Texture analysis of cone-beam computed tomography images assists the detection of furcal lesion

Abstract

Background: The aim of this study was to apply texture analysis (TA) to cone-beam computed tomography (CBCT) scans of patients with grade C periodontitis for detection of non-visible changes in the image. Methods: TA was performed on CBCT scans of 34 patients with grade C periodontitis. Axial sections of CBCT were divided into three groups as follows: Group L (lesion) in which there is a furcal lesion with periodontal bone loss; Group I (intermediate) in which the border of the furcal lesion has normal characteristics; and Group C (control) in which the area is healthy. Eleven texture parameters were extracted from the region of interest. Mann-Whitney U test was used to assess the differences in the texture between the three groups as follows: L versus I; L versus C, and I versus C. Results: Statistically significant differences (P <0.05) were observed in almost all parameters in the intergroup analyses (i.e., L versus I and L versus C). However, statistical differences were smaller in groups I versus C in which only entropy of sum, entropy of difference, mean of sum, and variance of difference were statistically different (P < 0.05). Conclusion: TA can potentially provide prognostic information to improve the diagnostic accuracy in the grading of the tissue around the furcal lesion, thus potentially accelerating the treatment decision-making process.