Purpose: To develop a QA phantom that can quantitatively evaluate stereotactic spaces defined among MRI, CT, and Cone Beam CT (CBCT). Variations in stereotactic space definition have been an issue when acquiring imaging with different modalities for gamma knife SRS patients. These discrepancies dispute the systematic integrity and are especially significant since variation in object localization may result in lack of prescribed dose to the target with additional radiation being absorbed by parenchyma.
Methods: A Leksell Gamma Knife dosimetry phantom was placed in the immobilized stereotactic frame so it can be locked in position when being imaged across multiple modalities. A total of 12 CT and 4 MRI spot markers were placed inside and around the phantom. These markers were used as the means of comparison among different modalities. The phantom was imaged in CT and MRI with dedicated localizer box and a CBCT image was also taken using on board imager from a Gamma Knife ICON machine. All coordinates for the markers were determined in Gamma plan with different stereotactic space definition accordingly. The coordinates from CT and MRI were compared with that of CBCT and a translational shift was calculated by a taking vector magnitude between the two stereotactic space coordinates.
Results: Minimum, median, and maximum total translational shifts of 0.65, 1.24, and 1.85mm, respectively were obtained for CT compared to CBCT. Interior markers were more susceptible to shifts compared to exterior. The MRI to CBCT comparison resulted in similar total translational shifts, with minimum, median, and maximum values of 0.64, 1.17, and 1.41mm. The 3 markers used for this comparison were on the phantom exterior.
Conclusion: With the magnitude of most of the shifts being close to 1mm, this experiment suggests that further investigation may need for systematic error when co-registering images from different modalities.