Diagnostic reproducibility of the 2018 Classification of Gingival Recessions: Comparing photographic and in‐person diagnoses

Background: To assess how the diagnostic reproducibility of the 2018 Classification of Gingival Recession Defects (GRD) could be applied when comparing in-person chairside measurements with photographic measurements. Methods: Thirty-four GRD were photographed and evaluated by 4 masked operators. For each case, the operators measured twice recession type (RT), recession depth (RD), keratinized tissue width (KTW), gingival thickness (GT), detectability of the cemento-enamel junction (CEJ), and presence of root steps (RSs), chairside, and on photographs. Intraclass correlation coefficient (ICC) with 95% confidence intervals (CI) was calculated for RD and KTW; Kappa with 95% CI was used for GT, CEJ, and RS; quadratic weighted Kappa with 95% CI was used for RT. Results: RD, KTW, and RT showed excellent overall intra-operator agreement (> 0.93), and from good to excellent overall inter-operator agreement (> 0.80), for both clinical and photographic measurements. Agreements were lower for GT, CEJ, and RS. Overall clinical and photographic agreements were within 0.1 difference for every variable, except for inter-operator agreement for RS which was 0.72 for clinical measurements and 0.45 for photographic measurements. The lowest overall agreement between clinical versus photographic measurements existed for CEJ (0.28) and RS (0.35). Conclusions: Variables composing the 2018 Classification of GRD are reproducible, both clinically and on photographs, with comparable agreements. The overall agreement was higher for KTW, RD, and RT, and lower for GT, CEJ, and RS, for both clinical and photographic measurements. The comparison between chairside and photographic evaluations indicated fair to excellent agreement for most variables, with CEJ and RS showing fair agreement.