Purpose: The conventional ocular proton treatment (OPT) for uveal melanoma uses eye-3D-model in the fixed proton beamline with a dedicated eye snout. The disadvantages of conventional method are inaccurate target segmentation and incorrect dose calculation because of not using CT-image. The aim of this study is to investigate the possibility of image-guided OPT through the development of the eyeball rotation system (ERS) based on CT-image.
Methods: ERS has been developed only to rotate the eyeball with the optimized azimuth and polar angles in acquired planning CT-image and exports a new eyeball-rotated CT-image in DICOM format. It enables that a treatment planner can decide the best eye-gaze direction to achieve OAR sparing. Our system requires two CT-image sets with two different eye-gaze directions to find the eyeball rotation center. This approach can be applied to both OPT and rotational gantry room (RGR). A simulated treatment plan using Pencil Beam Scanning (PBS) mode was tested for supine-position setup and gantry-angle 0ÌŠ in RGR.
Results: Ophthalmologist inserts four tantalum clips around eyeball for the 2D image setup-check of eyeball. During the patient setup procedure, clips position should be checked using anterior-posterior (AP) and lateral 2D images. Clips should not be located in proton beam path because high-Z material causes the range uncertainty. For this reason, ERS shows DRR in AP-direction, so that a treatment planner can decide azimuthal and polar angles using the DRR view window and the light gaze view window. The simulated PBS treatment plan shows that the possibility of image-guided OPT in terms of the good target coverage and OAR-sparing.
Conclusion: We proposed a novel image-guided OPT that can perform the correct target delineation and accurate dose calculation compared with conventional eye-3D-model method. It was confirmed that the image-guided OPT is applicable to uveal melanoma treatment by using ERS.
Funding Support, Disclosures, and Conflict of Interest: This study was supported by Institutional Inherent Research Project of Korea National Cancer Center grant (1810272-1).