Reproducibility of subtraction radiography in monitoring changes in approximal carious lesions in children

Background: Laboratory studies have shown that digital subtraction radiography (DSR) can be a more effective tool, than conventional radiography, in monitoring changes in carious lesions. The clinical performance of the technique, however, has not been sufficiently researched. Aim: To compare the reproducibility of DSR to that of bitewing radiographs, in monitoring changes in approximal caries in the mixed dentition. A secondary aim was to determine whether assessment outcomes differed as a function of the method used. Design: Six assessors evaluated 310 lesions first on bitewings, then with DSR. The overall reproducibility was evaluated via intra-class correlation coefficient (ICC). Intra- and inter-rater reproducibility were assessed using weighted Kappa. Paired t test was used to assess differences in the reproducibility across methods. Results: The overall reproducibility for DSR was (ICC = 0.47, 95% CI = 0.31-0.56). Intra- and inter-rater reproducibility were 0.65 and 0.44, respectively. The overall reproducibility for bitewings was ICC = 0.45, 95% CI = 0.42-0.57. Intra- and Inter-rater reproducibility were 0.71 and 0.46, respectively. Differences in the reproducibility across methods were not statistically significant. Significantly more lesions were scored as progressed using DSR. Conclusions: The reproducibility of DSR in monitoring changes in approximal caries is comparable to that of bitewings. Additionally, DSR detected higher proportion of progression than bitewing assessments