Fractional Anisotropy is a More Sensitive Diagnostic Biomarker Than Mean Kurtosis for Patients with Parkinson Disease with Cognitive Dysfunction: A Diffusional Kurtosis Map Tract-Based Spatial Statistics Study

DISCUSSION

This study detected the white matter microstructural change between patients with PD-MCI and PDD using the DKI sequence and voxel-based Tract-Based Spatial Statistics analyses. The diagnostic efficiency of DKI metrics was calculated. The correlation between the DKI metrics and cognitive tests was also explored. The main results of this work are as follows: 1) The FA, MK, and MD were all reduced in patients with PDD compared with those with PD-MCI; 2) the diagnostic efficiency between the PD-MCI and PDD groups was great AUC > 0.8 in cluster 1 (forceps minor) and cluster 2 (inferior fronto-occipital fasciculus and inferior longitudinal fasciculus) of the FA map, and the combination applies of FA and MK clusters performed better AUC > 0.9; 3) the mean FA values in cluster 1 (forceps minor) were significantly correlated with the MoCA score, representing the global cognitive status. According to the results, the FA and MK derived from the diffusion kurtosis estimation model were more sensitive to cognitive-related microstructure changes than MD.

Many studies have detected the neuroanatomic substrates of white matter in patients with early PD with cognitive decline using diffusion imaging, mostly DTI. The primary metrics of DTI are FA and MD. Decreased FA represents the break of axonal integrity, axonal loss, and demyelination. While the increased FA has been suggested to be related to the compensatory responses or excitatory reactions.⁸ In comparison, the FA value cannot reveal the white matter damage of neurodegenerative disease well. In previous studies, the FA values of patients with PD with cognitive decline varied a lot, and the conclusions were inconsistent. A study involving 109 patients with PD showed that FA derived from DTI did not differ significantly between the PD-MCI and PDD groups.¹⁸ Another study using voxel-based analysis demonstrated that compared with controls, the FA values in patients with PDD was significantly lower in the bilateral frontal, left temporal, and left parietal white matter.¹⁹ In a large longitudinal study (The Incidence of Cognitive Impairment in Cohorts with Longitudinal Evaluation-PD, ICICLE-PD), no FA difference was reported at baseline or follow-up between healthy controls and patients with PD with normal cognition or MCI.²⁰ Thus, we attempted to reflect better the white matter diffusion changes with DKI in the present study. There were 4 independent clusters of significantly decreased FA areas of the PDD group compared with the PD-MCI group. Three of these clusters had good diagnostic efficiency. It suggests that FA derived from DKI could be a sensitive radiomic marker to detect the microstructure damage of white matter.

In addition, in this study, univariate and multiple linear regression analyses showed that education and mean FA values in the forceps minor were influential factors in the cognitive status of patients with PD. FA in the forceps minor positively correlated with MoCA. Patients with PDD have a more severe FA decrease in the minor forceps of the corpus callosum, compared with the PD-MCI group. The corpus callosum, the largest bundle of commissural fibers, is located at the base of the longitudinal fissure of the brain and is broad, thick white matter composed of transverse fibers connecting the left and right cerebral hemispheres. After entering both hemispheres, the corpus callosum fibers spread out and project to the entire hemispherical cortex. They connect the corresponding parts of the 2 hemispheres, making the brain a functional whole. They play an essential role in coordinating activities between the 2 hemispheres. The fibers passing through the genu of the corpus callosum curve forward into the frontal lobes on each side and are called minor forceps. An ROI-based study found that FA values of the right parietal and left occipital lobes were significantly lower in the PD-MCI group, compared with the healthy controls.²¹ The posterior cingulate cortex–right medial temporal lobe functional connectivity strength of PD significantly correlates with the MoCA score; the PDD group shows a decreased functional connectivity of the posterior cingulate cortex right parahippocampal gyrus compared with normal cognition in the PD group, while PDD group shows a widespread decreased posterior cingulate cortex functional connectivity compared with healthy controls.²² The corpus collosum white matter damage break of axonal integrity, axonal loss, and demyelination induce cognitive dysfunction with different clinical manifestations, mainly because the damaged fibers are projected to different functional cortices.

MK reflects differences in the diffusion of both intracellular and extracellular compartments and indirectly reacts with the non-Gaussian distribution of water molecules or tissue complexity, which is generally higher for more complex or heterogeneous microstructures.²³ A voxelwise DKI study of cortical gray matter showed reduced MK in the frontal, temporal, limbic, and paralimbic regions and parts of the parietal lobe in the PD group, compared with health controls; meanwhile, the ROI analysis found a significant MK decrease in the anterior division of the left superior temporal gyrus and the left frontal operculum cortex.²⁴ In this study, compared with the PD-MCI group, the MK of the PDD group significantly decreased in white matter regions located in the forceps major, inferior fronto-occipital fasciculus, inferior longitudinal fasciculus, and superior longitudinal fasciculus. These white matter fibers are connected to the cortex where MK decreases.^23⇓-25 It suggested that the decrease of MK increases with the severity of cognitive status. The brain tissue complexity of the PD-MCI group was more inclined to non-Gaussian distribution than that of the PDD group.

As reported, the characteristic white matter abnormalities associated with cognitive impairment in PD predominantly appear as the diffusivity increases but not FA.⁸ It did not verify in the present study. The DKI-derived MD increased slightly more than 50,000 voxels in the forceps minor, suggesting more severe cellular damages and cell necrosis in the PD-MCI group. The present study had no healthy control group, and the patients’ ages at PD-MCI and PDD diagnoses were significantly different due to the natural history of disease progression. It cannot, therefore, be inferred that the decrease in MD in the PDD group was due to disease progression. DKI studies on healthy controls and patients with PD-MCI and PDD with larger cohorts were needed.

Pathologic studies of PD suggest that α-synuclein accumulates in the whole brain following a topographical sequence, beginning in the olfactory bulb, followed by the caudal brainstem, substantia nigra, limbic cortex, and neocortex.^7,26⇓-28 A longitudinal DTI study of patients with early PD with 18 months of follow-up showed that the accelerated white matter change in the rostral brainstem and subcortical regions fit with this theory of topographic disease progression.²⁹ In patients with PD-MCI and PDD, the Lewy-related pathology gradually extends from the brainstem to the cortex through white matter fibers. Except for this, Alzheimer disease-related pathology also detected, approximately 50% of patients with PDD have β-amyloid plaques in the cortex, and two-thirds of these patients also have phosphorylated τ deposition in cortical tangles.⁷ The β-amyloid plaques and τ enhance the complexity of brain tissue and are associated with cognitive decline in patients with PD. A study showed that τ deposition facilitates α-synuclein aggregation and propagation in Parkinson disease.³⁰ Studies have shown that one-third of patients with PDD with τ pathology will show more severe or faster disease progression. Most interesting, there is a longitudinal DTI study of patients with PD compared with controls. It finds significant gray and white matter DTI alterations of other cortical and white matter regions at the baseline (diagnosis of PD) and 1-year-follow-up, suggesting that the increased spatial spread of α-synuclein neuropathology is the crucial mechanism of PD progression.³¹ As for the present cross-sectional study, the significant difference in both diffusivity and kurtosis changes of white matter in patients with PD-MCI and PDD may reveal that the α-synuclein pathology inside the white matter differed. The reasons would be the following: 1) different regions of α-synuclein pathology invasion, 2) whether accompanied by β-amyloid plaques and/or τ pathology, 3) remyelination or neuroinflammation, and 4) genetic variations.

In the present work, compared with the PD-MCI group, the decrease of FA values in the PDD group suggests that the integrity of white matter is extensively damaged, and it is positively correlated with the cognitive status of patients. The decrease in MK values in the PDD group suggests the existence of a relatively Gaussian distribution in the white matter. The decreased MD in PDD implies less neuron edema of the white matter. The cognitive impairment spectrum of PD consists of executive function, language, attention, memory, and visuospatial skills. The previous study has revealed that different patterns of white matter diffusivity underlie impairment in distinct cognitive domains of patients with PD.³² Here, we demonstrated different patterns of white matter damage between patients with PD-MCI and PDD. For a further study, longer longitudinal follow-up of patients with PD-MCI to see if there is a difference between groups who do or do not progress to PDD should considered.

Several limitations of the study should be acknowledged. First, the small sample size of this study may limit the generalizability of the results. The study may fail to reach statistical power; therefore, the findings should not be considered as definitive evidence. More research with larger samples is still needed. Second, we included patients with PD-MCI and PDD, and none of the control group or patients with PD with normal cognitive function were included; therefore, the generalization of our findings should be considered with caution. Third, the ages of patients in the PD-MCI group in this study did not match those of patients in the PDD group. With the development of the natural history of the disease, some patients with PD-MCI gradually convert to PDD, so the age of patients with PD-MCI is usually younger than that of patients with PDD. Also, the cognitive scores were not measured until just before this admission. Fourth, the study assessed only the global cognitive status. DKI detection of each aspect of the spectrum of cognitive impairment still needs further study. Fifth, the study is limited by drug treatment heterogeneity and potential confounding factors, such as biochemical indicators.