Site‐Specific Olfactory Cleft Opacification Predicts Olfactory Function and Olfactory Training Outcome in Persistent Postinfection Olfactory Dysfunction

BACKGROUND: Persistent postinfectious olfactory dysfunction (PIOD) is a prevalent and often refractory condition, with low recovery rates following olfactory training (OT), primarily due to sustained localized inflammation and opacification within the olfactory cleft (OC). However, conventional radiological assessments of OC opacification overlook the vertical distribution of olfactory neuroepithelium, limiting prognostic accuracy. This study introduces a modified subregion computed tomography (CT) scoring system integrated with computational fluid dynamics (CFD) to evaluate OC opacification's role in baseline olfaction and OT efficacy. METHODS: In this prospective study, 58 adults with persistent PIOD of more than 3 months' duration were enrolled and completed CT imaging, psychophysical olfactory testing (Sniffin' Sticks TDI score), and questionnaire assessments. Using a modified scoring system, the OC was segmented on coronal sections into upper (superior one-third) and lower (inferior two-thirds) portions, with three coronal sections assessed in both the anterior and posterior OC. Each section was scored dichotomously (0 for no opacification, 1 for opacification), and bilateral scores were recorded. Unsupervised K-means clustering identified phenotypes, random forest modeling predicted functional anosmia, and CFD analyzed airflow dynamics. Thirty-one patients completed 3-month OT with follow-up assessments. RESULTS: OC opacification was present in 69.0% of persistent PIOD patients, with opacification in the upper posterior OC notably serving as the primary determinant of baseline olfactory function. Random forest analysis further confirmed that upper OC opacification ranked as the top predictor for functional anosmia, with the combined model achieving an area under the curve of 0.920. Cluster analysis revealed three distinct phenotypes, among which the diffuse severe opacification cluster involving upper OC regions was associated with the poorest olfactory function. Although posterior OC opacification was positively correlated with increased airflow velocity ratios (r = 0.570, p < 0.01) and these ratios were negatively correlated with olfactory scores, OC airflow itself did not significantly mediate the relationship between structure and olfactory function. Only 25.8% of patients achieved clinically meaningful olfactory improvement. Significant improvements were observed in TDI (p = 0.001), olfactory threshold (p = 0.010), olfactory discrimination (p = 0.028), and olfactory VAS scores (p < 0.001) after OT. Baseline opacification in posterior-middle section of OC significantly predicted non-response. CONCLUSIONS: The modified subregion OC opacification scoring system provides a more refined assessment of site‑specific OC obstruction for patients with persistent PIOD, underscoring the critical role of OC opacification in mediating olfactory impairment and resistance to OT. CLINICAL IMPLICATION: For patients with persistent PIOD, this modified olfactory cleft opacification scoring system allows for olfactory assessment and prognosis by identifying specific opacification patterns associated with olfactory training outcomes.