Researchers develop new computer program to predict Alzheimer's risk

A new, brain scan-based computer program that quickly and accurately measures metabolic activity in a key region of brain can be applied to predict the development of Alzheimer's disease, US researchers reported on Sunday.

The computer program, called "HipMask," is an image analysis technique that allows researchers to standardize and computer automate the sampling of PET (Positron emission tomography )brain scans.

Scientists at the New York University, who developed this program, hope the technique will enable doctors to measure the metabolic rate of the hippocampus and detect below-normal metabolic activity linked with Alzheimer's.

Earlier studies have demonstrated the hippocampus, a sea-horse shaped area of the brain associated with memory and learning, diminishes in size as Alzheimer's disease progresses from mild cognitive impairment to full-blown dementia.

Yet until now there has been no reliable way to accurately and quickly measure the hippocampal area of the brain on a PET scan. The hippocampus is small and its size and shape are affected greatly in individuals with Alzheimer's, making it difficult to sample this region. HipMask is a sampling technique that uses MRI (Magnetic Resonance Imaging) to anatomically probe the PET scan.

MRI relies on electromagnetic energy to excite water molecules in the brain to create an anatomical map of the brain. The MRI was used in the study to determine the total volume of the hippocampus and then to define the portion that was shared by all persons regardless of their disease status. PET employs radioactively labeled glucose to show the brain at work and the HipMask was applied to these scans to derive estimates of the hippocampal glucose metabolism.

"This is the first demonstration that reduced metabolic activity in the hippocampus may be used to help predict future Alzheimer's disease," said Lisa Mosconi, the research scientist who developed the computer program.

"Although our findings need to be replicated in other studies, " she said, "our technique offers the possibility that we will be able to screen for Alzheimer's in individuals who aren't cognitively impaired."

Mosconi and colleagues have published the technical details of the program in the June 2005 issue of the journal Neurology. She will present the new findings on June 20 at the Alzheimer's Association International Conference on Prevention of Dementia held in Washington.

The researchers followed 53 healthy, normal subjects between the ages of 54 and 80 for at least 9 years and in some cases for as long as 24 years. All subjects received two PET scans, one at baseline and a follow-up after 3 years. Thirty individuals had a second follow-up scan after another seven years. Altogether there were 136 PET scans.

The researchers applied the program to all 136 scans. The results showed that hippocampal glucose metabolism, as determined by the HipMask, was significantly reduced 15 percent to 40 percent on the first scan, compared to controls, of those 25 individuals who would later experience cognitive decline related to either mild cognitive impairment or to Alzheimer's.

The researchers found that the baseline hippocampal glucose metabolism was the only brain or clinical measure that predicted the future cognitive decline. They can show with great accuracy who will develop Alzheimer's nine years in advance of symptoms.

"We're confident this is a strong beginning, demonstrating accurate detection of early Alzheimer's disease, and we anticipate this might open some doors to prevention treatment strategies,"

said Mony de Leon, professor at the New York University and senior author of the paper.

Source: Xinhua