Only one study has reported progression of cognitively normal subjects to more advanced disease. Morris and colleagues [70] performed 11C-PIB scans in 159 cognitively normal (Clinical Dementia Rating kinase inhibitor U0126 (CDR) 0) subjects that were part of a longitudinal aging study and reported that the relative risk of conversion from CDR 0 to AD (nine subjects) was increased almost five-fold in the presence of a positive 11C-PIB amyloid scan. A lesser, non-significant increase in risk was reported for conversion from CDR 0 to CDR 0.5 (n = 23). The primary weakness of studies using conversion/stage change as an endpoint is that the rate of conversion, particularly from healthy to MCI or AD, may be low and variable across subjects and studies, depending on recruiting centers and entry criteria.

Hence, three studies in cognitively normal aging elderly have looked instead at the relationship between PET amyloid binding and continuous measures, that is, change in objectively measured cognitive performance. Storandt and colleagues [62], working with essentially the same subject population as Morris and colleagues [70], found that concurrent cognitive performance was unrelated to 11C-PIB binding, but the estimated annual rate of cognitive deterioration, as evidenced by change in visuospatial and working memory performance composite scores, was significantly greater in subjects with an amyloid-positive 11C-PIB PET scan than in subjects with an amyloid-negative 11C-PIB scan. High amyloid binding on 11C-PIB scans was also associated with reduced regional brain volume on MRI, further suggesting that even in cognitively normal subjects (CDR 0) amyloid accumulation is not benign.

Villemagne and colleagues [71] imaged 34 elderly subjects that had been previously followed longitudinally for 6 to 10 years. On average, subjects with memory decline over the observation period had higher 11C-PIB retention; 7 of 11 subjects with elevated 11C-PIB retention showed memory decline, versus 4 of 23 subjects with normal 11C-PIB retention. Finally, Resnick and colleagues [67] obtained 11C- PIB PET images on 57 subjects who had been followed for an average of 10.8 years as part of the Baltimore Longitudinal Study of Aging and found a significant correlation between 11C-PIB binding (DVR) and Mini Mental State Exam and verbal memory (California Verbal Learning Test).

One weakness of the Storandt and Drug_discovery colleagues [62], Villemagne and colleagues [71] and Resnick and colleagues [67] studies is that they rely primarily on retrospective sellekchem analysis of cognitive decline. Although several groups have now reported that change in 11C-PIB binding is relatively slow, particularly in amyloid-positive subjects [40,51], it is difficult to judge from a retrospective analysis how early the 11C-PIB PET could have predicted subjects likely to show cognitive decline.