Digitization in Dentistry: Clinical Applications

© Springer Nature Switzerland AG 2021

P. Jain, M. Gupta (eds.)Digitization in Dentistryhttps://doi.org/10.1007/978-3-030-65169-5_1

1. Artificial Intelligence and Big Data in Dentistry

Priyanka Jain¹   and Catherine Wynne²

(1)

Professor, Department of Endodontics, National University College of Dentistry, Manila, Philippines

(2)

University of Dundee, Dundee, Ireland

Keywords

Digital DentistryArtificial IntelligenceBig DataRapid Prototyping

1.1 Introduction

The profession of dentistry continuously demonstrates innovation and improvement on many fronts such as clinical applications, research, training, and education. This new era includes the broad array of technologies that brings the communication, documentation, manufacture, and delivery of dental therapy under the umbrella of computer-based algorithms. Digital dentistry can be defined as the use of any dental-related technology or device that is built-in digital or computer-controlled elements rather than operated electrically and/or mechanically alone.

Historically, digital advances have three foci: CAD/CAM systems, imaging, and practice/patient management systems. CEREC™, the first commercially available in-office CAD/CAM system made possible, the delivery of same-day restorations. Early drivers in imaging include both the intraoral imaging systems integral to the CERECTM system and evolutions in digital radiography. First introduced in the late 1980s, digital radiography has transformed the dental field, enhancing image quality, evolving from phosphor plates to solid-state detectors, cone beam computed tomography (CBCT), and new generations of digital impression taken with intraoral scanners. The development of cone beam computed tomography (CBCT) heralded a second wave of excitement as three-dimensional images of the craniofacial region offered new advantages in diagnostics and therapy. When iterative improvements in hardware, software, and materials merged in the early 2000s, new accomplishments in clinical dentistry were realized.

Practice management software makes possible capturing patient demographics, scheduling appointments, and generating reports. In parallel, electronic patient records, a digital version of patients’ clinical information, motivated changes in tracking patients’ health, facilitating quality of care assessments, and proved a resource for research, including evaluation of efficiency and efficacy of clinical procedures. In parallel, other technologies influenced and enabled innovations in digital dentistry, often at a remarkable pace. While not a comprehensive list, these technologies undoubtedly include sensor miniaturization, artificial intelligence, augmented and virtual reality, robotics, 3D printing, telehealth, Big Data, internet of things, nanotechnology, quantum computing, biomedical engineering, cost of data storage, connectivity, and others.

This chapter is an introduction to this book. As stated above, the list of digital technologies is not exhaustive, and every attempt has been made to include the most recent innovations to keep the reader up to date. However, technologies change and new innovations may already be on the market by the time this book goes into printing.

The chapter will review how digital dentistry has influenced different aspects of dentistry in brief. A major part of this introductory chapter will deal with the scope of Big Data and Artificial Intelligence, the two most recent advancements in digital dentistry.

While the scope of digital systems is immense, those discussed in this book focus primarily on those that have implications and opportunities for developments and innovations in clinical dentistry. Digitization, being so significant in the dental practice, needs to be understood well. The book serves as a contemporary reference considering the pros and cons of various technological progresses. The impact of digital technology in dentistry is immense, hence for ease its influence is considered under sections—the clinical aspect, laboratory procedures, training of students, patient motivation, practice management, and dental research (Table 1.1).

Table 1.1

Applications of digitization in dentistry

1.2 Recent Trends in Digitization

The digital transformation in dental medicine is recognized as one of the major innovation of the twenty-first century. The implementation of mobile technologies into the medical sector is fundamentally altering the ways in which healthcare is perceived, delivered, and consumed. The most recent trends and innovations of this new digital era, with potential to influence the direction of dentistry in the future, are: (1) rapid prototyping (RP), (2) augmented and virtual reality (AR/VR), (3) artificial intelligence (AI) and machine learning (ML) for diagnostic analysis, (4) Big Data and analytics (personalized (dental) medicine and linkage of health data, mHealth, teledentistry, and development of electronic health records).

The advancement and increased use of Internet and Communication Technologies (ICT), the rise of Big Data and algorithmic analysis, and the origin of the Internet of Things (IOT) are a multitude of interconnected innovations that are having a significant impact on today’s society and affecting almost all spheres of our lives, including medical science [1, 2]. Dentistry, as a branch of medicine, has not remained unaffected by the digital revolution. Research in the dental field on the effect of digitization is focused on both dental clinical practice and research.

Linkages of population health data could prove beneficial for research, such as assisting with the identification of unknown correlations of oral diseases with suspected and new contributing factors and furthering the creation of new treatment concepts [3]. Digital imaging can promote accurate tracking of the distribution and prevalence of oral diseases to improve healthcare service provisions [4]. New possibilities have opened up for automated processing in radiological imaging using artificial intelligence (AI) and machine learning (ML). AI can also be used to enhance the analysis of the relationship between prevention and treatment techniques in the field of oral health [5]. A digital or virtual dental patient, via 3D cone beam computed tomography (CBCT) and 3D-printed models, could be used for precise preoperative clinical assessment and simulation of treatment planning in dental practice [6, 7]. This topic is dealt with in great detail later in the book. With the use of smartphones and wearable technologies, mobile health (mHealth) applications are increasingly gaining popularity and being explored by healthcare providers and companies’ provision of healthcare services [8].

It is important to keep in mind that since these technologies are still at the early phases of implementation and trials, technical issues and certain drawbacks might emerge. Basic science, clinical trials, and subsequently derived knowledge for innovative therapy protocols need to be redirected toward patient-centered outcomes, enabling the linkage of oral and general health instead of merely industry-oriented investigations [5]. For example, use of Big Data analytics and its applications and AI must be done systematically according to harmonized and inter-linkable data standards, otherwise issues of data managing and garbage data accumulation might arise [9]. AI for diagnostic purposes is still in the very early phases, and analysis of diverse and huge amounts of EHR (electronic health records) data still remains challenging [10]. In the simulation of a 3D virtual dental patient, dataset superimposition techniques are still in their experimental phase.

It needs to be understood that digital smart data and other technologies will not replace humans providing dental expertise and the capacity for patient empathy. The dental team that controls digital applications remains the key and will continue to play the central role in treating patients. In this context, the latest trend word created is augmented intelligence, that is, the meaningful combination of digital applications paired with human qualities and abilities in order to achieve improved dental and oral healthcare ensuring quality of life [11].

1.2.1 Rapid Prototyping (RP)

In dental medicine, several digital workflows for production processing have already been integrated into treatment protocols, especially in the rapidly growing branch of computer-aided design/computer-aided manufacturing (CAD/CAM), and rapid prototyping (RP) [12]. RP is a type of computer-aided manufacturing (CAM) and is one of the components of rapid manufacturing. RP is a technique to quickly and automatically construct three-dimensional (3D) models of a final product or a part of a whole using 3D printers. These 3D physical structures are known as rapid prototypes. RP technique allows visualization and testing of objects [13].

Rapid prototyping may be categorized into additive method (widely used) and subtractive (less effective). The additive manufacturing process allows inexpensive production of complex 3D geometries from various materials and involves minimal material wastage [14]. The frequent technologies that are adopted in dental practice are selective laser sintering (SLS), stereolithography, inkjet-based system (3D printers [3DP]), and fused deposition modelling (FDM).

CAD and RP technologies are being used in various fields of medicine and dentistry. They have had a considerable impact especially on the rehabilitation of patients with head and neck defects, in fabrication of implant surgical guides, zirconia prosthesis and molds for metal castings, maxillofacial prosthesis and frameworks for fixed and removable partial dentures, and wax patterns for the dental prosthesis and complete denture.

In dentistry, one of the main difficulties today is the choice of materials. Commercially available materials (wax, plastic, ceramic, metal) commonly used for RP are currently permitted for short to medium-term intraoral retention only and not yet intended for definitive dental reconstructions [15].

RP has the potential for mass production of dental models, and also for the fabrication of implant surgical guides [16]. Production in large quantities at the same time in a reproducible and standardized way is an advantage in reducing the costs. It can also be used in 3D-printed models for dental education based on CBCT or micro CT. An initial study, however, has revealed that 3D-printed dental models can show changes in dimensional accuracy over periods of 4 weeks and longer. Further investigations comparing different 3D printers and material combinations are however still required [17]. The limitations of RP technology include complicated machinery and dependency on expertise to run the machinery during production in addition to the high cost of the tools.

Further research is currently focused on the development of printable materials for dental reconstructions, such as zirconium dioxide (ZrO2) [18]. This different mode of fabrication of ZrO2 structures could allow us to realize totally innovative geometries with hollow bodies that might be used, for example, for time-dependent low-dose release of anti-inflammatory agents in implant dentistry [19].

From a futuristic point, synthesis of biomaterials to artificially create lost tooth structures using RP technology could prove to be revolutionary [20]. Instead of using a preformed dental tooth databank, a patient-specific digital dental dataset could be acquired at the time of growth completion and used for future dental reconstructions. Furthermore, the entire tooth can be duplicated to serve as an individualized implant. RP technology can provide customized and tailored solutions to suit the specific needs of each patient. While the future looks promising from a technical and scientific point of view, it is not yet clear how RP and its products will be regulated.

1.2.2 Big Data

Information is the key for better practice and new innovations. The more information we have, the more we can arm and organize ourselves to deliver the best outcomes. Data collection plays an important role in this regard. In today’s modern age, we produce and collect data about almost everything in our lives such as social activities, science, work, health etc., and this is increasing at a rapid pace. In a way, we can compare the present situation to a data deluge. The technological advances have helped us in generating more and more data, but we have seemed to reach a level where it has become unmanageable with currently available technologies. However, as the data has been getting bigger, our ability to transform and translate the data has also improved, and allowed us to move from reporting what has happened in the past (data reports or descriptive analytics) to learning what is going to happen in the future (data science or predictive analytics) [21].

This has led to the creation of the term big data to describe data that is large and unmanageable. In order to meet our present and future social needs, we need to develop new strategies to organize this data and to extract the information contained and transform the raw data into knowledge and derive meaningful information. One such need is healthcare. Like every other industry, healthcare is producing data at a tremendous rate that presents many advantages and challenges at the same time. Big Data begins to form when a group of data sets brought together become so large and complex that it begins to challenge contemporary data processing and analytical approaches. More data does not let us see more of the same but it allows us to see better, to see different, and to see something new.

Health data can be gathered from routine care and other sources such as social determinants of health, posts by patients on internet forums, surveys and questionnaires from patient support groups and patient diaries. Big Data collaborations involve interactions between a diverse range of stakeholders with varying analytical, technical, and political capabilities. Medical data has its uses in many areas of applications in healthcare, such as prognostic analysis and predictive modelling, identification of unknown correlations of diseases, clinical decision support, treatment concepts, public health surveys, and population-based clinical research, as well as the evaluation of healthcare systems [22].

Even though a number of definitions exist for Big Data, the most accepted is given by Douglas Laney who observed that (big) data was growing in three different dimensions, namely, volume (the amount of collected data), velocity (the speed of generated data), and variety (the source and type of data) (known as the 3 Vs) [23]. The big part of big data is indicative of its large volume. In addition to volume, the Big Data description also includes velocity and variety. These three Vs have become the standard definition of Big Data. Another accepted fourth V is veracity (the quality of incoming data) [24] (Table 1.2).

Table 1.2

What creates Big Data?

Biomedical Big Data amasses from different sources such as electronic health records, health research, wearable devices, and social media. The delay in reaping the benefits of biomedical Big Data in dentistry is mainly due to the slow adoption of electronic health record systems, unstructured clinical records, tattered communication between data silos, and perceiving oral health as a separate entity from general health. Recent recognition of the complex interaction between oral and general health has acknowledged the power of oral health Big Data on disease prevention and management.

This section will introduce the reader about the basics of Big Data and data analytics in dentistry related to different applications such as population data linkage, personalized medicine and electronic health records (EHRs), and mobile health (mHealth) and teledentistry.

1.2.2.1 Electronic Health Records (EHR) and Data Analytics

The electronic health record is a rich source of data that helps dentists monitor the health data of their patients and promote the sharing of information between various members of the healthcare team. EHRs have introduced many advantages for handling modern healthcare-related data. Dental professionals have access to the medical and dental history of the patient. This enables an improved care coordination and communication among healthcare providers and patients. Healthcare professionals have also found access over web-based and electronic platforms to improve their practices significantly using automatic reminders and prompts regarding follow-ups and appointments, and other periodic checkups. EHR enables faster data retrieval and helps provide access to millions of health-related medical information. EHR is further covered in Chap. 13. Table 1.3 gives its uses in dental practices.

Table 1.3

Uses of EHR

National Institutes of Health (NIH) recently announced the All of Us initiative that aims to collect one million or more patients’ data such as EHR, including medical imaging, socio-behavioral, and environmental data over the next few years [26].

Similar to EHR, an electronic dental record (EDR) stores the standard medical, dental, and clinical data gathered from the patients. EHRs, EDRs, personal health record (PHR), medical practice management software (MPM), and many other healthcare data components collectively have the potential to improve the quality, service efficiency, and costs of healthcare. The Big Data in healthcare includes the healthcare payer-provider data (such as EHR, EDR, pharmacy prescription, insurance records) along with the gene expression data and other data acquired from the smart web of Internet of things (IoT). The management and usage of such healthcare data is, however, dependent on information technology.

By analyzing Big Data, dentists can help patients improve health, diagnose the disease at an early stage, and also provide them with personalized dental care. But analyzing huge amounts of patient data manually is impossible. Big Data analytics help in analyzing this data, including personal patient data and demographic data, to identify which oral health problems recur repeatedly, thus helping the dental professionals in diagnosing and treatment planning. Also, by examining the medical and dental records, Big Data analytics can give dentists accurate insights into the oral health problems that are likely to occur in the future. In a nutshell, by analyzing real-time data, Big Data analytics help in revolutionizing the oral health of the population, thereby paving the way to precision medicine. Falling costs (per record) of digital data storage and the spread of low-cost and powerful statistic tools and techniques to extract patterns, correlations and interactions, are also making data analytics more usable and valuable in dental medicine. However, there remains barriers to its universal adoption and integration that still needs to be overcome.

Recent research has focused on the implementation of EHRs both in private practices and in dental education [27–29]. Cederberg and Valenza [28] argue that the use of digital records might compromise the doctor–patient relationship in the future, as easy access to all relevant information through digital means and forced focus on the computer screen could accustom both dentists and students to becoming more detached from patients. Big Data in the field of biomedical research is also useful as researchers analyze a large amount of data obtained from multiple experiments to gain novel insights. However, this poses issues of informed consent for both patients and research participants [30, 31].

Other ethical issues that arise from its use are data security, resulting in a breach of patient privacy and confidentiality [28, 32]. The legal issues surrounding health privacy, for example, sharing of data across national borders, creates hurdles for both individuals trying to access their own personal information as well as for biomedical researchers attempting to establish randomized controlled clinical trials. Therefore, dental biobanks with sensitive patient material, such as saliva, blood, and teeth, must be clarified, as these samples could be used for genetic analysis [33]. Data anonymization is a type of information sanitization where privacy protection is the single most intent. It is the process of either encoding or removing personally identifiable information from datasets. Anonymization methods include encryption, hashing, generalization, and pseudonymization. De-anonymization is the reverse engineering process used to detect the source data. The most common technique of de-anonymization is cross-referencing data from multiple sources [34].

The concept of blockchain is also gaining popularity in data sharing. It is a distributed ledger technology implemented in a decentralized manner used to record transactions [35]. Therefore, dentists can store their patients’ records on a decentralized ledger, helping them save their money and time in having paper-based records. The records are kept across many computers such that data cannot be changed retroactively without the alteration of all subsequent blocks and collusion with the entire network [36]. Additionally, dentists and patients can reap the advantage of blockchain being immutable, helping their records to be secure by offering distributed database, peer-to-peer transmission, transparency with pseudonymity, irreversibility of records, and computational logic [37].

Challenges associated with Big Data should also be considered at the same time, apart from data sharing. Storing large volume of data is one of the main challenges, and an on-site server network can be expensive to scale and difficult to maintain. The data needs to cleansed or scrubbed to ensure the accuracy, correctness, consistency, relevancy, and purity after acquisition. This cleaning process can be manual or automatized. Patients produce a huge volume of data that is not easy to capture with traditional HER format, as it is not easily manageable. It is too difficult to handle Big Data especially when it comes without a perfect data organization. A need to code all the clinically relevant information is required. As a result medical coding systems like International Classification of Diseases (ICD) code sets were developed. However, these come with their own limitations. Studies have observed various physical factors that can lead to altered data quality and misinterpretations from existing medical records [38]. Images often suffer technical barriers that involve multiple types of noise and artefacts. Improper handling of medical images can also cause tampering of images which may lead to delineation of anatomical structures.

1.2.2.2 Personalized Medicine and Data Linkages

Personalized medicine can change how dental research is conducted. Genomic sequencing and recent developments in medical imaging and regenerative technology have redefined personalized medicine to perform patient-specific precision healthcare [39, 40]. An interdisciplinary approach to dental patient sample analysis in which dentists, physicians, and nurses can collaborate to understand the interconnectivity of disease in a cost-effective way can be made possible [41]. Examining large population-based patient groups could detect unidentified correlations of diseases and create prognostic models for new treatment. Linkage of population-based data has changed the way epidemiological surveys in public health are conducted and will play a predominant role in future dental research. The linkage of individual patient data gathered from various sources enables the diagnosis of rare diseases, and completely novel strategies for research [32] helps to identify unknown correlations of diseases, prognostic factors, and newer treatment concepts, and to evaluate healthcare systems [42].

Register-based controlled (clinical) trials (RC(C)T) is a relatively new approach in dental research. These trials can provide comprehensive information on hard-to-reach populations and allow observations with minimal loss to follow-up. However, they require large sample sizes and generate high level of external validity. In the context of data linkage in dental practices and personalized medicine, research has shown that consent might be a significant issue concerning data usage as the patient cannot be completely informed about the ways in which the collected data is/will be used [43]. Data anonymization [44] and patient confidentiality [45] are other issues of data linkage. Therefore, the use of linked biomedical data to support register-based research presents the challenge of disclosing sensitive information about individuals whose consent cannot be easily obtained in retrospect.

Individual dental disciplines (Prosthodontics, Restorative Dentistry, Periodontology, Oral Surgery, and Orthodontics) usually tend to work in isolation in academic dental institutions or large dental service providers. This can result in non-standardized diagnostics within the different departments. An integrated approach with standardized dental diagnostic protocols could enhance better patient flows and reduced overall treatment time and support interdisciplinary linked-therapy planning with improved quality, higher efficiency, and increased patient satisfaction [46]. Additionally, linking this standardized dental diagnostics with biomedical patient-level data could provide the information needed to better understand the epidemiology and etiological pathogenic pathways of oral diseases [47].

The collected data and information of population-based register-based controlled clinical trials RC(C)Ts can also provide an understanding of current and future applications of personalized dental medicine and help to improve prevention and rehabilitation concepts of oral diseases. In the future, private dental professionals together with academic dental institutions will increasingly generate digital data. A major challenge will be the compliance with quality standards in data acquisition, storage, and safe transfer. This will have an impact on the daily routine for the general dental practitioner [46].

1.2.2.3 Mobile Health (mHealth) and Teledentistry

Digital technologies are altering the ways in which healthcare is delivered and consumed. Tele-healthcare enables a convenient way for patients to increase self-care while potentially reducing office visits and travel time [48]. Considering the growing number of the elderly population with reduced mobility and/or nursing homestay, special-care patients, as well as people living in rural areas, these patient groups benefit significantly from teledentistry [49, 50]. Internet is the basis of modern systems of teledentistry, being able to transport large amounts of data. All new systems of teledentistry are internet-based. Changes within the past decade in the speed and method of data transfer have prompted clinicians and information technology experts to re-evaluate teledentistry as a highly valuable healthcare tool.

The practice of medicine using digital mobile devices, known as mHealth or mobile health, pervades different degrees of healthcare by finding ways to utilize mobile technologies for remotely measuring health and delivering healthcare and preventive health services. Newer mHealth technologies with embedded sensors require little attentional effort from the user and allow the unobtrusive collection of objective, high-resolution data on real-world health indicators and health behaviors [51, 52].

The capabilities of mHealth have led to the development of personalized healthcare delivery models that shift the responsibility for personal health away from health systems toward the individuals. By allowing individuals to conveniently track and manage everything about their health, from blood pressure to glucose, the mHealth technology encourages individuals to be actively responsible for their own health, helps them understand their health status, and engages them in preventive behaviors while being guided by input from their health professionals. Furthermore, the remote monitoring abilities of mHealth technologies allows the professionals to proactively identify those at risk for an adverse health event and intervene in a timely manner.

The application of mHealth technology is of great relevance to dentistry. Most often, a patient’s non-adherence to toothbrushing techniques recommended by dental professionals is misunderstood, forgotten, or even completely ignored [53]. The variety of brushing techniques recommended by dentists and dental associations also adds to the confusion among patients [54]. The gap between quality oral hygiene routines and what is actually practiced by individuals is further increased by the dentist’s inability to monitor actual brushing behaviors and good oral hygiene practices at home. Newer mHealth-based technology platforms being developed allow unobtrusive, remote monitoring of toothbrushing behaviors in real-world settings and provide customized, titrated feedback (Fig. 1.1).

Fig. 1.1

Smart toothbrush and its working

The Remote Oral Behaviors Assessment System (ROBAS) utilizes commercially available electronic toothbrushes and/or smart watches as data collection devices and captures key details of toothbrushing behaviors (when used, for how long, pressure applied, dental quadrants covered) in the home setting. The ecologically accurate data is collected and securely transmitted to a cloud server for subsequent analyses by appropriate statistical tools. Such a mHealth platform could serve as the basis of a scalable, interactive ecosystem that passively monitors OHRs, infers and predicts improper OHRs, and delivers engaging and timely personalized feedback to support quality OHRs by individuals [21].

Currently, brushing and flossing behaviors are recorded by measuring traditional oral hygiene indicators (i.e., dental plaque, periodontal inflammation, and caries) during a clinic visit in addition to patient self-reports of their toothbrushing practices. However, they become difficult when involving larger groups or populations, particularly those without regular access to dental services. Low-touch mHealth systems could help clarify the precise relationships between toothbrushing behaviors captured in the home environment and the health outcome (i.e., plaque and dental disease) assessed in the dental setting. By utilizing mHealth’s real-time monitoring and feedback capacities, dental professionals would be better armed to stress upon the importance of correct and long-term adherence to oral hygiene practices and understand the determinants/predictors of why individuals do or do not engage in the prescribed oral hygiene practices [21]. Using mHealth systems in combination with oral hygiene practice measurement and feedback devices (i.e., electronic toothbrushes, smartphones) and back-ended by risk prediction and personalized intervention algorithms, digitally engaged patients would exert more control of their own oral health while providers would be able to provide quality patient-centered and value-based care.

As discussed previously for other Big Data clinical applications, issues of data security, patient anonymity [55, 56] and confidentiality [57] are primary concerns, as networked transfer through unsecure means could enable unwarranted third parties to obtain easier access to sensitive patient data. Cvkrel [56] argued that mHealth creates additional vulnerability as smartphones gather additional data that are usually not collected by healthcare practitioners (e.g., fitness data, sleep patterns), and, as it is an object of everyday use, it might be easily accessible to unauthorized people. Also, easy access through the smartphone to raw data including data related to dental care could be counterproductive and harmful for patients who might self-adjust the prescription given by the practitioner. mHealth can also have an impact on the patient’s consent if not appropriately informed about all of the risks that teledentistry implies [57]. A 2015 WHO global survey revealed affordability, legal issues, evidence of cost-effectiveness, and lack of legislation or regulation as key barriers to mHealth adoption [58]. Beyond digital literacy and affordability, the ability of digital approaches to improve oral health practices and self-care ultimately depends on patient adoption and sustained engagement.

1.2.3 Artificial Intelligence (AI)

What is AI? AI can be seen as the intelligence of machines, that is, work performed by machine, in difference with the natural intelligence performed by humans [59, 60]. The term AI is mostly associated with robotics. It describes how intelligent systems (i.e., artificial intelligence) technology is used to develop a software or a machine that can easily mimic human intelligence and cognitive skills and perform specific activities like problem solving and learning. John McCarthy, a mathematician who coined the term artificial intelligence in 1955, is known as the father of artificial intelligence. He explained the potential of machines to perform tasks that can fall in the range of intelligent activities [61]. Therefore, AI can be defined as a field of science and engineering concerned with the computational understanding of what is commonly called intelligent behavior, and with the creation of artefacts that exhibit such behaviour [62]. Hence, this field deals with computational models that can think and act intelligently, like the human brain, and construct algorithms that can learn from data to make predictions.

With a significant increase in the patient data, electronic health records, its documentation and introduction of Big Data, AI in dentistry or medicine has started gaining popularity with the advent of data computing as well as cloud computing ability and availability of vast amount of data collected.

1.2.3.1 How Are Artificial Intelligence (AI) and Big Data Intertwined?

One of the advantages of AI is the ability for computers to read and analyze large amounts of data in a fraction of the time as opposed to a human. Having more reliable and current data to reference, dentists will soon be able to streamline decision-making in treatment while reducing errors. Specifically, AI can scan all records in a single practice to look for trends in the patient population and can further be integrated with electronic health systems already available.

In the field of dentistry, artificial intelligence is becoming important in radiology due to its ability to detect abnormalities in radiographic images that are unnoticed by the naked human eye, with more emphasis on diagnostic records in terms of digital IOPAs (Intraoral periapical), three-dimensional (3D) scans, and cone beam computed tomography (CBCT). Information can be gathered and computed to create an AI for aiding quick diagnosis and treatment planning. The existing literature reveals that intelligent systems have served as useful adjunctive tools for radiologists in the analysis of large quantities of diagnostic images with improved speed and accuracy [63–70]. They have the ability to detect abnormalities within images that may go unnoticed by the naked eye or to solve problems not resolved by human cognition.

To understand AI, it is important to know few of these key terminologies (Fig. 1.2). Basic introduction will be given to the principle behind AI. For detailed understanding of the mechanisms of AI, the reader may refer to a previously published literature.

Fig. 1.2

Important aspects of intelligent systems

1.

Machine learning is a subfield of AI, which depends on algorithms to predict outcomes based on unseen data and without actually being programmed. Machine learning refers to the ability of machines to gain human-like intelligence without explicit programming; that is, decisions are mostly data-driven rather than being strategically programmed for a certain task [71]. This helps to resolve issues without human input. For example, a machine learning (ML) algorithm can recognize or detect a lymph node in the head-and-neck image as normal or abnormal by analyzing thousands of such images which are labelled as normal or abnormal [72, 73].

2.

A popular type of ML model are neural networks (NNs), artificial neural networks (ANNs). Neural networks are a set of algorithms that compute signals via artificial neurons. The purpose of neural networks is to create neural networks that function like the human brain, but with networks which are created on a computer. Hence, they can be engineered to solve a specific task like radiographic image showing a decayed tooth (Fig. 1.3).

3.

Deep learning is a component of machine learning that utilizes the network with different and multiple layers (hence deep) in a deep neural network to analyze the input data. The purpose of deep learning is to construct a neural network that automatically identifies patterns to improve feature detection [74]. In a more recent innovation, deep learning programs comprise convolutional neural networks that utilize self-learning back-propagation algorithms to learn directly from the data via end-to-end processes to make predictions [75]. These programs are mainly used for processing large and complex images such as 2D radiographs or 3D CT. This differentiates deep learning from machine learning, which builds algorithms directed by the data [76].

Fig. 1.3

Application of deep learning to detect caries on a radiographic image. The x-ray is imported to the computer as the input, and through the hidden layers the algorithms will classify the x-ray image as carious/sound tooth (as the output), based on AI

1.2.3.2 AI and its Applications in Dentistry

In diagnosis and treatment, neural networks (NNs) can be significant, particularly in diseases and conditions with multifactorial cause or without any precise etiology. An example is recurrent aphthous ulceration where clinical diagnosis is usually made on the basis of exclusion of other factors. In a study, data from 86 participants were used to construct and train a neural network to predict the factors appearing to be related to the occurrence of recurrent aphthous ulcers such as gender, hemoglobin, serum Vitamin B12, serum ferritin, red cell folate, salivary candidal colony count, and further retested with untrained data of 10 participants, to test the predictions [77]. AI can also be used for screening and classifying suspicious altered mucosa undergoing premalignant and malignant changes. Minute changes at single pixel level which might go unnoticed by the naked eye can also be detected.

Artificial intelligence can be used to predict a genetic predisposition for oral cancer for a large population [78]. Artificial neural networks (ANNs) are a promising tool for predicting the sizes of unerupted canines and premolars with greater accuracy in the mixed dentition period [79] and can also be optimized for predicting the tooth surface loss, which is a universal problem that involves an irreversible, multifactorial, non-carious, physiologic, pathologic, or functional loss of dental hard tissues [80].

In orthodontics, AI models have been used to assess the craniofacial skeletal and dental abnormalities in cephalometry followed by comparison with an expert opinion. The assessments were found to be equivalent. In addition, the model pointed out contradictions presented in the data that were not noticed by the orthodontists [81]. It can also be used to provide orthodontic consultations to general practitioners for the alignment of crowded lower teeth [82].

A study was conducted to construct an artificial intelligence expert system for the diagnosis of extractions using neural network machine learning and to evaluate the performance of this mode. This study suggested that artificial intelligence expert systems with neural network machine learning could be useful in orthodontics and can be used as a tool for making decisions in clinical practice [83].

Other studies have also demonstrated the application of AI technologies in interpretation of cephalometric radiographs and identifying the landmarks. Studies conducted by Kunz et al. [84] and Hwang et al. [85] showed great accuracy in identifying the landmarks similar to the trained human eye using a specialized artificial intelligence (AI) algorithm and deep learning-based automated identification system, respectively. Furthermore, Choi et al. reported the use of new artificial intelligence model to decide the case for surgery/non-surgery using the lateral cephalometric radiographs. The results showed that the AI system was effective with 96% success rate in diagnosing the surgery/non-surgery cases [86].

In prosthetic dentistry, use of AI can guide the dental professional during the procedure of making a digital impression [87]. Factors such as facial measurements, ethnicity, and patient preferences need to be taken into account while creating an esthetic prosthesis. All these factors (and more) can be integrated by the use of a design assistant, which uses AI (RaPid). RaPiD integrates computer-aided design (CAD), knowledge-based systems and databases, and recruits a logic-based information as a unifying medium [88].

In the field of periodontics, Lee et al. developed an architecture and predicted that the accuracy of detecting periodontitis in premolars and molars was 81.0% and 76.7%, respectively [63]. A model which successfully distinguishes between inflamed and healthy gingiva has also been demonstrated [83]. ANN can also effectively be used in classifying patients into aggressive periodontitis and chronic periodontitis group based on their immune response profile by using parameters, like leukocyte counts in peripheral blood [89]. Yauney et al. used an AI-based system based on CNNs for correlating poor periodontal health with systemic health outcomes and reported that AI can be used for automated diagnoses and for screening other diseases [90]. Figure 1.4 shows the use of AI in the estimation of bone loss.