Chronic rhinosinusitis symptoms and computed tomography staging: improved correlation by incorporating radiographic density

Abstract

Traditional methods for staging of chronic rhinosinusitis (CRS) by computed tomography (CT) have not proven to be predictive of subjective reports of patients' symptoms. An objective measure of CRS severity that correlates well with patients' symptoms, particularly if available from CT data, would be a valuable tool in assessment of disease status and outcomes after surgery. Retrospective chart review of CRS patients with symptom data from the Rhinosinusitis Symptom Inventory (RSI) and objective data from a sinus CT scan. CRS disease severity on CT imaging was measured according to the traditional Lund-Mackay scoring, as well as by raw measures of the densities of sinus opacities (in Hounsfield units [HU]) and density-weighted Lund-Mackay scoring. These data were related to symptom severity scores using a multivariate regression model. There was no significant correlation between either raw density values of sinus opacities or weighted Lund-Mackay scores with facial or total symptom scores. Oropharyngeal symptoms scores were negatively correlated with the sum of "average HU" values (p = 0.036, β = -1.120) but were positively correlated with the sum of "maximum HU" values (p = 0.047, β = 1.221). There was a significant negative correlation between the systemic symptoms score and mean of "average HU" values (p = 0.010, β = -0.272). Finally, there was a positive correlation between "maximum HU" value-weighted Lund-Mackay score with nasal symptom scores (p = 0.016, β = 0.241), systemic symptom scores (p = 0.008, β = 0.605), and total symptom scores (p = 0.078, β = 0.179). Incorporation of radiographic characteristics of sinus opacification with Lund-Mackay scores offers greater predictive power of patients' subjective symptom severity.