New multispectral MRI data fusion technique for white matter lesion segmentation: method and comparison with thresholding in FLAIR images

Abstract

ObjectiveBrain tissue segmentation by conventional threshold-based techniques may have limited accuracy and repeatability in older subjects. We present a new multispectral magnetic resonance (MR) image analysis approach for segmenting normal and abnormal brain tissue, including white matter lesions (WMLs).MethodsWe modulated two 1.5T MR sequences in the red/green colour space and calculated the tissue volumes using minimum variance quantisation. We tested it on 14 subjects, mean age 73.3 ± 10 years, representing the full range of WMLs and atrophy. We compared the results of WML segmentation with those using FLAIR-derived thresholds, examined the effect of sampling location, WML amount and field inhomogeneities, and tested observer reliability and accuracy.ResultsFLAIR-derived thresholds were significantly affected by the location used to derive the threshold (P = 0.0004) and by WML volume (P = 0.0003), and had higher intra-rater variability than the multispectral technique (mean difference ± SD: 759 ± 733 versus 69 ± 326 voxels respectively). The multispectral technique misclassified 16 times fewer WMLs.ConclusionInitial testing suggests that the multispectral technique is highly reproducible and accurate with the potential to be applied to routinely collected clinical MRI data.Electronic supplementary materialThe online version of this article (doi:10.1007/s00330-010-1718-6) contains supplementary material, which is available to authorized users.

Morphologic, Distributional, Volumetric, and Intensity Characterization of Periventricular Hyperintensities

“…We used coregistered FLAIR and T2*-weighted images and multispectral data fusion to segment the WMH. 21 Combinations of multiple sequences have proved to largely remove the artifactual confound. 30,31 The study also has limitations.…”

“…WMH were segmented semiautomatically on the quantized color image obtained after fusing coregistered FLAIR and T2*-weighted sequences, mapping them in green and red respectively, and applying minimum variance quantization. This technique is described in Valdés Hernández et al, 20 validated elsewhere, 21 and…”

Brain volumetric changes and cognitive ageing during the eighth decade of life

“…Briefly, all sequences were co‐registered and tissue volumes quantified using Gaussian mixture modeling cluster analysis [Valdés Hernández et al, 2010; Wardlaw et al, 2011]. According to Standards for Reporting Vascular changes on nEuroimaging (STRIVE), we explicitly defined WMH as punctate, focal, or diffuse lesions in all subcortical regions [Wardlaw et al, 2013].…”

Interhemispheric characterization of small vessel disease imaging markers after subcortical infarct

“…The RSSI, lacunes, WMH, and old cortical infarcts were all delineated using a multispectral thresholding‐based method (Hernandez, Ferguson, Chappell, & Wardlaw, 2010) implemented on a freely available tool (www.sourceforge.net/projects/bric1936) for their automatic identification, followed by manual editing using the region‐growing thresholding algorithm implemented in Analyze ™ 10.0 (http://www.analyzedirect.com/Analyze/), as per Valdes Hernandez, Armitage, et al. (2015).…”

“…(2015). To differentiate the lacunes and infarcts (new and old) from the surrounding brain on the FLAIR image, we considered a threshold equal to 3.5 SD in pixel intensity, respectively, below or above the mean of the fitted intensity distribution for the normal‐appearing white matter (Hernandez et al., 2010). …”