Abstract #1780

Simonetti A, Van Huffel S, Laudadio T, Heerschap A, De Vos M

A fast, accurate and robust method is developed for nosologic imaging of the brain. To incorporate both spatial information and prior knowledge, the method is based on canonical correlation analysis (CCA). We apply CCA two times: once to detect the tumour type and once to show the heterogeneity of the tissue. We also show how the result on MRSI spectra can be improved by using MRSI and MRI information together. The final result gives a reliable segmentation and classification of the segments.