Abstract
PURPOSE: The study's purpose was to use a simple geometry phantom to validate the deformable image registration (DIR) accuracy and dose warping accuracy in carbon ion radiotherapy (CIRT) and to provide an index for dosimetry in CIRT. MATERIALS AND METHODS: We used geometric and anatomical phantoms provided by AAPM TG-132. The DIRs of 3 different settings were performed between reference and translational images for each phantom. CIRT and photon therapy (3-dimensional conformal radiotherapy and volumetric modulated radiotherapy) treatment plans were transformed by the use of deformation vector fields calculated from the DIR of each setting. The dose distribution calculated on the basis of rigid registration between images was used as the ground truth and compared with the warped dose determined by DIR to evaluate the error. RESULTS: The photon therapy treatment plans showed a dose warping error of <2% for a DIR error of <2 mm, whereas CIRT showed a dose warping error >10% for the same DIR accuracy. From this, even with similar DIR accuracy, the errors tended to be larger for CIRT than for photon therapy dose distributions. Due to the steepness of the CIRT dose gradient, the dose difference increased by about 5% for a DIR error of 5 mm, which was larger than the 3% dose difference generated for a 5 mm DIR error in photon therapy. CONCLUSION: We evaluated the relationship between DIR accuracy and dose warping accuracy in the CIRT dose distribution. Due to the steepness of the dose gradient in CIRT, we concluded that dose warping based on DIR accuracy should be required to be sufficiently higher than that in photon therapy.