This study clarified the relationship between the integrity of the cholinergic tract and responsiveness to donepezil. At baseline, responders had significantly lower CHIPS scores than nonresponders (Table 4). The proportion of APOE carriers is higher in responders (42.3% vs. 11.8%). Overall, nonresponders exhibited greater clinical regression in cognitive function than responders (Table 3). We conducted logistic regression analysis to predict responses to AChEI treatment by using scores on the CHIPS, Fazekas scale, and MTA scales as variables. Considering that age, education level, and baseline dementia severity may influence the rate of cognitive decline, adjustments were made for these factors. Our results suggest that the severity of WMHs in the cholinergic pathway helps predict responsiveness to AChEI treatment.
A substantial body of evidence indicates that diffuse WMHs in the brain are associated with the clinical severity of amnestic MCI  and AD , as well as with cognitive decline in these conditions. However, the link between cholinergic modulation and AChEI response remains to be established . Herein, the severity of WMHs in the cholinergic pathways was the only difference between the responder and nonresponder groups. Moreover, adjusting for the effect of global cognitive function (CDR-SB scores), more WMHs in the cholinergic pathways were associated with poorer responsiveness to AChEI treatment. Our results demonstrate that WMHs in the cholinergic pathway play a more decisive role in determining responsiveness to AChEI treatment than do WMHs in general. Similar observations have been presented in the literature, with one study reporting that white matter lesions in the frontal lobe and basal ganglia significantly reduced responsiveness to AChEI treatment . Mounting evidence suggests links between the distribution or location of WMHs in the brain and cognition [20, 22]. One investigation noted that specific white matter tracts in the brain contributed more significantly than others to the conversion of MCI to AD and the progression of AD . Overall, our findings align with those of other studies. Tracking the reduction of specific white matter tracts may be a promising direction to pursue in research on the monitoring of AD progression and responses to AChEI treatment.
The specific mechanism underlying the correlation between WMHs in the brain and the progression of MCI to AD or the progression of AD has yet to be elucidated. Several hypotheses have been advanced. Most studies predominantly regarded WMHs as a presentation of small vessel disease [21,24]. However, WMH is sometimes considered a vascular form of amyloid deposition, not necessarily associated with vascular risk factors such as hypertension or stroke. In one study, the severity and topographic pattern of WMHs was correlated with amyloid load and amyloid distribution in the brain . On the other hand, WMHs could be driven by neurodegeneration. WMHs may damage the subcortical neurological circuit and cause cognitive decline. Herein, the mean age of the nonresponders was slightly higher (79.3 ± 7.1 years vs. 76.6 ± 7.4 years). Nonresponder groups may exhibit more neurodegenerative pathological changes than responder groups. It seems higher CHIPS scores in nonresponders was more likely due to neurodegeneration instead of small vessel disease. Cohort studies employing neuroimaging modalities such as amyloid PET to examine cerebral microbleeds as a marker of diffuse vascular and neurodegenerative brain damage may provide further insight into the pathophysiological implications of WMHs. Some earlier studies predicted responses to AChEI treatment by considering hippocampal size [10, 11]. In one investigation, poor response was associated with younger age at AD onset and more severe hippocampal atrophy, but WMHs did not contribute to the prediction of this response . Notably, because the median follow-up duration in that study was 46.6 months, hippocampal atrophy possibly reflected the clinical progression of AD rather than poor responses to AChEI treatment. Given that hippocampal volume has been established as meaningful predictor of MCI conversion and cognitive decline [10, 26]. However, evidence supporting its role in predicting responses to AChEI is inconsistent (Table 1). Most of our participants had mild dementia (CDR score: 0.5 to 1); in other words, these patients were in a relatively early stage of the clinical course. Therefore, we believe that the integrity of the cholinergic system (in particular the cholinergic pathway) was the primary determinant of responsiveness to AChEI treatment in this group.
As many articles indicated, the APOE4 allele has been correlated with amyloid accumulation and the severity of WMHs in AD . APOE4 is not only the most notable genetic risk factor for AD; it also increases the risk of cardiovascular disease, stroke, and other neurogenetic disorders . The responder group in our study had a higher proportion of APOE4 positivity than the nonresponder group (42.3% vs. 11.8%). APOE4 has been implicated in enhancing the effects of AChEIs . However, a meta-analysis  indicated that APOE4 does not significantly affect responses to AChEI treatment. Whether APOE4 status influences the effects of AChEI treatment may need more work and a larger cohort.
This study has some limitations. First, changes in cognitive function were only evaluated through MMSE and CDR scores. However, responses to AChEI treatment may differ between specific cognitive domains. Prospective studies can compare the difference between specific cognitive domains before and after AChEI treatment. Second, only some of the participants had APOE genotyping results. Thus, an association between APOE and responses to AChEI cannot be established. Third, we examined cholinergic pathways, general WMHs, and hippocampal atrophy by using visual rating scale. Future studies can employ neuroimaging techniques such as diffusion tensor imaging or automated volumetry.