Related Experiment Videos
Application of computerised image analysis to monitoring endodontic therapy: reproducibility and comparison with
1Department of Cariology, Institute of Dentistry, University of Helsinki, Finland.
Objective:
To evaluate the intra- and interexaminer reproducibility of computerised image analysis for the assessment of periapical bone changes based on gray value ratio when compared with a visual method.
Methods:
353 radiographs obtained in a clinical endodontic study were analysed. Densitometric analyses were used to evaluate changes in bone density following endodontic treatment. The diseased area of apical periodontitis' (AP) and 'normal bone area' (AN) were defined and ratios of the gray values (AP/AN) were calculated for each radiograph. Periapical index (PAI) was used as the 'gold standard'. Correlation coefficient (r), Kappa (kappa), and t-test for matched pairs were used for the statistics.
Results:
The method proyed to be reproducible for intra-examiner re-examinations, r ranging from 0.7 to 0.84 (P < 0.01). Kappa showed fair to good agreement for data divided into four or three categories (kappa = 0.25-0.64). Dichotomised data showed fair to moderate agreement (kappa = 0.33-0.4) when the whole process was repeated, and good to very good agreement (kappa = 0.64-0.85) for the delineation process alone. The interexaminer comparisons showed fair to moderate agreement (kappa = 0.3-0.49). Comparison with PAI showed a 61% sensitivity and 79% specificity in detecting severe lesions (corresponding PAI scores 4 and 5) and the kappa value of 0.40 showed moderate agreement. The image analysis was sensitive enough to detect healing one month after treatment in a group of teeth with substantial lesions (P < 0.05).
Conclusions:
The computerised image analysis system seems to be an appropriate tool to detect the periapical bone changes in dental radiographs at a group level.