RESUMEN
PURPOSE: There is no consensus on the ideal treatment of odontogenic keratocysts (OKCs). Certain risk factors may modify the disease recurrence rate. The purpose of this study was: 1) to estimate the incidence of OKC recurrence in a statistically rigorous manner; and 2) to identify risk factors associated with OKC recurrence. METHODS: The investigators designed and implemented a retrospective cohort study and enrolled a sample derived from the population of patients presenting to the Department of Oral and Maxillofacial Surgery outpatient clinics at the University of Washington School of Dentistry and Harborview Medical Center for evaluation and management of OKCs between January 1, 2010, and December 31, 2020. Predictor variables included demographics (age, gender), radiographic characteristics (location of lesion, locularity, size, and cortical perforation), and type of operation performed (decompression + cystectomy, enucleation ± adjuvant therapy, or resection). The primary outcome variable was time to recurrent disease, defined as the time from treatment to radiographic or clinical evidence of a new lesion (recorded in months). Kaplan-Meier analysis was used to estimate median time to recurrence, and Cox proportional hazards models were used to identify covariates statistically associated with recurrent disease (P ≤ .05). RESULTS: The sample was composed of 60 subjects with 63 previously untreated lesions. Eight subjects (13%) developed a recurrent lesion during the study interval with a median time to recurrence of 31 (interquartile range, 24 to 48) months. One of the 8 recurrences occurred within 12 months of treatment and 7 of the 8 recurrences occurred more than 21 months after treatment. Based on using Kaplan-Meier analysis, the 5-year incidence of disease recurrence was estimated to be 34%. Lesions with cortical perforation were 8.3 times more likely to recur (95% confidence interval [1.7, 41.3]; P-value = .01), and multilocular lesions were 10.6 times more likely to recur (95% confidence interval [1.3, 86.9]; P-value = .03). The sample size was the limiting factor to performing regression analyses. CONCLUSIONS: Virtually every publication on OKCs to date reports frequencies of disease recurrence rather than applying appropriate survival analyses commonly used to estimate outcomes in cancer research. Failure to use the appropriate statistical analyses underestimates the risk of disease recurrence. Our study is no exception. The estimated frequency of disease recurrence during the study interval was 13% (8/60). When survival analyses are applied that account for varying months of follow-up, the incidence of disease recurrence is 34 per 100 per 5 years. We recommend the application of time-to-event analyses in the study of disease entities with the potential for recurrence.
