Researchers at University College London have created a new machine-learning algorithm to examine MRI scans from patients with dementia. When they applied the algorithm, it identified three separate subtypes of Alzheimer’s disease, as well as several different subtypes of frontotemporal dementia. Called SuStaIn (Subtype and Stage Inference), the algorithm looks at the build-up of proteins in specific locations and deduces which parts of the brain are degenerating. This new tool will make it easier to distinguish different patterns of disease progression, using a non-invasive technology. As a result, clinicians may be better able to prescribe interventions that match the patient’s particular form of dementia. The technology can also improve testing of potential new interventions, by allowing researchers to know which subgroups respond best to which treatments.
SOURCE: Nature Communications, October 15, 2018
DATE Retrieved: November 1, 2018