Abstract
Automatic brain-lesion segmentation has the potential to greatly expand the analysis of the relationships between brain function and lesion locations in large-scale epidemiologic studies, such as the ACCORD-MIND study. In this manuscript we describe the design and evaluation of a Bayesian lesion-segmentation method, with the expectation that our approach would segment white-matter brain lesions in MR images without user intervention. Each ACCORD-MIND subject has T1-weighted, T2-weighted, spin-density-weighted, and FLAIR sequences. The training portion of our algorithm first registers training images to a standard coordinate space; then, it collects statistics that capture signal-intensity information, and residual spatial variability of normal structures and lesions. The classification portion of our algorithm then uses these statistics to segment lesions in images from new subjects, without the need for user intervention. We evaluated this algorithm using 42 subjects with primarily white-matter lesions from the ACCORD-MIND project. Our experiments demonstrated high classification accuracy, using an expert neuroradiologist as a standard. A Bayesian lesion-segmentation algorithm that collects multi-channel signal-intensity and spatial information from MR images of the brain shows potential for accurately segmenting brain lesions in images obtained from subjects not used in training.
Highlights
White-matter lesions (WMLs) are common findings in magnetic-resonance (MR) brain examinations of elderly subjects [1,2,3]; they are usually caused by diseases, such as hypertension and diabetes [4]
We propose a novel segmentation algorithm based on Bayesian methods for combining multivariate signal-intensity and spatial information
Note that some subjects have much higher similarity index (SI) values than others; as described previously [11,15], larger total lesion volumes tend to result in larger SI values, as seen in Fig. 6, which shows the relationship between SI and the relative lesion volume of each subject, which we define as the fraction of total brain volume occupied by manually segmented lesions
Summary
White-matter lesions (WMLs) are common findings in magnetic-resonance (MR) brain examinations of elderly subjects [1,2,3]; they are usually caused by diseases, such as hypertension and diabetes [4]. On T1-weighted images, lesions appear hypointense or isointense relative to normal brain, and on T2-weighted, spin-density (SD)-weighted, and FLAIR images, these lesions appear hyperintense. Because of this high tissue contrast, largescale epidemiological studies of cardiovascular risk factors have increasingly relied on MR examination to determine the effects of these conditions on the brain. Segmentation results may differ among raters, or even for the same rater at different times; automated techniques promise greater reliability [6,7] Such considerations are relevant to longitudinal studies involving hundreds, or perhaps thousands of subjects, such as the Cardiovascular Health Study [2,5] or the ACCORD-MIND study [8]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
