A study of accuracy and reliability of intraoral ultrasound using ex vivo and in vivo data

Abstract

Objective

This study aimed to evaluate the reliability and accuracy of intraoral ultrasound (US) imaging for assessing periodontal parameters.

Study Design

Measurements from human cadavers and orthodontic patients were used to evaluate the accuracy and reliability of high-frequency (20 MHz) intraoral ultrasonography. First, measurement accuracy was assessed on US and micro-computed tomography (μCT) images for the maxillary and mandibular teeth from 3 cadavers (aged 62, 70, and 75 years), which were donated by the Department of Anatomy at the University of Alberta. Fifty pairs of US and corresponding μCT images were selected for measurements. Two trained raters measured the alveolar bone level (ABL) between the cementum-enamel junction and the alveolar bone crest (ABC), and the thickness of the alveolar bone near its crest (ABT) on both types of images, following a 2-week calibration exercise. The accuracy was evaluated using t tests, and Bland-Altman plots. Second, reliability evaluation was conducted on clinical US images of 134 teeth including central incisors, lateral incisors, canines, and premolars from 19 adolescent orthodontic patients. Three raters performed measurements of 4 parameters twice on the clinical US images: ABL, ABT, the gingiva thickness close to the ABC (GT1), and the gingiva thickness 2 mm away from the gingival margin (GT2). The intra- and inter-rater reliabilities were evaluated using the intraclass correlation coefficient.

Results

Accuracy evaluation revealed low bias in Bland-Altman analysis, and the t tests had P values > .2. Reliability evaluation demonstrated excellent intraclass correlation coefficient scores (>0.9) for ABL, GC1, and GC2, and good scores (>0.8) for ABT measurements in both inter- and intra-rater analyses on US.

Conclusion

The accuracy analysis on the cadavers showed no statistically significant difference in measurements between US and μCT. The reliability analysis on clinical data demonstrated high intra- and inter-rater reliabilities for the periodontal measurements.