Identifying a sphenoid sinus fungus ball using a nomogram model

Abstract

Background: Sphenoid sinus fungus ball (SSFB) is a rare entity and usually presents with non-specific symptoms. SSFB could potentially lead to serious orbital and intracranial complications. Computed tomography (CT) scan is usually the first imaging test of the diagnostic workup in patients with specific clinical symptoms. This study aimed to compare the clinical characteristics and CT features between SSFB and unilateral (non-fungus ball) chronic sphenoid rhinosinusitis (USRS) and help differentiate between these two most common inflammatory diseases of the sphenoid sinus. Methods: By retrospective database review, 66 patients with a histopathologic diagnosis of isolated SSFB were recruited for analysis. Fifty-four patients who underwent endoscopic sinus surgery with clinical and histopathological diagnoses of USRS were enrolled as the control group. Clinical characteristics and CT features were evaluated. Results: Headache, rhinorrhoea, nasal obstruction, postnasal dripping, and hyposmia were the most common symptoms in both groups. In the univariate analysis, older age, lower white blood cell counts, irregular surface, bony dehiscence, lateral wall sclerosis, and intralesional hyperdensity (IH) were significant predictors for SSFB. Older age, irregular surface, and IH remained statistically significant in the multivariate analysis. Based on the results of the regression analysis, a nomogram for predicting the probability of SSFB was plotted. Conclusions: We developed a nomogram model as a novel preoperative diagnostic tool for identifying SSFB according to the predictors both in clinical characteristics and on CT features. This could help the clinicians in predicting the probability of SSFB, to reduce ineffective or delayed treatment and occurrence of complications.