Contributions
Abstract: P575
Type: Poster
Abstract Category: Pathology and pathogenesis of MS - 21 Imaging
Background: Central vein sign (CVS) is a promising diagnostic biomarker for multiple sclerosis (MS), but its use is limited by the potential for inter-rater differences in the judgment of CVS, and the time and effort required to adjudicate CVS for patients with heavy lesion loads. The goal of this study was to develop an automated technique for the detection of CVS in subcortical and deep white matter lesions.
Methods: Magnetic resonance imaging (MRI) was performed on 39 patients: ten had MS and no comorbidities for MRI white matter abnormalities, ten had MS and comorbidities for MRI white matter abnormalities, ten had migraine with an MRI showing white matter abnormalities and no additional comorbidities, and nine were previously incorrectly diagnosed with MS. 3D T1-weighted, T2-weighted fluid attenuated inversion recovery (FLAIR), and high resolution segmented echo-planar imaging T2*-weighted sequences were acquired. A multi-stage algorithm was used to detect veins, segment white matter lesions, partition confluent lesions, remove periventricular lesions, and determine the centrality of veins inside candidate lesions.
Validation: Lesion segmentation and vein quantification techniques were calibrated using data from one participant with MS and no comorbidities and one participant with migraine. These participants were then removed from the sample, and differences in the proportion of lesions with CVS were studied among the remaining 37 subjects.
Results: Among the lesions segmented (n = 381), CVS was identified in a greater proportion of the lesions from all participants with MS compared to lesions from all participants without MS (propms+msc = 65%, propmig+mis = 53%, p < .03). Additionally, within-person proportions of CVS were higher in participants with MS compared to participants who were misdiagnosed with MS (meanms+msc = 70%, meanmis = 52%, p < .04).
Conclusion: This study introduces an automated method for detecting central vein sign in white matter lesions and provides preliminary evidence for its validity and potential diagnostic utility. After further study and refinement, the ability to detect the presence of this biomarker using automated methodology could prove instrumental for future clinical application.
Disclosure:
Jordan D. Dworkin: Nothing to disclose
Andrew J. Solomon: Consulting and advisory boards for Biogen Idec, EMD Serono, Teva Pharmaceuticals, and Genentech
Pascal Sati: Nothing to disclose
Dzung Pham: Nothing to disclose
Richard Watts: Nothing to disclose
Matthew K. Schindler: Nothing to disclose
Daniel Ontaneda: Grant support from Genzyme, Genentech, Novartis, consulting from Biogen Idec and Genentech
Daniel S. Reich: Research funding from Vertex Pharmaceuticals and the Myelin Repair Foundation
Russell T. Shinohara: Consulting and advisory boards for Roche and Genentech
Abstract: P575
Type: Poster
Abstract Category: Pathology and pathogenesis of MS - 21 Imaging
Background: Central vein sign (CVS) is a promising diagnostic biomarker for multiple sclerosis (MS), but its use is limited by the potential for inter-rater differences in the judgment of CVS, and the time and effort required to adjudicate CVS for patients with heavy lesion loads. The goal of this study was to develop an automated technique for the detection of CVS in subcortical and deep white matter lesions.
Methods: Magnetic resonance imaging (MRI) was performed on 39 patients: ten had MS and no comorbidities for MRI white matter abnormalities, ten had MS and comorbidities for MRI white matter abnormalities, ten had migraine with an MRI showing white matter abnormalities and no additional comorbidities, and nine were previously incorrectly diagnosed with MS. 3D T1-weighted, T2-weighted fluid attenuated inversion recovery (FLAIR), and high resolution segmented echo-planar imaging T2*-weighted sequences were acquired. A multi-stage algorithm was used to detect veins, segment white matter lesions, partition confluent lesions, remove periventricular lesions, and determine the centrality of veins inside candidate lesions.
Validation: Lesion segmentation and vein quantification techniques were calibrated using data from one participant with MS and no comorbidities and one participant with migraine. These participants were then removed from the sample, and differences in the proportion of lesions with CVS were studied among the remaining 37 subjects.
Results: Among the lesions segmented (n = 381), CVS was identified in a greater proportion of the lesions from all participants with MS compared to lesions from all participants without MS (propms+msc = 65%, propmig+mis = 53%, p < .03). Additionally, within-person proportions of CVS were higher in participants with MS compared to participants who were misdiagnosed with MS (meanms+msc = 70%, meanmis = 52%, p < .04).
Conclusion: This study introduces an automated method for detecting central vein sign in white matter lesions and provides preliminary evidence for its validity and potential diagnostic utility. After further study and refinement, the ability to detect the presence of this biomarker using automated methodology could prove instrumental for future clinical application.
Disclosure:
Jordan D. Dworkin: Nothing to disclose
Andrew J. Solomon: Consulting and advisory boards for Biogen Idec, EMD Serono, Teva Pharmaceuticals, and Genentech
Pascal Sati: Nothing to disclose
Dzung Pham: Nothing to disclose
Richard Watts: Nothing to disclose
Matthew K. Schindler: Nothing to disclose
Daniel Ontaneda: Grant support from Genzyme, Genentech, Novartis, consulting from Biogen Idec and Genentech
Daniel S. Reich: Research funding from Vertex Pharmaceuticals and the Myelin Repair Foundation
Russell T. Shinohara: Consulting and advisory boards for Roche and Genentech