Can Neuroimaging Predict Dementia in Parkinson's Disease?

Juliette H. Lanskey; Peter McColgan; Anette E. Schrag; Julio Acosta-Cabronero; Geraint Rees; Huw R. Morris; Rimona S. Weil

Disclosures

Brain. 2018;141(9):2545-2560. 

In This Article

Grey Matter Measurements

Grey matter atrophy may represent neuronal cell death and has long been associated with cognitive decline in Alzheimer's disease (Fox and Schott, 2004). This observation led to a large number of neuroimaging studies comparing patients with Parkinson's disease without cognitive impairment and people with PD-MCI. These showed varying atrophy patterns in frontal (Song et al., 2011; Melzer et al., 2012; Mak et al., 2013; Hanganu et al., 2014; Gao et al., 2017), temporal (Melzer et al., 2012; Hu et al., 2013; Mak et al., 2013; Pagonabarraga et al., 2013; Hanganu et al., 2014; Noh et al., 2014; Gao et al., 2017), occipital (Melzer et al., 2012; Pagonabarraga et al., 2013), parietal (Melzer et al., 2012; Pereira et al., 2014) and insular cortices (Mak et al., 2013; Hanganu et al., 2014), as well as subcortical atrophy (Melzer et al., 2013; Hanganu et al., 2014; Foo et al., 2017; Schneider et al., 2017), including hippocampal (Melzer et al., 2012; Schneider et al., 2017), amygdala (Melzer et al., 2012; Hanganu et al., 2014) and nucleus accumbens volume loss (Hanganu et al., 2014; Foo et al., 2017) (Figure 1C). Such wide variability may reflect sensitivity differences across grey matter volume estimation methods [voxel-based morphometry (Song et al., 2011; Melzer et al., 2012; Mak et al., 2013), cortical surface-based analyses (Pagonabarraga et al., 2012; Hanganu et al., 2014; Pereira et al., 2014) and region of interest analyses (Choi et al., 2012)], as well as varying definitions of PD-MCI. Studies conducted prior to recent 2012 guidelines for definitions of PD-MCI (Litvan et al., 2012) used differing selection criteria (Song et al., 2011; Pagonabarraga et al., 2013) to studies using the new criteria (Mak et al., 2013; Foo et al., 2017).

Specific cognitive tests also differ between studies (Mak et al., 2013, 2015; Pereira et al., 2014), as well as the weight given to each cognitive domain. Overall, there is lack of evidence to ascertain which cognitive domains are most affected in Parkinson's disease. Despite methodological heterogeneities, there is some consistency in regions where atrophy correlates with cognitive involvement in Parkinson's disease. For example, precuneus (Pereira et al., 2014; Segura et al., 2014) and lingual gyrus (Pagonabarraga et al., 2013; Segura et al., 2014) thinning correlates with loss of semantic fluency and visuospatial performance (Figure 1D), while temporal thinning correlates with memory (Mak et al., 2013; Pagonabarraga et al., 2013; Pereira et al., 2014).

Longitudinal grey matter studies are similarly affected by methodological discrepancies, including different assumptions for the correction of serial data and power, and few studies include large numbers. Despite this, some consistency is emerging. For example, hippocampal thinning is prominent in several studies (Aybek et al., 2009; Weintraub et al., 2012; Morales et al., 2013; Kandiah et al., 2014; Mak et al., 2015; Gasca-Salas et al., 2017; Gee et al., 2017). Another longitudinal study found frontal and cingulate thinning in patients that progressed to PDD, but it was the combination of biomarkers, including CSF, neuropsychological measures and grey matter volume that was most predictive for Parkinson's dementia (Compta et al., 2013). Ultimately, large prospective studies will be needed to determine the earliest neuroimaging correlates of Parkinson's dementia. This will be most effectively achieved by large-scale collaboration programmes such as the Parkinson's Progression Markers Initiative (PPMI, http://www.ppmi-info.org/).

Variations in findings also likely reflect the low sensitivity of grey matter atrophy as a neural correlate of cognitive involvement in Parkinson's disease. Neuronal cell death, indexed by grey matter atrophy (Rossor et al., 1997), is a relatively late event in the pathogenesis of Parkinson's dementia (Kurowska et al., 2016). Axonal and synaptic accumulation of pathogenic proteins occurs at an earlier stage, before neuronal loss (Hattori et al., 2013). Therefore, neuroimaging techniques sensitive to changes in axonal microstructure might be better suited to detect the earliest stages of cognitive involvement in Parkinson's disease.

processing....