Automatic Brain and Tumor Segmentation

Abstract

Abstract. Combining image segmentation based on statistical classification with a geometric prior has been shown to significantly increase robustness and reproducibility. Using a probabilistic geometric model of sought structures and image registration serves both initialization of probability density functions and definition of spatial constraints. A strong spatial prior, however, prevents segmentation of structures that are not part of the model. In practical applications, we encounter either the presentation of new objects that cannot be modeled with a spatial prior or regional intensity changes of existing structures not explained by the model. Our driving application is the segmentation of brain tissue and tumors from three-dimensional magnetic resonance imaging (MRI). Our goal is a high-quality segmentation of healthy tissue and a precise delineation of tumor boundaries. We present an extension to an existing expectation maximization (EM) segmentation algorithm that modifies a probabilistic brain atlas with an individual subject's information about tumor location obtained from subtraction of post-and pre-contrast MRI. The new method handles various types of pathology, space-occupying mass tumors and infiltrating changes like edema. Preliminary results on five cases presenting tumor types with very different characteristics demonstrate the potential of the new technique for clinical routine use for planning and monitoring in neurosurgery, radiation oncology, and radiology.

Automatic brain tumor segmentation by subject specific modification of atlas priors1

“…This method builds on the previously published work done by Van Leemput, et al (13,14). Additionally, tumor and edema classes are added to the segmentation as was done by Moon et al (16). The spatial atlas that is used as a prior in the classification is modified to include prior probabilities for tumor and edema.…”

Non-rigid Atlas to Subject Registration with Pathologies for Conformal Brain Radiotherapy

“…An alternative is to consider tumor intensities as outliers in this mixture of Gaussians, or to add some specific classes to model the tumor and edema intensities [9]. As this was often not sufficient, some anatomical knowledge was added, either by combining geometric priors given by the non-rigid registration of an atlas to a tissue classification [10], or by using Markov Random Fields [11].…”

Multi-class Posterior Atlas Formation via Unbiased Kullback-Leibler Template Estimation

“…Again, this method is independent of the choice of the number of images comprising each image set. These class posterior densities are produced using the expectation maximization method described in [14,15]. Following [15,16], for each class c j , the associated data likelihood, p(Ī i (x)|c j (x), µ j , Σ j ), is modeled as a normal distribution with mean, µ j , and covariance, Σ j .…”