AIM: To define the prevalence and anatomical patterns of paranasal sinus abnormalities (PSA) in thyroid-associated ophthalmopathy (TAO) and to test the hypothesis that TAO is partially driven by contiguous orbital inflammation rather than systemic autoimmunity or generalized orbital pressure. METHODS: Data included ophthalmic assessments and a panel of thyroid function and autoimmune biomarkers. Blinded radiological analysis of orbital computed tomography (CT) scans was performed to quantify sinus abnormalities and extraocular muscles (EOMs) involvement. Patients were categorized into two groups based on CT findings, those with no radiological evidence of sinus abnormalities (non-PSA control group) and those with identifiable PSA. Furthermore, ethmoid sinus mucosal biopsies from a subset of TAO patients and non-inflammatory controls were subjected to histopathological analysis. RESULTS: Totally 121 TAO patients (mean age 42.4±12.8y, range 10–78y), male:female=42:79, were included. PSA was identified in 44.6% (n=54) of patients, with a distribution anatomically restricted to the maxillary (50.0% isolated) and ethmoid sinuses (18.5% isolated; 29.6% combined). Compared to the non-PSA group (n=67), patients with PSA were significantly older (45.1±11.8 vs 40.3±13.2y; P=0.040) and were more likely to be male (55.6% vs 17.9%; P<0.001). They also had significantly higher proptosis (22.1±3.2 vs 20.7±2.9 mm; P<0.001). Medial/inferior rectus involvement was most frequent (88.4% vs 89.3%). Histopathological analysis of sinus mucosa from PSA patients provided direct evidence of pathology, revealing a dense, chronic lymphoplasmacytic infiltrate and submucosal edema, validating the radiological findings as a true inflammatory process. No significant correlation was found with systemic autoimmune markers, including thyroid-stimulating hormone (TSH) receptor antibodies (TRAb, median 4.86 vs 2.71 IU/L, P=0.104). CONCLUSION: TAO is associated with a high prevalence of PSA in a pattern consistent with the orbital anatomy. The correlation with ipsilateral muscle thickening combined with the lack of association with proptosis laterality or systemic biomarkers lend strong support to a model of contiguous inflammation over systemic autoimmunity, a hypothesis that warrants further validation through longitudinal and mechanistic studies.