07 Nov 2001

Numerous studies have shown that positron emission tomography (PET) can reveal small metabolic changes in the brains of patients with neurodegenerative disorders, particularly Alzheimer's disease. However, the paucity of cases examined to date has made it difficult to assess the diagnostic sensitivity of this technique and its usefulness as a prognostic tool. Both of these issues have now been addressed by an international team of researchers, which released its findings in today's Journal of the American Medical Association.

The collaboration, led by Daniel Silverman at University of California, Los Angeles, used PET analysis to examine glucose metabolism in 284 patients who presented with mild cognitive impairment between 1991 and 1998. Two hundred ten patients were predicted by PET to have progressive dementia. This was subsequently confirmed in 91 percent of cases either by a follow-up clinical evaluation conducted at least two years after PET scans (78 patients), or by a neuropathological examination (113 patients) performed, on average, 2.9 years after PET. The analysis also categorized 74 patients as being non-progressive and only 15 of these turned out to be false negatives. "PET, as demonstrated by this study, shows a specificity and sensitivity that makes it the premier imaging modality for diagnosis," said coauthor Mony de Leon at New York University School of Medicine.

In addition, diagnosis by PET removes some of the subjectivity associated with memory tests, which can be complicated by language difficulties or the patient's level of education. The study leaves unsolved the question of precisely how favorably PET diagnosis compares with diagnosis by an experienced neurologist, nor did it address the question of how early PET can predict future AD. But could PET improve monitoring of disease progression, or track the efficacy of potential treatments? "Absolutely," said de Leon.—Tom Fagan

See USA Today Story