Unencrypted login | home

Program Information

Real Time Estimation of Prostate Tumor Rotation and Translation with a KV Imaging System Based On An Iterative Closest Point Algorithm

P Keall

P Keall1*, j nasehi1, R O'Brien1, P Poulsen2, (1) The University of Sydney, Camperdown, NSW, (2) Department of Oncology, Aarhus, Denmark

TU-G-141-9 Tuesday 4:30PM - 6:00PM Room: 141

Purpose: During cancer radiotherapy a small translation or rotation of the target could be important and change the delivered dose of the treatment. In this work, we implement and test an algorithm to estimate prostate cancer rotation and translation in real-time using a kV imaging system.
Methods:We developed a method based on the iterative closest point (ICP) algorithm and the three dimensional target positions for estimating rotation and shape deformation of prostate in real time. For evaluating our algorithm, 11748 kV images acquired from ten patients during intensity modulated arc therapy treatment (one fraction for each patient were used). The three fiducial markers coordinates were calculated from two dimensional kV images and used as input to the ICP algorithm for investigation of the rotation around three axes.
Results:The root mean square error (RMSE) of the rotation matrix was calculated for ICP algorithm and the value was less than 0.16mm. Correlation between all six degrees of freedom shows that the highest correlation belongs to the anterior-posterior (AP), and superior-inferior translation with 0.67, and Roll and Yaw rotation with cor=-0.33 has the second highest correlation in our study. The maximum standard deviation (as expected and previously reported) belongs to AP and Roll.
Conclusion:a 2D to 6D estimation method has been developed and applied to a 10-patient dataset. Given the widespread availability of cancer radiotherapy systems with single x-ray imagers, this method could have a major impact on prostate IGRT, and potentially other cancer sites where three or more implanted markers are routinely placed.
Acknowledgement: This work is supported by an NHMRC Australia Fellowship. Jin-Aun Ng, Jeremy Booth, Thomas Eade, Andrew Kneebone and Walther Fledelius are acknowledged for clinical implementation of the kV intra-treatment monitoring program.

Contact Email: