Concerns raised about AI in clear aligner therapy

Search Dental Tribune

Study raises concerns about AI in remote monitoring of clear aligner therapy

A recent study has identified concerns with the consistency of remote monitoring instructions in clear aligner therapy related to gauge compatibility and discrepancies in tooth position. (Image: greenbutterfly/Shutterstock)
Dental Tribune International

Dental Tribune International

Tue. 5. December 2023


BOSTON, US: The use of artificial intelligence (AI) in remote monitoring for clear aligner therapy is becoming increasingly popular and promising greater clinical control and an improved patient experience. In a recent study, researchers at Harvard School of Dental Medicine in the US evaluated AI-generated decisions about whether a patient is ready to move to the next aligner stage (Go or No-Go decisions) and measured what the systems identified as an unseat or tracking discrepancy. According to the researchers, the findings raise concerns about the consistency of remote monitoring instructions.

In addition to assisting with Go or No-Go decisions, remote monitoring technology identifies any misalignment between the tooth position and the aligners. The study aimed to evaluate the consistency of Go or No-Go recommendations made by the software and to quantify the 3D discrepancies that indicate an aligner misfit.

A total of 30 patients undergoing clear aligner treatment were scanned twice using a smartphone-based remote monitoring app, and the data was analysed for consistency. The study also included 24 clear aligner patients who had completed their treatment. These patients underwent both intra-oral scanning and remote monitoring scanning using their final aligners. The study compared the post-treatment intra-oral scans with the STL files representing the planned tooth positions at the final aligner stage in order to assess the maximum deviations between the actual and intended tooth positions.

The study observed a gauge compatibility of 44.7%. There was an 83.3% agreement in patient instructions between the first and second scans; however, there was no complete consensus on which teeth were or the number of teeth experiencing tracking issues. For patients advised to proceed with the next aligner via a Go instruction, the mean maximum discrepancies observed were 1.997 mm mesiodistally, 1.901 mm buccolingually, 0.530 mm occlusogingivally, and 8.911°, 7.827° and 7.049° for tipping, torque and rotation, respectively. These discrepancies were statistically similar to those in patients advised against proceeding (No-Go instruction), who showed discrepancies of 1.771 mm, 1.808 mm, 0.606 mm, 8.673°, 8.134° and 6.719° in the same respective dimensions.

“Despite the study’s limitations, these findings suggest concerns with the consistency of remote monitoring instructions because of gauge compatibility over the industry standard. Similarly, large discrepancies in tooth position for patients receiving Go and No-Go instruction suggest that artificial intelligence decisions were inconsistent with quantitative findings,” the authors wrote.

The study, titled “Assessment of artificial intelligence–based remote monitoring of clear aligner therapy: A prospective study”, was published in the August 2023 issue of the American Journal of Orthodontics and Dentofacial Orthopedics.

To post a reply please login or register