S3 EP04: The AI Revolution in Dentistry

Ioannidou: [00:00:00] Technological advances have always been at the heart of improving the dental care experience. And the hot topic right now, wherever we go and whatever we do, actually, I just came back from the American Academy of Periodontology annual meeting. Everybody's talking about AI. So let's join this conversation.

My name is Effie Ioannidou.

Wright: [00:00:20] And I'm Dr. ArNelle Wright, and this is Dental Sound Bites. So today's episode, we're going to be talking about the intersection of artificial intelligence and dentistry and its impact on our profession.

Announcer: [00:00:34] From the American Dental Association, this is Dental Sound Bites. Created for dentists by dentists.

Ready? Let's dive right into Real Talk on Dentistry's Daily Wins and Sticky Situations.

Ioannidou: [00:00:49] So there's no denying, ArNelle, that we are living at the age of AI, right? I mean, I can see how comfortably my kids use chat GPT because it makes their life way easier, right? And we see it in all aspects of our daily life, in our society, in security, in education, in healthcare, even in advocacy. So it's changing our world.

And it's already changing the way we practice dentistry. I think that this will be a very interesting conversation.

Wright: [00:01:19] It's going to be very, very interesting, especially because who we have on the show today. We have invited Dr. Mina Ghorbani, who is a leader in the world of dentistry AI to tell us more about it.

Hey, Dr. Mina, can you tell our listeners about yourself, what you do in the AI realm?

Ghorbani: [00:01:38] Hi everyone. Thank you so much. I'm honored to be here with you guys today. My name is Dr. Mina Ghorbanifarajzadeh, if you really want the whole thing, Dr. Ghorbani or Dr. Mina. I'm actually an ADA Wellness Ambassador and very excited and honored to have that position, but also I wear a hat as a Senior Clinical Manager at Overjet, working on building these models to be used in clinical practice for practicing clinicians, just like yourselves. And maybe in the future, in education for dental students, and hygiene students, and auxiliary students. It's so fascinating to be at this intersection and watch it grow. So thank you so much for having me on this journey with you guys.

Wright: [00:02:17] Oh my gosh.

Ioannidou: [00:02:19] So tell us a little bit, it's used in the dental practices. And how can we, one, integrate this technology into the practice?

Ghorbani: [00:02:29] I love it. First, I have a quick question you can answer. Do you have any digital workflows already? Like, do you use a scanner?

Ioannidou: [00:02:36] So I'm in academics, so I'm not in practice. And actually going from one school to the other. So I just left UConn to go to UCSF. I can tell you that both schools have digital workflow, scanners, 3D printers. We even have SIM labs, which is, you know, another form, I guess, for AI in education.

Right? So, yeah, but I think ArNelle is more qualified to speak about the way of practicing with AI. I mean, I don't know what your experience was or has been, uh, ArNelle.

Wright: [00:03:06] Well, so I don't have a scanner right now, but I've been in practices that do have, like, the iTero scanner. In dental school, we had the CEREC machines and the E4D.

So. I have that experience, but right now I am not scanning in my practice right now.

Ghorbani: [00:03:21] Understandable. I just wanted to set the stage though that AI has actually already been in dentistry for a couple of decades. So all of the scanning technologies are AI in the infrastructure and the foundations because you give it an input, it goes through an algorithm and it gives you a computer generated crown that it's going to mill.

As an example, the CBCT evaluation of the airway. AI driven, right? So we're already impacting patient outcomes. Clear aligner therapies with their treatment planning and staging those trays, AI generated. Smile design with photographs and facial analysis. All of these are AI based. But what we're really talking about, when we're looking at Overjet, is the application of computer vision on 2d radiography, something that hasn't been done in dentistry before.

So supporting that detection to diagnosis, clinical accuracy in a way that we haven't done. We've looked at it from a restorative section, but now we're looking at it at, as the crux of general dental pathology, that bread and butter dentistry. And since, Dr. Effie, you're in academia, you may or may not have heard the statistics around how accurately dentists identify caries on a radiograph or periodontal disease.

But I can tell you, I'm shocked at how great we are at our statistical levels. It's not that high. Have you heard any of these statistics before?

Ioannidou: [00:04:44] No, I haven't. I'm a periodontist. So, especially with a complex, a slightly complex new classification of periodontal diseases, it's intimidating. It's not that difficult, but intimidating.

So I'm sure that there are errors and I can guarantee those statistics may be depressing.

Ghorbani: [00:04:59] We have a study that shows that regardless of the experience level of the dentist interpreting the radiograph, they're about 65% accurate on interpreting caries on a radiograph. 65%.

Wright: So we're all failing.

Ghorbani: But we knew this, like, I spoke to radiologists at different academic institutions, and this is something we're aware of.

And there's compounding variables, from how you visualize, to being in the appropriate room, to cognitive fatigue and the like, right? So all we can do is look at the opportunity for augmenting – and that's why augmented intelligence comes into play – augmenting the clinician to be a better clinician, detecting the areas of concern, leading to a more accurate diagnosis and better treatment outcomes.

Wright: [00:05:47] I think we're off to a good start.

Ioannidou: [00:05:49] Yeah, let's just get a better idea, Dr. Mina. I would like to get a sense of what dentists think about the AI's role in patient care and in general in dentistry's future. So what we did, we, we asked and here is what they told us.

Dental Professionals: [00:06:05] I think Artificial Intelligence is very important in patient care and actually will lead to the future and will be an expectation that patients will actually have.

I think AI's role in patient care will only augment what we're doing right now. It's more of, like, a second opinion type of thing.

As long as we're utilizing AI in a productive way to supplement what we're doing, it's going to be excellent. As soon as we start trying to have AI substitute for real human brilliance, I don't really like that as much and I feel like that could get us in more trouble.

I think one of the pros is that it'll enable diagnosing of caries lesions and other, and periodontal disease, and other issues with patients.

It's a good tool, but we have to find a balance. Like, the computer doesn't think like us. And we have to find a balance between Artificial Intelligence and what we like to do. I think if it's going to help us, of course, I will try it.

I think it's something that should be well developed, well managed, and well taught to current dentists and the dentists of the future.

It can make our lives easier or harder. It'll make our lives harder if we don't address it.

Things are so different now in the technology than it was when I was in dental school. So the students that are coming out are definitely utilizing that technology with their digital platforms. And that's definitely a way of the future.

I think it has two sides. It has a good face and a bad face. I find it helpful for 3D scanners, 3D printing, but that also makes us also lack personnel. Maybe having those tools will help more people in the community.

Our patients are using it. We should be using it too.

Wright: [00:07:42] Dr. Mina, so we hear some people that clearly think that it's very promising. Others think that it's scary. So can we start with the basics and then maybe we can just build on some of the feedback that we've heard from some of the guests that we asked this question to.

Ghorbani: [00:07:57] I was very happy to hear. that there's some alignment that there's good and there's bad. We have to be critical of these new technologies, right? The car was a new technology at some time. My favorite one to talk about is the toaster or your washing machine or your dishwasher. All of these were new technologies, right?

And there was some hesitation, some general concern about is it going to be better than the human, right? And we willingly get in cars every day to transport us from point A to point B. And we've had a lot of advances there, but at some point it was very taboo to get into a car. People were worried and concerned about their safety.

And let's be honest. If you eat toast in the morning, I'm pretty sure you use a toaster. You don't go start an open flame, take your bread outside that you made in a Dutch oven, fermented that dough for, like, two or three days. No, you put in a toaster and I bet you your toaster is so fancy that it has infrared and it knows exactly what the doneness is, but you don't question any of that.

And you're washing machine. This is something that people don't think about. The sensing option. If you have a new washing machine, it's an input, right? It does a measurement of how heavy the load is. It fills the water accordingly. You don't think twice about that.

Wright: Not currently.

Ghorbani: Bless the Lord that we have washing machines because we don't spend two days washing clothes anymore.

Right? It was a technology that solved the pain point. And we've adopted it and we couldn't imagine life without it. We're at the beginnings of the same thing right now with AI, right? So a little bit of hesitation is totally, totally normal, but it's understanding how it works and how do we incorporate it into our life.

And the thing that's important about AI within dentistry, specifically in 2D radiography, is that the way the models are trained is, imagine that your peers have annotated this information. It's supervised learning. So other dentists have helped develop these models. There's regulations that go into it. There's heavy testing that goes into it. And it's very regulated in the way that it's developed in the way that it's put into practice. So take that into consideration.

Ioannidou: [00:10:02] I really like the comment that you made and I think this is really very important, the supervised learning.

Wright: [00:10:08] I just wrote that down. Oh my gosh.

Ioannidou: [00:10:11] One of the comments specifically addressed the fact that how scary it is to let the computer think. And not a human think. So the discrepancy between human and computer mind. But I think it's really important for people to understand that the computer doesn't think, right? We guide the computer, or AI responses are supervised by humans.

So it's whatever you feed, then you get out. I mean, I'm by no means a coder, but we do know that, right? It's a multidimensional analysis of the data points that you feed to the whatever platform. I think this is a scary notion, the notion that your mental capacity is on the side, and then you are guided by a unit that is not human.

Ghorbani: [00:10:56] If I may expand on it a little bit.

Wright: [00:10:57] Please, yeah, yeah, sure.

Ghorbani: [00:10:58] Just to make it even more crystal clear. So supervised learning, we go to dental school for four years, right? So it's a specialized body of knowledge. We didn't say Model, go review every image on Google and tell us what dental decay is. Absolutely not.

That like sacrilegious decay. Like the disease process for caries is multifactorial. It's longitudinal over time. It's very, very specific. So that supervised learning is such a specific recipe. Like we tell what it needs to annotate. There's guidelines and the way that we annotate, we have to reduce the amount of noise. There's lots of specificity that goes into it. And then when we go through FDA regulations to make sure that we can produce and show that it's of value, the thing that's most impressive to me is that human plus machine. If you never ever look at a radiograph that's annotated or has predictions from an AI model, it's like it never did it.

It's like a hammer, right? If you don't pick up the hammer and hit the nail, the hammer just sits there. Same thing with this. It's a tool. It's a tool. So the human plus the hammer can build a house.

Wright: [00:12:08] I love these analogies. I'm like living for this. Like I love this. Yeah.

Ghorbani: [00:12:12] Right. So if you don't open the platform, if you don't review the radiograph with the predictions, it's like it was never there. The human is in charge. But the human is augmented to be a better clinician because now they have a little bit more ease and the cognitive load is a little bit lower and there's a little bit more focus, and there's standardization that goes into it. And then we augment the clinician to be the superhero.

Wright: I love it.

Ghorbani: Think Tony Stark. Like, think Avengers. Tony Stark is a mere mortal. He knows how to use technology. He knows how to build technology. He knows how to wear his suit. If you take the suit away, he dies. Like, dead. But if you give him a suit, he saves the universe. So let's wear our AI suits as clinicians and save our patients from dental decay and periodontal disease and all of the other ailments that fall under our purview and our scope.

Ioannidou: [00:13:07] I wish that it was that easy, right? Like, use AI and then save the patients from periodontitis. Not today. Yeah. You know, but this route, it's a great opportunity, but it has a few threats. A lot of stakeholders have to partner together to be able to give the best outcomes for a patient, right?

Ghorbani: [00:13:25] Exactly. And that's what I'm hoping to encourage.

Wright: [00:13:29] But we actually will be going through some of the concerns, actually, later on in the episode, so are there any additional ways that you can tell us how AI is being used in dental offices? How can one integrate AI technologies into their practices to enhance patient care or practice management, if there's anything that we haven't touched on?

Ghorbani: [00:13:46] The first thing I'll ask, it's kind of like that personal awareness first. Find out what you want to make better about yourself. Like what, where do you want to enhance your practice? So if you don't have digital workflows already, CAD Cam scanning, introduce that. Those are AI technologies. Maybe you want to go into the sleep dentistry realm and you want to buy a CBCT, start analyzing airway. That's also AI. 

Maybe you want to get an AI-enhanced practice management system, has the chatbots, helps manage your administrative team. That's AI. I don't know if you've heard about this one, Dr. Effie, AI supported voice recording of Perio charting.

Ioannidou: [00:14:25] Oh, that will be, I have the dream come true for me.

Ghorbani: [00:14:30] So that's out on the market.

And then I, personally, am very excited about TD radiography. There are multiple companies out there, just like how you buy anything being digitally literate, understanding the landscape, what are the products, what do they do? And do they do what you need them to do? So why do you buy one car company X versus car company Y's SUV? That's your, like, you have to do a little bit of due diligence, figure out which one works best for you and go from there. I'm not going to sit here and tell you that you should buy one product over the other. It's clinically what's valuable to you and your practice. 

That's the first one. And then does it enhance care? Absolutely. We know that there's an increase in case acceptance. We know that there's an improvement in patient communication because now they're able to see the radiographs in a way that are colored so that that's communicated with the patient. It brings trust.

Ioannidou: [00:15:20] Oh, a hundred percent. Transparency builds trust and especially to the level that the patient really understands the finding, right?

So I think that's, that's very much supported by AI.

Ghorbani: [00:15:30] Right.

Ioannidou: [00:15:30] And, you know, if you think about this, even the applications that, electronic health records, some systems at least, have for your cell phone, like MyChart app that Epic offers, that's AI, right? So you, this empowers the patient to have all the information in their hands. To call for an appointment without actually making the call, requesting the appointment. I mean, I love this as a patient. I feel well served.

Ghorbani: [00:15:59] Exactly. So like, I think when the patient feels at ease that the care that they're giving is for them – so patient centric care – we're headed there. It's amazing. 

So. Yes, you can look into many technologies. If you think that you would be supported with having a technology like this to help support your patients and empower your patients to be part of their care and achieve those amazing outcomes that you know that they can have, this is a technology for you. Also supporting your staff and those insurance claims that we've already talked about. And talking about how you can empower your hygiene staff to be a little bit more diligent on their Periodontal diagnosis if they're not already being on top of medical record keeping and making sure the periodontal charts are up to date. We can go on and on and on.

Ioannidou: [00:16:46] Yeah.

Wright: [00:16:46] Our patients are more savvy now anyway. So I feel like, I know my patients, they come armed with certain information, even though sometimes they're a little off. They've done a quick Google search, that's a different story. But utilizing AI not only helps us help them better, it also, like you said, invites them to be a part of their care, if that makes sense.

Can you talk to us a little bit about how accurate AI is? Does it diagnose? I think our listeners would appreciate hearing from you. As it relates to those things.

Ghorbani: [00:17:17] So how accurate is AI? This is a profound question to me because I don't know how accurate I am as a human and Artificial Intelligence is like human intelligence, but by a machine. So I can only expect it to be as perfect as I am, which is not very perfect. But I can hope that it makes me better, right? 

So now to more pointedly respond to your question, accuracy is a metric. So sensitivity and specificity, if you're truly, truly interested, we have to submit studies to the FDA. Those metrics are reported to the FDA. The FDA studies are out for public review. If you want to know the exact sensitivity and specificity for a specific pathology. So, Dental decay. We'll have one for the different products that are out there. And it's, you're getting into, like, the super deep details if you do that, because the way that the studies are done are different.

You have to read the indications for use and how they develop them. And product A is slightly different from product B. And Perio, for example, periodontal disease. If you're looking at what is the accuracy? So it'd be like CEJ to crustal bone and what's the accuracy there. And the overall way those interact is a very different conversation is dependent upon how good is the raw input?

So you give it a diagnostic quality radiograph in this realm of 2D radiography, it gives you diagnostic quality predictions. You give it an underexposed, overexposed, cone cut, foreshortened bite wing with all the contact overlap. What can you expect? It's not magic.

Ioannidou: [00:19:01] That's my question. The platforms that you are talking about, obviously you, you're right.

If you give it well-taken, accurate diagnostic x-ray, you get the measurements that you wish. As accurate as possible, right? But then it's not only this. Periodontitis cannot be diagnosed based on an x-ray, right? How does the clinical exam and the clinical findings are incorporated and how do we reach really an accurate and clinically relevant diagnosis?

Ghorbani: [00:19:30] Just like how you were taught to do. You take into consideration the clinical factors, right? So in periodontal disease, bleeding on probing is not visible on a radiograph. It's just not. So AI doesn't diagnose. The only person who's licensed to diagnose – notice I said person – is a licensed dentist, right?

So diagnosis remains in the hands of the clinician. Do not relinquish that superpower, that privilege. This is supposed to help detect the factors that you need from a radiograph specifically. To help you diagnose. So detection is the first step and caries, and like, caries diagnosis and you have to understand, if the other risk factors, time, contact with sugars, demineralization, things like that.

For periodontal disease, we're looking at radiographic bone loss as only one component. How quickly? Is it due to trauma? Is there a calculus? Is the calculus on enamel or on root surface? What is the bacterial load? What bacteria are present?

Ioannidou: [00:20:34] I don't think anybody knows this answer. Like, I don't think there is any clinician that would look at bacterial load and what bacteria are present.

So that's, that's nice in theory, but it doesn't happen in practice, right?

Ghorbani: [00:20:45] It doesn't happen in practice, but it's to remember that those, there are more factors than just the radiograph that help you determine.

Ioannidou: [00:20:52] So wouldn't it be amazing to have a platform where all these questions are answered, and all these data points are collected and then the platform is calibrated. To, you know, assess all the cutoffs and suggest a diagnosis.

Ghorbani: [00:21:10] Slowly. Right. We got to take baby steps.

Ioannidou: [00:21:14] We're getting there.

Wright: [00:21:15] We got to crawl before we walk.

Ghorbani: [00:21:17] Yeah. First, we got to make sure the models are accurate. Right? Yeah. We have to make sure that the accuracy is accepted and then we say, okay, well step two is, and we'll go staging and grading, right?

Radiographic bone loss, A. B, then we have to make sure we have pocket depths and probing depths and clinical attachment loss. How is the soft tissue? So we have to build these algorithms. We're only at the beginning.

Ioannidou: [00:21:39] And I'll tell you if this was a requirement by insurance companies to have an accurate diagnosis in, and if the, for AI to function probing and bleeding points were required, then all dentists would probe.

Wright: [00:21:53] I was literally just going to say, like, this could be… I was just literally about to say, this could be the thing that gets everybody to be doing all of those things anyway, to feed that algorithm. Like, so, it wouldn't go against us.

It actually would help us get more positive outcomes and more accuracy for our patients anyway.

Ioannidou: [00:22:11] Yes.

Wright: [00:22:11] Which seems like it's the whole point of the situation, right? 

Ghorbani: [00:22:15] And wouldn't it be nice if the AI said, hey, your Perio chart is out of date. Please take a new one.

Ioannidou: [00:22:20] Oh yeah. Uh oh. Good point. Uh oh.

Announcer Ad: [00:22:25] Need a snack to pair with your podcast? Grab something from your new fridge. ADA members get member-only savings on select GE appliances including Profile™, Cafe™, GE®, Monogram® Haier and Hotpoint® appliances. Visit ADA.org/GE to get the special authorization code to access your savings.

We see you listening to Dental Sound Bites in your car. Maybe it's time for you to upgrade. Take advantage of special ADA member savings on a new or pre-owned Volvo fully electric, plug-in hybrid, or mild hybrid car. ADA members receive $1000 off a purchase or lease of any vehicle. Visit ADA.org/volvo to access your discount.

Ioannidou: [00:23:13] So as we are talking about all this, and it's so exciting in that you open all these windows for us to think about, Dr. Mina, I wonder what are your concerns or what are the concerns that have been already identified and what are you hearing from dentists and the industry about some immediate worries about AI?

Ghorbani: [00:23:32] Well, the favorite one is, it's going to take my job. And I said, let me know when the robot can cut the Class 2 prep. Or let me know when the robot can scale. Because that's not happening anytime soon. Until the robot can pick up the hammer to build the house.. we’re just not there yet.

Wright: [00:23:49] So Dr. Mina, is it accurate to say, like, it's not an either or?

Ghorbani: [00:23:53] It's not an either or. 

Wright: It's a both and. 

Ghorbani: It's an and. We're in the medical and health related industry. We're not just going to let the machine learn on its own. That's irresponsible.

Wright: [00:24:04] Right. The first one that came to my mind when I think about concerns is that I would become too dependent on it, that I would lose my skills, that I would be replaced and I would no longer be necessary.

Like that's the immediate thing that comes to my mind. Although we are very, very far away from that. I think that's probably a common thread for everybody. Like when you hear AI, like, when I'm in Publix and I see self checkout or something like that. I automatically think, oh man, two more registers are taken away or two more lanes are taken away. You know, there's not anybody manning those and actually checking me out. You know what I mean? 

Like they were, they've been replaced by people who just want to get in and get out and do it themselves, kind of thing like that. And then in dentistry, there's been this issue with mid level providers, right?

So I was thinking, like, of how close could this come to us being replaced at some sort of level, you know what I mean? But I guess hearing this conversation and having this talk today, I'm almost eased. And just in talking to you. But I would say I think that those are pretty popular concerns that people would have listening to it.

What about you, Dr. Effie?

Ioannidou: [00:25:06] I agree. I have heard people talking about this. 

The other thing I was thinking about is, you know, when I said all the stakeholders have to be on the same page just for these platforms to wear appropriately. I was thinking about the utilization of AI by insurance companies and, you know, helping out the speed of claims. How you can really speed up reimbursement. How you can really accurately apply diagnosis and the proposed treatment planning by bringing all these factors together very quickly.

Ghorbani: [00:25:39] So I'll talk to both of these. So, Dr. Wright, let's talk about that general concern that you, that you have about, like, replacement. So that's where I'm going to challenge you. Actually, I'm going to take away some of your ease. ‘Cause I want you, I'm going to challenge you to think a little bit bigger. I’m going to challenge you to think about how would you modify or improve or develop or allow yourself to be creative as a dentist, utilizing AI?

So let, like, let it replace some of the stuff you don't want to do. I'm very prepared for someone to give me some backlash here, but imagine that you have mid level providers who are enhanced in rural areas that don't have access to care, right?

So you're in a big city and you have a satellite clinic with a mid level provider and you're helping review the radiographs. Is that a bad thing? You don't have to answer. I'm not, I'm not positioning you to do that, but I want people to think outside of the box. We have serious care issues. So that's one version.

Second version is, how do you reach your clinical excellence? How do you get to the next level of dentistry? If you're still doing the things that were done, the same Class 2 prep that we've been doing for 50 to 60 years, give us the space for creativity. And if AI can do some things that we've been doing for so long, maybe we need to look at it differently and change that. So it's being uncomfortable, like being comfortable with being uncomfortable.

Wright: Or with the unknown too.

Ghorbani:  And with the unknown. Like, where do you want this to go? And we have, for that, that means we have to be present in our industry and kind of guide those conversations to be very, very important. 

Now then, Dr. Effie. Oh you hit the nail on the head. So insurance companies are already using AI to help adjudicate claims. This is, the ADA has a paper on it. It's mentioned. And I can say that AI is being utilized. And it's beautiful because if we can get to a point of care adjudication day, where you don't have to wait eight weeks for a pre-authorization for an SRP, we're doing good.

And let's talk about the subjectivity and claims review, that's reduced. Because now it's measurable, quantifiable, and transparent, like you'll get to a level of transparency because we know what the measurements are on the radiograph. if that's what's being used to approve the claim.

Ioannidou: [00:28:05] That's a good point.

That's great. This is really a good point.

Ghorbani: [00:28:08] It's great. It's not against, it's not like insurance companies want patients to not get care that they need, right? It's to beat against the fraudulent claims or an over utilization. That's what we need to understand here.

Ioannidou: [00:28:22] And the other important thing would be at some point, you know, again, with the use of AI, when we are able to bring electronic records together, merge databases and be really able, in oral health, to promote disease surveillance, right?

So we know what is out there. Because we just now speculate, we have the, you know, the NHANES databases, we have some large databases and initiatives from the CDC, but wouldn't it be amazing if we had real time understanding of what's happening to our patients? Like, bring Axiom data and Epic data and merge it together and know the medical profile and know all these confounders that affect disease prevalence and management.

Just saying I know I'm thinking probably 5 to 10 years ahead.

Ghorbani: [00:29:12] but I love it.

Wright: [00:29:13] That's outside the box

Ghorbani: [00:29:14] I mean, I don't know how to, like yes, I was taught how to manage someone who's going through chemotherapy. I was taught how to take care of someone and manage a patient going through radiation therapy. How many patients did I actually treat?

Wouldn't it be nice to have some support the first time I'm doing it, and not have to call my professor back from dental school. Or try to go find the lecture that's now dated. And it takes me too long to go do the research.

Ioannidou: [00:29:43] Yeah, you're absolutely right. 

Wright: [00:29:49] Tell everybody what it is that you're holding up, which made us laugh.

Ghorbani: [00:29:51] I need to bring this in just because I thought, some comic relief. I have a dentiform on my desk and I have this really, really character, like character set of really, really large lips with the brightest white smile on two little feet.

So, you know, the people who just like their smiles, their whole face, but it's artificial. I have one of those. I pulled that up. I needed ArNelle to laugh because…

Ioannidou: [00:30:15] And you, you succeeded.

Ghorbani: [00:30:16] It's incredible.

Wright: [00:30:17] Yes. Mission accomplished. I had no idea that that little thing was going to show up.

Ghorbani: [00:30:19] But you did like, you were saying your patients come in with what they want. Sometimes you get the like, I want my lips like this. I want my teeth like this. Yes. This blue, not this white, this blue.

Ioannidou: [00:30:29] Dr. Mina, you mentioned about the regulation. So who sets the standards, who regulates, who protects the professional, the dentist, the consumer from the AI application in dentistry?

Ghorbani: [00:30:41] So right now the FDA is doing the best it can and is learning lots and lots about what AI can and can't do in the health realm and medicine realm.

And same thing with the applications to dentistry. So, anything that's related to patient care and patient facing, and is used in clinical decision support is the actual term, is regulated by the FDA, so there has to be used to admit for your medical device to be cleared for patient use, just like you would a pacemaker, just not that level of device, right?

Or your blood pressure cuff, which is a lower level device. You go through the same process. So the FDA is part of that regulation, but in terms of setting standards, that's the conversation that's happening today, right? From, you can look at a global perspective or even within the ADA, there's a standards committee.

There's a white paper on augmented and Artificial Intelligence. It's very important to make the distinction here. The human is augmented by the Artificial Intelligence. It's not a replacement of our profession or our abilities. It's there to augment us and make us better. It's an assistive tool. And if you are interested, like we should be having these conversations. I'm engaging people to make sure that these standards are well set.

And we, as a, I mean, I say we as the Royal We, within the company that we set our set standards very highly because we want to perform as good as a dentist. Right. So how do we do that? So it's a respect. It's a respect thing for the industry. I mean, you want to talk about company and industry standards, making sure that we have clinicians helping build a product that's going to be used by clinicians.

We have to advocate for ourselves.

Ioannidou: [00:32:23] Many residences have 3D technologies. Many residences have scanners. Do you think that the way we train the dental students, I'm not talking about, you know, advanced training, but I'm talking about basic training in dental school. Do you think AI has infiltrated this training or we are a little bit behind in dental education?

I'm talking about, again, pre doc before you get your DDS or DMD.

Ghorbani: [00:32:46] If the school has a digital dentistry workflow, they have AI, they just don't know it. We've talked about that, right? If we're talking about this 2D application, there are schools that are interested. They're thinking about it. They're the early adopters, right?

That's amazing. I think we'll hear a lot about it at the next ADA conference in 2024 about more applications. They're doing the research right now to see how to best integrate it. But I will, I'm going to pose another, another question here. And we don't do gold foil fillings anymore. We just don't.

Right? So there, there's some, there's some skills that are going to become obsolete. We just have to decide how we're going to decide that we need to archive them so we make space for new skills. And I have two thoughts here. I don't know how long we're going to be still doing 2D radiography.

I'm really hoping for dental MRI to come out because I think it's amazing. But these skills will change that we need. And the curriculum needs to be as agile as the technology is being developed, because the marketing campaigns for these companies developing the technology don't wait for curriculum to update.

Ioannidou: [00:33:57] Good point.

Wright: [00:33:57] But then you're tasked with the fact that the faculty, who's going to be teaching and instructing on this curriculum, they have to, they have to be brought in and then they have to also have the training. So it's kind of like, are we always going to be behind or do we ever have the opportunity to get ahead?

Ioannidou: [00:34:13] I love it.

Ghorbani: [00:34:14] You mean that we just have to have a mindset that's open minded enough to enter, like welcome in new technology. That's the idea.

Wright: [00:34:22] The shift is in the mindset.

Ioannidou: [00:34:23] Yeah.

Ghorbani: [00:34:23] The shift is in the mindset. It's not that the new technologies are bad. But it's that, just how new startups say, fail fast, right?

So look at these new technologies, evaluate quickly whether or not it's going to impact the industry the right way.

Wright: [00:34:37] Love it.

Ioannidou: [00:34:38] What can dental professionals expect in the future for AI and dentistry? We know for fact that the dentists are not going to be replaced. We solved this problem. So what else do you see?

Ghorbani: [00:34:49] The thing that I'm really looking forward to is evidence-based consensus-optimized treatment planning.

Ioannidou: [00:34:58] That's a big, big statement.

Ghorbani: [00:35:00] I mean, think about it. Evidence based first, right? How do we make it evidence based? We do the research, we do the randomized control studies. We're seeing that there's value in the treatment. But then it's consensus based.

So is it used out in practice readily? Because now that we have large data and we can see treatment done over a nation, or internationally, we can see those outcomes. So where's the consensus actually falling, not just randomized control studies. So I think that's most profound for me. And then really taking into consideration the factors that we don't know about. Right?

Because as humans, we only have so much space to process. We're not going to be able to take in everyone's 12 page medical history. Let the machine do it. And by the way, the patient's not going to do it because they're poor historians anyway.

Wright: [00:35:53] Very true.

Ioannidou: [00:35:54] Yeah, yeah, yeah. Exactly. I mean, we all, we all are. We all forget.

Ghorbani: [00:35:58] Exactly. So let the machine tell the machine, hey, this is my medical history. This is my dental history. You clinically review what's in front of you in that day. You try to optimize. You give the patient the emotional care that they need as well. 

Wright: Because they really do need that. 

Ioannidou: I like this.

Ghorbani: We all need the emotional wellness. 

Wright: And that, that wasn't sarcasm. I'm being so serious. Like, I feel like our patients really do need that. 

Ghorbani: That's that future. That's the future I'm hoping for. Because that's how you optimize for overall oral health.

Ioannidou: [00:36:28] It's, well, I like what you said. It's, let the machine do what it knows. How to do well, and then, you know, still the dentist controls the decision making. Still the dentist controls the treatment. So it's not a threat for us. It's an assistance. It makes our life easier.

Announcer: [00:36:50] On the next Dental Sound Bites.

Wright: [00:36:52] Join us as we explore the fine line between being great at what we do and taking perfectionism too far.

We're talking about the pressures we put on ourselves, how that pressure impacts our mental health, and the strategies we can use to fight the perfectionism problem.

Wright: Oh, thank you so much for being here, Dr. Mina. This was such an amazing conversation. I think all of our listeners are going to be able to pull so much from all of the wisdom and the knowledge that you shared with us.

Ioannidou: [00:37:20] Thank you. Thank you so much. So much opportunity. We have a lot to learn and we have a lot to be looking forward to.

Ghorbani: [00:37:26] Well, I thank you both so much, Dr. Wright and Dr. Effie for a lovely conversation and allowing me to take you on this journey of artificial and augmented intelligence.  And we're on it together.

Wright: [00:37:36] Dr. Mina, can you please tell all of our listeners where can we find you? Where can they connect with you and chat a little bit more about AI and dentistry?

Ghorbani: [00:37:45] Well, I've added the Overjet link down at the bottom of the show notes, so that's one way to connect with me. You can find me on LinkedIn or you can send me an email at mina@overjet.ai.

Wright: [00:37:57] Awesome. It was wonderful. Thank you so much for being here, Dr. Mina.

Ghorbani: [00:38:00] Thank you for having me on Dental Sound Bites. It was a pleasure.

Wright: [00:38:03] We will have all the resources and information mentioned in this episode linked for you in the show notes on ADA.org/podcasts.

Ioannidou: [00:38:11] And don't forget as usual, if you like this episode, share it with a friend. Then be sure to subscribe to this podcast, wherever you're listening. So you can get the latest episodes.

Wright: [00:38:23] You can also rate and write a review and follow us on social media.

Ioannidou: [00:38:27] And don't forget the conversation continues on the ADA Member App. Catch this episode's bonus content, what you didn't hear on the show.

Announcer: [00:38:36] Thank you for joining us. Dental Sound Bites is an American Dental Association podcast.

You can also find this show resources and more on the ADA Member App and online at ADA.org/podcast.