(Sunday, 7/14/2019)
Room: ePoster Forums
Purpose: The AAPM radiation therapy committee task group 132 explored the fact that there is no known universal standard to validate the uncertainty for software programs used in image registration. This is due to mathematically unpredictable physical problems such as noise, anatomical changes, and the lack of documentation given by the software providers. Using their validation suggestions, different software platforms uncertainty values were compared to the values provided by TG 132.
Methods: Digital phantom data sets were registered in each of the software platforms according to the suggestions given by the TG 132. The translations and rotations were then compared to that of the TG 132 report and their allowed error uncertainties.
Results: Comparing the absolute error with that of the allowed voxel error (0.5 voxels) provided by TG 132, software A proved to be the most accurate for both the geometric and the anatomic phantom datasets. None of the absolute voxel error calculations were greater than 0.5 voxel while in software B, sixteen of the individual absolute error calculations were greater than the allowed 0.5 voxel error.
Conclusion: Ensuring images are registered correctly and within a known universal uncertainty is important to the future of patient safety and quality assurance. It is vital to continue analyzing different software platforms to find acceptable universal registration translation and rotation error values.
Keywords
Quality Assurance, Registration, Validation
Taxonomy
IM/TH- Image registration : General (Most aspects)
Contact Email
