P2-422: vMRI and hippocampal volume in patients with suspected predementia Alzheimer's disease

Clinical assessments of Alzheimer's disease (AD) progression rely on subjective evaluation of worsening function and memory loss. Research suggests that hippocampal volumetric magnetic resonance imaging (vMRI) in subjects with suspected predementia AD (PDAD [i.e., mild cognitive impairment (MCI)]) offers a complementary objective tool to diagnose AD progression, as well as to track disease course and potentially, treatment response. A systematic literature review was conducted to evaluate vMRI performance in subjects with suspected PDAD to predict risk of progression to AD-dementia and longitudinally track volume loss. Studies published in English (January 1995-October 2010) were considered. Data were extracted from studies reporting appropriate data for meta-analytical pooling, and random-effects meta-analyses of odds-ratios were performed. Of a total of 1164 citations screened, 29 studies examined hippocampal vMRI in MCI (suspected PDAD), with 14 suitable for the meta-analysis. Follow-up ranged from 1 to 2 years. Cox proportional hazards models showed that increased hippocampal atrophy correlated to a significantly higher risk of progression from MCI to AD-dementia. Sensitivity of vMRI to identify baseline MCI progressors ranged from 61%-93%; specificity ranged from 20%-80%. Altogether, 14 whole-hippocampus studies (1039 subjects) showed that subjects who progressed from MCI to AD-dementia showed significantly lower hippocampal volumes at baseline (odds-ratio 2.83; 95% CI: 2.14-3.74). Four studies (173 subjects) separated left- and right-hippocampus volume and data were meta-analyzed. While a slightly higher odds ratio was found for right-hippocampal volume (2.6 vs 2.0), the difference did not approach statistical significance. Longitudinal change in hippocampal volume, comparing MCI progressors to subjects stable in MCI, was also assessed. In all studies, subjects who progressed to AD-dementia demonstrated substantially higher annual hippocampal volume loss compared to non-progressors; progressors showed a 2-3 fold increase in annual volume loss compared to non-progressors. Total hippocampal volume has been shown to be a good predictor of progression from MCI to AD-dementia in many individual studies. Further work is needed to define optimal methods for identifying, a priori, those subjects at greatest risk for progressing to AD-dementia.