Comparison of two classification systems for tissue eosinophilia in patients with chronic rhinosinusitis with nasal polyps.

BACKGROUND

Eosinophilic chronic rhinosinusitis with nasal polyps (ECRSwNP) is commonly defined by absolute tissue eosinophil count (≥ 10 per high-power field; Def1) or eosinophil percentage of total inflammatory cells (≥ 10% per high-power field; Def2). However, these two criteria sometimes yield discordant classifications, and no consensus exists on an appropriate range of total inflammatory cells for consistent diagnosis. This study compared the consistency of these two definitions and explored whether a specific range of total inflammatory cell counts could improve their diagnostic agreement.

METHODS

Nasal polyp tissues from 104 patients with chronic rhinosinusitis with nasal polyps (CRSwNP) and inferior turbinate samples from 19 healthy controls were analyzed. Tissue eosinophil count, eosinophil percentage, and total inflammatory cell counts were assessed histologically. Receiver operating characteristic analysis, correlation analysis, positive and negative predictive values (PPVs and NPVs), peripheral blood eosinophil indices, and local type 2 cytokine expression (IL-4, IL-5, and IL-13) were used to explore and internally assess the proposed interval.

RESULTS

Def1 and Def2 were concordant in 88 of 104 patients, with discrepancies mainly observed outside the proposed 385-1236 total inflammatory cell-count interval. Within this interval, the correlation between tissue eosinophil count and percentage was stronger (r = 0.93), and PPVs and NPVs for ECRSwNP classification were higher (PPV: 100.00%/91.89%; NPV: 89.29%/100.00%). Peripheral blood eosinophil indices and IL-4, IL-5, and IL-13 expression also showed stronger associations with eosinophil-based classification within this interval than outside it.

CONCLUSION

We identified an exploratory total inflammatory cell-count interval of 385-1236, within which Def1 and Def2 showed more consistent classification of ECRSwNP in the present cohort.