a fiducial marker system using digital techniques

The individual kilo voltage (kV)projection images acquired via the CBCT have increased image contrast and signal to noise ratio over mega voltage (MV) portal images, as well as allowing a 3D volume reconstruction.Īdditionally the kV beam can be used during treatment whilst switching between irradiation angles, giving a number of projections over a limited rotation of the gantry. Since the study by Nederveen (2000) is over five years old, new techniques are likely to be available.Ī further important reason to look into other extraction techniques is provided by new types of accelerators, which provide a Cone Beam CT (CBCT), instead of flat-panel imagers. Hence the primary goal of this study is to analyse and compare techniques currently available for marker detection using portal images. Furthermore it would reduce radiation exposure to healthy tissue. By using such a method, a higher degree of confidence on organ location in between treatment fractions can be obtained. Aubin et al (2003) suggest a different marker extraction technique, which might yield better results. However, the marker detection success published by Nederveen (2000) can be difficult to reproduce in clinical practice at the UMC Utrecht, due to slight setup differences. The pixel giving maximal response (marker likelihood) on a marker-like shape identifies the location of the marker. This kernel is convolved with every pixel in the region of the predicted marker location. A digital marker extraction kernel (MEK) is designed to detect the specific shape. The positional verification techniques clinically used at the UMC Utrecht for intensity modulated radiotherapy (IMRT) prostate treatment are based on fiducial gold markers, detected using template matching techniques applied on the portal image (Nederveen et al, 2000).

A fiducial marker system using digital techniques