Neural networks-based prediction of gingival recession in camouflage-based orthodontic treatment

• Author Details:   
• Daphane Anishya
• Prasanna Arvind
• Pradeep Kumar ^*

Background and Aim: Recent advances in Artificial Intelligence (AI) and machine learning have provided dental professionals with excellent tools for forecasting the possibility of gingival recession during or after skeletal discrepancy concealment treatment. Machine learning is an effective decision-support tool for estimating the likelihood of gingival recession in treating certain skeletal malocclusions. The current study used artificial intelligence to predict the gingival recession during camouflage orthodontic treatment to disguise skeletal disparities.

Materials and Methods: Sixty-five patients with minor skeletal discrepancies treated by camouflage orthodontic treatment were selected for the study. Four factors were considered -crowding of lower anteriors,proclination of lower anteriors, spacing due to extraction of lower interior’s, and canines (high risk). YES or NO were given whether the previously mentioned characteristics were present in the subjects. Orange, a machine learning tool that uses neural networks, was used to assess prediction accuracy. Test data and training and were split 80/20. Cross-validation, confusion matrix, and ROC analysis assessed model performance. This study examined precision and recall.

Results: The accuracy of the neural networks is 92%. CA (Classification Accuracy) rate of 87.5% implies that predictions were correct in 87.5% of situations.

Conclusion: Artificial intelligence solutions are intended to increase orthodontic performance and care quality. Current applications include recognizing cephalometric landmarks, categorizing skeletal components, and deciding on tooth extractions. Artificial intelligence solutions for anticipating periodontal difficulties in disguised orthodontic treatment are presently in development but will be successful shortly.