A Single-Center Retrospective Study on Noninvasive Prediction of Terson Syndrome in Aneurysmal Subarachnoid Hemorrhage (aSAH) Patients: The Role of CT-Measured Posterior Globe Thickness and Age.

Background Terson syndrome (TS), an intraocular hemorrhage secondary to aneurysmal subarachnoid hemorrhage (aSAH), has a high incidence rate. Clinically, patients with aSAH often present with concomitant TS; however, owing to the difficulty in performing ophthalmic examinations in critically ill patients, many cases may be missed. This study aimed to develop and evaluate a CT-based diagnostic model incorporating posterior globe thickness to predict TS in patients with aSAH. Materials and methods This was a retrospective study on patients who underwent direct surgery or endovascular treatment for ruptured cerebral aneurysms at our institution between January 1, 2018, and August 31, 2025 (analyzed by eye). We extracted data from eyes definitively diagnosed with TS via ophthalmic examination. In addition to collecting epidemiological and clinical data, posterior globe thickness was measured for each eye. Statistical analyses included the Mann-Whitney U test, chi-square test, generalized estimating equation (GEE) logistic regression analysis, and receiver operating characteristic (ROC) analysis. Statistical significance was set at p < 0.05. Results A total of 177 patients (354 eyes) received aSAH treatment, of whom 26 individuals (52 eyes) underwent ophthalmic examination, and within this subgroup, 11 patients (17 eyes) were diagnosed with TS. In the univariate GEE logistic regression analysis, the presence of TS was significantly correlated with age (p=0.005), World Federation of Neurosurgical Societies (WFNS) grade (p=0.021), complaints of visual and visual field impairment (p=0.021), and posterior globe thickness (p=0.038). The multivariate GEE logistic regression analysis demonstrated that age and posterior globe thickness significantly influenced the risk of developing TS. In this final multivariate model, the odds of having TS decreased by a factor of 0.85 for every one-year increase in age (p=0.007), whereas the odds increased by a factor of 13.74 for every 1 mm increase in posterior globe thickness (p=0.027). ROC analysis, performed using this final multivariate model, yielded a calculation to determine the age-dependent cutoff for posterior globe thickness: Cutoff(mm)≈-1.295+0.0637×Age (years), which showed a sensitivity and specificity of 82.4% and 82.9%, respectively. Conclusion This study proposes a noninvasive prediction model for estimating TS based on CT measurements of posterior globe thickness. Serving as a practical triage tool, these findings suggest that incorporating age significantly enhances the diagnostic utility. To ensure broad generalizability and facilitate its application in clinical practice, prospective multicenter trials are necessary to validate these results.