Perivascular spaces relate to the course and cognition of Huntington’s disease

Xiaoyan Li, Juan‐Juan Xie, Jinhong Wang, Yu‐Feng Bao, Yi Dong, Bin Gao, Ting Shen, Peiyu Huang, Haochao Ying, Xu Han, Anna Wang Roe, Hsin‐Yi Lai, Zhi‐Ying WuTranslational 2023 · DOI · OpenAlex

This study, published in Translational Neurodegeneration, investigates the role of the brain’s "waste clearance" system in the progression of Huntington’s disease (HD). By leveraging ultra-high-field 7.0 Tesla MRI and deep learning, researchers discovered that enlarged perivascular spaces in the basal ganglia serve as an early biomarker for the disease, appearing even in the pre-symptomatic stage and correlating strongly with cognitive decline and brain atrophy.

TL;DR

Researchers used 7.0T MRI and deep learning to show that enlarged perivascular spaces (PVS) in the basal ganglia are early indicators of Huntington’s disease, present even before symptoms appear. These structural changes are directly linked to the severity of cognitive impairment and the shrinking of key brain regions like the caudate and putamen.

What Problem Does This Paper Address?

Huntington’s disease (HD) is a devastating genetic disorder characterized by motor, cognitive, and psychiatric symptoms. While we know it is caused by a specific genetic mutation (CAG expansion in the HTT gene), we are still mapping out the early physiological changes that occur before the onset of motor symptoms.

The authors focused on perivascular spaces (PVS)—fluid-filled channels that surround cerebral blood vessels. These spaces are critical components of the glymphatic system, the brain’s "sewer system" responsible for draining interstitial fluid and clearing metabolic waste. When PVS become enlarged and visible on MRI, it typically signals glymphatic dysfunction. While previous studies noted PVS changes in patients with manifest HD, this paper addresses a critical gap: Do these changes occur in the pre-manifest (pre-HD) stage, and do they actually relate to a patient's cognitive performance?

Methods

The study employed several high-tech approaches to quantify PVS with high precision:

  • 7.0 Tesla MRI: Most clinical MRIs use 1.5T or 3.0T magnets. The 7.0T "ultra-high-field" strength provides significantly higher resolution and signal-to-noise ratios, allowing for the visualization of tiny PVS (diameter < 3mm) that might be missed at lower strengths.
  • Deep Learning Segmentation: Manually counting PVS is slow and prone to human error. The team used a U-shaped network (U-net), a class of deep learning architecture, to automatically segment and quantify PVS volume across the brain.
  • Cognitive Battery: Participants underwent a series of tests focusing on executive function, including the Symbol Digit Modality Test (SDMT) and various Stroop tests.
  • Statistical Modeling: The researchers used multiple linear regression and partial correlation analysis to control for the effect of age, ensuring that the observed PVS changes were due to the disease rather than just getting older.

Datasets

The study involved a cohort of 106 participants:

  • 49 Healthy Controls
  • 32 Pre-HD individuals (gene carriers without motor symptoms)
  • 25 Manifest HD patients

For the U-net model, a training subset of 30 subjects was used to refine the segmentation algorithm before applying it to the full dataset.

Key Insights and Ideas

  • Regional Specificity Matters: While "global" brain PVS didn't show a strong link to the disease, PVS specifically in the Basal Ganglia (BG)—the area of the brain most affected by HD—was highly significant. This suggests that glymphatic failure in HD is localized to the regions undergoing the most intense neurodegeneration.
  • An Early Warning Sign: The discovery that BG-pPVS (Basal Ganglia PVS volume proportion) is elevated in pre-HD individuals is a major "aha moment." It suggests that waste clearance issues begin long before the "triad" of HD symptoms (motor, cognitive, psychiatric) becomes clinically apparent.
  • The "Waste-Atrophy" Cycle: The findings suggest a potential feedback loop: as the brain's waste clearance system fails (enlarged PVS), toxic proteins may accumulate more easily, leading to the death of neurons (atrophy), which in turn might further disrupt the structural integrity of the perivascular pathways.

Key Results

  • Segmentation Accuracy: The U-net performed well, achieving a Dice Similarity Coefficient (DSC) of 0.85 in the training set and 0.76 in the validation set.
  • Disease Progression: After adjusting for age, both pre-HD and manifest HD groups showed significantly higher BG-pPVS than healthy controls ($P=0.001$ and $P<0.0001$, respectively).
  • Brain Atrophy Link: In mutation carriers, higher BG-pPVS was inversely correlated with the volume of the caudate ($r = -0.51$) and putamen ($r = -0.39$). Essentially, more enlarged spaces equaled more brain shrinkage.
  • Cognitive Correlation: Higher BG-pPVS scores were associated with lower scores on the Stroop Word Reading Test ($r = -0.35$) and the Symbol Digit Modality Test ($r = -0.39$), linking structural "plumbing" issues to actual mental performance.

Limitations and Future Work

The authors acknowledge several constraints:

  1. Narrow Cognitive Scope: The tests used primarily measured executive function. Future studies should use broader batteries (like CANTAB) to assess other cognitive domains.
  2. Selection Bias: Patients in the late stages of HD were excluded because their involuntary movements make high-resolution MRI scanning impossible.
  3. Cross-sectional Design: This study provides a "snapshot" in time. Longitudinal studies are needed to confirm if PVS enlargement can predict the rate of future decline.
  4. Functional vs. Structural: While enlarged PVS suggest poor drainage, they don't measure flow directly. The authors suggest using Diffusion Tensor Imaging (DTI) along the perivascular space in future work to more accurately evaluate glymphatic function.

Summary updated

BibTeX
@inproceedings{li2023perivascular,
  author = {Xiaoyan Li and Juan‐Juan Xie and Jinhong Wang and Yu‐Feng Bao and Yi Dong and Bin Gao and Ting Shen and Peiyu Huang and Haochao Ying and Xu Han and Anna Wang Roe and Hsin‐Yi Lai and Zhi‐Ying Wu},
  title = {{Perivascular spaces relate to the course and cognition of Huntington’s disease}},
  booktitle = {Translational},
  year = {2023},
  doi = {10.1186/s40035-023-00359-9},
}

Loading related content...

Figure