Long-term follow up on recurrence of 305 ameloblastoma cases.

The aim of this study was to determine the appropriate treatment for ameloblastoma by considering the factors associated with recurrence, and to make a quantitative prediction of the risk factors for recurrence. Data on age and gender distribution, location of the tumour, histopathological findings, treatment method, and whether or not patients had a preoperative biopsy confirmation report were collected in 305 cases (239 patients; M: 139, F: 100) of ameloblastoma diagnosed and treated in 1985-2002. After initial statistical evaluation (chi(2)-test and Fisher's exact test), logistic regression analysis was performed to check relative significance and predict recurrence. The disease-free survival function curves of the patients with or without recurrence were obtained by the Kaplan-Meier method and compared using univariate regression analysis. The correlation between recurrence and the treatment method or histopathological type was significant. The differences between the 'conservative' and 'resection with bone margin' and between the 'conservative' and 'segmental resection or maxillectomy' groups in terms of disease-free survival were highly significant. The difference between the 'resection with bone margin' and 'segmental resection or maxillectomy' groups was not significant. A resection with safety margin is the best method to treat most proven ameloblastomas, and conservative treatment is reasonable for patients in their first decade or with unicystic or plexiform ameloblastoma.