M A Mora1, A Mol, D A Tyndall, E M Rivera. 1. Department of Diagnostic Sciences and General Dentistry, UNC School of Dentistry, CB 7450, Chapel Hill, NC 27599-7450, USA.
Abstract
OBJECTIVES: To determine the effect of the number of basis images on the accuracy of local CT (LCT) in detecting longitudinal fractures and test the accuracy of terminal point assessment. METHODS: Longitudinal fractures were induced in 30 of 60 posterior teeth. LCT volumes were generated from 180, 60, 36 and 20 basis images. Ten observers determined the presence of a fracture and its terminal point. Receiver operating characteristic (ROC) analysis was used to quantify detection accuracy, and kappa statistics were used to quantify the concordance between actual and observed terminal point locations. RESULTS: A(z)-values were 0.91, 0.84, 0.74, 0.57 and 0.70 for LCT180, LCT60, LCT36, LCT20 and conventional radiography, respectively (ANOVA: P<0.0001). LCT180 and LCT60 outperformed the other modalities. The respective kappa values for terminal point assessment were 0.52, 0.40, 0.37, 0.16 and 0.32 (ANOVA: P=0.00). LCT180 and LCT60 provided better agreement. CONCLUSIONS: LCT maintained its efficacy for detecting longitudinal fractures with 60 instead of 180 basis images. Agreement between actual and observed terminal point locations was moderate for LCT60 and LCT180.
OBJECTIVES: To determine the effect of the number of basis images on the accuracy of local CT (LCT) in detecting longitudinal fractures and test the accuracy of terminal point assessment. METHODS:Longitudinal fractures were induced in 30 of 60 posterior teeth. LCT volumes were generated from 180, 60, 36 and 20 basis images. Ten observers determined the presence of a fracture and its terminal point. Receiver operating characteristic (ROC) analysis was used to quantify detection accuracy, and kappa statistics were used to quantify the concordance between actual and observed terminal point locations. RESULTS: A(z)-values were 0.91, 0.84, 0.74, 0.57 and 0.70 for LCT180, LCT60, LCT36, LCT20 and conventional radiography, respectively (ANOVA: P<0.0001). LCT180 and LCT60 outperformed the other modalities. The respective kappa values for terminal point assessment were 0.52, 0.40, 0.37, 0.16 and 0.32 (ANOVA: P=0.00). LCT180 and LCT60 provided better agreement. CONCLUSIONS: LCT maintained its efficacy for detecting longitudinal fractures with 60 instead of 180 basis images. Agreement between actual and observed terminal point locations was moderate for LCT60 and LCT180.