Related Experiment Video
Updated: May 9, 2026
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Reliability of direct and indirect clinical attachment level measurements
Priscila Corraini1, Vibeke Baelum, Rodrigo Lopez
1Department of Dentistry, Section of Periodontology, Health, Aarhus University, 8000 Aarhus C, Denmark. p.corraini@odontologi.au.dk
Aims:
To investigate the influence of subject-, tooth- and site-level characteristics on intra-examiner reproducibility of direct and indirect clinical attachment level (CALDIR and CALIND ) recordings, and to quantify the impact of less-than-perfect reliability on our ability to assess periodontitis progression.
Materials And Methods:
Within a 1-week interval, a single examiner performed duplicate probing depth (PD), CALDIR and gingival recession (GR) recordings in six sites of all teeth present in 148 periodontitis patients. CALIND was calculated on the basis of the PD and GR.
Results:
Agreement was observed in 65%, 62%, 69% and 84% of the duplicate CALDIR , CALIND , PD and GR recordings, respectively, and >95% of the differences were within ±1 mm. This study identified multiple predictors for CAL measurement errors ≥1 mm, including tooth and site location, presence of supra- and subgingival calculus, bleeding on probing and suppuration. Measurement errors were more likely in patients diagnosed with "extensive" rather than "less extensive" periodontitis. In over half of the patients, measurement error frequencies were too high to allow for detection of possible CAL changes ≥2 mm with a false-positive rate ≤5%. Detection of CAL changes with low false-positives was more likely using recordings obtained by the direct method.
Conclusions:
From a measurement error point of view, CALDIR recordings are preferable over CALIND .
