Dynamic Follow-up of Factors Influencing Implant Success and Models for Predicting Implant Outcomes

Nowadays, artificial intelligence technology with machine learning as the main means has been increasingly applied to the oral field, and has played an increasingly important role in the examination, diagnosis, treatment and prognosis assessment of oral diseases. Among them, machine learning is an important branch of artificial intelligence, which refers to the system learning specific statistical patterns in a given data set to predict the behavior of new data samples \[8\]. Machine learning is divided into two main categories: Supervised learning and Unsupervised learning. Whether there is supervision depends on whether the data entered is labeled or not. If the input data is labeled, it is supervised learning. Unlabeled learning is unsupervised. Supervised learning is a kind of learning algorithm when the correct output of the data set is known. Because the input and output are known, it means that there is a relationship between the input and output, and the supervised learning algorithm is to discover and summarize this "relationship". Unsupervised learning refers to a class of learning algorithms for unlabeled data. The absence of label information means that patterns or structures need to be discovered and summarized from the data set.

Starting from different data types, researchers built a variety of models to mine the data itself and predict the prognosis of the implant. Machine learning is often more impressive and intuitive in terms of images. In the field of oral implantology, researchers analyze preoperative image data based on machine learning to identify important anatomical structures (such as maxillary sinus, mandibular neural tube, etc.) and analyze alveolar bone quality. Large-scale imaging data is also used to identify the different implant systems on the market. Machine learning also plays an important role in the development of implant surgery plans, which is conducive to more accurate and efficient implantation surgery. The evaluation of implant retention rate and individual bone level is also one of the key clinical concerns. Most methods to study such issues are: Kaplan-Meier survival analysis, Cox survival analysis, etc., to study implant retention rate and influencing factors. Linear (mixed) model and multiple logistic regression were used to study the changes and influencing factors of bone absorption at implant edge. However, in daily clinical practice, there may be some practical problems such as lost follow-up and partial data missing. As the clinical scenarios of research become more and more clear, even partial data missing often leads to results that cannot be accurately evaluated and predicted. Therefore, in terms of supervised learning, this study aims to establish a predictive model of implant bone level change and evaluate the accuracy of the model through machine learning of implant edge bone level (MBL) with large amounts of data. In terms of unsupervised learning, the aim is to identify susceptibility phenotypes to implant failure through: clustering of individual-related information about implants.