Analysis of Intrafractional Mass Variabilities Using Deformable Image Registration Program

The aim of this study is to compare the geometric characteristics of the lung tumor, such as tumor centroid, HU change relative to breath phase, depending on tumor location and adhesion using 4DCT and deformable image registration program (MIMVista). The Y axis change was most significant and the mean Y axis centroid fluctuation was 7.32±6.88 in lower lung tumor. The mean HU variation in lower lung mass has changed more than other locations, and its mean HU variation was 7.7±4.97 and non-adhered mass was more changed. Correlation for the mass volume between 3DCT and MIP was very high and its coefficient was 0.998. The effect of tumor location, adhesion and diaphragm excursion to geometric uncertainties was analyzed by linear regression model, it was influenced to mass deformation and geometrical variation so much except diaphragm excursion. But intra-fractional and inter-patient's uncertainties were great, so it couldn't find any exact deformation trend.