MRI-based prediction of conversion from clinically isolated syndrome to clinically definite multiple sclerosis using SVM and lesion geometry
Neuroanatomical pattern classification using support vector machines (SVMs) has shown promising results in classifying Multiple Sclerosis (MS) patients based on individual structural magnetic resonance images (MRI). To determine whether pattern classification using SVMs facilitates predicting conversion to clinically definite multiple sclerosis (CDMS) from clinically isolated syndrome (CIS). We used baseline MRI data from 364 patients with CIS, randomised to interferon beta-1b or placebo. Non-linear SVMs and 10-fold cross-validation were applied to predict converters/non-converters (175/189) at two years follow-up based on clinical and demographic data, lesion-specific quantitative geometric features and grey-matter-to-whole-brain volume ratios. We applied linear SVM analysis and leave-one-out cross-validation to subgroups of converters (n = 25) and non-converters (n = 44) based on cortical grey matter segmentations. Highest prediction accuracies of 70.4% (p = 8e-5) were reached with a combination of lesion-specific geometric (image-based) and demographic/clinical features. Cortical grey matter was informative for the placebo group (acc.: 64.6%, p = 0.002) but not for the interferon group. Classification based on demographic/clinical covariates only resulted in an accuracy of 56% (p = 0.05). Overall, lesion geometry was more informative in the interferon group, EDSS and sex were more important for the placebo cohort. Alongside standard demographic and clinical measures, both lesion geometry and grey matter based information can aid prediction of conversion to CDMS.
KeywordsClinically isolated syndrome Multiple sclerosis Support vector machine MRI Classification Lesion geometry
Compliance with ethical standards
Conflict of interest
Kerstin Bendfeldt declares that she has no conflict of interest.
Bernd Taschler declares that he has no conflict of interest.
Laura Gaetano declares that she has no conflict of interest.
Philip Madoerin declares that he has no conflict of interest.
Pascal Kuster declares that he has no conflict of interest.
Nicole Mueller-Lenke declares that she has no conflict of interest.
Michael Amann declares that he has no conflict of interest.
Hugo Vrenken declares that he has no conflict of interest.
Viktor Wottschel declares that he has no conflict of interest.
Frederik Barkhof declares that he has no conflict of interest.
Stefan Borgwardt declares that he has no conflict of interest.
Stefan Klöppel declares that he has no conflict of interest.
Eva-Maria Wicklein declares that she has no conflict of interest.
Ludwig Kappos declares that he has no conflict of interest.
Gilles Edan declares that he has no conflict of interest.
Mark S. Freedman declares that he has no conflict of interest.
Xavier Montalbán declares that he has no conflict of interest.
Hans-Peter Hartung declares that he has no conflict of interest.
Christoph Pohl ✝ - no conflict of interest.
Rupert Sandbrink declares that he has no conflict of interest.
Rupert Sandbrink declares that he has no conflict of interest.
Bernd Taschler declares that he has no conflict of interest.
Till Sprenger has received research grants from the Swiss MS Society, Swiss National Research Foundation, EFIC-Grünenthal and Novartis Pharmaceuticals Switzerland. Till Sprengers current and/or previous employers have received compensation for consultation and speaking activities from Mitsubishi Pharma, Eli Lilly, Sanofi Genzyme, Novartis, ATI, Actelion, Electrocore, Biogen Idec, Teva and Allergan.
Ernst-Wilhelm Radue declares that he has no conflict of interest.
Jens Wuerfel declares that he has no conflict of interest.
Thomas E. Nichols declares that he has no conflict of interest.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.
- Arns, C. H., Knackstedt, M. A., Pinczewski, W. V., & Mecke, K. R. (2001). Euler-Poincaré characteristics of classes of disordered media. Physical Review E, 63(3), 031112.Google Scholar
- Bakshi, R., Dandamudi, V. S., Neema, M., De, C., & Bermel, R. A. (2005). Measurement of brain and spinal cord atrophy by magnetic resonance imaging as a tool to monitor multiple sclerosis. [research support, non-U.S. Gov’tReview]. Journal Neuroimaging, 15(4 Suppl), 30S–45S. https://doi.org/10.1177/1051228405283901.Google Scholar
- Barkhof, F., Polman, C. H., Radue, E.-W., Kappos, L., Freedman, M. S., Edan, G., Hartung, H. P., Miller, D. H., Montalbán, X., Poppe, P., de Vos, M., Lasri, F., Bauer, L., Dahms, S., Wagner, K., Pohl, C., & Sandbrink, R. (2007). Magnetic resonance imaging effects of interferon Beta-1b in the BENEFIT study: Integrated 2-year results. Archives of Neurology, 64(9), 1292–1298. https://doi.org/10.1001/archneur.64.9.1292.Google Scholar
- Bergsland, N., Horakova, D., Dwyer, M. G., Dolezal, O., Seidl, Z. K., Vaneckova, M., Krasensky, J., Havrdova, E., & Zivadinov, R. (2012). Subcortical and cortical gray matter atrophy in a large sample of patients with clinically isolated syndrome and early relapsing-remitting multiple sclerosis. American Journal of Neuroradiology, 33, 1573–1578. https://doi.org/10.3174/ajnr.A3086.Google Scholar
- Brex, P. A., Ciccarelli, O., O'Riordan, J. I., Sailer, M., Thompson, A. J., & Miller, D. H. (2002a). A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. [research support, non-U.S. Gov't]. The New England Journal of Medicine, 346(3), 158–164. https://doi.org/10.1056/NEJMoa011341.Google Scholar
- Ceccarelli, A., Rocca, M. A., Pagani, E., Colombo, B., Martinelli, V., Comi, G., & Filippi, M. (2008). A voxel-based morphometry study of grey matter loss in MS patients with different clinical phenotypes. NeuroImage, 42(1), 315–322.Google Scholar
- Dalton, C. M., Chard, D. T., Davies, G. R., Miszkiel, K. A., Altmann, D. R., Fernando, K., et al. (2004). Early development of multiple sclerosis is associated with progressive grey matter atrophy in patients presenting with clinically isolated syndromes. Brain, 127(5), 1101–1107. https://doi.org/10.1093/brain/awh126.Google Scholar
- Filli, L., Hofstetter, L., Kuster, P., Traud, S., Mueller-Lenke, N., Naegelin, Y., Kappos, L., Gass, A., Sprenger, T., Nichols, T. E., Vrenken, H., Barkhof, F., Polman, C., Radue, E. W., Borgwardt, S. J., & Bendfeldt, K. (2012). Spatiotemporal distribution of white matter lesions in relapsing-remitting and secondary progressive multiple sclerosis. Multiple Sclerosis, 18(11), 1577–1584. https://doi.org/10.1177/1352458512442756.Google Scholar
- Fisniku, L. K., Brex, P. A., Altmann, D. R., Miszkiel, K. A., Benton, C. E., Lanyon, R., Thompson, A. J., & Miller, D. H. (2008). Disability and T2 MRI lesions: A 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain, 131(Pt 3), 808–817. https://doi.org/10.1093/brain/awm329.Google Scholar
- Ge, T., Muller-Lenke, N., Bendfeldt, K., Nichols, T. E., & Johnson, T. D. (2014). Analysis of multiple sclerosis lesions via spatially varying coefficients. Annals of Applied Statistics, 8(2), 1095–1118.Google Scholar
- Goldberg-Zimring, D., Achiron, A., Guttmann, C. R. G., & Azhari, H. (2003). Three-dimensional analysis of the geometry of individual multiple sclerosis lesions: Detection of shape changes over time using spherical harmonics. Journal of Magnetic Resonance Imaging, 18(3), 291–301. https://doi.org/10.1002/jmri.10365.Google Scholar
- Gourraud, P. A., Sdika, M., Khankhanian, P., Henry, R. G., Beheshtian, A., Matthews, P. M., Hauser, S. L., Oksenberg, J. R., Pelletier, D., & Baranzini, S. E. (2013). A genome-wide association study of brain lesion distribution in multiple sclerosis. Brain, 136, 1012–1024. https://doi.org/10.1093/brain/aws363.Google Scholar
- Jenkins, T. M., Ciccarelli, O., Atzori, M., Wheeler-Kingshott, C. A. M., Miller, D. H., Thompson, A. J., & Toosy, A. T. (2011). Early pericalcarine atrophy in acute optic neuritis is associated with conversion to multiple sclerosis. Journal of Neurology Neurosurgery and Psychiatry, 82(9), 1017–1021. https://doi.org/10.1136/jnnp.2010.239715.Google Scholar
- Kappos, L., Freedman, M. S., Polman, C. H., Edan, G., Hartung, H. P., Miller, D. H., Montalbán, X., Barkhof, F., Radü, E. W., Bauer, L., Dahms, S., Lanius, V., Pohl, C., & Sandbrink, R. (2007). Effect of early versus delayed interferon beta-1b treatment on disability after a first clinical event suggestive of multiple sclerosis: A 3-year follow-up analysis of the BENEFIT study. Lancet, 370(9585), 389–397.Google Scholar
- Kappos, L., Freedman, M. S., Polman, C. H., Edan, G., Hartung, H. P., Miller, D. H., Montalbán, X., Barkhof, F., Radü, E. W., Metzig, C., Bauer, L., Lanius, V., Sandbrink, R., Pohl, C., & Benefit Study Group. (2009). Long-term effect of early treatment with interferon beta-1b after a first clinical event suggestive of multiple sclerosis: 5-year active treatment extension of the phase 3 BENEFIT trial. Lancet Neurology, 8(11), 987–997. https://doi.org/10.1016/S1474-4422(09)70237-6.Google Scholar
- Kappos, L., Polman, C. H., Freedman, M. S., Edan, G., Hartung, H. P., Miller, D. H., Montalban, X., Barkhof, F., Bauer, L., Jakobs, P., Pohl, C., Sandbrink, R., & for the BENEFIT Study Group. (2006). Treatment with interferon beta-1b delays conversion to clinically definite and McDonald MS in patients with clinically isolated syndromes. Neurology, 67(7), 1242–1249.Google Scholar
- Kelly, M. A., Cavan, D. A., Penny, M. A., Mijovic, C. H., Jenkins, D., Morrissey, S., Miller, D. H., Barnett, A. H., & Francis, D. A. (1993). The influence of HLA-DR and-DQ alleles on progression to multiple sclerosis following a clinically isolated syndrome. Human Immunology, 37(3), 185–191.Google Scholar
- Koutsouleris, N., Patschurek-Kliche, K., Scheuerecker, J., Decker, P., Bottlender, R., Schmitt, G., Rujescu, D., Giegling, I., Gaser, C., Reiser, M., Möller, H. J., & Meisenzahl, E. M. (2010). Neuroanatomical correlates of executive dysfunction in the at-risk mental state for psychosis. [research support, non-U.S. Gov't]. Schizophrenia Research, 123(2–3), 160–174. https://doi.org/10.1016/j.schres.2010.08.026.Google Scholar
- Lang, C., Ohser, J., & Hilfer, R. (2001). On the analysis of spatial binary images. Journal of Microscopy, 203(3), 303–313.Google Scholar
- Legland, D., Kiêu, K., & Devaux, M.-F. (2007). Computation of Minkowski measures on 2D and 3D binary images. Image Analysis & Stereology, 26(2), 83–92.Google Scholar
- Locatelli, L., Zivadinov, R., Grop, A., & Zorzon, M. (2004). Frontal parenchymal atrophy measures in multiple sclerosis. Multiple Sclerosis, 10(5), 562–568.Google Scholar
- Lovblad, K. O., Anzalone, N., Dorfler, A., Essig, M., Hurwitz, B., Kappos, L., et al. (2010). MR imaging in multiple sclerosis: Review and recommendations for current practice. [research support, non-U.S. Gov't. Review]. AJNR. American journal of neuroradiology, 31(6), 983–989. https://doi.org/10.3174/ajnr.A1906.Google Scholar
- McDonald, W. I., Compston, A., Edan, G., Goodkin, D., Hartung, H. P., Lublin, F. D., McFarland, H. F., Paty, D. W., Polman, C. H., Reingold, S. C., Sandberg-Wollheim, M., Sibley, W., Thompson, A., van den Noort, S., Weinshenker, B. Y., & Wolinsky, J. S. (2001). Recommended diagnostic criteria for multiple sclerosis: Guidelines from the international panel on the diagnosis of multiple sclerosis. Annals of Neurology, 50(1), 121–127.Google Scholar
- Morrissey, S. P., Miller, D. H., Kendall, B. E., Kingsley, D. P. E., Kelly, M. A., Francis, D. A., et al. (1993). The significance of brain magnetic-resonance-imaging abnormalities at presentation with clinically isolated syndromes suggestive of multiple-sclerosis - a 5-year follow-up-study. Brain, 116, 135–146. https://doi.org/10.1093/brain/116.1.135.Google Scholar
- Newton, B. D., Wright, K., Winkler, M. D., Bovis, F., Takahashi, M., Dimitrov, I. E., Sormani, M. P., Pinho, M. C., & Okuda, D. T. (2017). Three-dimensional shape and surface features distinguish multiple sclerosis lesions from nonspecific white matter disease. Journal of Neuroimaging, 27(6), 613–619. https://doi.org/10.1111/jon.12449.Google Scholar
- Orru, G., Pettersson-Yeo, W., Marquand, A. F., Sartori, G., & Mechelli, A. (2012). Using support vector machine to identify imaging biomarkers of neurological and psychiatric disease: A critical review. [research support, non-U.S. Gov't. Review]. Neuroscience and biobehavioral reviews, 36(4), 1140–1152. https://doi.org/10.1016/j.neubiorev.2012.01.004.Google Scholar
- Polman, C., Kappos, L., Freedman, M., Edan, G., Hartung, H. P., Miller, D., et al. (2008). Subgroups of the BENEFIT study: Risk of developing MS and treatment effect of interferon beta-1b. Journal of Neurology, 255(4), 480–487.Google Scholar
- Polman, C. H., Reingold, S. C., Edan, G., Filippi, M., Hartung, H.-P., Kappos, L., Lublin, F. D., Metz, L. M., McFarland, H. F., O'Connor, P. W., Sandberg-Wollheim, M., Thompson, A. J., Weinshenker, B. G., & Wolinsky, J. S. (2005). Diagnostic criteria for multiple sclerosis: 2005 revisions to the ldquoMcDonald Criteriardquo. Annals of Neurology, 58(6), 840–846.Google Scholar
- Popescu, B. F. G., Pirko, I., & Lucchinetti, C. F. (2013). Pathology of multiple sclerosis: Where do we stand? CONTINUUM: Lifelong Learning in Neurology, 19(4, multiple sclerosis), 901–921.Google Scholar
- Ramasamy, D. P., Benedict, R. H. B., Cox, J. L., Fritz, D., Abdelrahman, N., Hussein, S., Minagar, A., Dwyer, M. G., & Zivadinov, R. (2009). Extent of cerebellum, subcortical and cortical atrophy in patients with MS: A case-control study. Journal of the Neurological Sciences, 282(1–2), 47–54.Google Scholar
- Richert, N. D., Howard, T., Frank, J. A., Stone, R., Ostuni, J., Ohayon, J., Bash, C., & McFarland, H. F. (2006). Relationship between inflammatory lesions and cerebral atrophy in multiple sclerosis. [research support, N.I.H., intramural]. Neurology, 66(4), 551–556. https://doi.org/10.1212/01.wnl.0000197982.78063.06.Google Scholar
- Schölkopf, B., Smola, A.J. (2001) Learning with Kernels. MIT Press.Google Scholar
- Tintore, M., Rovira, A., Brieva, L., Grive, E., Jardi, R., Borras, C., et al. (2001). Isolated demyelinating syndromes: Comparison of CSF oligoclonal bands and different MR imaging criteria to predict conversion to CDMS. Multiple Sclerosis, 7(6), 359–363. https://doi.org/10.1177/135245850100700603.Google Scholar
- Tintore, M., Rovira, A., Rio, J., Nos, C., Grive, E., Tellez, N., et al. (2006). Baseline MRI predicts future attacks and disability in clinically isolated syndromes. [comparative study Research Support, Non-U.S. Gov't]. Neurology, 67(6), 968–972. https://doi.org/10.1212/01.wnl.0000237354.10144.ec.Google Scholar
- Tintore, M., Rovira, A., Rio, J., Tur, C., Pelayo, R., Nos, C., et al. (2008). Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurology, 70(13 part 2), 1079–1083.Google Scholar
- Wei, X., Guttmann, C. R., Warfield, S. K., Eliasziw, M., & Mitchell, J. R. (2004). Has your patient's multiple sclerosis lesion burden or brain atrophy actually changed? [research support, Non-U.S. Gov't Research Support, U.S. Gov't, P.H.S.]. Mult Scler, 10(4), 402–406.Google Scholar
- Young, J., Modat, M., Cardoso, M. J., Mendelson, A., Cash, D., Ourselin, S., Alzheimer’s Disease Neuroimaging Initiative. (2013). Accurate multimodal probabilistic prediction of conversion to Alzheimer's disease in patients with mild cognitive impairment. Neuroimage Clin, 2, 735–745, https://doi.org/10.1016/j.nicl.2013.05.004.
- Zivadinov, R., Grop, A., Sharma, J., Bratina, A., Tjoa, C. W., Dwyer, M., & Zorzon, M. (2005). Reproducibility and accuracy of quantitative magnetic resonance imaging techniques of whole-brain atrophy measurement in multiple sclerosis. Journal of Neuroimaging, 15(1), 27–36. https://doi.org/10.1177/1051228404271010.Google Scholar