Use of Spherical Segmentation Radiomic Analysis for Differentiation of Odontogenic Keratocysts and Radicular Cysts

Aim: The primary objective of this investigation was to assess the capability of radiomics in distinguishing between odontogenic keratocyst (OKC), characterized
by aggressive clinical behavior and a heightened recurrence rate, and radicular cyst (RC), an inflammatory odontogenic cyst. Additionally, this study employed an experimental spherical segmentation technique, strategically designed to unveil the inherent characteristics of the lesions under scrutiny.

Material and Methods: This retrospective group comparative research evaluated histopathologically verified cases of OKC and RC which were retrieved from archives of the Department of Oral and Maxillofacial Radiology at Kocaeli University Faculty of Dentistry. Cases possessing lesions with intact osseous boundaries and with no evidence of any perforation were included in the study. In 3D Slicer software, a total of 93 radiomic features were extracted using the SlicerRadiomics plugin for the segmented area of the lesions.

Results: The CBCT data of a total of 10 OKC and 10 RC cases were analysed. The radiomics scores obtained from the calculations had a sensitivity value of 80%,and a specificity value of 80% in predicting radicular cysts. The model was found to accurately detect radicular cysts at a rate of 80% and OKC at a rate of 80%. The radiomics score was higher in patients with OKC to those with RC (OR: 25.251; p=0.018).

Conclusion: The favorable outcomes of this study substantiate the potential utility of radiomics which may be useful before surgical intervention of RC and OKC lesions with similar radiologic features.