High-resolution ultrasonography of gingival biomarkers for periodontal diagnosis in healthy and diseased subjects

ABSTRACT Periodontal disease affects nearly 50% of Americans but diagnostic methods have remained the same for decades. Periodontal examination via physical probing provides critical metrics such as pocket depth, clinical attachment level, and gingival recession; however, this practice is time consuming, variable, and often painful. In this study, we investigated high-frequency ultrasound (40 MHz) for the image-based measurement of periodontal metrics. Imaging was performed at midbuccal sites for a set of periodontally healthy (n = 10) and diseased (n = 6) subjects and image-based measurements were compared to gold-standard physical probing measurements. Human operators identified relevant markers (e.g., cementoenamel junction, gingival margin, alveolar bone crest) in B-mode ultrasound images from 66 teeth to calculate gingival height and alveolar bone level. These metrics were correlated to clinical measurements of probing pocket depth and clinical attachment level for disease staging (1.57-mm bias and 0.25-mm bias, respectively). Interoperator bias was negligible (<0.1 mm) for gingival height measurements and 0.45 mm for alveolar bone level measurements. The ultrasonographic measurements of gingival height and alveolar bone level served as effective diagnostic surrogates for clinical probing measurements while offering more detailed anatomical information and painless operation.