Patterns of olfactory dysfunction in chronic rhinosinusitis identified by hierarchical cluster analysis and machine...

Area of Science:

• Immunology
• Otorhinolaryngology
• Data Science

Background:

• Olfactory dysfunction is a frequent symptom in chronic rhinosinusitis (CRS).
• Previous research suggests inflammatory mediators may underlie CRS-associated smell loss.
• This study aimed to validate these inflammatory patterns in CRS using advanced analytical methods.

Purpose of the Study:

• To validate olfactory dysfunction patterns in CRS patients.
• To identify inflammatory mediators associated with smell loss in CRS.
• To determine predictors of olfactory dysfunction in CRS.

Main Methods:

• Functional endoscopic sinus surgery patients (n=110) underwent preoperative Smell Identification Test (SIT).
• Mucus inflammatory mediators were quantified using multiplexed flow-cytometric bead assay.
• Hierarchical cluster analysis, machine learning (random forest), and multivariate regression were employed.

Main Results:

• Five distinct CRS clusters emerged, with significant differences in SIT scores (p < 0.001).
• A cluster with severe nasal polyposis, allergic fungal rhinosinusitis, and aspirin-exacerbated respiratory disease (AERD) predominantly housed anosmic patients.
• Interleukin (IL)-5 and IL-13 were strongly associated with olfaction; AERD, CT score, and IL-2 were most predictive of smell loss.

Conclusions:

• Specific CRS endotypes, marked by severe nasal polyposis, eosinophilia, and AERD, are linked to olfactory dysfunction.
• Mucus IL-2 levels, CT score, and AERD independently predict smell loss in CRS patients.