Content-based Image Retrieval of Glioblastoma Multiforme

We propose a Content Based Image Retrieval system, for patients diagnosed with Glioblastoma Multiforme that will predict of time to survival and allow a neuroradiologist to modify treatment procedures based on quantified features explicitly extracted from segmented regions of the tumor. Our proposed system has two components: a preprocessing scheme to improve the image quality and provide consistency and a multivariate linear model. The multivariate linear model applied to the training data had a correlation coefficient of 0.848, which indicated a strong association of the selected features to time to survival. Future work will involve expanding the training set and incorporating additional features not explicitly extracted from the segmented tumor regions.