Room: ePoster Forums
Purpose: MRI images suffer geometric distortion originating from both patient- and MRI system-specific factors. We assessed the geometric distortion inherent in MR images used in stereotactic radiosurgery (SRS) treatment planning and attempted to evaluate changes in target volume contour and dose delivery based on target location.
Methods: Geometric distortion of MR images was evaluated on a phantom and ten patients with targets from 0.13-0.13 cc. We evaluated the geometrical volume comparison between gross target volumes contoured on corrected and uncorrected MRI using simple volume assessment (%) and concordance index (CI%) with its location on a derived distortion map. To study dosimetry, their corresponding center of mass shift (both absolute and in all directions (X, Y, Z)) was evaluated for both contours. We evaluated differences in target dose coverage as a dose volume histogram after geometric distortion correction based on the target location with respect to the MRI distortion map, volume difference after correction, and center of mass shift.
Results: The target volume of corrected MRI images on average is 11.3% is smaller than non-corrected images, with a CI% of 0.41% and 0.45-mm distortion. The average change in center of mass between contours was 0.04 mm, with the maximum in the Z direction. The average under-dose for corrected MRI images was 1.5%. As we expected, the maximum under-dose was in the highest distortion area, where the lesions located near periphery region and boundaries such as cerebrum and brainstem with 0.7â€“0.8 mm distortion as high as 8.98%. The average center of change on targets for corrected MR images was 0.2 mm.
Conclusion: The changes in contours using the corrected MRI images is definitive. This technique will offer new opportunities to evaluate its effect on SRS tumor control probability and normal tissue complications.