Methods: Thirty-six cases of ameloblastomas and forty-seven cases of odontogenic keratocysts were selected. All the cases were confirmed histopathologically and taken panorama and CT. Six oral and maxillofacial radiologists diagnosed the lesions by 3 methods: using panorama, using CT, and using panorama and CT. The observers were divided 3 groups, having experience over 10 years in Department of Oral and Maxillofacial Radiology (group 1), having experience several years (group 2), and in the process of resident (group 3). After over 2 weeks, the observers diagnosed them by same manners. Receiver operating characteristic (ROC) curves and kappa statistics were analyzed for diagnostic accuracy and intraobserver agreement (SPSS, ver10.0).
Results: The ROC value except for group 3 showed the highest with interpretation using panorama and CT, followed by interpretation using CT only, and was the lowest with interpretation using panorama only. The overall difference was not found in diagnostic accuracy among groups in using panorama only, but there was difference in diagnostic accuracy of group 1 and 2 vs. 3 in using CT only, and combination panorama and CT.
The intraobserver agreement of group 1 and 2 were moderate-to-good, but those of group 3 were fair-to-moderate. The intraobserver agreement of using panorama only and using CT only were fair-to-good, but those of using panorama and CT were moderate-to-good.
Conclusion: To differentiate between ameloblastoma and odontogenic keratocyst more accurately, the experienced interpreters in Department of Oral and Maxillofacial Radiology should diagnose with combination of panorama and CT.