Using algorithms that can pick out textures and subtle structural features in the brain affected by Alzheimer's can improve the information we get with standard imaging modalities, said Dr. Paresh Malhotra of Imperial College London.

According to a study from Imperial College London, MRI scans can be used to diagnose Alzheimer's disease - a progressive disease that slowly destroys memory and thinking abilities - even in its early stages. Researchers have found that an algorithm trained to recognize changes in the brain can accurately predict 98% of the time whether a person has the disease. And can distinguish between early and late-stage Alzheimer's with 79 percent accuracy.

The researchers modified the algorithm to classify cancerous tumors - they divided the brain into 115 regions and used 660 characteristics such as size, shape, and texture to assess them. Algorithms were then trained to recognize changes in these traits and predict Alzheimer's disease. The researchers then applied the algorithm to scan the brains of 400 patients with early and late-stage Alzheimer's, health controls, and people with other neurological disorders, including other forms of dementia.

Alzheimer's disease is the most common form of dementia and mainly affects people over 65, but it can sometimes affect younger patients. This disease disrupts the connections between neurons (the basic unit of the brain) and their eventual death over time. Genetics, lifestyle, and environment are believed to play a role when a person is exposed to the disease.

Doctors usually look for unusual protein clumps, tangled bundles of fibers, and contractions in the part of the brain associated with memory called the hippocampus to diagnose patients with the disease. Cognitive tests and brain scans are currently needed to confirm the diagnosis. The new approach can identify it only with an MRI scan, which is readily available in most major hospitals.