Only 6% of memory services currently have access to the specialised tests needed to confirm Alzheimer's disease. Without evidence of amyloid-β protein buildup in the brain—the hallmark of Alzheimer's—patients cannot receive a reliable diagnosis or become eligible for new disease-modifying treatments. The two existing diagnostic methods, PET scans and lumbar punctures, are either expensive and involve radiation or are invasive, and most NHS clinics simply do not offer them. A company called AINOSTICS has developed a generative AI algorithm that can identify amyloid patterns from standard MRI scans—which are already routinely performed—and synthesise an accurate virtual PET scan without requiring the actual scan or a spinal tap. This project will test the technology on real-world clinic patients and asymptomatic individuals, assess its accuracy, and evaluate whether it can reduce the need for costly PET scans and lumbar punctures while improving geographical equity in diagnosis. If successful, the tool could be integrated into NHS pathways, allowing far more people to receive a specific Alzheimer's diagnosis and timely access to treatment without invasive procedures or radiation exposure.
View original technical description
Multiple different conditions can cause dementia, each carrying different prognoses and requiring different treatments. It is therefore crucial that people are not just given a non-specific diagnosis of "dementia", but that they undergo appropriate assessments to identify the specific sub-type. The most common condition that leads to dementia is Alzheimer s disease (AD). To receive a reliable diagnosis of AD, and to be eligible for new disease modifying treatments (DMTs), it is essential to have evidence of amyloid-β deposition – the core pathological feature of AD. Additionally, assessing tau spread can help stage disease and may identify those most likely to benefit from treatment. However, currently, identifying amyloid-β is only possible with either a PET scan, which is expensive and involves radiation, or a lumbar puncture (LP), which is invasive. Only 6% of memory services currently have access to these specialised tests, leading to huge geographical disparities in timely and specific diagnosis, and meaning the NHS is ill-equipped for the increased demand that will come with the roll-out of DMTs. AINOSTICS has developed a novel generative AI algorithm that provides an efficient, cost-effective, and scalable solution to the current limited availability of AD diagnostic tests. By applying the novel technology to widely available MRI scans, which are already acquired as part of the standard clinical pathway, we are able to identify patterns specific to amyloid, allowing the accurate synthesis of amyloid PET scans without the need for an actual PET scan or LP. This project will obtain efficacy data for this technology in a real-world setting and develop a software prototype for further validation and incorporation into NHS pathways. Aims of the project include: To perform clinical validation of our technology, assessing the accuracy of the amyloid PET synthesis and prediction of amyloid status. Assess the potential clinical benefit, in terms of improved diagnostic accuracy in settings where LP and PET are unavailable. Evaluate the cost-effectiveness of this technology for use in the NHS, including 1) earlier diagnosis and access to treatment, 2) lower LP and PET usage, and 3) fewer appointments and reduced burden from undiagnosed AD. Estimate potential improvements in the geographical equity of access to tests able to determine eligibility for DMTs. To adapt the technology for additional synthesis of tau PET, assessing its accuracy in staging tau (based on spatial distribution) with actual tau PET. We will apply the technology to two separate cohorts: 1) 250 "real-world" clinic patients, and 2) a "trial ready" cohort of 250 asymptomatic individuals. Data collection from these cohorts is already in process, ensuring the current project can be delivered in a timely manner and providing excellent value for money. We will map the future regulatory and commercial strategy for the technology in the NHS. We will co-produce the research with members of the public and clinicians involved in routine care, working with them to understand the potential value of the technology to patients and carers, and how best to communicate the outputs in clinical settings.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know