The purpose of this study was to construct a risk score for glioblastomas based on magnetic resonance imaging (MRI) data. Tumor identification requires multimodal voxel-based imaging data that are highly dimensional, and multivariate models with dimension reduction are desirable for their analysis. We propose a two-step dimension-reduction method using a radial basis function-supervised multi-block sparse principal component analysis (SMS-PCA) method. The method is first implemented through the basis expansion of spatial brain images, and the scores are then reduced through regularized matrix decomposition in order to produce simultaneous data-driven selections of related brain regions supervised by univariate composite scores representing linear combinations of covariates such as age and tumor location. An advantage of the proposed method is that it identifies the associations of brain regions at the voxel level, and supervision is helpful in the interpretation.
CITATION STYLE
Kawaguchi, A. (2017). Supervised dimension-reduction methods for brain tumor image data analysis. In Frontiers of Biostatistical Methods and Applications in Clinical Oncology (pp. 401–411). Springer Singapore. https://doi.org/10.1007/978-981-10-0126-0_24
Mendeley helps you to discover research relevant for your work.