Radiologic Assessment of Periapical Health: Comparison of 3 Different Index Systems

IntroductionThe aim of this study was to evaluate observer variations and observer reproducibility in the interpretation of periapical health using the Strindberg system, the periapical index (PAI), and the probability index for radiologic assessment of periapical health.MethodsFour observers read digital periapical images to evaluate the periapical health of 200 root-filled teeth. Each observer evaluated a tooth twice by using the Strindberg system, twice by using the PAI, and twice by using the probability index. Each observation session was held at 1-month intervals. Three scores were used for the evaluations conducted by using the Strindberg system. Both the PAI and the probability index were considered on 5-point scales and were also dichotomized. SPSS for Windows 15.0 software (SPSS Inc, Chicago, IL) was used for data analysis, and intra- and interobserver agreements were described by using kappa statistics.ResultsThe average kappa values of intra- and interobserver agreement calculated for the Strindberg system were 0.53 and 0.36, respectively. The corresponding average kappa values were calculated as 0.48 and 0.39 for the PAI and 0.45 and 0.30 for the probability index. The highest mean proportion values of intra- and interobserver agreement were observed for the dichotomization of the PAI (88.9% and 87.0%, respectively) followed by the dichotomization of the probability index (86.8% and 82.9%, respectively).ConclusionsThe Strindberg system has lower intraobserver variation compared with the PAI and the probability index, whereas PAI has lower interobserver variation compared with the Strindberg system and the probability index. The dichotomization of the PAI and the probability index provided higher intra- and interobserver agreement values in the radiologic assessment of periapical health.