Dynamic contrast enhancement (DCE) is the leading technique in magnetic resonance imaging for cancer detection and
diagnosis. However, there are large variations in the reported sensitivity and specificity of this method that result from
the wide range of contrast-enhanced MRI sequences and protocols, image processing methods, and interpretation
criteria. Analysis methods can be divided to physiological based models that take into account the vascular and tissue
specific features that influence tracer perfusion, and to model free algorithms that decompose enhancement patterns in
order to segment and classify different tissue types. Inhere we present a general hybrid method for analyzing dynamic
contrast enhanced images integrating a mathematical, model-free technique with a model derived approach that
characterizes tissue microvasculature function. We demonstrate the application of the method for breast cancer
diagnosis. A brief description of this approach was recently presented for the diagnosis of prostate cancer. The model
free method employed principal component analysis and yielded eigen-vectors of which two were relevant for
characterizing breast malignancy. The physiological relevance of the two eigen-vectors was revealed by a quantitative
correlation with the model based three time point technique. Projection maps of the eigen-vector that specifically related
to the wash-out rate of the contrast agent depicted with high accuracy breast cancer. Overall, this hybrid method is fast,
standardized, and yields parametric images characterizing tissue microvascular function. It can improve breast cancer
detection and be potentially extended as a computer-aided tool for the detection and diagnosis of other cancers.
We introduce an automatic 3D multiscale automatic segmentation algorithm for delineating specific organs in Magnetic Resonance images (MRI). The algorithm can process several modalities simultaneously, and handle both isotropic and anisotropic data in only linear time complexity. It produces a hierarchical decomposition of MRI scans. During this segmentation process a rich set of features describing the segments in terms of intensity, shape and location are calculated, reflecting the formation of the hierarchical decomposition. We show that this method can delineate the entire uterus of the rat abdomen in 3D MR images utilizing a combination of scanning protocols that jointly achieve high contrast between the uterus and other abdominal organs and between inner structures of the rat uterus. Both single and multi-channel automatic segmentation demonstrate high correlation to a manual segmentation. While the focus here is on the rat uterus, the general approach can be applied to recognition in 2D, 3D and multi-channel medical images.