Development and evaluation of an automatic tumor segmentation tool: a comparison between automatic, semi-automatic and manual segmentation of mandibular odontogenic cysts and tumors.

INTRODUCTION: In the treatment of cancer in the head and neck region, computer-assisted surgery can be used to estimate location and extent by segmentation of the tumor. This article presents a new tool (Smartbrush), which allows for faster automated segmentation of the tumor. METHODS: This new method was compared with other well-known techniques of segmentation. Thirty-eight patients with keratocystic odontogenic tumors were included in this study. The tumors were segmented using manual segmentation, threshold-based segmentation and segmentation using Smartbrush. All three methods were compared concerning usability, time expenditure and accuracy. RESULTS: The results suggest that segmentation using Smartbrush is significantly faster with comparable accuracy. CONCLUSIONS: After a period of adjustment to the program, one can comfortably get reliable results that, compared with other methods, are not as dependent on the user's experience. Smartbrush segmentation is a reliable and fast method of segmentation in tumor surgery.