Artificial Intelligence 101

The term "artificial intelligence" was coined in the 1950s, and it refers to the development of machines that are able to perform tasks that previously required human involvement. Machine learning is a component of AI in which algorithms are applied to learn the intrinsic statistical patterns and structures of data to enable predictions regarding unseen data. "AI and machine learning can sound like elements of a scary science fiction movie to people who don't understand how integrated they already are in our day-to-day lives," says Seth Gibree, DMD, senior director of clinical advocacy at Heartland Dental. "Of course, it's important to maintain clinical autonomy. It's not about machines telling clinicians what to do; it's about machines providing clinicians with more information and assisting with care."

When it comes to clinical decision-making, ideally, AI and machine learning models in dentistry apply the same logical processes as dentists; they just execute them at higher speeds and with greater accuracy. The identification of decision-making patterns that deviate from the norm, such as by confounding or artifacts, can reveal the presence of bias. This helps to increase the trust in AI of its users and recipients.¹ "There is a lot of noise that influences our decision-making as humans," explains Eric Pulver, DDS, an oral and maxillofacial surgeon with a private practice in Highland, Indiana, and chief dental officer at Denti.AI. "AI doesn't tell us what we must do; it says, ‘hey, this is what you may not be seeing. Please take a look at it.' AI doesn't get tired, and it doesn't get upset by friends-it's the same every day. That consistency can help to keep us on track and become better at what we do."

Advances in AI are occurring throughout all of the subfields of dentistry. For example, robotic protocols for implant planning and placement and root canal therapy are gaining momentum. AI is also being used to enhance radiography. "AI applications that are designed to address radiologic assessments seek to identify caries, periodontal bone loss, and periapical infections," says Sanjay Mallya, BDS, MDS, PhD, associate professor and chair of the Section of Oral and Maxillofacial Radiology at the University of California, Los Angeles, School of Dentistry. "Such approaches have the potential to enhance efficiency and overall accuracy. In addition, AI is being applied to cephalometric analysis. When performed conventionally, this is an arduous manual task, but now, it can be accomplished within a few seconds."

Diagnostic Support

According to Kyle Stanley, DDS, a cosmetic dentist in Beverly Hills, California, and the chief clinical officer at Pearl, AI-powered radiography is at the center of Pearl's offerings. "I see Second Opinion^® as our flagship product because it delivers—in the simplest, most straightforward package—the core technology underlying a majority of our other solutions. It is the only real-time patient-facing pathology detection product on the market that has been cleared by the US Food and Drug Administration (FDA) to detect 9 different conditions on permanent teeth captured in bitewing and periapical radiographs of patients who are 12 and older," he explains. "We often talk about AI elevating the standard of care in dentistry by bringing more accuracy and consistency to diagnosis, but the scope of that impact depends on the breadth of the AI's utility. Second Opinion has the breadth to really have a profound impact on quality of care and the patient health outcomes that follow." Pearl offers a robust collection of AI-powered capabilities, many of which can be used in tandem. For example, Practice Intelligence^® comes with Second Opinion included. It provides real-time feedback during patient visits while also tapping into the data from practice management software applications and synthesizing that data in conjunction with a review of patients' past radiographs to help maximize scheduling, optimize treatment planning, and reduce practice liability.

Although imaging is essential to diagnosis in dentistry, reading radiographs can be a highly subjective process. AI is especially suited to overcome the inherent variability in the individual clinical examination. "In dentistry, there is a saying that if you show the same set of radiographs to two different providers, you'll get two different interpretations," explains Teresa Dolan, DDS, MPH, chief dental officer for Overjet. "Artificial intelligence can help make dentists' radiographic interpretations more precise, consistent, and objective." Overjet's Clinical Intelligence Platform is the only AI solution that is FDA-cleared for both decay detection and bone level quantification. It adds a layer of real-time data to traditional radiographs, color-coding the images to make it easy to see carious lesions, bone loss, and prior dental work, including restorations, implants, and root canal treatments. Overjet's all-in-one technology integrates with the practice management system, analyzes radiographs from prior visits, and helps identify patients who may have undiagnosed treatment needs. "The AI not only provides data to help with clinical decision-making but also serves as a powerful patient education tool. When patients see the AI findings, they better understand the need for treatment," says Dolan. As machine learning algorithms get increasingly sophisticated, they will be able to support broader applications in dentistry. To assist doctors in their diagnoses, Overjet is currently working on additional models that are capable of identifying more pathoses and others designed specifically for pediatric dentistry.

The ability to accurately identify the development of caries as early as possible is a major component of diagnosis in dentistry. As part of a partnership between Henry Schein One and VideaHealth, Dentrix has launched Dentrix^® Detect AI, an AI-powered x-ray analysis tool for the Dentrix Ascend Imaging system that provides real-time clinical decision support to help dentists detect caries earlier, avoid misdiagnosed or missed caries, and provide a higher level of patient care. In addition to detecting caries, Dentrix Detect AI also identifies and communicates conditions that are at risk of developing caries, such as overhanging restorations. It is currently FDA-cleared for the analysis of bitewing images for patients aged 22 years and older.

Smile and Restoration Design

As dental design software continues to evolve, companies are adding more and more tools to automate various aspects of the process, further improving the efficiency and accessibility of digital dentistry. Some offerings now include AI-powered features to assist clinicians, and smile design is no exception. For exocad's CharisideCAD 3.0 and newer versions, Smile Creator has been updated to enable the AI-based detection of facial features, which automates the process of smile design and helps clinicians get to esthetic proposals faster. The design of final restorations is being empowered by AI as well. With 3Shape Automate, clinicians can upload cases and select design parameters and turnaround times, then the platform's advanced AI technology delivers designs for single and multi-unit final restorations as well as nightguards. Clinicians with 3Shape Dental System are given the opportunity to modify and edit the designs before accepting and downloading the files.

Charting Solutions

"There are a lot of incomplete charts. It happens to all of us," says Pulver. "Therefore, at Denti.AI, we have created a series of algorithms that emulate the entire process to support dentists, hygienists, and assistants from beginning to end, combining pathologic and nonpathologic findings to streamline the flow from treatment options to suggested codes for billing." This auto-charting solution recently received FDA-clearance for clinical use. AI evolves with demand, and during the COVID-19 pandemic, Denti.AI responded to the demand for voice-enabled software for periodontal charting. The system, which works with the company's auto-charting technology to provide a complete solution, uses relatively off-the-shelf microphones that require minimal training, and the plan is to seamlessly incorporate imaging into that system to make it even more robust. Beyond patient charting, Denti.AI has also partnered with P&R Dental Strategies, a premier dental informatics company, to implement their knowledge into risk assessments, which will help build efficiencies into the payer/ insurance domain.

In addition to automating the charting process and mitigating the risk of infection associated with manual entry, AI-enabled voice recognition and recording software can increase the accuracy of charting and offers other benefits as well. Using affordable recording devices and advanced AI technology from BOLA AI, Dentrix Voice seamlessly captures dictated periodontal results as well as clinical notes and other practice correspondence. Front office staff can use it to dictate insurance claims, patient communications, billing notes, and more.

Patient Conversion

AI is also being used in dentistry to power a number of practice management solutions. For example, a lack of awareness or training among team members regarding patient engagement and follow-up can result in practices missing out on opportunities to appoint new patients. "That's where Patient Prism^® comes in," says Amol Nirgudkar, chief executive officer of Patient Prism. "We leverage AI to understand those conversations so that we can move that patient forward."² Patient Prism records and scores new patient calls, which allows dentists to track the performance of their team members, discover missed opportunities, and train their teams to win them back, focusing on the callers with the potential to generate the highest revenue. With AI, dentists can provide their teams with data-driven guidelines to book high-value patients, reduce marketing costs, and boost profits.

Education

In addition to working its way into modern practice, AI is also finding its way into the curricula of dental schools and into patient education. In the spring of 2021, New York University became the third dental school in the United States to acquire a robotic system for dental implant surgery. The robot-assisted surgical device, Yomi^®, was developed by Miami-based healthcare start-up Neocis. A year later, NYU became the first dental school in the United States to perform a student-led dental implant surgery using Yomi from start to finish.³ Many dental schools are increasingly emphasizing the value of AI by providing students with exposure and access to training early in their careers. At UCLA, Mallya is exploring options to use AI to enhance student learning, particularly for caries evaluation. Regarding patient education, systems such as Oral-B's Genius X toothbrush use AI to recognize brushing styles and coach users so that they can achieve the best results from at-home hygiene.

Data Management

So much of AI in dentistry is aimed at reducing the siloing of information to create better workflows. Across disciplines, the collection of digital healthcare data is ubiquitous, and although these data are relatively heterogeneous at this time, more and more organizations are striving to produce cleaned, curated, and structured data. With AI, dental professionals are enabled to best utilize multilevel data, such as medical and dental history data, sociodemographic and clinical data, imagery data, biomolecular data, social network data, and more.⁴

"Data generated from AI-based applications will have the potential to yield insights into dentists' diagnostic and treatment decisions and patient outcomes," says Mallya. "Hopefully, such data can be channeled to develop quality measures that can be applied for practice and provider improvement and education." Furthermore, the use of AI to integrate data management systems can help to bridge the medical-dental divide and provide an opportunity for a more comprehensive, preventive, and holistic approach to healthcare.

Improved algorithms are the future of AI in dentistry. "Denti.AI is currently working on a complex algorithm that will analyze through the identification processes to provide treatment planning options," says Pulver. In addition, the company is working to support providers with automation and improved accuracy and efficiency for intraoral scanners and orthodontics.

Stanley believes that the development of improved algorithms has the potential to affect every aspect of dentistry. "Historically, most of our focus at Pearl has been directed at pathology detection and assisting in a diagnostic capacity; however, AI-assisted treatment planning is an extremely exciting opportunity," he explains. "Orthodontics, pediatric dentistry, oral surgery, and other major specialties are getting a lot of our attention now that we have a solid handle on many of the key fundamentals of general dentistry."

Whatever the application, the reach and potential of AI seem limitless. "There's a real dynamic movement happening in the industry, and it's all about connecting the dots and identifying innovators and ambassadors to develop deeper inroads and implement technological solutions," says Pulver.

"When we look at what people were considering AI for in dentistry 10 years ago, reality has already outpaced those dreams," notes Stanley, "which isn't surprising because technological progress tends to accelerate at an exponential rate." Although it may sound like science fiction, AI technology developed by a multidisciplinary team at the University of Pennsylvania may soon be able to brush and floss your teeth for you. They've created microrobots composed of iron oxide nanoparticles that demonstrate both catalytic and magnetic activity. With a magnetic field, the microrobots' motion can be directed, and they can be configured to form either bristle-like structures to sweep dental plaque from the broad surfaces of teeth or elongated strings that can pass into interproximal spaces like dental floss. The microrobots also have the potential to produce antimicrobials that kill harmful oral bacteria on contact.⁵

Dolan emphasizes that innovators in the field of AI are looking for ways to connect every aspect of the dental experience. "As more software integrations take place, auto-charting and other features will become readily available," she says. "I also predict that AI will be integrated with dental payer software to enable immediate adjudication for pre-authorizations and claim submissions, which will take place while patients are still in the dental chair."

In addition to streamlining the practice of dentistry, AI can improve the delivery of care by using data to individualize the dental experience. "One thing that I hope to see in 10 years is a world where AI assists us in finding the lesion, selecting the surface, and applying the materials best suited to the individual patient's biologic and lifestyle profile," says Stanley. "That may sound like a pie-in-the-sky notion, but if we can develop processes for tracking the materials that we use in prophylactic and restorative care over the next 5 years, AI can offer us a much clearer picture of which materials will work best in specific situations to deliver the best long-term outcomes for individual patients. With that picture, we can approach dental care on a patient-by-patient basis. If that comes to fruition, then I expect that the practice of dental medicine will become much more about delivering preventive care than it is at present."

Innovations in AI have the potential to not only further improve the patient experience but also the efficiency and predictability of treatment protocols. "There are so many steps involved in placing implants," says Pulver. "One day, I think that we'll be able to just press a button and say, ‘place six implants in the mandible,' and the planning software will place them in the correct positions and with the proper angulation to support a prosthesis. The technology will not only design a guide and print a model but also integrate with my practice management system to ensure that I have all of the components that I need on hand for the entire process." As an oral surgeon who works in hospitals, Pulver lives in the space between medicine and dentistry, and in that vein, he is excited about the future of AI in salivary diagnostics. "To be able to look into salivary diagnostics and identify which patients are at greater risk for various issues and then take a deeper dive into the expression of genes will be very exciting," he says.

The power of AI is enabling the organization, categorization, investigation, and depiction of healthcare data, and its influential algorithms for diagnosis, charting, treatment planning, robotics, education, and more are aiding in the growth and development of dentistry as well as the overall provision of healthcare.⁶ Indeed, the worldwide usage of AI in the medical sector has been projected to grow from $1.3 billion to $10 billion by 2024, which represents a 40% annual growth rate.⁷

With the breakneck pace of AI evolution, the options for integration appear to be endless. However, before incorporating AI-based technology into routine clinical operations, it is important to consider its cost-effectiveness, dependability, and applicability.⁸ "In terms of when to integrate new technology, it's a fine line," explains Gibree. "We want to be on the cutting edge of technology, not at the bleeding edge. At Heartland Dental, we pilot new services and tools for our supported clinicians, and we look for ones that will help our supported doctors and their teams provide their patients with a higher level of care." AI-based technology will continue to streamline oral healthcare, relieving dental professionals of laborious routine tasks and improving the delivery and affordability of care for more patients, and eventually, it may even facilitate the development of fully personalized, predictive, preventive, and participatory dentistry.