Tooth Color Measurement Reproducibility with Digital Imaging

Objective: This research evaluated clinical tooth color measurement reproducibility of a digital imaging system. Methods: After training, a new operator collected repeat tooth color measures from 20 healthy dentate adults over a two-day period. These images were collected using a common method using a high resolution digital camera (JVC CCD), zoom lens and fixed lighting conditions. Subject orientation was achieved using a chin rest, and paired images were captured over a two-day period in order to introduce subject repositioning as part of the reproducibility assessment. For each image, maxillary anterior tooth pixels were classified and counted, and average L*a*b* tooth colors were derived using standard formulas. Intra-class correlations (ICC) and 95% lower confidence bounds (LCB) were calculated using a 0-to-1 scale, where 0 represented no agreement and 1 represented perfect agreement. Results: All 20 subjects (mean age=30 years) attended both visits, and all data were included in the analysis. Tooth pixel count differed by less than 0.022% between days. For color, the Day 1 and Day 2 image pairs had means (SD) b* of 14.7 (1.46) and 14.7 (1.49). Mean (SD) L* was 76.1 (1.84) on Day 1, and 76.1 (1.92) on Day 2. The measurements exhibited considerable reproducibility across visits. Pixel count ICC (95% LCB) was 0.985 (0.967). For color, the ICC (95% LCB) was 0.981 (0.960) for b* yellowness and 0.985 (0.968) for L* lightness. Similar reproducibility was demonstrated for a* (redness). Conclusion: This research demonstrates that digital image analysis yields highly reproducible clinical measurement of tooth pixels and tooth color, in this case, with a newly trained system operator.