The study adhered to the Declaration of Helsinki and is approved by the Hospital’s Ethics Committee of Farabi Eye Hospital. All study participants provided written informed consent before enrollment.
Study Participants
In this prospective observational study, we evaluated patients clinically diagnosed with TED who presented at the orbit and oculoplastic clinic of Farabi Eye Hospital in Tehran, Iran. The study spanned 14 months, from January 2023 to February 2024. Patients with proven TED who planned to undergo decompression surgery were consecutively enrolled in the study. The demographic data, ophthalmic examinations, and vascular features were documented one week before and 6 weeks after decompression surgery. The exclusion criteria included inability to undergo OCT and OCT-A imaging, concurrent ocular surface disease (e.g. Glaucoma, diabetic retinopathy, and any optic neuropathy unrelated to TEDD), active thyroid eye disease, orbital decompression, strabismus, or eyelid surgery, and uncontrolled thyroid function tests.
The demographic data and past medical history (history of smoking, history of thyroidectomy, and treatment with radioactive iodine) were documented via interviews. All enrolled patients underwent comprehensive ophthalmic examination, including best corrected visual acuity (BCVA), Hertel exophthalmometry, margin-to-reflex distance (MRD) measurements of 1 and 2, and intraocular pressure (IOP) measurements. Disease activity was evaluated using the CAS, which includes spontaneous retrobulbar pain, pain in vertical eye movement, eyelid erythema, eyelid edema, conjunctival injection, conjunctival chemosis, and caruncle inflammation. A score was considered for each of these criteria. Patients with CASs ≥ 3 were classified as having active TED and excluded from the study.
Surgical technique:
To decompress the medial orbital wall, a medial transcaruncular orbitotomy was performed first after general anesthesia and the placement of a blepharostat. For this purpose, a conjunctival incision was made in the medial part and the tissue under the caruncle using Westcut scissors, and then the dissection was performed to the periorbita level. Then, the medial orbital wall was exposed by incising the exposed periorbita. After that, the bone was removed posterior to the posterior lacrimal crest and inferior to the ethmoidal arteries. Finally, the conjunctiva was repaired with an 8/0 Vicryl suture.
If inferior wall decompression is needed due to severe proptosis and requires double-wall decompression, inferior wall decompression was also performed in the same session. The periosteum was incised, and the orbital rim was exposed. The medial portion of the orbital floor was removed medial to the infraorbital nerve canal and extended posteriorly to the posterior wall of the maxillary sinus. The bone was preserved one centimeter behind the inferior orbital rim and at the junction of the lower and medial walls of the orbit. Finally, the conjunctiva in the inferior fornix was repaired with an 8/0 Vicryl suture.
Anterior Segment Optical Coherence Tomography Angiography
All study participants underwent AS-OCTA imaging at the designated imaging site. Experienced photographers conducted the imaging sessions using the AngioVue OCTA system manufactured by Optovue, CA, USA. The total scan acquisition time of AngioVue OCTA was less than 3 seconds. The device operated at a central wavelength of 840 nm and had a beam width of 22 µm. It achieved a scanning speed of 70,000 A-scans/second and provided an optical axial resolution of approximately 5 µm. The device acquired 304 × 304 A-scans, capturing two consecutive B-scans at each position to distinguish between static tissues and structures with high signal fluctuations. Blood flow was detected by combining the two images at the exact location using proprietary angiography algorithms and motion correction techniques. The split spectrum amplitude decorrelation angiography (SSADA) algorithm in AngioVue enhances the signal-to-noise ratio and improves flow detection.
Image acquisition and segmentation
We employed the Angioretina mode of the AngioVue OCTA system to capture the vasculature of the anterior segment of the eye. The image of the temporal and nasal halves of each eye is captured separately. The vascular parameters of each scanned image, including vessel density (VD), vessel length density (VLD), and vessel diameter index (VDI), were assessed at two distinct depths: 1) Superficial, from the conjunctival epithelium to a depth of 200 µm. 2) Deep, from a depth of 200 µm to a depth of 1000 µm, indicating the intrascleral layer. The conjunctival area within the image was manually identified by aligning it with the inner limiting membrane as indicated by the device. Additionally, each image was subdivided into temporal and nasal halves.
Statistical analysis
Descriptive statistics were calculated for the study variables, including the mean and standard deviation. We compared the means of the parameters before and after the operation. Given the inclusion of data from both eyes of the same individuals, we employed a generalized estimating equation model with robust estimators to compare the preoperative and postoperative data. We constructed a generalized linear model in which vascular features were the dependent variables, while age, sex, demographic data, and ophthalmic data (including LogMAR BCVA, MRD1, MRD2, CAS, and IOP) were considered independent variables. We excluded missing data from the analysis. To determine the effect of each independent variable on vascular features, we used multivariable analysis. Statistical significance was defined as a p value less than 0.05. All statistical analyses were performed using IBM SPSS Statistics software, version 27 (IBM Corp, New York, USA).