Experimental Study On the Performance of the OMAR CT Artifact Correction Algorithm Near Titanium and Stainless Steel
A Schoenfeld1,2*, R Crilly3, B Poppe1,2, W Laub3, (1) WG Medical Radiation Physics, Carl von Ossietzky University, Oldenburg, Germany (2) Pius Hospital, Oldenburg, Germany (3) Oregon Health & Science University, Portland, ORSU-E-I-39 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: Implants with high atomic numbers produce star-like artifacts in CT-images. When coupled in pairs, a black window appears in between. The Philips OMAR algorithm approaches these artifacts by post processing the images. In this study, measurements were performed with radiochromic EBT2 films in order to verify, whether the OMAR algorithm contributes to the accuracy of treatment plans in implant regions.
Methods: Two sets of titanium and stainless steel rods with diameters of 1inch and 0.5inch were centered in a water phantom either single or in pairs. The phantom was scanned in a Philips Bigbore Brilliance scanner, which provides the OMAR correction algorithm. A set of uncorrected CT data was kept in addition to the post processed data.
The treatment plans were calculated with AAA in Eclipse. The Hounsfield units in the volume of the metal rods were set to match the relative electron density with the conversion table extended accordingly. A 10x12 field at 90 degrees gantry rotation and 400MU was calculated using on a 1mm grid. The plans were irradiated using a Varian Novalis-TX linac with EBT3 films placed on the sides of the single rod setup or between the rods in the two rod setups. The FilmQA Pro Software was used for evaluation of the films.
Results: With one inhomogeneity present in the phantom, the difference between the plans based on the uncorrected data and the plans based on the post processed data is small. With paired inhomogeneities and the associated increased artifact, the difference becomes more significant and the OMAR algorithm improves the accuracy.
High Z scattering effects cause considerable deviation from the planned dose profiles near the rods.
Conclusion: The OMAR-algorithm improves the accuracy of the dose calculations. The dose profiles across implant regions are calculated accurately, if the Hounsfield units are defined accordingly.