Alzheimer's Imaging Consortium.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Computational Neuroscience

Background:

• Manual brain segmentation is the gold standard but is time-consuming and unreliable.
• Automated segmentation pipelines offer faster and more consistent results, but performance varies.
• QyScore® is a novel medical imaging platform for brain segmentation.

Purpose of the Study:

• To compare the performance of QyScore® against state-of-the-art segmentation pipelines (FreeSurfer, FSL, ANTs).
• To evaluate segmentation accuracy and robustness across multiple brain regions.

Main Methods:

• 54 T1-weighted MRI scans were segmented using QyScore® v1.13, FreeSurfer v7.4.1, FSL v6.0.6.2, and ANTs v2.5.
• Segmentations included whole brain grey matter, white matter, hippocampi, amygdalae, brainstem, cerebellum, caudate, putamen, thalamus, globus pallidus, and lateral ventricles.
• Dice Similarity Coefficients (DSC) were calculated against consensus manual segmentations from three expert neuroradiologists.

Main Results:

• QyScore® achieved the highest accuracy for whole brain grey matter, white matter, brainstem, amygdala, putamen, and lateral ventricles.
• FSL showed the best performance for hippocampal segmentation, followed closely by QyScore®.
• QyScore® and ANTs processed all images without preprocessing errors, unlike FreeSurfer and FSL which required parameter optimization for a significant percentage of subjects.

Conclusions:

• QyScore® exhibits excellent segmentation accuracy and robustness, outperforming other leading pipelines in most brain regions.
• Its reliability in processing diverse medical images makes it suitable for clinical diagnosis and monitoring in healthcare and clinical trials.