Machine learning pre-screening from MRI
The Neuroimaging Research group has developed a family of machine learning algorithms able to predict the presence of abnormal levels of Alzheimer’s biomarkers in the brain of cognitively unimpaired individuals by analysing derived data from Magnetic Resonance Imaging (MRI).
The aim of the project is to use this technology as a pre-screening tool for Alzheimer’s disease clinical trials, before current invasive and expensive techniques are applied.
Clinical trials in this field typically require assessing the presence of abnormal Alzheimer’s disease biomarkers in candidate participants before inclusion. The current tests for obtaining these biomarkers (CSF analysis or PET scans) are invasive, expensive, and not very accessible. In addition, the rate of positivity is typically low, ranging from 10 to 50%. Therefore, recruitment costs typically amount up to 50% of the total clinical trial costs in AD.
In the short term, the use of this technology will avoid 63% of unnecessary cerebrospinal fluid (CSF) and positron emission tomography (PET) procedures, associated to 40% reduction of costs. Alzheimer's disease phase III trials typically recruit around 1,500 patients (20k€/patient). Therefore, the solution will be linked to 15M€ savings per clinical trial allowing pharmaceutical companies to improve their efficiency. In the long term, the solution will bring us closer to an effective preventive therapy for Azheimer's disease, which now has a cost of 32k€/patient/year.
Alzheimer's disease is a major global health threat, which represents 1.25% of GDP and whose incidence is continuously increasing, expecting to reach 150M patients in 2050. The capability to pre-screen patients ensuring an optimal recruitment in clinical trials will accelerate finding an effective treatment for this devastating disease. Moreover, we must keep in mind the huge impact on the life quality of parents’ family members and caregivers who must deal with the dependency they develop.
This project has been awarded highly competitive grants from CaixaImpulse, supported by “la Caixa” Foundation, and from EIT Health Digital, supported by the European Institute of Innovation and Technology (EIT), a body of the European Union.
Principal investigator of the study
Juan Domingo Gispert, head of the Neuroimaging Research Group.
Juan Domingo Gispert is the project leader for this project. He has over 20-years’ experience in working with medical imaging in diagnostic and research settings, including clinical trials. He has co-authored more than 100 peer-reviewed publications and co-directed 5 clinical trials. Currently, he is the Group Leader in Neuroimaging of the Alzheimer’s Prevention Program and Scientific Consultant of the Neuroimaging Platform of BBRC / Pasqual Maragall Foundation. Juan Domingo Gispert is currently the principal investigator of the European project AMYPAD to develop the use of amyloid PET imaging in clinical trials and in the management of patients with suspected AD. In addition, he serves in the Imaging Scientific Advisory Board of the European Prevention of Alzheimer’s Dementia (EPAD).
