The precision of gingival recession measurements is increased by an automated curvature analysis method

• Main Authors: Marko Kuralt, Rok Gašperšič, Aleš Fidler
• Format: Article
• Language: English
• Published: BMC 2021-10-01
• Series: BMC Oral Health
• Subjects: Gingival recession; Computer-assisted image processing; Dental models; Cementoenamel junction; Observer variation; Computer-assisted image interpretation reproducibility of results
• Online Access: https://doi.org/10.1186/s12903-021-01858-9

Abstract Background The extent of gingival recession represents one of the most important measures determining outcome of periodontal plastic surgery. The accurate measurements are, thus, critical for optimal treatment planning and outcome evaluation. Present study aimed to introduce automated curvature-based digital gingival recession measurements, evaluate the agreement and reliability of manual measurements, and identify sources of manual variability. Methods Measurement of gingival recessions was performed manually by three examiners and automatically using curvature analysis on representative cross-sections (n = 60). Cemento-enamel junction (CEJ) and gingival margin (GM) measurement points selection was the only variable. Agreement and reliability of measurements were analysed using intra- and inter-examiner correlations and Bland–Altman plots. Measurement point selection variability was evaluated with manual point distance deviation from an automatic point. The effect of curvature on manual point selection was evaluated with scatter plots. Results Bland–Altman plots revealed a high variability of examiner’s recession measurements indicated by high 95% limits of agreement range of approximately 1 mm and several outliers beyond the limits of agreement. CEJ point selection was the main source of examiner’s variability due to smaller curvature values than GM, i.e., median values of − 0.98 mm− 1 and − 4.39 mm− 1, respectively, indicating straighter profile for CEJ point. Scatter plots revealed inverse relationship between curvature and examiner deviation for CEJ point, indicating a threshold curvature value around 1 mm− 1. Conclusions Automated curvature-based approach increases the precision of recession measurements by reproducible measurement point selection. Proposed approach allows evaluation of teeth with indistinguishable CEJ that could be not be included in the previous studies.