Diagnosis of Tooth Prognosis Using Artificial Intelligence.

Area of Science:

• Dental Prognosis
• Artificial Intelligence in Dentistry
• Machine Learning Applications

Background:

• Accurate tooth prognosis is crucial for comprehensive dental treatment planning.
• Existing methods for prognosis assessment can be subjective and vary among clinicians.
• The Harvard School of Dental Medicine Comprehensive Treatment Planning Curriculum (HSDM-CTPC) provides a framework for evaluation.

Purpose of the Study:

• To develop an effective artificial intelligence (AI)-based module for accurate tooth prognosis prediction.
• To integrate AI decision-making into the dental treatment planning process.
• To validate AI algorithms against expert consensus using established clinical factors.

Main Methods:

• Evaluated prognosis for 2359 teeth from 94 cases using a 1-5 level scale.
• Utilized 17 clinical factors derived from the HSDM-CTPC.
• Developed and compared three AI machine learning classifiers: gradient boosting, decision tree, and random forest.
• Validated AI models against a gold standard established by three experienced prosthodontists.

Main Results:

• The decision tree classifier achieved the highest accuracy: 0.8413 (Model-A) and 0.7523 (Model-B).
• Gradient boosting classifier accuracy was 0.6896 (Model-A) and 0.6687 (Model-B).
• Random forest classifier accuracy was 0.8413 (Model-A) and 0.7523 (Model-B).

Conclusions:

• The decision tree classifier demonstrated superior performance in predicting tooth prognosis.
• AI implementation, particularly using decision trees, can enhance the objectivity and accuracy of tooth prognosis assessment.
• This AI module supports informed dental treatment planning by providing reliable prognosis predictions.