PT - JOURNAL ARTICLE AU - Bedda Rosario AU - William Klunk AU - Chet Mathis AU - Lisa Weissfeld AU - Wenzhu Bi AU - Brian Lopresti AU - Steven DeKosky AU - Julie Price TI - Use of partial least squares to relate yearly PiB DVR to clinical covariates in non-demented subjects DP - 2009 May 01 TA - Journal of Nuclear Medicine PG - 1412--1412 VI - 50 IP - supplement 2 4099 - http://jnm.snmjournals.org/content/50/supplement_2/1412.short 4100 - http://jnm.snmjournals.org/content/50/supplement_2/1412.full SO - J Nucl Med2009 May 01; 50 AB - 1412 Objectives Examine yearly PiB retention relative to clinical covariates of Alzheimer’s disease in 25 control (CN) (70±10 yr, MMSE 29±1) and 10 mild cognitive impairment (MCI) (70±7 yr, MMSE 28±1) subjects at 1 year and 20 CN (75±9 yr, MMSE 29±1) and 4 MCI (70±6 yr, MMSE 28±1) subjects at 2 years. Methods PiB PET data (HR+, 15 mCi, 90 min) were collected over 2 years. MRI was used for region definition and atrophy correction. Yearly changes in Logan distribution volume ratio measures (DVR, cerebellum reference) of PiB retention were computed: DIFFyr = DVRyr – DVRbase. Partial least squares (PLS) was applied to covariates (PiB+/-, age, MMSE, APOE status, yrs of education: Ed) and yearly PiB DVRs (9 areas: global mean [Gm: cingulate ACG, precuneus PRC, frontal FRC, temporal LTC, parietal PAR, striatum AVS] and 3 white matter rich areas). A PLS latent variable (LV) weight ≥0.3 was empirical indicator of important variable. Results Gm DIFFyr for year 1 was 0.04±0.07 for CNs (3%) and 0.06±0.12 for MCIs (4%) and at year 2 was 0.06±0.10 for CNs (4%) and 0.10±0.12 for MCIs (5%). A PLS analysis applied to a subset scanned at each time (12CN, 4MCI) yielded 1 component (>96% variation explained). The covariates more related to the LV were Ed (0.7), PiB+/- (0.6) for year 1 and PiB+/- (0.8), MMSE (0.5) for year 2. PiB DVRs for global areas were related (0.3-0.4) to Ed and PIB+/- (year 1) and PIB+/- and MMSE (year 2). Conclusions Yearly changes were greater for MCIs than controls. PiB DVR values were related to clinical variables with Ed most strongly related at year 1 and MMSE at year 2. Research Support NIH, Dana Foundation, Alzheimer’s Association