Rectal tumor registration based on virtual templates
Michael Höfer, Stephen L. Keeling, Gernot Reishofer, Rudolf Stollberger · International Symposium on Image and Signal Processing and Analysis · 2011
Rectal cancer is the second most common form of cancer in developed countries. The decision of an appropriate therapeutic approach like radiation therapy or surgical intervention largely depends on the quality of structural and functional imaging methods. Dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) is a widely accepted method for the non-invasive assessment of rectal tumors, suffering from artifacts due to bowel movement, respiratory motion and rectal coil movement. Giving the fact that MRI contrast media change the signal intensities over time, classical image registration strategies fail. To overcome these problems, the algorithm Virtual Template Registration (VTR), originally designed for the registration of DCE-MRI Renography, is adapted to register DCE-MRI data from rectal tumors. VTR derives a smooth virtual template image series from the DCE-MRI data, where motion artifacts are minimized but intensity changes due to contrast media uptake are preserved. The original dynamic time series (source images) is then registered by applying elastic registration to the generated virtual template. Our results indicate, that motion artifacts, visible in evaluated time courses, are significantly reduced and allow for a more differentiated analysis of different rectal tumor types.