Quantifying movement of optical structures during radiotherapy treatment for tumors near the orbita.

Background and purpose: Movement of optical structures during radiotherapy for tumors near the orbita might affect the amount of radiation given and consequently the risk for side effects. The aim of this study was therefore to quantify motion of optical structures during radiotherapy. Materials and methods: Twenty brain tumor patients were retrospectively included, with planning computed tomography (CT)-scan (pCT) and five repeat-CT-scans (reCTs) without gazing instructions. Six optical structures were delineated bilaterally: lens, cornea, retina, lacrimal glands, macula, and optic nerves (ON). The ON was split in three subregions. The dice similarity coefficient (DSC), absolute distance (AD), and difference in 3D midpoint (ΔMP) were calculated between pCT and reCT. Planning risk volume (PRV)-margins and isotropic expansions to cover 95 % volume for 90 % patients were calculated. A dose-volume proof-of-principle was performed for a neurological and nasopharyngeal tumor. Results: Highest median ΔMP was found for the cornea: 1.9 mm. For ON subregions, highest median and 95th-percentile ΔMP was found for proximal intra-orbital ON: 1.3 mm and 3.1 mm. ON showed highest median AD: 3.0 mm (negative Z-direction). Open eyelid status resulted in statistically significant lower DSC for ON, intra-cranial ON, and proximal intra-orbital ON, and higher ΔMP for proximal intra-orbital ON. 1-4 mm isotropic expansions were needed for separate structures, dependent on typical movement range. Higher dose-differences were found for the neurological than the nasopharyngeal plan. Conclusions: The observed movement of optical structures indicated that a PRV-margin should be considered in clinical practice. Asking the patient to close their eyes during the treatment could decrease the movement.