Side-Specific Concordance Between Nasal Endoscopy and Computed Tomography Severity in Pediatric Chronic...

Background/Objectives: The European Position Paper on Rhinosinusitis (EPOS) diagnoses chronic rhinosinusitis (CRS) from symptoms with nasal endoscopy and/or computed tomography (CT), but how closely a graded endoscopic score reflects CT severity in children-and whether it does so side-specifically-is poorly defined. We quantified this concordance to confirm when CT might be deferred. Methods: Retrospective analysis of 73 children (33 girls, 40 boys; aged 5-17 years) treated surgically for CRS after failed medical therapy. CT was scored with the Lund-Mackay scale (LMS); a modified score (MLMS) accounting for sinus underdevelopment was also computed. Nasal endoscopy, performed by the same examiner the day before surgery, was scored separately for each side with the Lund-Kennedy system (discharge, edema and polyps, each 0-2; side score 0-6; total 0-12). Concordance was assessed with Spearman correlations (cluster bootstrap for side-level data), Cohen's kappa and sensitivity/specificity and Bland-Altman analysis after normalization to percentage of maximum. Results: Endoscopy correlated strongly with CT severity at the patient level (endoscopy total vs. LMS ρ = 0.76, p < 0.0001) and side-specifically (146 sides; ρ = 0.63, p < 0.0001); mucosal edema was the strongest individual sign. Bland-Altman analysis showed a systematic bias (endoscopy -10.6% relative to CT) and wide 95% limits of agreement (-32.6% to +11.4%), indicating that endoscopy ranks disease severity well but is not quantitatively interchangeable with CT at the individual level. Dichotomized side-specific agreement was more modest and improved as the CT threshold for a diseased side was raised (κ 0.21-0.50). Symptom burden correlated moderately with CT severity (ρ = 0.41). Children with polyps had higher LMS than those without (p = 0.007). Adenoid size, asthma and allergy were unrelated to CT severity. Only the agger nasi cell correlated with ipsilateral frontal disease (adjusted p = 0.017). Conclusions: A graded nasal endoscopic score is a strong ordinal correlate of CT severity in pediatric CRS, but Bland-Altman analysis shows it cannot quantitatively substitute for CT at the individual level. Endoscopy is therefore best used as a non-ionizing surrogate to grade severity, monitor disease and select which children need CT, while CT remains necessary for surgical planning. Because the cohort comprised children with advanced CRS, generalizability to milder disease requires confirmation.