Perivascular Spaces Segmentation in Brain MRI Using Optimal 3D Filtering

Abstract

Perivascular Spaces (PVS) are a feature of Small Vessel Disease (SVD), and are an important part of the brain’s circulation and glymphatic drainage system. Quantitative analysis of PVS on Magnetic Resonance Images (MRI) is important for understanding their relationship with neurological diseases. In this work, we propose a segmentation technique based on the 3D Frangi filtering for extraction of PVS from MRI. We used ordered logit models and visual rating scales as alternative ground truth for Frangi filter parameter optimization and evaluation. We optimized and validated our proposed models on two independent cohorts, a dementia sample (N = 20) and patients who previously had mild to moderate stroke (N = 48). Results demonstrate the robustness and generalisability of our segmentation method. Segmentation-based PVS burden estimates correlated well with neuroradiological assessments (Spearman’s ρ = 0.74, p < 0.001), supporting the potential of our proposed method.

Intensity‐curvature functional‐based filtering in image space and k‐space: Applications in magnetic resonance imaging of the human brain

“…Consistently, the fourth column from left reports the difference between the mode value of the edge image of the MRI and the mode value of the edge image of the five imaging modalities. The results that were closest to those obtained running the edge finder on the MRI were attained with the following two imaging techniques: De Martino et al, 2018) and the other one is the large body of methodologies that through applied computational neuroimaging had contributed to vessels segmentation (Moccia, De Momi, El Hadji, & Mattos, 2018) and also to perivascular segmentation (Ballerini et al, 2018). The methodologies are currently being developed to the purpose of addressing major human brain diseases and affections in the pathological state.…”

“…Human brain vessel imaging with MRI has been influenced in the past decade by the advent of two main novelties. One is the strength of the magnet used for MRI recording (De Cocker et al., ; De Martino et al., ) and the other one is the large body of methodologies that through applied computational neuroimaging had contributed to vessels segmentation (Moccia, De Momi, El Hadji, & Mattos, ) and also to perivascular segmentation (Ballerini et al., ). The methodologies are currently being developed to the purpose of addressing major human brain diseases and affections in the pathological state.…”

Structural neuroimaging differentiates vulnerability from disease manifestation in colombian families with Huntington's disease

“…Despite recent computational methods for assessing PVS have shown promising for being applied in clinical research and practice (Ballerini et al, 2018), still the gold-standard method for assessing PVS is neuroradiological (i.e., visual) rating. A trained observer (XQ) rated the PVS using the Potter scale (Potter, Chappell, Morris, & Wardlaw, 2015).…”

Quantitative Measurements of Enlarged Perivascular Spaces in the Brain are Associated with Retinal Microvascular Parameters in Older Community-Dwelling Subjects

“…Whilst shown to provide valuable information about PVS in aetiological studies to date, these scales are inherently insensitive to small details due to their limited number of categories, floor and ceiling effects, and may be affected by observer bias 16 . Computational tools for PVS quantification have been developed in the last five years [17][18][19][20] . The method by Ballerini et al 20,21 was able to segment PVS in the centrum semiovale and enabled quantification of several PVS features, including the total count and total volume per individual subject's brain, plus the size, length, width, shape and direction of each individual PVS.…”

“…Computational tools for PVS quantification have been developed in the last five years [17][18][19][20] . The method by Ballerini et al 20,21 was able to segment PVS in the centrum semiovale and enabled quantification of several PVS features, including the total count and total volume per individual subject's brain, plus the size, length, width, shape and direction of each individual PVS. All these can then be analysed as mean or median per individual subject 20 or indeed per brain region.…”