Degree Name

Master of Science by Research


Faculty of Engineering


Respiratory gating enables breathing synchronised activation of CT image acquisition and linear accelerator radiation output. Two commercially available respiratory gating systems used for planning and treatment of thoracic and abdominal cancer are investigated. The strain gauged AZ-733V respiratory gating system (Anzai Medical Systems, Tokyo, Japan) was used concurrently with the infrared Real-time Position Management system (Varian Medical Systems, Palo Alto, CA) to measure the respiratory cycle of 15 volunteers. Correlation between systems was measured in six locations and the optimum position of the external surrogates determined based on signal amplitude, reproducibility of breathing waveforms and the coefficient of determination between Anzai and RPM signals. The mean value of R2 between the two systems was found to be 0.611, 0.788 and 0.925 when both markers were positioned at the xiphoid, midway between the xiphoid process and umbilicus, and at the umbilicus respectively. When positioned in separate locations results were varied, R2 values ranging from 0.345-0.965. Results highlighted the importance of external surrogate position to the respiratory signal obtained, and indicated that the external marker position on the chest wall needs to be reproducible between 4D CT scanning and treatment. Recommendations are made that external surrogates must always be positioned at the umbilicus for the most clinically useful scans.

Image distortion and artifacts were studied using the Anzai AZ-733V respiratory gating system in combination with the Siemens Sensation Open CT scanner. A moving respiratory phantom was constructed and the volumetric accuracy of retrospectively reconstructed 4D CT images for three moving test objects, across five frequencies and four amplitudes of movement was compared. Volumetric accuracy was found to be within 10% for retrospectively reconstructed gated objects moving with a period of 4 s, amplitude 1 cm. Large deviations of 19.4-51.6% from the static volume of the objects were observed in gated images for periods of 3 s or less. Significant distortion and under sampling was observed in gated images of the objects moving with a period of 10 s. Artifacts were related to the partial projection effect and data sufficiency conditions outlined in literature (Keall 2004, Pan 2004, Dinkel 2007).

02Whole.pdf (7304 kB)