MR guided therapy has a specific requirement for dynamic MRI methods. First, the image field-of-view (FOV) may be selected in a continuously varying fashion by physician using, for example, a pointing device. Second, independent of the selection of the imaging plane, the objects in the image FOV may themselves be changing physically (e.g., as a biopsy needle enters) or chemically (e.g., during laser or rf ablation, or as a bolus of contrast agent is injected) therefore altering the resultant MR image. All of these circumstances require a dynamic MR imaging method. Dynamic MRI differs from standard MRI in that a sequence of time-ordered images is obtained by continually updating or reacquiring image data. Recent investigations of our laboratory have focused on new dynamically adaptive methods for imaging the processes involved in MR guided therapy that evolve so rapidly that they cannot be adequately resolved both spatially and temporally using current dynamic approaches. Dynamically adaptive methods are distinguished as those in which the image data acquisition strategy is modified dynamically depending on information obtained after processing the most recently acquired image(s). Our approach is founded on the hypothesis that significant improvement in temporal resolution, without loss of image quality, can be obtained in dynamic MRI if an adaptive method is used to minimize redundancy in image encoding and data acquisition. The MRI methods that are under study in our laboratory, are efficient encoding and acquisition methods. They can be implemented in the context of fast imaging techniques (e.g., echo- planar or fast gradient-echo), which do not exploit redundancy in data acquisition. We will focus on implementing and improving three main dynamic MRI approaches: a. Wavelet transform encoding techniques; b. SVD (Singular Value Decomposition) encoding techniques; c. Specialized Fourier transform encoding techniques. We will also evaluate each method in terms of standard performance criteria (signal-to-noise ratios, spatial and temporal resolution). Finally, we will study the physics of selective rf excitation methods for MRI (the basis of each of the imaging methods) and the diagnostic utility of new methods using standard ROC methods modified to assess the detection and characterization of time-varying information in series of images. This method development project will have an impact on the program by providing improved spatial and temporal resolution, faster (dynamic) scanning, and quantitation of spatially dependent information tailored to the specific goals of each of the other four projects.
Showing the most recent 10 out of 187 publications
