Evaluation of artificial intelligence based prosthetic innovation on dental implants: A database research

Accurate radiographic identification of dental implant systems remains challenging in clinical practice when implant documentation is unavailable and manual interpretation of panoramic radiographs is time-consuming and prone to error. Participants and methods deep learning performance was retrospectively evaluated in a database study for automated dental implant system recognition on panoramic radiographs using a labelled institutional image set. Model evaluation was based on accuracy, precision, recall and F1-score on a held-out test set. The top model reached a high diagnostic precision and only low rate of confounding between visually similar implant systems. Thus, we show the potential of using AI-supported implant recognition in prosthetic rehabilitations, especially when no implant documentation is available.