Ultrasound-Based Estimation of Periodontal Probing Depth Using Category-Specific Supracrestal Tissue Attachment Calibration

OBJECTIVE: To determine category-specific supracrestal tissue attachment (STA) dimensions and apply these values to calibrate high-frequency ultrasound-derived measurements to match clinical probing depth (PD), and to examine the reliability of ultrasound landmark annotation in healthy and diseased cohorts. METHODS: Intraoral high-frequency ultrasound (40 MHz) imaging was performed to acquire images from 34 subjects (20 healthy and 14 with stage II-IV periodontitis). Four trained raters annotated 955 teeth for the gingival margin (GM), cemento-enamel junction (CEJ), and alveolar bone crest (ABC). The image-based gingival height (iGH) was defined as the distance from the gingival margin to the average bone crest. Ultrasound assessment is in terms of distances to the bone crest, whereas probing depth is measured to the supracrestal tissue attachment. Ultrasound measurements were calibrated using category-specific STA values to estimate probing depth. Method agreement between iPD and clinical PD was evaluated via Bland-Altman analysis using a ±1 mm clinical acceptance limit. This prospective calibration-validation study (34 subjects; 955 annotated tooth-surface observations) compared high-frequency ultrasound-derived measurements with conventional clinical periodontal probing depth measurements. The dataset was partitioned into a calibration set (721 observations, 27 subjects) and an independent validation set (234 observations, 7 subjects). RESULTS: Inter- and intra-rater reliability was high, with correlation coefficients ranging from 0.894 to 0.998. Calibrated STA values ranged from 1.349 mm for molar-buccal to 2.970 mm for anterior-buccal teeth, yielding an overall mean of 2.162±0.549 mm that falls within established histological reference ranges. Validation analysis (n = 234 teeth) demonstrated 80.8% agreement between iPD and clinical PD within ±1 mm, with a mean bias of -0.137 mm and limits of agreement of -1.718 to + 1.443 mm. CONCLUSIONS: Category-specific STA-calibrated high-frequency intraoral ultrasound enables clinically acceptable PD estimation across tooth types and surfaces without ionizing radiation or probe-induced discomfort. The observed level of agreement lies within the reported inter-examiner variability of conventional periodontal probing. ADVANCES IN KNOWLEDGE: This study represents the initial comprehensive characterization of supracrestal tissue attachment for all tooth types and surfaces by means of high-frequency ultrasound, which illustrates that calibration for each of these tooth type groups is essential for accurate estimation of probing depth. This approach may provide a non-invasive adjunct to periodontal probing assessment.