Medical Visualization for Orthopedic Applications

This dissertation discusses visualization techniques of articular cartilage for both quantitative and qualitative assessment. Articular cartilage is a very thin structure covering the endings of human bones. Thus, even slight changes in its thickness and inner structure may indicate an occurrence or progress of degeneration. The early detection of these factors is crucial for diagnosis and treatment of cartilage degeneration. Research to find treatments to stop or even reverse these degenerative changes is well in progress. Magnetic resonance imaging is currently the best and most used non-invasive technique for acquisition of soft tissue structures like articular cartilage. In this work we use two types of data: a high-resolution anatomical scan of the cartilage and a T2 map, which is computed from a set of sequences with varying parameters. While the thickness of the data can be precisely assessed from the anatomical scan, the T2 map offers information on the inner structures of the cartilage. Since the femoral cartilage is a curved thin-wall structure with a relatively small extent in one direction, it is very difficult to estimate its thickness from a stack of slices or even from a three-dimensional reconstruction of its surface. We discuss inspection of the tissue by unfolding and, thus, representing the tissue as a two-dimensional height field. Such a transformation of the object enables the application of 2D geometrical operations in order to observe subtle details in the thickness of the tissue. Nowadays scanners allow a quality assessment checking disruptions in the pattern of the T2 map of the patellar cartilage. The T2 map illustrates the quality of the cartilage and changes in the pattern of T2 map indicate defects before changes in the thickness itself occur. We propose the Profile Flags an intuitive interface for probing of the T2 maps by browsing the reconstructed surface of the cartilage. The Profile Flag is positioned on the reconstructed surface of the tissue, and can be moved along it. The Profile Flags can be applied to annotate local as well as global characteristics of the underlying data in a single still image. Furthermore, we present a set of extensions to Profile Flags for selection, classification and automatic positioning. Profile Flags can also be used to measure time-varying dynamic contrast enhanced magnetic resonance imaging data.