MAVEN: An Algorithm for Multi-Parametric Automated Segmentation of Brain Veins From Gradient Echo Acquisitions.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Neuroscience

Background:

• Cerebral vein analysis offers insights into neurodegenerative disorders and traumatic brain injuries.
• Manual segmentation of vascular anatomy is time-consuming and observer-dependent.
• Automated methods are needed for improved reproducibility and efficiency.

Purpose of the Study:

• To propose a novel, fully automated algorithm for segmenting the entire cerebral venous system from MR images.
• To enhance the accuracy and reproducibility of cerebral vein analysis.

Main Methods:

• Development of the multi-parametric automated segmentation of brain VEiNs (MAVEN) algorithm.
• Utilizing structural, morphological, and relaxometric information for segmentation.
• Testing on gradient echo brain data sets acquired at 1.5, 3, and 7 T.

Main Results:

• MAVEN demonstrated high accuracy and reproducibility in segmenting the cerebral venous system.
• The algorithm successfully rejected false positives and detected thin vessels.
• Performance was consistent across different magnetic field strengths (1.5, 3, and 7 T) without parameter adjustments.

Conclusions:

• MAVEN outperforms previous methods in both quantitative and qualitative analyses.
• The algorithm is a promising tool for characterizing venous tree topology.
• MAVEN offers a reliable and efficient approach for cerebral vein analysis in clinical research.