In this work, we propose a novel probabilistic reference-region-based segmentation method to automatically distinguish various pathological tissue regions within soft tissue sarcoma, including high cellularity, high T2 and necrosis. The classification is based on a calculation of the probability that a tumour voxel belongs to a given class using the quantitative diffusion and T2 information when compared to a reference tissue. The probabilistic approach provides a more realistic classification of the complex tumour microenvironment compared to the previous proposed binary classification method.
This abstract and the presentation materials are available to members only; a login is required.