Differentiation of developmental and post-orthodontic white lesions using image analysis

The aim of this study was to investigate differences in shape and size characteristics between developmental opacities and post-orthodontic white enamel lesions using computerized image analysis. Material, in the form of 35 mm slides, was obtained from the archive of photographic patient records in the orthodontic clinic at the Charles Clifford Dental Hospital. Images of 30 teeth with developmental white lesions and 30 teeth with post-orthodontic white lesions were selected using strict inclusion and exclusion criteria. The slides were converted to a digital format, coded, placed in a random order and analysed blindly using a computerized image analysis system by one clinician. After a 2 week interval, the images were recoded, placed in a new random order and the measurements repeated. The outcome measures were: area and luminance proportionality, and the shape of the perimeter line (expressed as the mathematical factor, roundness). Reproducibility was assessed by a paired samples t-test for systematic error and the intra-class correlation coefficient (ICC) for random error. Differences between groups were tested using the Mann-Whitney U-test for non-parametric data. Reproducibility was substantial for all measurements except for developmental white lesion roundness, which was moderate. There was a statistically significant difference between developmental white opacities and post-orthodontic white lesions for measurements of luminance intensity, proportionality (P = 0.002) and roundness (P = 0.001). Developmental white opacities had a higher luminance (i.e. were whiter) and the boundaries were more circular in shape than the post-orthodontic lesions. Roundness is a useful measure when distinguishing developmental and post-orthodontic demineralization.