Effect of image standardization on FLAIR MRI for brain extraction

Fluid attenuation inversion recovery (FLAIR) magnetic resonance images (MRI) are being used by physicians to identify and analyze white matter lesions in the brain to determine whether patients are at risk of stroke. Pipelines used to analyze these images require a preprocessing step of brain extraction in order to be robust and to be applied to multicenter, large-scale studies. This paper proposes a novel brain extraction tool solely for the FLAIR modality, as well as a robust standardization pipeline that eliminates variability between datasets by reducing image noise, intensity inhomogeneity, patient movement, and the nonstandardness of tissue intensities, which are inherent in MRI. Feature extraction is performed on the standardized dataset, and a brain segmentation is produced by a random forest classifier. The effects of the standardization steps are evaluated using objective metrics, and the resultant segmentations produced by the unstandardized and standardized images are compared. By implementing a robust standardization pipeline, images acquired from different scanners at different centers can be processed automatically and accurately, allowing for the fast processing of large-scale, multicenter data.

[1]  Paul M. Thompson,et al.  Random forest classification of depression status based on subcortical brain morphometry following electroconvulsive therapy , 2015, 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI).

[2]  Erik Reinhard,et al.  Color Transfer between Images , 2001, IEEE Computer Graphics and Applications.

[3]  April Khademi,et al.  Robust White Matter Lesion Segmentation in FLAIR MRI , 2012, IEEE Transactions on Biomedical Engineering.

[4]  Ying Wang,et al.  Automated White Matter Hyperintensity Detection in Multiple Sclerosis Using 3D T2 FLAIR , 2014, Int. J. Biomed. Imaging.

[5]  Paul M. Thompson,et al.  Robust Brain Extraction Across Datasets and Comparison With Publicly Available Methods , 2011, IEEE Transactions on Medical Imaging.

[6]  Hui Tang,et al.  Tumor segmentation from single contrast MR images of human brain , 2015, 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI).

[7]  L G Nyúl,et al.  On standardizing the MR image intensity scale , 1999, Magnetic resonance in medicine.

[8]  J Debus,et al.  Fast fluid‐attenuated inversion‐recovery (FLAIR) MRI in the assessment of intraaxial brain tumors , 1998, Journal of magnetic resonance imaging : JMRI.

[9]  Qidan Zhu,et al.  Exploration and improvement of Ostu threshold segmentation algorithm , 2010, 2010 8th World Congress on Intelligent Control and Automation.

[10]  M. Beauchemin,et al.  On the Hausdorff Distance Used for the Evaluation of Segmentation Results , 1998 .

[11]  Zhenhua Li,et al.  Registration of the Cone Beam CT and Blue-Ray Scanned Dental Model Based on the Improved ICP Algorithm , 2014, Int. J. Biomed. Imaging.

[12]  Anant Madabhushi,et al.  Interplay between bias field correction, intensity standardization, and noise filtering for T2-weighted MRI , 2011, 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[13]  Stephen M Smith,et al.  Fast robust automated brain extraction , 2002, Human brain mapping.

[14]  Akinobu Shimizu,et al.  Medical image analysis of 3D CT images based on extension of Haralick texture features , 2008, Comput. Medical Imaging Graph..

[15]  Jin Young Kim,et al.  White Matter Hyperintensities Extraction Based T2-FLAIR MRI Using Non-Local Means Filter and Nearest Neighbor Algorithm , 2014, 2014 International Conference on IT Convergence and Security (ICITCS).