Can artificial intelligence training improve clinical decision-making during deep caries excavation?

Area of Science:

• Dental Education
• Artificial Intelligence in Dentistry
• Clinical Decision Support Systems

Background:

• Deep carious lesions pose a risk of pulp exposure during excavation.
• Accurate prediction of pulp exposure is crucial for appropriate treatment planning.
• Artificial intelligence (AI) platforms offer potential decision support for predicting pulp exposure.

Purpose of the Study:

• To evaluate the impact of a structured educational intervention on dental students' use of an AI platform for pulp exposure prediction.
• To assess whether training improves students' interaction and agreement with AI recommendations before deep caries excavation.

Main Methods:

• A randomized controlled trial (RCT) involving 18 dental students.
• Experimental group received personalized AI platform training; control group received a video.
• Participants assessed deep carious lesions using radiography and the AI system.

Main Results:

• Trained students showed significantly greater agreement with AI recommendations compared to controls.
• No meaningful differences were observed in diagnostic accuracy, sensitivity, or specificity between groups.
• Response times were slightly reduced in the trained group.

Conclusions:

• Short, personalized AI training can increase dental students' reliance on AI predictions for pulp exposure.
• The intervention did not substantially improve objective diagnostic performance.
• Further research is needed to explore AI training's impact on clinical decision-making and patient outcomes in managing deep caries.