Background A staging system is essential for determining the optimal surgical approach and predicting postoperative outcomes for inverted papilloma (IP). Although staging systems based on the extent to which the location is occupied by an IP have been widely used, an origin site-based classification of IP using unsupervised machine learning algorithms has recently been reported. Objective To determine the most appropriate of five staging systems for sinonasal IP by comparing recurrence rates for each stage according to each of those systems. Methods Eighty-seven patients with sinonasal IP were enrolled in the study. Their tumors were retrospectively categorized according to the Krouse, Oikawa, Cannady, and Han staging systems, which are based on the extent of IP, and the Meng system, which is based on the site of origin. The rates of recurrence for each stage of the five systems were compared. Results Seven of the 87 patients (8.0%) had recurrences during an average 45.5 months (12–138 months) of follow-up. There were significant differences in disease-free survival between the stages specified by Han and Meng (p = 0.027 and p < 0.001, respectively), but not between the stages specified by Krouse, Oikawa, and Cannady (p = 0.236, 0.062, and 0.130, respectively). Cox proportional hazard models revealed that Meng system (adjusted hazard ratio [aHR] 4.32, 95% confidence interval [CI] 1.10–17.04) and presence of dysplasia (aHR 7.42, 95% CI 1.15–47.85) were significantly associated with recurrence. Conclusion The staging systems proposed by Han and Meng were found to be accurate in terms of tumor recurrence. We recommend use of the Han staging system before surgery and the Meng system after intraoperative identification of the origin of the tumor.
Clinical Analysis of Sinonasal Inverted Papilloma according to the Staging System
