A Preoperative Grading System for Endoscopic Sinus Surgery

Importance Chronic rhinosinusitis (CRS) is common, yet one-third of patients do not achieve clinically meaningful improvement after endoscopic sinus surgery (ESS). Existing preoperative measures capture only a portion of the prognostic landscape, and no tool integrates patient-reported phenotype, symptom severity, and objective imaging into a single ordinal index for surgical prognosis. Objective To develop a CRS Severity Index (CRS-SI) that integrates patient-reported symptom phenotypes, characterized across symptom duration categories, with objective imaging measures to estimate the probability of clinically meaningful improvement after ESS. Design, Setting, and Participants Prospective cohort study with longitudinal data collected from academic tertiary care rhinology clinics between January 2020 and May 2025; follow-up was assessed at 3, 6, and 12 months after surgery. Participants were consecutive patients with CRS aged 16 to 95 years for whom medical management failed and who underwent ESS. Exposures Baseline 22-item Sino-Nasal Outcome Test (SNOT-22) measures, symptom phenotype dimensions derived from multiple correspondence analysis, polyp status, Lund-Mackay, and Lund-Kennedy scores. Main Outcomes and Measures The primary outcome was clinically meaningful improvement, defined as SNOT-22 improvement of at least 8.9 points from baseline to follow-up. Seven candidate predictors were organized into clinical (baseline SNOT-22, latent symptom phenotype dimensions, polyp status) and imaging (Lund-Mackay, Lund-Kennedy, frontal opacification) arms. CRS-SI was derived using sequential sequestration and conjunctive consolidation. Results Of 488 eligible patients with at least 1 postoperative SNOT-22, 450 with complete clinical and imaging grading data comprised the analytic cohort. Among 450 patients (mean [SD] age, 55.5 [16.3] years; 54.7% [246 of 450] male), 64% achieved clinically meaningful improvement. Of 7 candidate predictors, 6 met bivariate screening criteria (rank-biserial |r|≥0.1 with 95% CI excluding 0). Sequential sequestration selected baseline SNOT-22 severity and a headache and/or chemosensory–dominant symptom phenotype as clinical classifiers, and Lund-Mackay and Lund-Kennedy scores as imaging classifiers. The 4-grade CRS-SI produced a monotonic improvement gradient (grade 1, 81% [95% CI, 72%-88%] to grade 4, 29% [95% CI, 19%-40%]), with a risk difference of 52.5 (95% CI, 39.3-65.7) percentage points between best and worst grades. Discrimination was moderate (C statistic, 0.692; 95% CI, 0.643-0.740) and exceeded imaging grading alone; bootstrap calibration slope, 1.00 (95% CI, 0.77-1.31). Performance was consistent across CRS phenotypes and time points. Conclusions and Relevance In this study, a 4-level CRS Severity Index integrating patient-reported symptom profiles with computed tomography and endoscopic imaging stratified improvement after ESS across a 52 percentage–point gradient. If validated externally, the CRS-SI may support preoperative counseling and clinical trial stratification.