WASHINGTON, DC—A new technique—the HipMask—has provided the first evidence that hippocampal glucose metabolism reductions detected in the normal stages of cognition can predict future cognitive decline and a diagnosis of Alzheimer’s disease. “This is the first brain imaging study to demonstrate detection of sporadic Alzheimer’s disease in normal elderly subjects,” Lisa Mosconi, PhD, said in a presentation during the Alzheimer’s Association International Conference on Prevention of Dementia. “These data suggest that the recognition of future Alzheimer’s disease in a person who is cognitively normal is a realistic goal.”

Previously, there was no reliable way to measure the hippocampal area of the brain accurately and quickly using positron emission tomography (PET). In the June 14 Neurology, Dr. Mosconi and colleagues noted that paradoxically, more than 200 studies using 2-[¹⁸F]fluoro-2-deoxy-d-glucose (FDG)-PET found abnormal reductions of glucose metabolism in large cortical regions, compared with only five reporting hippocampal abnormalities. They believed that this discrepancy was associated with features of the imaging itself; that is, studies that detected hippocampal hypometabolism used FDG-PET with a manual region-of-interest guided by co-registered MRI, and those that used FDG-PET with automated voxel-based analysis had negative findings “due to a failed spatial alignment of relatively small structures like the hippocampus that are prone to high anatomic variability with aging and neurodegeneration.”

However, it is known that damage to hippocampal formation is critical for cognitive decline, said Dr. Mosconi, a researcher at the Center for Brain Health, New York University School of Medicine, New York City, who helped develop the HipMask technique. To demonstrate the clinical utility of the HipMask procedure in determining reductions in hippocampal glucose metabolism in mild cognitive impairment and Alzheimer’s disease and to test its equivalence to the gold standard, the manual region-of-interest, the investigators performed a longitudinal FDG-PET study of normal aging. They followed 53 healthy subjects (mean age, 67; range, 50 to 84) for a mean of nine years (range, eight to 14 years). All participants had two FDG-PET scans—one at baseline and one after three years. After an additional seven years, 30 individuals had a second follow-up scan, for a total of 136 scans. All FDG-PET scans were performed using the same scanner.

THE HIPMASK TECHNIQUE

The HipMask, which uses MRI to probe the PET scan anatomically, was applied to the 136 scans. First, MRI was used to determine total volume of the hippocampus and then to define that portion (namely, the HipMask) shared by all persons regardless of their disease status. The HipMask was then applied to the PET scan to derive estimates of hippocampal glucose metabolism. Results showed that hippocampal glucose metabolism was significantly reduced by 15% to 40% on the first scan versus control scans of those 25 individuals who would later experience cognitive decline related to either mild cognitive impairment or Alzheimer’s disease. At the end of the study, 28 subjects remained healthy, while 19 demonstrated decline to mild cognitive impairment, and six declined to Alzheimer’s disease.

Compared with results in those who remained healthy, baseline hippocampal glucose metabolism was reduced by 11% in those who progressed to mild cognitive impairment and by 22% in those who developed Alzheimer’s disease; in addition, those who progressed to Alzheimer’s disease had an additional 13% reduction compared with those who had mild cognitive impairment.

At baseline, no cortical glucose metabolism differences were found between groups, Dr. Mosconi said. Longitudinal interaction effects were restricted to the left superior temporal cortex, with both those progressing to mild cognitive impairment and those who developed Alzheimer’s disease showing increased cortical glucose metabolism over time compared with the healthy group. At the third follow-up, those with Alzheimer’s disease showed parietotemporal, posterior cingulate, and hippocampal glucose metabolism reductions as compared with those who remained healthy. Two of these received a primary pathologic diagnosis of Alzheimer’s disease. Dr. Mosconi said that hippocampal glucose metabolism, the only baseline brain or clinical measure that predicted future cognitive decline, yielded an overall prediction accuracy of 81%, with individual group accuracies of 85% for the group who developed Alzheimer’s disease, 75% for those developing mild cognitive impairment, and 89% for those who remained cognitively healthy.

Dr. Mosconi added that none of the neuropsychological tests administered at baseline were of value in predicting mild cognitive impairment or Alzheimer’s disease.

PREDICTING ALZHEIMER'S DISEASE

“Right now, we can show with great accuracy who will develop Alzheimer’s [disease] nine years in advance of symptoms, and our project suggests we might be able to take that out as far as 15 years,” said lead investigator Mony de Leon, EdD, Professor of Psychiatry and Director of the Center for Brain Health at the New York University Medical Center. “Our basic results will need to be replicated in other studies and expanded to include PET data from diverse patient groups,” Dr. de Leon added. “But we’re confident this is a strong beginning, demonstrating accurate detection of early Alzheimer’s disease. Now we have a better tool to examine disease progression, and we anticipate this might open some doors to prevention treatment strategies.”

NR

—Debra Hughes

Suggested Reading
Mosconi L, Tsui W-H, De Santi S, et al. Reduced hippocampal metabolism in MCI and AD: automated FDG-PET image analysis. Neurology. 2005;64:1860-1867.

Return to table of contents