This study was approved by the institutional review board, and the requirement of informed consent was waived considering its retrospective design. This study adheres to the guidelines of the Declaration of Helsinki. Research data are stored in an institutional repository and will be shared upon request to the corresponding author.
We reviewed the medical records of 62 consecutive GO patients treated with RT between December 2015 and December 2018. Patients who were diagnosed with moderate-to-severe active GO, treated with RT, and followed up for at least 6 months were included in this study. These patients were refractory to high-dose IV steroids or were not eligible for treatment with high-dose IV steroids. The activity and severity of GO were accessed according to the standardized criteria recommended by the European Group on Graves’ Orbitopathy (EUGOGO)17. Patients who underwent ophthalmic surgery within three months prior to RT, pregnant patients, patients under 20 or over 80 years old, and patients with a history of other eye diseases such as glaucoma, diabetic retinopathy, or maculopathy, were excluded. Information on age, sex, duration of GD and GO prior to RT, smoking, family history of GO, and treatment history was collected. Serum concentrations of thyroid stimulating hormone (TSH), free thyroxine (FT4), and thyroxine binding inhibitory immunoglobulin (TBII) were evaluated prior to RT.
Comprehensive ophthalmologic examinations including the degree of proptosis, margin reflex distance 1 (MRD1, the vertical distance between the center of the pupil and the center of the upper eyelid), palpebral fissure height (PFH, the vertical distance between the center of the lower eyelid to the center of the upper eyelid), intraocular pressure (IOP), diplopia, EOM movement, visual acuity (VA), existence of compressive optic neuropathy (CON), and clinical activity score (CAS) were carried out before RT and at 3 and 6 months afterward. The degree of proptosis was measured using a Hertel exophthalmometer (Oculus, Arlington, VA, US). MRD1 and PFH were measured using custom-made PC-based EAS software (Eyelid Analysis Software, Biomedical Research Institute, Seoul, Korea) using photographs taken in primary position18. Diplopia was evaluated based on Gorman score19. EOM movement was measured using the Light Reflex Method, where a score of 45° was given when the light reflex was at the limbus, 30° at halfway between the limbus and pupil edge, and 15° at the pupil edge. Improvement in the EOM movement was defined if the variation was at least 15°, as previously described20. VA measured using a Snellen chart was converted to the logarithm of the minimum angle of resolution (logMAR). CON was judged based on the presence of decreased VA, plus one or more of the following findings: relative afferent pupillary defect, color vision deficit, or visual field defect21. A modified CAS was assessed using a seven-point modified formulation21. When differences in scores or measures existed for the two eyes of one patient, data from the worse eye were considered for analysis. All ophthalmic examinations were performed by one specialist (J.K.L).
Computed tomography (CT) was performed before and again at 3–6 months after RT at convenience. Orbital CT scans were obtained with 1 mm sections. Volumetric measurements of EOM and orbital fat were carried out as per Regensburg et al.22 The volumes of the superior rectus, inferior rectus, medical rectus, lateral rectus, and orbital fat were calculated with the use of manual segmentation of CT scans with the Eclipse treatment planning system (TPS) (ver. 13.7, Varian Medical Systems Inc., Palo Alto, CA, USA). The orbital soft tissue CT numbers referenced for volume measurements were set at ‒200 to ‒30 Hounsfield units (HU) for fat tissue and ‒30 to +100 HU for muscle tissue. Tissues of interest with the chosen CT number in all slices were delineated and reconstructed in three dimensions and tissue volumes were measured.
All patients were treated with retroorbital irradiation using a linear accelerator with a three-dimensional conformal technique. Patients were immobilized with a custom-made thermoplastic case, and CT scanning with a slice thickness of 2.5 mm without contrast was performed for image acquisition and target contouring. The clinical target volume (CTV) included the EOMs and retroorbital fat. The lenses, optic nerves and lacrimal glands were zoned as organs-at-risk. A 2 mm margin around the CTV was generated as the planning target volume. The planned irradiation dose was 24 Gy in 12 fractions, and treatment planning was carried out using Eclipse TPS. Irradiation was performed using a 6 MeV photon beam. Concurrent low-dose oral steroids of 20mg prednisone daily were administered to patients with active high CAS score for 2~4 weeks.
Response to orbital RT was defined as an improvement in at least two of the following six parameters: 1) CAS improvement in at least two points, 2) reduced proptosis by at least 2 mm, 3) improvement in diplopia (lessening of Gorman score), 4) improvement in EOM movement, 5) improvement in VA , and 6) disappearance of CON.
To verify the effects of timing of orbital RT on the treatment response, the subjects were divided into two groups based on the GO symptom duration from onset. The timing of RT according to GO duration was classified as the early-active phase until 24 months, and after 24 months as the late-active phase. Orbital RT in the late-active phase was performed in patients who recently experienced a distinct deterioration of symptoms or patients resistant to long-term steroid treatment. Response to orbital RT was analyzed and symptoms improved after RT was identified in each group.
All statistical analyses were performed using R version 3.4.0. Data are expressed as mean ± SD or median (interquartile range) for continuous variables. Categorical variables were reported as sample numbers and percentages. Changes in ophthalmic findings before and after treatment were compared using a linear mixed effect model or a generalized linear mixed effect model to analyze a fixed-time effect by controlling for random effects. For ordinal response variables, repeated measures proportional odds logistic regression was used. Changes in orbital fat and EOM volumes were compared using paired t-tests. Changes in ophthalmic findings according to the treatment phase were compared using the Wilcoxon signed-rank test or McNemar test. Univariable and multivariable logistic regression analyses were performed to analyze the effect of each clinical measurement on the binary response to RT. Variables showing significance levels of 0.1 in the univariable analysis were used in the multivariable analysis. All statistical analyses were considered significant at two-tailed P < 0.05.