Table 7.

AI outcomes in implantology.

Target	AI Model	Sample	Results	Study
Decision making in edentulous maxillary implant prostheses	ANN	Implant cases	Within a learning rate of 0.005, the network functioned admirably. The network’s accuracy for the new instances was 83.3%.	Sadighpour et al., (2014) [175]
To fabricate implant-supported monolithic zirconia crowns	ANN	Quality of the fabrication of the individual (zirconia abutment) and clinical parameters in subjects	AI appears to be a dependable solution for the restoration of single implants with zirconia crowns cemented on customized hybrid abutments using a fully digital process.	Lerner et al., (2020) [176]
Implant planning	CNN	75 CBCT images	There were statistically significant differences in bone thickness measurements between AI and manual measurements in all locations of the maxilla and mandible (p < 0.001). In addition, the proportion of correct recognition for canals was 72.2%, 66.4% for sinuses/fossae, and 95.3% for missing tooth areas.	Bayrakdar et al., (2021) [177]
Fractured dental implant detection and classification	CNN	Radiographic images of 251 intact and 194 fractured dental implants	When compared to fine-tuned and pre-trained VGGNet-19 and Google Net Inception-v3 architectures, automated DCNN architecture using periapical images demonstrated the highest and most reliable detection with an AUC of 0.984 [CI, 0.9–1.0] and classification performance AUC of 0.869 [CI, 0.778–0.929].	D. W. Lee et al., (2021) [178]

AI, artificial intelligence; ANN, artificial neural network; AUC, area under the curve; CBCT, cone-beam computed tomography; CI, confidence interval; CNN, convolutional neural network.