you mentioned in your submission was the diagnostic-to-therapeutic gap. Earlier you mentioned genomics and disease modelling—well, you mentioned genomics. I'm interested in what the biggest barriers are to you for translating genomic or genetic diagnosis into the clinical interventions in Australia at the moment. Prof. Maljevic : Obviously, genomic diagnosis is kind of naming what is the cause. We now know this is actually because there is a genetic variant in a gene—let's say SCN2A because this is the one that we provide as an example. It's really interesting because, once you have that diagnosis, you actually still don't know much more about what the treatment should be. If you then talk to people who work on SCN2A, or if you do the functional analysis of a variant, you can actually see that there is increased function or decreased function. Depending on what you discover for each of the variants, they will have completely different presentations in the clinic but they will also have completely different responses to drugs. You would obviously need to have a laboratory model, or at least a computational model, to predict how a specific variant in a child will behave. Is there going to be a gain of function? If that is the case, you can use specific antiseizure medications that block a sodium channel. Is there going to be a loss of function? In that case, if you apply that drug, you can make things worse. So that is the first step and distinction. The next one would be—if you are developing novel therapies, you would want to know the disease mechanisms. We now have access to technologies like stem cell technologies, so we can have patient-derived cells, and you can study the pathways that are affected. I will also mention that it's not only seizures we are treating here; these variants are disrupting development. We want to understand to what extent they are and what else we need to repair apart from the very specific mutations that these patients carry. That obviously involves years-long research; it is not a thing that can happen overnight. The thing that is missing on the other end are our translational capabilities, because it would be really helpful to have things like platforms for quicker screening—maybe biorepositories where we know that all of the patient lines that are in animal models or other models are centralised and that everybody in Australia working on DEE can access them. Those are things that would make the path to the clinic faster. But, yes, one barrier is making disease models and laboratory predictions, because those are important. Even if we go with ultrarare diseases, we do not want to end up with something that is unsafe and something that is not working going into patients. Senator RUSTON: Yes. Thank you, Chair. CHAIR: Senator Ananda-Rajah. Senator ANANDA-RAJAH: To the Australian Epilepsy Project, we've seen throughout all these submissions this concept of the diagnostic odyssey—which I think is a little bit euphemistic, actually; it sounds like 'diagnostic drift' to me—and, of course, the tyranny of distance and the postcode lottery. Can you talk us through how the AEP is trying to solve these problems? Mr de Weger : I'll start, and then I might hand over to Rebecca. We only have one in-person component of our testing, which is the MRI. We use research MRIs at the moment, and we want to get into commercial MRIs to give us more capability. Everything else is done remotely. The cognitive testing is done via a telehealth tool that we've developed ourselves. The genetics testing involves a saliva sample that you post in. We have online capabilities for the anxiety and depression surveys and other surveys that the patients do. We collect medical histories from their clinics and hospitals and use those as well. That means that it is much more capable of being used externally from people—and we have had people from every state in Australia come through our program, including some who have had to travel extraordinary distances to get to the MRI. But we feel that this really does open it up as something that gives you the same result, irrespective of your postcode. That is really important because that's where we're seeing a real difference. There is an imbalance in the access to information. If you saw one of the very capable epilepsy experts in Melbourne or Sydney, they might have the capability to do some or most of what we do. If they wanted to spend the many weeks it would take to do the kind of processing that we've now automated, then you would get the perfect result, but most people don't have access to that. What we want to do is bring up the quality of care from whomever you see to a much higher level. I think we can do that through technology and embedding the knowledge about the conditions specifically in an automated process. Senator ANANDA-RAJAH: Could you step us through the patient journey, from referral, and how it all happens? Mr de Weger : Rebecca, do you want to take that? Ms Forgasz : Sure. Patients currently have to be referred through a neurologist. They see their neurologist; the neurologist assesses whether they meet our eligibility requirement. We've built an online clinician portal, and the clinician can make a single referral through to the AEP. We then get in touch with the patients, and we coordinate all of their bookings for the MRI—whichever is their closest one. We book in the collection of all their medical history data, the surveys that we do with them, the interviews and their cognitive testing. We have a central coordination role with the patient. We've got pretty quick turnaround times at the moment in that we're able to get the patient off to the scan, collect up all of the data centrally, put it through our processing and return the report back to the neurologist. The idea is that, in the normal workflows for a neurologist, by when they would next be seeing their patient at an appointment, they will have the AEP report back to them in time for that next appointment. We also currently do follow-ups at six, 12 and 24 months with our participants so we can get that longitudinal data on their outcomes. Senator ANANDA-RAJAH: It would be helpful if you could supply us, on notice, with some information on the timestamps between these steps. We've also had submissions talking about the delay in seeing neurologists. That's a key issue. It's one to two years in some places. What is it that a neurologist brings that a GP can't bring to the table? I just want to cut the red tape and get care to patients—that's the bottom line. I know we can, because I'm a doctor. What do we need to do to upskill an army of GPs to cut the delay? Dr Tailby : I'll speak first to the question of the neurologists. Through the mechanism that AEP has developed, all these diagnostic tests occur early and concurrently. That adds a level of efficiency that should enable the neurology workforce to work quicker. They're getting all that information in one bundle at one point in time, so there are not repeated bookings and follow-up bookings to book the next investigation. That's one way of enabling the neurological workforce to operate more efficiently. The question about how we can work with GPs—I think this was touched on before with the advocacy groups as well. Is there some level of education that we can engage with? Through our AEP platform, we've got mechanisms by which anyone who's a referrer can log in, they can access the report and they can see all the key information there, and then we can use that portal as a means to link in with appropriate services as need be, along with an outreach service from us that upskills people. Senator ANANDA-RAJAH: So you're telling me that this could be done? Obviously, the neurologist would have to be there as a key decision-maker, but it is conceivable to train up and educate GPs. GPs do this sort of stuff all the time. I mean, it's not rocket science to identify a patient who has epilepsy. All doctors are trained in this. Where it's difficult for GPs is in that specialised management and diagnostic pathway. But is it conceivable to bring this forward and actually have GPs be the referral point and then get them into the AEP? Dr Tailby : Ideally, we'd still like to have links through a neurologist for complex cases. Senator ANANDA-RAJAH: No question. But I'm talking about the first step. I don't think it's acceptable for patients to be waiting one to two years to see a neurologist. That's crazy. Ms Forgasz : I know it's certainly a discussion that has been happening with AEP. I'm speaking as a non-specialist. It's a question that we do ask: whether that is something that at some point we could consider introducing, that a referral to the AEP comes through a GP. I think there are a lot of questions to work through. Senator ANANDA-RAJAH: Talk us through what the barrier is. Where is the resistance to having the referral come through a GP? Dr Tailby : It's making sure that patients have access to specialist care where a trigger is required—if a complex case comes through, where it needs to be linked in. Is that answering your question? Senator ANANDA-RAJAH: Not really. I need a little bit more detail and clarity. What is the barrier to us making a recommendation that a GP could be the referrer into the AEP pathway? You'd still have your neurologist and all the other bits and pieces. Mr de Weger : We have somewhat addressed that by making an expression of interest process, so that the patient can express interest and then we can talk to their neurologist directly about that. Obviously, that doesn't solve the problem if they don't already have a neurologist, but we can recommend the people that do refer through as a potential option there. But, yes, I think there is an opportunity for us to consider that, although that hasn't been part of what we've looked at so far. Ms Forgasz : I think we should take that on notice, perhaps, to come back with a considered view on that. Senator ANANDA-RAJAH: I think that would be helpful. Could you talk us through the difference between your research scanners and standard MRI scanners that are found in tertiary hospitals, for example? Dr Tailby : The modern scanners have different capabilities, so the hardware that's on board can deliver different types of scans. An example would be our diffusion images. You need a certain field strength to be able to drive the scanner to acquire the images that we collect. But we're actively working now towards a clinically deployable protocol that would be more generalisable across scanners that are available more broadly. Senator ANANDA-RAJAH: And how many research scanners are there in the program? Ms Forgasz : We've got eight sites currently. Senator ANANDA-RAJAH: Dotted around the country? Ms Forgasz : Yes. Senator ANANDA-RAJAH: That's helpful. With regard to model bias, where is the AI used? Is it image analysis of the scan or is it integrating all this multimodal data? And what are you doing—how has bias been mitigated or how is it being mitigated? Obviously, we're a multicultural nation and very diverse. Mr de Weger : Yes. I'll get on to the multicultural side in a second. We use AI—or, more accurately, machine learning and statistical methods—to analyse the data. In the imaging itself, we are using AI trained capabilities to detect the size and segmentation of different parts of the brain that are worthwhile measuring. We look at different activations of the brain for things like language laterality. They're all based around different types of models. We also use different models to analyse the different biomarkers for correlations across them that might point to things. Our approach has always been to use much smaller and more statistical approaches that are much more resistant to bias, rather than one large neural net that solves all of our problems. The real benefit for us is not in the prediction that we can get out of a machine-learning model; it's also in pointing back to what the main drivers of that were for this individual, because that's really what the clinician needs to see to be able to trust and question those results. We also try to make sure that the healthy volunteer cohort that we're using to validate our approaches and normalise the data has the same kind of distribution as our epilepsy cohorts in the group. We try to get people from everywhere that we can. Regarding the diversity here, we do a number of adjustments for different demographics, but actually the diversity is one of our big benefits here—in Australia, we have got a good ethnicity range. That means that tools that we can build here in Australia are fairly generally applicable globally. So that is the— Senator ANANDA-RAJAH: I guess my next question, on commercialisation, follows. Given the strength of the diverse dataset that you're building, this tool could then be applicable globally, in other markets and other settings. What are the steps needed to go from this research project to commercialisation? Where do you go to raise capital, what next steps are you looking at and what do we need to do to make that easier? Mr de Weger : There are a number questions there. There is an opportunity for Australia to move into a world-leading position in medtech exporting. We've got all of the capabilities that we need here. We've got access to the right technology, we've got the right medical devices, we have a fantastic research section and we have very good medical people here. What we need is a way that supports the development of these technology enabled tools so that they are appropriately regulated and appropriately controlled. But also there must be a way for us to get cost recovery for that. At the moment, we pay for everybody that's been through our process from our grant so that we can collect that data to build these kinds of systems. If we had more support in building these kinds of capabilities in a trusted and appropriate way, that is something that other countries will be very keen to use. I think we're in a really good position for the future to move in that direction. Ms Forgasz : To add to that, the new 10-year National Health and Medical Research Strategy points to challenges in bringing Australian innovations to market and potentially making them exportable internationally. The two challenges that we've talked about that are really important for us—the regulatory pathway and the reimbursement pathway—are specifically singled out in that strategy as key areas and priorities that we need to work towards. I know that the department is appearing later and perhaps they're points that they could speak to later today. Senator ANANDA-RAJAH: On the trustworthy AI component, it sounds like you're developing trustworthy AI. Can you briefly talk us through the cyber protections around this dataset? Mr de Weger : Yes. It is very important. We are providing a service, as well as collecting data, so we do have personally identifiable information in our dataset. It is extremely important to us that we keep this data secure. We do that using the best practices for cloud computing. We use AWS as the host for our staff. We encrypt all of our data at rest and we hold the keys for doing that. We also build in a way that is very resistant to external access. We build via microservices that are serverless and leave less attack vectors for external access to that data. We also de-identify all of our data before we provide that to any external parties for research or for other actions. Senator ANANDA-RAJAH: Thank you. CHAIR: I thank all of you for appearing before the committee today and providing your evidence, as well as your submissions. The committee will report to the Senate on 10 September 2026 and has agreed that responses to questions on notice—and there were a couple—should be provided by close of business on Friday 10 July 2026. Proceedings suspended from 10:51 to 11:07 .