Florian Hillen is the CEO & Founder of VideaHealth, a platform on the mission to improve patient health through the power of AI. By leveraging algorithms, analytics and health data, Videa is unlocking health and economic value across the dental value chain – making dentistry more transparent for dentists, insurers and patients.
When you were at MIT, you conducted extensive research on the intersection of AI, healthcare and oral health. What were some of the biggest takeaways from this research?
I studied medicine in Germany, worked with several healthcare startups (including founding one), and always knew I wanted to do something impactful, so my focus at MIT and Harvard Research was looking at where AI can have the biggest impact on health. I researched AI in radiology, dermatology, ophthalmology and biology. Initially I thought AI might have the most impact in mammography, so I worked with five different companies in that space. However, I ultimately moved away from mammography because I realized AI had greater potential in dentistry. The biggest reason for this is that dentistry has one of the highest – if not the highest – rates of missed instances of disease. Dentists miss nearly 50% of decay, on average. This can have major implications as missed cavities, abscesses, lesions and oral diseases can evolve into larger problems when untreated. A tool that enables a data-driven approach has immense potential to help dentists and their patients.
My research and work at MIT led to founding VideaHealth, as well as two master’s degrees in Electrical Engineering & Computer Science and Technology & Policy.
How did this research then transition into the launch of VideaHealth?
MIT recognized the potential for my research project, which was called Delta V at the time, and provided the first $50,000 in funding the summer after I graduated. Well before dental AI was commercially viable, I designed the technology encapsulated in Patent no. 11,553,874, which covers dental image feature detection powered by machine learning that helps dentists detect diseases on images and matches those findings to patient records. It is the critical concept that drives today’s dental AI technologies.
I wrote the first algorithm for the minimum viable product (MVP) on Nov. 19, 2018, and from there I was able to raise $1.1 million in angel investment. With that, I hired my first team of five people, built my MVP to the first small product, and secured increasing interest from venture capital (VC) firms and the dental community. I then raised $5.4 million from Zeta and Pillar, two Boston-based VCs focused on healthcare and early innovation. This allowed me to build out a larger team, further develop the platform, submit for and ultimately gain FDA clearance for our medical algorithms and onboard our first customers. We then raised an additional $21 million from Spark Capital. It’s incredible to see how far we’ve already come with our mission-driven team, amazing clients and visionary partners, including Henry Schein One, the world’s largest software and services company for the dental industry.
One big initial challenge, which is common in medtech, is the lack of user feedback throughout all the years of product development. To mitigate this, we included dental practices as pilot programs prior to submitting for FDA clearance, and gathered an enormous amount of user feedback to help us optimize the platform.
Why does dentistry have one of the highest rates of diagnostic errors in healthcare?
This can be largely attributed to dentists having such a wide range of experience levels and diagnostic capabilities that can be influenced by external factors such as equipment quality and room lighting – so much so that in many cases, 10 different dentists will provide 10 different diagnoses.
The widespread prevalence of dental problems is another contributing factor. More than 90% of American adults have had at least one cavity and 25% have at least one currently. More than 3.5 billion people globally suffer from oral disease – 44% of the population – and dentists miss detecting more than 40% of them.
Most oral issues are asymptomatic, including most gum diseases and periodontal lesions. Therefore, dentists who aren’t as well equipped are likely to overlook unreported issues that patients don’t know about, particularly incipient decay, lesions and diseases. For example, a dentist might focus on one side of a patient’s mouth where they’re having a toothache, while missing a more significant problem on the other side of their mouth.
While it goes without question that dentists want to be as effective as possible, they are constantly dividing their time. Dentists are highly specialized, but they essentially have four jobs: radiologist, primary care physician, surgeon and practice manager. It’s a lot to carry and this constant juggle can give way to instances of oversight.
What are some examples of how AI improves accuracy of diagnosis?
Our AI-powered platform improves diagnostic accuracy by significantly aiding dentists in their analysis of radiographs and providing predictive analysis and diagnosis/treatment recommendations in near-real time, based on an algorithm that leverages hundreds of millions of data points.
In the FDA clinical study for our dental caries detection platform, Videa Caries Assist, the platform enabled dentists to reduce the number of missed caries (cavities) by 43% and reduce the number of erroneous detections (false positives) of caries lesions by 15%. This aligns closely with the results of our joint study with Heartland Dental, in which the platform helped clinicians detect 46% more caries lesions and reduced errors by 10%.
U.S. Food and Drug Administration (FDA) most recently issued 510(k) clearance for Videa Perio Assist, our AI-powered interproximal bone level measurement tool for periodontal disease assessment. With Videa Perio Assist, dentists and hygienists can easily measure interproximal bone level changes over time for patients 12 and older, resulting in faster and more accurate treatment recommendations and improved patient outcomes.
Beyond improving diagnostic accuracy, our AI platform enhances preventative care, which is critical in dentistry. VideaHealth enables dentists to catch decay, abscesses and lesions at their infancy so that tiny restorations can be made pre-emptively, rather than more invasive treatments later, like crowns and root canals. For example, Videa Caries Assist can identify incipient lesions up to five years before a dentist can on their own, and then suggest a pre-emptive strike such as remineralization to strengthen enamel. In cases of existing disease, our technology allows dentists to track and monitor disease progression with side-by-side image comparison and AI annotation overlay, enhancing a vital step in the clinical diagnosis and treatment process.
What are some other ways that AI can be used in dentistry?
AI provides dentists with an objective analysis and/or validation, along with a credible, detailed and objective report to reference when communicating issues to patients. These findings build a compelling visual presentation that improves communication, trust and accountability, strengthening the overall patient-dentist relationship. Improvements in trust and understanding directly correlate with more patients accepting treatment plan recommendations which helps enable better outcomes for the patient and bolsters dental practice profitability.
In addition to improving diagnostic accuracy and preventative care, AI can be used to customize crowns, bridges and implants much more quickly and accurately than traditional procedures.
Ultimately, by leveraging algorithms, analytics and health data, dental AI like VideaHealth’s not only improves patient care and experience but also unlocks health and economic value across the dental value chain – making dentistry more transparent for dentists, insurers and patients.
Could you share some details on the datasets that were used to train the algorithms that are used at VideaHealth?
Our AI algorithms leverage the VideaFactory, the industry’s most diverse dataset, which includes over 100 million data points from leading dental service organizations (DSOs), hospitals, insurance companies, clearinghouses and universities. For context, the 100 million data points represent approximately 50 times more X-rays than a single dentist sees in their entire lifetime. Intentionally aggregating data from a wide range of other sources helps us ensure that our algorithms remain impartial, accurate and representative of the diversity of the population. We then vetted thousands of dentists and tapped the top 5% to annotate X-rays used to train our AI. These expert dentists label images to identify disease and provide high quality data to the system, enabling our AI to learn from the best “teachers.”
Our algorithms are continually trained by a team of data scientists to ensure accuracy and ongoing optimization. Armed with this information, the VideaFactory produces extremely accurate models that can identify and measure clinical indicators on dental X-rays with ease.
What are some of the reasons that AI has not been more widely adopted by the dental industry?
Dentistry is a highly specialized sector, and therefore relies on unique data points to provide effective service to patients. But it’s extremely difficult to build a sufficient database because of the fractured nature of most dental practices. The vast majority of dental practices are independently owned and operated, which makes data collection and analysis very difficult, especially at the scale needed to draw accurate conclusions. Despite this improving somewhat in recent years with the increase in DSOs and greater consolidation within the insurance industry, there has still been minimal progress overall in capturing broad dental data sets, which are largely subject to strict data privacy and protection laws.
Why do you believe the AI revolution will be led by dentistry?
I truly believe AI can have the biggest impact in dentistry. The industry has billions of patients worldwide and over a billion X-rays taken a year. When it comes to detecting one of the most prevalent health conditions in the world, dental caries (cavities), our FDA study demonstrated our AI helps reduce the number of caries missed by dentists by more than 40%*, which has tremendous implications.
The number of missed instances of disease is even greater for certain use cases. Infections of root tips, which require root canal as a primary treatment, are missed 50% of the time by dentists. 50% are false negatives and 30% are false positives. That is mind-boggling. There is clearly a huge problem here and AI can solve it.
The benefits of AI also extend across the industry: patients get better treatment faster and reap greater value from their treatment plans, while dentists streamline their operations and payments, and insurers reduce fraud and improve claims processing efficiency. It’s a win-win-win!
Private equity is also extremely strong in dentistry, which is why I believe dentistry will overtake healthcare in terms of AI adoption. It doesn’t rely on complex hospital systems, but rather private-equity-backed dental practices and streamlining software. And AI can be deployed across hundreds of offices at once!
What are some of the solutions that are offered at VideaHealth?
Our AI suite includes components that assist with charting, diagnosing, treatment planning and practice management:
1. Videa Assist
- Helps dental clinicians detect more diseases, faster with objective analysis.
- Includes Videa Caries Assist – FDA-cleared AI caries detection tool.
- Reduces false positives by 15% and improves diagnostic accuracy of 100% of clinicians who use the technology, regardless of years of experience.*
- Improves patient communication, building patient trust and increasing case acceptance rates.
- Videa Assist supports organic growth by increasing the treatment plan value by an average of 26% across all practices, effectively scaling business with better efficiencies and faster reimbursements
2. Videa Insights
- Helps DSOs manage and grow practices with actionable insights.
- Improves quality assurance and training for clinical staff.
- Provides data to support better-informed practice acquisition decisions
Is there anything else that you would like to share about VideaHealth?
Our mission is to help millions of patients receive more accurate, effective and transparent dental care – and we’re excited about it becoming increasingly clear to them what has already been known in the medical and dental communities for a long time: oral and medical health are inextricably linked and dental care directly impacts health in the rest of the body.
Meanwhile, we are honored to better equip dentists, who are expected to wear many hats – diagnostician, surgeon, primary care physician and specialist – with something that’s untenable without the use of advanced technologies like AI. With the right use of AI, dentists can manage all these specialized roles and realize incredible efficiency gains while still delivering superior care.
*According to results from the FDA clinical study of Videa Caries Assist.
Thank you for the great interview, readers who wish to learn more should visit VideaHealth.
- Decoding Opportunities and Challenges for LLM Agents in Generative AI
- Tailor Introduces ChatGPT Plugin Enabling Conversational Interface for ERP Operations
- Anthropic Introduces Paid Subscription for its AI Chatbot Platform Claude
- Top 10 LLM Vulnerabilities
- AI-Generated Drake Song Submitted for Grammys: A Pivotal Moment for Music and AI