Table 2.
Details of the studies that have used AI based models for orthodontic diagnosis, treatment planning, and predicting the prognosis.
| Serial no | Authors | Year of publication | Algorithm Architecture |
Objective of the study | No. of images/photographs for testing | Study factor | Modality | Comparison if any | Evaluation accuracy/average accuracy | Results (+)effective, (−)non effective (N) neutral |
Outcomes | Authors suggestions/conclusions |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Leonardi et al.10 | 2009 | CNNs | CCNs-based AI system for automatic location of cephalometric landmarks | 41 | Landmarks | Lateral cephalometric radiographs | 5 Experienced orthodontists | Not clear | (+)Effective | An acceptable level of accuracy was obtained by the CCNs based system designed for automatic landmark detection | Using soft copies of the digital x-rays is effective |
| 2 | Mario et al.11 | 2010 | PANNs | A paraconsistent artificial neural network (PANN) for analyzing the cephalometric variables for orthodontic diagnosis | 120 | Landmarks | Cephalometric radiographs | 3 Experienced orthodontists | Not clear, | (+)Effective | The performance of the model was equivalent to that of the specialist's | Can be used as auxiliary support for orthodontic decision making |
| 3 | Arik et al.12 | 2017 | CNNs | AI based deep (CNNs) for automated quantitative cephalometry | 250 | Landmarks | Cephalometric radiographs | 2 Trained experts | Accuracy of 76% | (+)Effective | This system demonstrated higher performance when compared with the top benchmarks in the literature | None |
| 4 | Park et al.13 | 2019 | CNNs | Comparing latest deep-CNN based systems for identifying cephalometric landmarks | 283 | Landmarks | Cephalometric radiographs | Single Shot Multibox Detector (SSD) | 5% higher accuracy with (YOLOv3) than Single (SSD) | (+)Effective | You-Only-Look-Once model outperformed in accuracy and computational time than the Shot Multibox Detector | This model can be used in clinical practice for identifying the cephalometric landmarks |
| 5 | Kunz et al.14 | 2020 | CNNs | An automated cephalometric X-ray analysis using a specialized (AI) algorithm | 50 | Landmarks | Cephalometric radiographs | 12 experienced examiners | Not clear | (+)Effective | AI algorithm was able to analyze unknown cephalometric X-rays similar to the quality level of the experienced human examiners | None |
| 6 | Hwang et al.15 | 2020 | CNNs | Deep -learning based automated system for detecting the patterns of 80 cephalometric landmarks | 283 | Landmarks | Cephalometric radiographs | Human examiners | Detection error <0.9 mm | (+)Effective | This system accuracy in identifying of cephalometric landmarks similar to the human examiners | This system might be a viable option when repeated identification of multiple cephalometric landmarks |
| 7 | Xie et al.16 | 2010 | ANNs | ANN based AI model for deciding if extractions are necessary prior to orthodontic treatment | 20 | Tooth malocclusion | Lateral cephalometric radiographs | Not mentioned | Accuracy of 80% | (+)Effective | ANN was effective in determining whether extraction or non-extraction treatment was best for malocclusion patients | None |
| 8 | Jung et al.17 | 2016 | ANNs | Artificial Intelligence expert system for orthodontic decision-making of required permanent tooth extraction | 156 | Tooth malocclusion | Lateral cephalometric radiographs | 1 Experienced orthodontists | Accuracy of 92% | (+)Effective | The success rates of the models were 92% for the system's recommendations for extraction vs non-extraction | AI expert systems with neural network machine learning could be useful in orthodontics |
| 9 | Choi et al.18 | 2019 | ANNs | ANN based model for deciding on surgery/non-surgery and determining extractions | 316 | Landmarks | Lateral cephalometric radiographs | 1 Experienced orthodontists | ICC value ranged from 0.97 to 0.99 | (+)Effective | This ANN based model demonstrated higher success rate in deciding on surgery/non-surgery and was also successful in deciding on the extractions. | This ANN based model will be useful in diagnosing of orthognathic surgery cases. |
| 10 | Kök et al.19 | 2019 | ANNs | AI algorithms for determining the stages of the growth and development by cervical vertebrae | 300 | Cervical vertebrae | Cephalometric radiographs | 1 orthodontists | Mean Accuracy of 77.02% | (+)Effective | ANN could be the preferred method for determining cervical vertebrae stages | None |
| 11 | Makaremi et al.6 | 2019 | CNNs | CCNs-based AI system for determining of the degree of maturation of the cervical vertebra | 300 | Cervical vertebrae | Lateral cephalometric radiographs | Not mentioned | Mean Accuracy lesser than 90% | (+)Effective | This proposed model is validated by cross validation and is of use for orthodontists | This is a validated software and can be readily used by orthodontists |
| 12 | Lu et al.20 | 2009 | ANNs | ANN based model for predicting post-orthognathic surgery image | 30 | Face | Profile images | 1 orthodontists | >80% accuracy in prediction | (+)Effective | The ANN based system demonstrated an improved accuracy and reliability in prediction | Can be used for clinical and treatment planning |
| 13 | Patcas et al.21 | 2019 | CNNs | AI system for describing the impact of orthognathic treatments on facial attractiveness and age appearance | 2164 | Facial landmarks | Facial photographs | Not mentioned | Not Clear | (+)Effective | This CNN based AI system can be used for scoring facial attractiveness and apparent age in patients under orthognathic treatments. | None |
| 14 | Patcas et al.22 | 2019 | CNNs | AI system for evaluating the facial attractiveness of patients who have undergone treatment for clefts and the facial attractiveness of controls and to compare these results with panel ratings performed by laypeople, orthodontists, and oral surgeons | 30 | Face | Frontal and profile images | 15 laypeople, 14 orthodontists, 10 oral surgeons | Cleft cases (all Ps ≥ 0.19), For Control group (all Ps ≤ 0.02) |
(−)Non Effective | AI system scores were comparable with the scores of the other groups for the cleft patients, but the scores were lower for the controls | There is a need for further refinement in this AI based system |
| 15 | Thanathornwong23 | 2018 | Bayesian network (BNs) | Bayesian Network (BN) for predicting the need for orthodontic treatment | 1000 | Tooth malocclusion | Data sets | 2 Experienced orthodontists | AUC (0.91) | (+)Effective | This BN based system; and demonstrated promising results with high degree of accuracy in the need for orthodontic treatment. | None |
| 16 | Li et al.24 | 2019 | ANNs | ANN based model for orthodontic treatment planning | 302 | Landmarks | Extraoral and intraoral photos, lateral cephalometric radiographs | 2 Experienced orthodontists | Accuracy of 94.0% for prediction of extraction-non-extraction, (AUC) of 0.982 | (+)Effective | The ANN based system demonstrated excellent accuracy levels in predicting for extraction-nonextraction, and also extraction and anchorage patterns |
Can be useful for guiding less-experienced Orthodontists for predicting orthodontic treatment. |
ANNs = Artificial Neural Networks, CNNs = Convolutional Neural Networks, DCNNs = Deep Convolutional Neural Networks, BN = Bayesian Network, BN = Bayesian Network
PANN = Paraconsistent Artificial Neural Network, ROC = Receiver Operating Characteristic curve, AUC = Area Under the Curve, ICC = Intraclass Correlation Coefficient.