Kıvanç Kamburoğlu1, Funda Yılmaz2, Elif Naz Yeta3, Doĝukan Özen4. 1. Associate Professor, Department of Dentomaxillofacial Radiology, Faculty of Dentistry, Ankara University, Ankara, Turkey. Electronic address: dtkivo@yahoo.com. 2. Research Assistant, Department of Endodontics, Faculty of Dentistry, Ankara University, Ankara, Turkey. 3. Research Associate, Department of Dentomaxillofacial Radiology, Faculty of Dentistry, Ankara University, Ankara, Turkey. 4. Research Associate, Department of Biostatistics, Faculty of Veterinary Medicine, Ankara University, Ankara, Turkey.
Abstract
OBJECTIVE: To investigate observer ability to diagnose ex vivo simulated endodontic furcal perforations in root-filled teeth from cone beam computed tomography (CBCT) images using different artifact reduction algorithms. STUDY DESIGN: Our study consisted of 135 first maxillary molar teeth. In 89 teeth, furcation perforations were created using dental burs. Forty-six teeth without artificial perforations were used as controls. MTA Fillapex, Activ GP, and AH Plus were used with or without metal posts. All teeth were imaged using Planmeca ProMax 3-D Max CBCT, and four image modes were obtained as without artifact reduction and with artifact reduction in low, medium, and high modes. Images were evaluated by three observers for the presence or absence of furcation perforation using a five-point scale. Weighted kappa coefficients were calculated to assess observer agreement. Receiver operating characteristic analysis was performed. Areas under the curve (AUCs) were calculated for each image mode, observer, treatment group, and reading and were compared using Χ(2) tests, with a significance level of α = 0.05. The effects on diagnosis were calculated using analysis of variance (ANOVA). RESULTS: Intraobserver agreements for all observers ranged from 0.857 to 0.945. Kappa coefficients among different observers ranged from 0.673 to 0.763. AUC values ranged from 0.83 to 0.92, and there were no statistically significant differences (P > .05) between different CBCT image modes. Ratings in Activ GP treatment groups with or without posts showed statistically significant differences (P < .001). CONCLUSIONS: All CBCT image modes performed similarly in detecting furcal perforations near different root canal sealers with or without posts.
OBJECTIVE: To investigate observer ability to diagnose ex vivo simulated endodontic furcal perforations in root-filled teeth from cone beam computed tomography (CBCT) images using different artifact reduction algorithms. STUDY DESIGN: Our study consisted of 135 first maxillary molar teeth. In 89 teeth, furcation perforations were created using dental burs. Forty-six teeth without artificial perforations were used as controls. MTA Fillapex, Activ GP, and AH Plus were used with or without metal posts. All teeth were imaged using Planmeca ProMax 3-D Max CBCT, and four image modes were obtained as without artifact reduction and with artifact reduction in low, medium, and high modes. Images were evaluated by three observers for the presence or absence of furcation perforation using a five-point scale. Weighted kappa coefficients were calculated to assess observer agreement. Receiver operating characteristic analysis was performed. Areas under the curve (AUCs) were calculated for each image mode, observer, treatment group, and reading and were compared using Χ(2) tests, with a significance level of α = 0.05. The effects on diagnosis were calculated using analysis of variance (ANOVA). RESULTS: Intraobserver agreements for all observers ranged from 0.857 to 0.945. Kappa coefficients among different observers ranged from 0.673 to 0.763. AUC values ranged from 0.83 to 0.92, and there were no statistically significant differences (P > .05) between different CBCT image modes. Ratings in Activ GP treatment groups with or without posts showed statistically significant differences (P < .001). CONCLUSIONS: All CBCT image modes performed similarly in detecting furcal perforations near different root canal sealers with or without posts.