Comparison of 3D cephalometric evaluations: current AI systems versus human providers

Cephalometric analyses are an important aspect of orthodontic and orthognathic treatment planning. Correct landmark placement and measurements are required to produce adequate treatment. This has traditionally been performed by practitioners manually or semi-automatically. Recently in dentistry, completely automated methods of 2 Dimensional and 3 Dimensional cephalometric analyses have become available leading to reduced treatment planning times and more consistent results. Many studies have been performed evaluating automated systems that were developed by the particular authors of the studies, however, more commercial software programs are integrating these features into their user interfaces. To date there have been no studies that the authors of this article have found that evaluate the accuracy and efficacy of the “Auto tracing” feature included the “3D analysis” addon in Invivo 7 (Osteoid a dental branch of Anatomage, Santa Clara, CA). So, 40 commonly utilized cephalometric landmarks were evaluated in 40 datasets by Two third year Oral and maxillofacial radiology residents, and the Auto tracing feature. X, Y, and Z coordinates for all the datapoints were collected and subjected to statistical analyses with one of the residents Rater A being set as the ground truth and Rater B and the Auto tracing being compared to the ground truth. The results of the study showed significant differences only in Three of the 40 landmarks evaluated when comparing Rater B and the Auto tracing. This suggests that there is no significant difference between human providers and the Auto tracing feature included in the 3D analysis add-on.