Digital Health

Healthy teeth thanks to AI

Evaluating X-rays, checking teeth, fighting gnashing: //next presents exciting applications that are already improving our dental health with artificial intelligence (AI). A trend that has only just begun.

On the topic of artificial intelligence (AI), the Helmholtz Association writes on its website: “This is the story of an unusual treasure hunt.” “Treasure” refers to the huge amount of data in medicine that is collected every day and offers a number of starting points for AI. This also includes dentistry, where this “treasure hunt” is in full progress: on the one hand, AI is to help with the interpretation of X-ray images on the basis of large amounts of data, and on the other hand, more extensive measurement data and the networking of production steps are to revolutionise the manufacture of dental products. The focus is on dental health. Early detection and even more precise treatment methods than before are the key to long-term healthy teeth.

Focus on time waster X-ray image

An X-ray of the jaw and teeth is usually the first step in any dental treatment. However, a competent and comprehensive analysis of the image material takes a lot of time – and its quality depends heavily on the experience of the treating doctors. AI should ensure more speed and accuracy here.

In Germany, the dentists Prof. Falk Schwendicke and Dr Joachim Krois from Berlin's Charité University Hospital are pioneers in the field of machine analysis of X-ray images. In 2017, they began to develop software that uses AI to evaluate and interpret large amounts of image data. Their solution is now available under the name dentalXrai Pro.

The software enables the automated analysis of dental X-ray images with the help of AI algorithms. The software reliably detects caries, infections or restorations such as crowns, implants or root fillings and highlights the findings in colour. The treating dentist receives a “preliminary result” that he only has to complete and check. This saves time and ensures the quality of the diagnosis.

Where does the AI get its knowledge from?

In the beginning, there is training. In order to develop an AI solution like dentalXrai Pro, the largest possible amount of data must first be available. For machine learning, thousands of X-ray images are fed into the system, which have been analysed and marked by competent medical professionals. In this way, the AI is provided with the necessary knowledge to make diagnoses and therapy suggestions.

The software recognises patterns in the large amounts of data and develops models and algorithms. It then applies these to new X-ray images. Within seconds, the AI system is able to “read” and interpret an X-ray image. It returns to the practice with appropriate colour markings and indications.

The doctor bears the responsibility

This does not make the doctor's work superfluous. The doctor still bears the responsibility for the patient. They must still be able to check the image interpreted by the machine and correct it if necessary.

The AI systems currently being developed in Germany for diagnosis and therapy in dentistry are considered so-called “assistance systems”. Their function is to support the doctor, not to replace him.

But internationally, people are thinking further ahead: fully automated "medical robots", for example for difficult operations, have been on the agenda of AI research worldwide for some time.     

For dentistry in Germany, the first task will be to establish and expand AI-supported data analysis on a large scale. The solutions that have been developed so far are only the beginning. The support of practice processes by means of machine learning systems is conceivable in many places. 

Many innovative approaches internationally

Digital support for data analysis in dentistry is applied both horizontally (comparative data) and vertically (comprehensive individual analyses). It is not only possible to accumulate thousands of individual findings and evaluate them for patterns, as is applied in the case of X-ray images. The individual patient can also be viewed and "measured" more comprehensively.

There are already a number of start-ups that target end users with their products and help them gain their own information on their dental health and improve their preventive care:

The app Pearlii, for example, offers to perform a free dental check-up using a mobile phone camera. The images of one's own teeth are evaluated on the basis of AI. In this way, initial indications of diseases or weak points can be found. 

Toothbrushes are also becoming intelligent and increasingly using artificial intelligence. There are already various toothbrush models that measure the pressure on the teeth, the duration of brushing and much more. They evaluate this data using AI and provide tips on how to brush properly. With some providers, the measured values are also transmitted to an app via Bluetooth. This prepares the information graphically so that the customer can present the results to his dentist. 

The product Mouth Guard aims to tackle teeth grinding using AI: patients who suffer from teeth grinding wear a splint with sensors that can measure and locate tooth movements. The oral splint is used at home, the collected data can be evaluated in the dental practice and used for a precisely fitting therapy. 

Dental implants or healthy teeth can also be equipped with sensors. There are various innovative approaches to this on the market, which aim, for example, to recognise eating habits by analysing saliva changes and to evaluate them for dental health.

If this "data movement" catches on, dental practices will in future have a range of additional information at their disposal that could be of relevance for possible treatment. Bringing together and reading the multitude of individual results is also difficult without AI. The corresponding practice solutions for this are likely to follow in the coming years.

Internationally, there are many innovative approaches

Digitale Digital support for data analysis in dentistry is applied both horizontally (comparative data) and vertically (comprehensive individual analyses). It is not only possible to accumulate thousands of individual findings and evaluate them for patterns, as is applied in the case of X-ray images. The individual patient can also be viewed and “measured” more comprehensively.

There are already a number of start-ups that target end users with their products and help them gain their own information on their dental health and improve their preventive care:

The app Pearlii, for example, offers to perform a free dental check-up using a mobile phone camera. The images of one's own teeth are evaluated on the basis of AI. In this way, initial indications of diseases or weak points can be found. 

Toothbrushes are also becoming intelligent and increasingly using artificial intelligence. There are already various toothbrush models that measure the pressure on the teeth, the duration of brushing and much more. They evaluate this data using AI and provide tips on how to brush properly. With some providers, the measured values are also transmitted to an app via Bluetooth. This prepares the information graphically so that the customer can present the results to their dentist. 

The product Mouth Guard aims to tackle teeth grinding using AI: patients who suffer from teeth grinding wear a splint with sensors that can measure and locate the tooth movements. The oral splint is used at home, the collected data can be evaluated in the dental practice and used for a precisely fitting therapy. 

Dental implants or healthy teeth can also be equipped with sensors. There are various innovative approaches to this on the market, which aim, for example, to recognise eating habits by analysing saliva changes and to evaluate them for dental health.

If this “data movement” catches on, dental practices will in future have a range of additional information at their disposal that could be of relevance for possible treatment. Bringing together and reading the multitude of individual results is also difficult without AI. The corresponding practice solutions for this are likely to follow in the coming years.

The more data, the better!

The core challenge for all AI solutions described is the database. The Helmholtz Association is certain: "To achieve breakthroughs, research needs enormous amounts of data."

In Germany, however, it is difficult to obtain this data in many places. According to Helmholtz, a strong fragmentation of the health system combined with very high data protection requirements often prevent a central consolidation of individual patient information. The electronic patient file, which will be introduced in 2021, represents a first important step in terms of data transparency. It could become a basis for the central provision of data.

It is still a long way to go and certainly not without controversy to bring together medical data from different sources centrally and make it accessible. The data collected by the patients themselves on their own smartphones pose an additional challenge.

These difficulties will have to be solved in order to be successful with AI solutions. The data basis represents an essential success factor for artificial intelligence; it cannot work without it. The rule is: the more, the better!


AI in the production of dental products

While AI in diagnosis and treatment still has many hurdles to overcome, the intelligent and networked production of dental products is already in the midst of digital upheaval: precision and cost savings are the drivers that will move this area forward with great momentum.

Networked production strives for autonomous work in which AI systems communicate with each other and the individual process steps are perfectly coordinated. The accuracy of computer-measured models, precision in automated manufacturing and the streamlining of production chains are the most important advantages of AI-supported production processes.

This applies to the production of dental tools as well as dental prosthesis products or implants. By means of comprehensive analysis of various measurement data, precise models are created, which can then be produced with a 3D printer, for example.

This makes production more efficient, but also more precise than is possible in manual production. The networking of the individual production steps is based on the Internet of Things (IoT). This type of artificial intelligence is currently making its way into the production of many industries and is steadily advancing.

Conclusion: AI has arrived in dentistry

Overall, it is already impossible to imagine dentistry without AI. The most urgent goal is to tap into the “treasure trove of data”: automated analyses have enormous potential for improving diagnosis, prevention and treatment. If AI solutions are used successfully, some dental diseases can be detected earlier or even avoided altogether. This would be a benefit for us as patients.

The future will show where these developments will lead. Scenarios such as the “transparent patient” or robots that replace doctors are certainly more of a theory in medical practices. Nevertheless, the topic of AI will always be about balancing economic interests and the well-being of patients.