Impact of Respiratory-Correlated CT Reconstruction Algorithms in the Choice of Margin Definition for Free Breathing Lung Treatment
S Thengumpallil1*, J-F Germond1, O Matzinger2, J Bourhis2, F Bochud1, R Moeckli1, (1) Institute of Radiation Physics, Lausanne University Hospital, (2) Radio-Oncology Departement, Lausanne University HospitalSU-C-141-5 Sunday 1:00PM - 1:55PM Room: 141
Purpose: To investigate the impact of phase- and amplitude-based reconstruction algorithms in 4D treatment margin strategies in presence of respiratory irregularities.
Methods: 4D CT images of Dynamic Thorax Phantom were acquired and the respiratory signal was tracked by the belt. Three irregular breathing patterns were simulated: amplitude, frequency, and combined amplitude and frequency variations. 4D CT image artifacts were analyzed for the phase-and amplitude based reconstruction algorithms. To assess the impact of the 4D CT image artifacts on the dose distribution, twelve plans (three simulated breathing models for the two reconstruction algorithms and for the two margins strategies) were calculated with a Monte Carlo based TPS. The lung tumor was delineated according to the mid ventilation and the ITV-based strategy. All the plans were evaluated for Dmin, Dmax, Dmean, homogeneity index and conformity index.
Results: The amplitude-based reconstruction shows a significant reduction of artifacts compared to the phase-based only for breathing amplitude variations. Most of the observed artifacts are image blurring and incomplete structures. The dosimetric analysis shows that the mid-ventilation strategy provides better tumor coverage than the ITV-based strategy. In the mid-ventilation strategy the selected phase is close to the exhale phase which is less subject to motion artifacts. In the ITV-based strategy the influence of motion artifacts is more pronounced due to the contribution of different phases and leads to a blurred dose distribution.
Conclusion: Amplitude-based reconstruction algorithm reduces image distortions in the particular case of amplitude variations but not when frequency variations are present. Based on the impact of image artifacts on the definition of margin strategies we showed that the mid-ventilation strategy is more robust than the ITV-based strategy. The ITV-based strategy is more prone to image artifacts which cause significant degradation in the dose distribution compare to the mid ventilation.