Investigators may have identified trends and opportunities in the use of artificial intelligence models to assist in the planning and preparation of dental prosthetics, according to a systematic review published in Medical Science Monitor.
Dental prosthodontics is often supported by computer-assisted technology. Machine learning models are often employed to recognize patterns in data sets for classification, regression, and clustering. However, there are currently gaps in the literature regarding the assessment of AI.
Investigators used January 2012 to January 2024 data from the Web of Science, ScienceDirect, PubMed and Cochrane Library to identify and analyze 30 in vitro and clinical studies reporting on the applications and performance of AI in dental prosthodontics in accordance with PRISMA guidelines. They examined the study findings for the efficacy of AI across prosthodontic diagnosis, treatment outcome prediction, and treatment planning.
The investigators then used the four domains of Quadas-2—patient selection, index test, reference standard, and flow and timing—to evaluate the quality and risk bias of each of the studies included in the analysis.
In some of the studies, AI was used to diagnose prosthodontic cases, assess panoramic radiographs for prosthesis detection, identify the type of restoration performed on molars, classify dental arches on the basis of edentulism status, and detect the presence or absence of teeth.
Investigators identified dental charting, tooth shade selection, automated restoration design, mapping of the preparation finishing line, manufacturing casting optimization, prediction of facial changes in patients with removable prostheses, and removable partial denture design as areas in which AI has already shown efficacy.
They found that 18.3% (n = 7/38) of the studies had a low risk of bias. However, because 52.6% (n = 20) and 28.9% (n = 11) of the studies had a respective high and unclear risk of bias, more rigorous studies may be needed to validate whether AI may be effectively utilized in these areas.
The investigators concluded that AI could help improve clinical service efficiency, assist dentists with clinical decision-making, and speed up the time spent interpreting clinical findings.
No conflicts of interest were disclosed.