Dental fluorosis in populations from Chiang Mai, Thailand with different fluoride exposures - paper 2: the ability of fluorescence imaging to detect differences in fluorosis prevalence and severity for different fluoride intakes from water

Abstract

Background: To assess the ability of fluorescence imaging to detect a dose response relationship between fluorosis severity and different levels of fluoride in water supplies compared to remote photographic scoring in selected populations participating in an observational, epidemiological survey in Chiang Mai, Thailand.

Methods: Subjects were male and female lifetime residents aged 8-13 years. For each child the fluoride content of cooking water samples (CWS) was assessed to create categorical intervals of water fluoride concentration. Fluorescence images were taken of the maxillary central incisors and analyzed for dental fluorosis using two different software techniques. Output metrics for the fluorescence imaging techniques were compared to TF scores from blinded photographic scores obtained from the survey.

Results: Data from 553 subjects were available. Both software analysis techniques demonstrated significant correlations with the photographic scores. The metrics for area effected by fluorosis and the overall fluorescence loss had the strongest association with the photographic TF score (Spearman's rho 0.664 and 0.652 respectively). Both software techniques performed well for comparison of repeat fluorescence images with ICC values of 0.95 and 0.85 respectively.

Conclusions: This study supports the potential use of fluorescence imaging for the objective quantification of dental fluorosis. Fluorescence imaging was able to discriminate between populations with different fluoride exposures on a comparable level to remote photographic scoring with acceptable levels of repeatability.

Figure 1

Images demonstrating fluorosis analysis. a. Conventional digital image of a subject presenting with fluorosis. b. Fluorescence image captured demonstrating fluorosis (areas of florescence loss). c. Output from analysis using existing technique. d. Output from analysis using convex hull technique. (Image adjusted for contrast for illustrative purposes).

Figure 2

The photographic score demonstrating separation of the intervals for cooking water fluoride content, suggestive of a dose response. TF scores of 4 or higher have been grouped together as 4+.

Figure 3

Boxplot with error bars (SD) for ΔQblur. Outliers (subject 837) highlighted.

Figure 4

Boxplot with error bars (SD) for ΔQch. Outliers (subjects 837 and 230) highlighted.

Figure 5

ROC curve for convex hull software.

Figure 6

ROC curve for existing technique software.

Figure 7

Images of subjects with confounding factors for QLF. a. clinical photograph subject 837 presenting with non-fluorotic hypomineralization and enamel loss on maxillary right central incisor. b. QLF image of subject 837. Note the pattern of fluorescence loss on the maxillary right central incisor typical of enamel loss with possible caries. The areas in red indicate presence of plaque stagnation. c. Clinical photograph of subject 230 presenting with confluent areas of fluorosis with pitting and staining. d. QLF image of subject 230. Areas of fluorosis with stain exhibit greater fluorescence loss. e. Clinical photograph subject 545 presenting with confluent fluorosis and enamel loss and possible caries. f. QLF image of subject 545. Note the loss of fluorescence in the areas of enamel loss.