TH-CD-303-09: Respiratory Motion Guided 4DCBCT On a Linear Accelerator with Lung Cancer Patient Breathing Traces




Image quality from Four Dimensional Cone Beam Computed Tomography (4DCBCT) is poor due to projection clustering and under sampling. We have developed a novel method called respiratory motion guided 4DCBCT (RMG-4DCBCT) in which the gantry speed and projection frequency are varied in response to the patients real-time respiratory signal. The method has been shown theoretically to improve image quality by 63% on average. The purpose of this study was to implement RMG-4DCBCT on a linear accelerator for a variety of lung cancer patient breathing rates.


In house software was developed to process RPM breathing traces from lung cancer patients and compute the optimal gantry trajectory and projection pulse rate schedule so that projections were uniformly distributed in each of 10 phase bins. The computed schedule was then programmed into developer mode of the Varian True Beam linear accelerator and the Scandidos Hexamotion phantom was modified to move the Catphan phantom according to the breathing trace. Four lung cancer patient breathing traces were selected with an average breathing period of 3, 4, 5 and 6 seconds respectively. We assess streaking artefacts by examining the variation in the angular separation between projections with a value of zero indicating uniformly space projects and larger values indicating more streaking in the reconstructed images.


RMG-4DCBCT with 60 projections per phase produces similar image quality to conventional 4DCBCT with 120 projections per phase. The image quality with RMG-4DCBCT is similar for all breathing rates with an average variation between projections of 1.3°. Conventional 4DCBCT acquiring 60 projections per phase had a variation of 5°, 6.1°, 7.3° and 8.1° for the 3, 4, 5 and 6 second breathing periods respectively.


RMG-4DCBCT halves the imaging dose over conventional 4DCBCT with consistent image quality for a range of breathing rates.

This project is supported by an NHMRC Australia Fellowship, NHMRC project grant 1034060 and Cancer Australia grant number 1084566.