Quantifying enamel demineralization from teeth with orthodontic brackets--a comparison of two methods. Part 1: repeatability and agreement

The aim of this investigation was to compare the repeatability of measuring enamel demineralization surrounding an orthodontic bracket using two techniques: computerized image analysis from digitally converted photographic slides and quantitative light-induced fluorescence (QLF). Fifteen human molars were halved and shaped to look like incisors. The teeth were individually numbered and orthodontic brackets bonded to the buccal surface. The crowns were covered with acid resistant varnish, except for windows approximately 1.5 x 3 mm adjacent to the gingival, occlusal, mesial, and distal edges of the bracket. The windows were variously exposed to a demineralizing gel for 0, 3, 7, or 14 days, and the acid resistant varnish was removed. Standardized photographic slides and QLF images of the teeth were taken. These were repeated after 1 week. The slides were converted to grey scale digital format and analysed using Image-Pro Plus 3.0. The QLF images were stored, processed, and analysed using customized software. All images were recoded for blind analysis. The four surfaces of the bracket were inspected and only areas of suspected demineralization were analysed. This was repeated after 1 week. The limits of agreement and mean difference between repeat readings of the area of demineralization were similar for both techniques (-0.04 +/- 0.43 for photographs and -0.10 +/- 0.63 for QLF). Mean grey level (photographs) and mean loss of fluorescence from that area (deltaF) (QLF) showed acceptable limits of agreement. The Intra Class Correlation (ICC) was below 0.81 for the measurement of area from QLF, suggesting that random error needs to be reduced. There was evidence of systematic bias for the repeat readings of the grey levels from the photographs (P < 0.001). Enamel demineralization surrounding an orthodontic bracket can be measured reproducibly using these two techniques.