Episode #516: Mike Buckner at ADA Smilecon

Your eye can see up to 50 shades of gray. But wouldn't it be great to see more in your X-rays?
Well, you can — with artificial intelligence! There are many things AI can do better than humans,
and Kirk Behrendt brings in Mike Buckner, senior vice president of partnerships at Pearl, to
explain what AI is, why it’s important for dentistry, and how a system like Pearl can elevate your
practice. AI is the future of dentistry! To learn more about Pearl and how to integrate it today,
listen to Episode 516 of The Best Practices Show!
Episode Resources:
• Mike’s email: [email protected]
• Pearl: https://www.hellopearl.com
• Subscribe to the Best Practices Show Podcast
• Join ACT’s To The Top Study Club
• Join ACT’s Master Class
• See our Live Events Schedule here
• Get the Best Practices Magazine for Free!
• Write a Review on iTunes
Links Mentioned in This Episode:
Oura Ring: https://ouraring.com
Main Takeaways:
AI does not replace, but it enhances human capabilities.
The average dentist picks up 70% to 75% of everything on an X-ray.
On average, AI’s accuracy for the same image is between 84% to 85%.
When dentists work with AI to analyze an image, accuracy increases to 90+%.
Pearl can provide a comprehensive report on all the undiagnosed opportunities.
Remember that AI is not perfect, and it can't detect 100% of everything, every time.
Quotes:
“You think about where artificial intelligence has been over the last few years in the medical
space, you think of mammograms and X-rays. Artificial intelligence has already been
incorporated in medical for quite some time now. And so, the fact that we are just now able to
bring this into dentistry in 2022 is a lot of fun to be on the front lines here.” (3:41—4:03)
“I don't think that we need to be worried about computers taking over dentistry at all. We know
that's not going to happen. But what we do know is there are some things that computers can
do that humans just straight-up can't.” (4:43—4:54)

“Are you ever in those situations where you're talking to a good friend and you're like, ‘Oh my
gosh, I've got to show you that picture of,’ fill in the blank, whatever that picture is, and you're
scrolling back, you're looking through your thousands of pictures that you take that you never do
anything with, and they're lost somewhere in your phone? Well, for example — and if you're at
home, you can do this — get your iPhone, go to the Photos app. Down in the search bar of the
Photos app, you can type in something very generic: bird, dog, tree. And AI will actually go in,
and within a second, find all of the pictures in your phone that have a dog, or a cat, or a tree in
those pictures. And that is, for an example, of how AI has been programmed to recognize what
looks like a dog.” (5:16—5:59)
“I didn't tag the photo with [my wife], but I have her picture in a contact, just her face shot, in my
phone as “Carly”. And so, the example that I gave you was I typed in “dog”, and I had 328
pictures of dogs that popped up on my phone. And I typed in “Carly” — thank goodness there
were more than 300 — there were like 3,000 pictures of her. But one of the pictures really
caught my eye. It was a picture of when we were in Tahiti, and it was a silhouette of her face. I
literally had to do a double take. I'm like, ‘Wait, is she even in this picture?’ And I look a little bit
closer and zoom in, and sure enough, she is. So, there are certain things that computers can
quickly analyze and see that are probably naked to the human eye.” (6:09—6:54)
“The human eye can see anywhere between 30 and 50 different shades of gray. Now, you think
of the X-rays that our doctors are viewing every day. The black and white, the grays, the human
eye can distinguish between 30 and 50 different shades. Computers and AI can come in and
now analyze and distinguish between 550 to 750 different shades of gray. So, we are able to go
in and, with the AI technology, pick up things that otherwise may have been missed, or catch
early-stage caries that oftentimes goes missed.” (6:57—7:35)
“[Pearl has] one of the largest collections of X-rays in the world. We started this process of
analyzing dental X-rays clear back in 2015, and really going in and highlighting, for example,
what caries looks like, widening of the periodontal ligament, calculus, etc., and programming
those things so that after millions of X-rays, the computer system can basically recognize and
learn and continue to improve. So, what's fascinating, even since I've been at Pearl from the
end of 2021 until now, even seeing the accuracy improve has been phenomenal.” (7:58—8:40)
“Looking at 2D images, bitewings, and periapicals, one of the studies we did was, basically, if
you have a dentist who is looking at a set of X-rays, the average dentist will pick up about 70%
to 75% of everything on that X-ray, as far as what's accurately depicted on that X-ray. Now,
when AI comes in and analyzes that image, now the accuracy is anywhere between 84%, 85%.
However, when AI analyzes the image and the dentist comes in and they work together, they
analyze that radiograph together, now the accuracy is upwards in the 90 percentiles of
everything that should be detected, picked up, and presented to the patient.” (9:02—9:48)
“Pearl has two platforms. The first one, we call Second Opinion. Second Opinion, all we need
there is to sync with the imaging software. So, the patient comes in, sits down, gets an X-ray
taken. And within a minute, minute-and-a-half, you now have all of the AI detections and
predictions up on that X-ray. So, you can see bone level measurements. You can see calculus,
caries, how far into the dentin, how far into the enamel the caries has protruded. You can see all
those details within a very short period of time after taking the X-ray. That's what we call Second
Opinion.” (10:05—10:44)

“We know that there's a lot of data there that we’re able to pick up [with Pearl]. So, if we can
take all of that data and all of the detections the AI found, and then we want to compare that
with what's in the practice management software — so, we’re actually looking to see, ‘All right, if
AI picked up this list of things here for this patient, what has actually been presented to the
patient? What's on the treatment plan?’” (10:46—11:08)
“For the first time, ever, a software is now able to provide a report, a very comprehensive report,
of all of the undiagnosed opportunity in a practice.” (11:16—11:28)
“When we work with a practice that is utilizing Practice Intelligence and we go and we analyze
the last 18 months of radiographs — and we’re choosing 18 months because that's generally
your active patients — it’s very rare that we find a practice that does not have over $1 million of
undiagnosed opportunity.” (11:55—12:16)
“How often do we find that we were looking at an X-ray, all of these things that we want to
present, but we know the patient in the chair, ‘Oh, they're going to freak out if we go through this
whole list of things. They're going to think that we’re overselling,’ so we maybe aim low? But

now, to be able to show, ‘Hey, here’s the X-ray. And all of our X-rays are backed by an FDA-
cleared AI platform.’ Let's use it as an educational experience. Let's talk about what AI has

picked up, what we’re looking at on the X-ray.” (12:40—13:08)
“[What dentists get wrong is thinking that] AI is perfect. AI is perfect, and it’s going to find
everything. Everything is going to be 100% accurate. It’s going to come pretty dang close. But
the example that I gave with the phone, like when I typed in “dog” into my phone, out of those
328 pictures of dogs, I had two that were deer. It got pretty dang close. And as I'm looking at it, I
can see, ‘Oh, I can see how the computer confused that.’ And we’re analyzing the X-ray. The
quality of X-rays, as you know, it ranges. And so, sometimes, we’ll come across a false positive,
for example. But the accuracy is continuing to improve, so much so that we can even distinguish
the difference between cervical burnout and caries.” (13:24—14:10)
“Most of our clients that come in and start utilizing Practice Intelligence — so, for us to analyze

the last 18 months of images, depending on the volume of images, we allow about two-and-a-
half weeks to really analyze all of those images and compile that report for the practice. And

going forward, we do a 40-minute training. And most of our clients will have this up in each
operatory as almost like their image reader, so that as soon as the X-rays come in, as soon as
the X-rays are taken, they can show this to the patient immediately. So, about a 40-minute
training.” (14:23—14:58)
“We really are making a difference in what that standard of care is and elevating the standard of
care. We did a case study with the Dental AI Council. We took 137 dentists and we had them all
review one FMX, and basically said, ‘Go in, analyze every tooth, how you would treat it, and
what the treatment plan would be.’ Out of 137 dentists, we had less than a 50% average
diagnostic concurrence. We had results all over the board. We had treatment plans ranging
from $300 to $36,000. It was all over the board. And yet, we are still relying on a technology that
is a doctor and the team reading black-and-white images.” (15:22—16:04)
Snippets:
0:00 Introduction.
2:54 Artificial Intelligence in the dental space.

4:03 AI, explained.
4:58 How AI is already part of your life.
6:00 Some advantages of using AI.
7:35 How Pearl helps with caries management.
9:48 Is Pearl easy to integrate with current systems?
10:45 Find your undiagnosed opportunity with Pearl.
13:09 What dentists get wrong about AI.
14:11 Other things AI can do, and the learning curve.
14:58 The problems that Pearl can solve.
16:06 More about Pearl and how to get in touch with Mike.
16:28 Q&A: What is the Oura Ring?
19:40 Q&A: Have you ever been mistaken for someone else?
21:23 Q&A: What do you do to blow off steam when you're fired up?
25:13 Q&A: Daily routine secrets.
Mike Buckner Bio:
Mike Buckner was most recently the director of business development for the dental division at
Weave, a communication platform for small business. He also served as the director of
business development at software company Dental Intelligence, and was a business
development executive at Solutionreach, a patient engagement platform for healthcare
practices.