SU-E-J-130: Study of the Image Quality Degradation in Phase-Based 4DCT Imaging for Radiation Oncology

Four-dimensional computed tomography (4DCT) integrated into radiation imaging system is a useful tool for accurate targeting. 4DCT has the ability to minimize breathing related artefacts compared to conventional CT but irregular breathing and large tumour motion may cause inappropriate reconstruction. Our aim is to estimate the image quality degradation of 4DCT and to assess the clinical consequences.

The performance of the respiratory gating system of the multi-slice CT-simulator Brilliance Big Bore was evaluated employing one-dimensional moving phantom. The binning algorithm was phased-based. 4DCT scans of the test-phantom were acquired applying periodic motion patterns characterized by amplitude and frequency spanning a clinically range and also irregular waveforms simulating realistic breathing cycles. The internal spherical objects of the moving phantom were contoured using semiautomatic segmentation for evaluating uncertainties in volume delineation and motion amplitude calculation. Motion amplitude was determined by a specially written MATLAB program. Afterwards we investigated the influence of the image distortions on the targeting and the consequences of the related uncertainties on the calculated dose distribution of the tumour and the lungs in a group of five lung-cancer patients.

The results obtained with the phantom show that 4DCT imaging is still affected by distortions due to residual motion. The inaccuracies are mainly related to the amplitude (R2=0.99, R2=0.97 for 20 mm-dia sphere and 10 mm-dia sphere, respectively) and to the oscillation frequency (R2=0.87 for 20 mm-dia sphere, R2=0.96 for 10 mm-dia sphere) of simulated respiratory cycle and can cause both underestimation and overestimation of the real tumour motion amplitude with an average difference between real and calculated amplitude of about 10%.

The resulting inaccuracies on the internal tumour volume delineated do not have significant clinical consequences on lung doses but they could be important on tumour dose distribution if they imply underestimation of the real tumour motion.