Artificial Intelligence is slowly turning out to be an integral part of our everyday life. Right from the simple Google search you do to complicated stuff such as Scientific, Medical and also Robots roaming the face of other planets, AI which was touted to be the future is now slowly becoming the present. And Dentistry is not far behind is integrating Artificial Intelligence into our field – Diagnosing and treatments.
Yomi a robot-assisted surgical device developed by a Start up named Neocis has helped clinicians place 10,000 dental implants in patients using the dental robotic platform.
Some terms which are useful when talking about AI are –
AI – Artificial Intelligence
ML – Machine Learning (It is a branch of AI which performs intelligent tasks without any prior knowledge or rules by just learning)
DL – Deep Learning
ANN – Artificial Neural Network
DNN – Deep Neural Network
CNN – Convolutional Neural Network
In Dentistry, the most commonly used AI technology is Convolutional neural network or CNN which is used in image recognition by taking in digital signals such as sound, image and video.
Application of AI in various Branches of Dentistry:
This is one branch of Dentistry where AI is already playing a vital role in determining the treatment plan and also aiding in designing the treatment protocol actively in the form of “Aligners treatment”. ANN is being used actively in planning Orthodontic treatments and predicting treatment outcomes for patients.
As some Orthodontic treatment requires extraction of teeth, clinical evaluation along with use of ANN can be a good measure in determining the need of tooth extraction. When ANNs were used to determine the need of tooth extraction considering clinical indices as well, it showed an accuracy of 80-93% to determine the need of extraction depending on the malocclusion.
CNN has been trained to identify Teeth in a Periapical Radiograph, on testing it has demonstrated a precision rate of 95.8% to 99.45% which is close to the clinical work of a person having 99.98% precision rate.
When the CNN algorithm was used in identifying Dental Caries from healthy tooth structure it has shown a precision rate of 75.5 to 93.3% in a sample size of 3000 radiographs of Posterior teeth. It showed a sensitivity rate of 74.5% to 97.1% in detecting Dental Caries using Radiograph alone. This is considerably higher when compared to clinicians who have a sensitivity rate of 19% to 94% when given only Radiographs to detect dental caries. With more refinement it can be a very useful tool in detecting Dental caries faster and making it easier for clinicians to process cases and come to a definitive diagnosis.
Smile Designing has gained importance with many giving importance to their smile and aesthetics, 3D smile designing software’s which use AI to show the end result of a patient’s smile is an important tool for the clinician and the patient to come to a definitive conclusion about the treatment plan.
Some software’s such as DSD (Digital Smile Design) or DTS Pro (Digital Treatment Simulation) or Planmeca Romexis Digital Smile Design, etc. These software’s make use of AI and Machine learning to help Dentist’s design smiles in minutes.
In Periodontics AI is used to diagnose between the two types of Periodontitis – Aggressive and Chronic. Diagnosing the type of Periodontitis can help in determining the treatment plan to a large extent. Papantanopoulos and colleagues have used ANN technology to distinguish between Aggressive Periodontitis and Chronic types by using immunologic parameters such as Leukocytes, IgG antibody titers and interleukins.
ANN was used to identify these parameters which helped in differentiating AgP from CP having 90 to 98% accuracy. Other parameters were also included such as monocytes, eosinophil, neutrophil counts and CD4+/CD8+ T-cell ratio as inputs helping in more accurate diagnosis.
Lee and coworkers have used CNN algorithm to determine the need for extraction of Periodontally Compromised teeth. With the help of CNN Algorithm, the accuracy of predicting the need for extraction was 73.4 to 82.8%. With 73.4% being Molar teeth with multiple roots while the Single rooted pre molars giving 82.8% accuracy.
It is useful is identifying between Precancerous and Cancerous lesions of the Oral cavity, head and neck region using clinical as well as other diagnostic aspects. CNN has been used to detect tumor tissues in tissue samples as well as radiographs, and they have given promising results. The specificity in diagnosing head and neck cancer lesions has been 78-81.8% and 80 to 83.3% which is almost similar to clinicians.
In a study which was conducted to distinguish between Ameloblastoma and OKC based on certain parameters, identifying which has become much easier and faster with AI integration.
With AI integration using CAD/CAM in fabricating crowns, onlays etc it has made it possible for single sitting dentistry along with reducing contamination, operating time and also the time spent on the dental chair. With the advancements happening at a fast pace, we can expect a lot more integration of AI into Dentistry which can make us more efficient.