Fully Automatic Segmentation of Brain Tumor Images using Support Vector Machine Classiffication in Combination with Hierarchical Conditional Random Field Regularization

Delineating brain tumor boundaries from magnetic resonance images is an essential task for the analysis of brain cancer. We propose a fully automatic method for brain tissue segmentation, which combines Support Vector Machine classification using multispectral intensities and textures with subsequent hierarchical regularization based on Conditional Random Fields. The CRF regularization introduces spatial constraints to the powerful SVM classification, which assumes voxels to be independent from their neighbors. The approach first separates healthy and tumor tissue before both regions are subclassified into cerebrospinal fluid, white matter, gray matter and necrotic, active, edema region respectively in a novel hierarchical way. The hierarchical approach adds robustness and speed by allowing to apply different levels of regularization at different stages. The method is fast and tailored to standard clinical acquisition protocols. It was assessed on 10 multispectral patient datasets with results outperforming previous methods in terms of segmentation detail and computation times.

A transgenic model for conditional induction and rescue of portal hypertension reveals a role of VEGF-mediated regulation of sinusoidal fenestrations

Portal hypertension (PH) is a common complication and a leading cause of death in patients with chronic liver diseases. PH is underlined by structural and functional derangement of liver sinusoid vessels and its fenestrated endothelium. Because in most clinical settings PH is accompanied by parenchymal injury, it has been difficult to determine the precise role of microvascular perturbations in causing PH. Reasoning that Vascular Endothelial Growth Factor (VEGF) is required to maintain functional integrity of the hepatic microcirculation, we developed a transgenic mouse system for a liver-specific-, reversible VEGF inhibition. The system is based on conditional induction and de-induction of a VEGF decoy receptor that sequesters VEGF and preclude signaling. VEGF blockade results in sinusoidal endothelial cells (SECs) fenestrations closure and in accumulation and transformation of the normally quiescent hepatic stellate cells, i.e. provoking the two processes underlying sinusoidal capillarization. Importantly, sinusoidal capillarization was sufficient to cause PH and its typical sequela, ascites, splenomegaly and venous collateralization without inflicting parenchymal damage or fibrosis. Remarkably, these dramatic phenotypes were fully reversed within few days from lifting-off VEGF blockade and resultant re-opening of SECs' fenestrations. This study not only uncovered an indispensible role for VEGF in maintaining structure and function of mature SECs, but also highlights the vasculo-centric nature of PH pathogenesis. Unprecedented ability to rescue PH and its secondary manifestations via manipulating a single vascular factor may also be harnessed for examining the potential utility of de-capillarization treatment modalities.

Predicting future risk of asthma exacerbations using individual conditional probabilities

Determination of future risk of exacerbations is a key issue in the management of asthma. We previously developed a method to calculate conditional probabilities (π) of future decreases in lung function by using the daily fluctuations in peak expiratory flow (PEF).