Dentist explains X-ray to female patient in examination chair.
An arrow pointing leftHome

Why the nation's largest dental group is embracing AI

  • Hope Reese

“AI [will be] in the room with us. It’s another set of diagnostic eyes.”

Dentists are now getting a new virtual assistant to help find your cavities –– artificial intelligence. In May 2022, Overjet, a Boston-based company, received FDA approval for its software, Caries Assist, that is designed to help detect cavities and bone loss. The Seattle Study Club, the world’s largest network of dentists, with over 250 clubs around the globe, then partnered with Overjet to incorporate AI into its curriculum.

Dr. Parag R. Kachalia, the director of educational development and industry relations at Seattle Study Club, taught at University of the Pacific for 17 years and maintains a private dental practice. PNW.AI spoke to Kachalia about the role of the Seattle Study Club in the field, and what it means to bring AI into the dental office.

This interview has been edited and condensed for clarity.

Hope Reese: What are the origins of the Seattle Study Club? And how did it become the largest dental network in the world?

Parag R. Kachalia: Dr. Michael Cohen, a gum tissue specialist, founded the SSC about 30 years ago. The intention was to get doctors of different disciplines within dentistry, together around case planning and treatment planning for patients. Where we can get out of our silos and look at the patient holistically to figure out the best treatment and approach. It started with local doctors, and little by little started to grow, close to 300 clubs globally.

The SSC is a university without walls. It’s happening locally. This is “you stay in your own area, understand the principles, we’ll send speakers and content and help you grow locally, but don’t worry about going somewhere centralized.” That’s what’s helped build this network out.

While there are many study clubs around, there is not always an overarching curriculum and philosophy they operate by.

Are dentists encouraged to join the study club after earning their license?

Historically, a local specialist, generally periodontist or oral surgeon, founds a study club, goes to a referring dentist and pulls them into what they’re doing.

In dental education today, it’s very difficult to learn everything you truly need to learn in order to be a successful clinician in the long run. You have great foundational information, but because there’s been more innovation and more procedures — yet the length of education has not changed — in order to move on, you need to continue education somewhere.

I’m happy to be vulnerable among people I know, and share that I’m not comfortable with a procedure but want to learn. It’s more intimidating to do it with people I don’t know.

How has the dental field come to adopt AI?

Part of the reason AI adoption has been late, relative to other industries, is that in dentistry, if you’re looking at the radiography space, each radiograph — the person taking it, the equipment, the environment — is different. If you go to a medical hospital for a chest X-ray, it’s a pretty standardized procedure. You walk in, stand here, there’s a target, etc. There’s very little movement that can occur. It’s occurring outside the body.

In dental, we’re working in a small space to position something with lots of movement, different mouth sizes, etc. And even though there’s an image that’s captured, five different practitioners or assistants or trained radiologists will take that image, and it will all look different, just because of the limitations. So, to apply an AI-algorithm across different data sets is, on a technical level, quite challenging.

The other piece is: I was trained to read things a certain way, so do I really need the help? How does AI provide better information than what I’m doing myself? Most radiographs in dental offices are read almost immediately. they’re taken, and in the next 20 minutes they’re probably read. Whereas in the medical space, there’s so much data, and it’s up to radiologists read all of it. There aren’t enough radiologist hours available today to actually read all the data being generated.

I’m a dentist, but I’m a techie. I enjoy innovation and technology. My role at university was to lead innovation, so AI has been on my radar for a long time. It led me to ask myself, “How is this going to help our patient base?”

How did the partnership happen with Overjet? Did they approach you?

We actually approached them. I was giving a presentation on AI for our members at our annual symposium, and I started doing a lot of research. I found Overjet and some others in that category.

What intrigued me about Overjet, specifically, is that yes, they’re looking at radiographs, but they’re also looking at, to some degree, the treatment planning. It’s virtually impossible to be 100% perfect at all times. When I sit down for a clinical examination, am I always seeing everything on the screen? If I do see something, do I document it? If I didn’t, now we have the ability to take our analog entry of our documentation along with the AI data set and correlate the two. Here’s what the AI found, here’s what the clinical treatment plan was, and do they match? I look at AI as a second associate doctor in my practice. I like the checks and balances, helping me along the way.

What does the curriculum plan look like now?

We’re going to talk to our doctors about how the AI is adding a specialist to our room. Historically it was the general dentist as a “quarterback,” the oral surgeon as the person who would deal with external oral surgery or implants, those things, the gum specialist, the root canal specialist — and now we have, almost a radiologist. To a degree — I don’t want to upset radiologists. But it’s almost like a virtual, or augmented reality, person that has AI in the room with us. It’s another set of diagnostic eyes.

Think of this as a virtual colleague, helping you see things that you may not see.

I’m also big on educating our clinicians on why. Why do we think we’re missing something? Well let’s say we had the ideal image that was captured. Now the perfect image every time is only as good as I can read. The perfect image may be on the screen, but the lighting is not correct. My monitor isn’t showing the right thing. I’m fatigued at the end of the day. All of those things are now impacting how I’m reviewing the data that’s in front of me. AI has no emotions. It has no fatigue. But I don’t think people realize that — that we’re only as good as we are at that given point, with the other obstacles in the room.

Do you see Overjet as a leader in the AI field for dentistry? Will other companies follow suit?

Do I see a commoditization of the space? Not necessarily. I do think there will be Overjet plus two or three other companies that really lead the space. Do I see this as a space of 25 options I can pick from with AI? No. Dental is so focused.

I do think companies will come and go, and I think we’ll go beyond radiology — that radiology will be a starting point, but that there could be surface scans. So we’re no longer doing impression materials, but when we do that, unlike putting a physical impression, there wasn’t a lot of data that came out. There was a model of your mouth, but if something changed a 1/2-millimeter position, which is a lot, from a dental standpoint, it’s hard to see that. So I have model #1 and model #2 — the human eye may not detect the 1/2 millimeter difference, but the AI does. The AI can detect change over a period of time. Then it’s helping you as a patient. One millimeter away over the course of life is normal, but over 5 years is not normal.

How do dentists feel about implementing this? Are they worried about losing their jobs? Do they feel comfortable using it?

Right now, we’re at the awareness stage. We have to educate our clinicians about the way that AI can do something for them. Here’s how it can help you, in actual patient care. The early adopters are all in. But it takes a while to move. Unlike other industries, in the healthcare space, at least, at the end of our AI algorithm is a real patient. So the first phase is, “let’s check this out.”

There’s also the realization that AI is not telling you what to do. It’s not replacing you. AI is augmenting what we’re doing.