Subscribe
Share
Share
Embed
Surgical robots can do all sorts of things better than humans. But human doctors have a breadth of experiential knowledge and instinct robots cannot replicate anytime soon. Understanding that difference is the key to designing a successful robotic platform. Galen Robotics is expanding the benefits of minimally invasive surgeries by enabling precise surgical maneuvers through human-machine cooperation.
Speed-to-data determines go-to-market success for medical devices. You need to inform critical decisions with user data, technical demonstration data, and clinical data. We interview med tech leaders about the critical data-driven decisions they make during their product development projects.
00:00:00:06 - 00:00:03:00
Welcome to MedTech Speed to Data a Key Tech podcast.
00:00:03:09 - 00:00:04:20
I'm your host, Andy Rogers,
00:00:04:20 - 00:00:06:23
VP of Business Development at Key Tech.
00:00:06:23 - 00:00:07:19
Each month, me
00:00:07:19 - 00:00:10:01
and a Key Techer are going to interview a MedTech leader
00:00:10:07 - 00:00:10:22
and talk to them
00:00:10:22 - 00:00:13:00
about the critical data driven decisions
00:00:13:06 - 00:00:16:08
they make in their programs.
00:00:17:18 - 00:00:18:18
All right, welcome everybody
00:00:18:18 - 00:00:21:02
to the next episode of MedTech Speed to Data
00:00:21:12 - 00:00:23:06
a KeyTech Podcast.
00:00:23:06 - 00:00:25:15
Today's guest, Dave Saunders,
00:00:25:23 - 00:00:29:18
CTO of Galen Robotics, a local Baltimore company.
00:00:30:06 - 00:00:31:13
Dave, welcome to the show.
00:00:31:13 - 00:00:33:08
Hey, thanks. It's great to be here.
00:00:33:08 - 00:00:34:12
Yeah, I mean, you're stirring up
00:00:34:12 - 00:00:37:06
a lot of excitement over there in the Carroll Camden area
00:00:37:12 - 00:00:39:01
with the with your startup company.
00:00:39:01 - 00:00:41:10
And you've been in Baltimore for a few years now.
00:00:42:21 - 00:00:43:04
Yeah.
00:00:43:12 - 00:00:46:00
Got a lot to cover here in this episode.
00:00:46:00 - 00:00:47:20
Very excited about it.
00:00:47:20 - 00:00:50:16
Usually we're interviewing maybe CEOs or folks
00:00:50:16 - 00:00:52:16
that are kind of kind of trying to run the company.
00:00:52:16 - 00:00:54:00
And I'm hoping to get into a little bit
00:00:54:00 - 00:00:55:22
more of the weeds with you today.
00:00:55:22 - 00:00:57:10
Not not too much, of course.
00:00:57:10 - 00:00:58:06
Okay.
00:00:58:06 - 00:00:59:11
My first question for you
00:00:59:11 - 00:01:02:16
related to Galen robotics is
00:01:02:16 - 00:01:07:02
what drew you to the company six years ago?
00:01:07:17 - 00:01:09:21
My co-founder and I, Bruce Lichorowic,
00:01:10:11 - 00:01:12:06
we had actually done
00:01:12:06 - 00:01:14:06
another medtech company together.
00:01:15:09 - 00:01:16:10
I actually
00:01:16:10 - 00:01:19:12
have known his family and worked with his brother
00:01:19:12 - 00:01:23:16
back in 1990 at a very early Internet startup.
00:01:23:16 - 00:01:26:11
So I've taken a lot of products to market,
00:01:26:11 - 00:01:31:06
I've done a lot of startups and in 2016
00:01:31:09 - 00:01:35:06
we were approached by the head of
00:01:35:06 - 00:01:38:18
Johns Hopkins Technology Ventures
00:01:38:18 - 00:01:40:16
actually at the
00:01:40:16 - 00:01:44:02
big medtech JPMorgan conference in San Francisco
00:01:44:13 - 00:01:47:10
and he said you know we've got said
00:01:47:22 - 00:01:51:00
we've got this new robotic technology,
00:01:51:00 - 00:01:53:16
we're looking for somebody to commercialize it.
00:01:54:02 - 00:01:56:00
It was developed by Russell Taylor
00:01:56:00 - 00:01:57:02
and I knew who he was.
00:01:57:02 - 00:02:00:11
He's basically, if you've got like,
00:02:00:11 - 00:02:02:23
baseball cards of like who's who in robotics
00:02:03:21 - 00:02:05:03
it all starts with Russ.
00:02:05:03 - 00:02:06:18
And so it was,
00:02:06:18 - 00:02:09:00
that right there got my interest.
00:02:09:00 - 00:02:11:11
And so we came out to Johns Hopkins
00:02:11:11 - 00:02:13:07
and we took a look at everything
00:02:13:07 - 00:02:15:02
and kicked the fairs gears for a few months[1][2]
00:02:15:02 - 00:02:20:15
and licensed the technology and began in San Jose
00:02:21:02 - 00:02:25:01
working out of our homes and we came out here
00:02:26:06 - 00:02:28:22
about one week out of every month
00:02:28:22 - 00:02:33:09
and just began to try to develop the technology.
00:02:33:09 - 00:02:34:03
But it was just
00:02:34:03 - 00:02:36:11
what was really interesting to me was
00:02:36:11 - 00:02:38:16
there are a lot of surgical robots
00:02:38:16 - 00:02:41:04
in the marketplace, like in spine alone,
00:02:41:14 - 00:02:43:10
there are seven surgical robots
00:02:43:10 - 00:02:45:18
just for doing pedicle screws and spine.
00:02:46:01 - 00:02:47:00
So there's actually a lot
00:02:47:00 - 00:02:48:15
more surgical robots out there
00:02:48:15 - 00:02:51:01
than a lot of people realize.
00:02:51:01 - 00:02:53:10
But in spite of how many there are,
00:02:53:10 - 00:02:56:19
there are huge, huge areas of unmet need
00:02:57:18 - 00:03:00:17
where surgeons were literally coming to
00:03:01:03 - 00:03:03:07
the university side of Johns Hopkins,
00:03:03:07 - 00:03:05:08
saying, Where are our tools?
00:03:05:08 - 00:03:07:12
Why don't we have the cool technology?
00:03:08:05 - 00:03:08:14
Right.
00:03:08:14 - 00:03:11:23
And and so this robot really checked
00:03:11:23 - 00:03:14:19
a lot of incredibly important boxes
00:03:16:02 - 00:03:17:07
for
00:03:17:23 - 00:03:19:19
what was needed in the industry.
00:03:19:19 - 00:03:21:00
And it just seemed like something
00:03:21:00 - 00:03:22:20
that could make a lot of impact
00:03:22:20 - 00:03:24:03
for a variety of reasons.
00:03:24:03 - 00:03:25:11
So I was I was really stoked.
00:03:25:11 - 00:03:29:18
And so, yeah, we started the company in 2016.
00:03:29:18 - 00:03:33:15
And then right before COVID we had actually come to
00:03:33:15 - 00:03:34:20
an agreement with
00:03:37:14 - 00:03:39:05
the state of Maryland
00:03:39:05 - 00:03:43:01
and we moved the company headquarters to Baltimore.
00:03:43:01 - 00:03:44:20
And now I'm here.
00:03:45:23 - 00:03:46:22
No, that's great.
00:03:46:22 - 00:03:47:13
It's always good
00:03:47:13 - 00:03:48:07
to hear
00:03:48:07 - 00:03:49:18
the screaming from the mountaintops
00:03:49:18 - 00:03:51:07
from your customers, right?
00:03:51:07 - 00:03:52:14
That's right. Yeah,
00:03:52:14 - 00:03:55:04
So, so as I understand, you know, your prototype,
00:03:55:14 - 00:03:58:06
you know, it's very much a collaborative robot
00:03:58:06 - 00:03:59:17
and we'll talk a little bit more towards the end
00:03:59:17 - 00:04:01:12
about the future of robotics.
00:04:01:12 - 00:04:02:21
But I want to make it clear to our audience
00:04:02:21 - 00:04:05:12
that this is very much a tool
00:04:06:00 - 00:04:08:21
that's collaborating with the human.
00:04:08:21 - 00:04:11:06
It's a machine collaborating with the human.
00:04:11:06 - 00:04:12:07
So why don't you just describe
00:04:12:07 - 00:04:13:09
to our audience real quick,
00:04:13:09 - 00:04:14:15
you know, the make up of the product,
00:04:14:15 - 00:04:16:11
kind of how the user interacts with it
00:04:16:11 - 00:04:18:20
and how it's used in microsurgery.
00:04:18:20 - 00:04:19:06
Sure.
00:04:19:06 - 00:04:19:21
You know,
00:04:19:21 - 00:04:21:08
when you look at surgical robotics,
00:04:21:08 - 00:04:22:06
just to kind of like lay
00:04:22:06 - 00:04:23:04
the groundwork,
00:04:23:04 - 00:04:27:01
you've got kind of a a continuum of types of robots.
00:04:27:01 - 00:04:30:12
And on one far side, you've got things like the robot
00:04:30:12 - 00:04:32:00
that does Lasik surgery,
00:04:32:00 - 00:04:33:16
you know, to improve your vision,
00:04:33:16 - 00:04:35:16
you set it up, you push a button
00:04:35:16 - 00:04:38:04
and then the ophthalmologist goes and drinks a soda,
00:04:38:04 - 00:04:38:20
something like that.
00:04:38:20 - 00:04:40:03
While it's doing its thing,
00:04:40:03 - 00:04:42:14
it's fully autonomous, it's navigated, it
00:04:42:15 - 00:04:43:21
runs by itself.
00:04:43:21 - 00:04:45:08
Once you've set it up
00:04:45:08 - 00:04:47:23
and there are other robots that do similar
00:04:48:00 - 00:04:51:18
kinds of things for knees and hips, and usually bony
00:04:51:18 - 00:04:53:16
robots are autonomous.
00:04:53:16 - 00:04:55:18
And then you've got a category of robots
00:04:55:22 - 00:04:58:14
like The da Vinci, the famous one.
00:04:58:14 - 00:05:00:10
Right. And it's
00:05:01:15 - 00:05:03:19
you could describe it as a collaborative robot,
00:05:03:19 - 00:05:04:05
but it's
00:05:04:05 - 00:05:06:02
but people are more likely
00:05:06:02 - 00:05:08:07
to recognize it as a remote controlled robot.
00:05:08:15 - 00:05:11:12
So you put your hands in these little special controls
00:05:11:12 - 00:05:14:15
and you actually move laparoscopic rods around,
00:05:15:04 - 00:05:19:12
but it is 100% guided by the surgeon in real time.
00:05:20:00 - 00:05:21:22
It has some additional abilities
00:05:21:22 - 00:05:23:03
that are under development
00:05:23:03 - 00:05:26:03
that will do little things on its own here
00:05:26:03 - 00:05:27:13
and there in a very collaborative
00:05:27:13 - 00:05:29:05
kind of back and forth way.
00:05:29:05 - 00:05:31:18
But for the most part, it's a passive robot
00:05:31:18 - 00:05:32:21
that is basically
00:05:32:21 - 00:05:34:23
just extending the hands of the surgeon
00:05:36:09 - 00:05:37:19
for what we're doing.
00:05:37:19 - 00:05:39:02
We had a challenge
00:05:39:02 - 00:05:41:12
where there are a lot of procedures,
00:05:41:12 - 00:05:43:21
especially up in the head and neck,
00:05:43:21 - 00:05:47:05
where a lap, a robot that's been optimized
00:05:47:05 - 00:05:48:02
for laparoscopic
00:05:48:02 - 00:05:51:03
surgery is just not the right solution.
00:05:51:07 - 00:05:54:15
It just won't fit into those kinds of environments.
00:05:54:15 - 00:05:56:23
And so the surgeons needed something
00:05:56:23 - 00:05:58:17
that really didn't.
00:05:58:17 - 00:05:59:19
It's not like they were looking for
00:05:59:19 - 00:06:01:12
a new way to do surgery,
00:06:01:12 - 00:06:04:09
but what they needed was either a third hand
00:06:04:09 - 00:06:05:06
or some way
00:06:05:06 - 00:06:07:13
to really stabilize the instrument
00:06:07:13 - 00:06:09:07
that they were using
00:06:09:07 - 00:06:12:15
and and support the surgeries as they were being done.
00:06:12:23 - 00:06:15:15
So what our robot is is
00:06:15:18 - 00:06:17:19
optimized for is basically
00:06:17:19 - 00:06:19:12
taking a surgeon's instrument
00:06:19:12 - 00:06:23:02
that they would already be using, holding onto it.
00:06:23:11 - 00:06:25:06
And then just like power steering,
00:06:25:06 - 00:06:26:05
you know, when you're in a car
00:06:26:05 - 00:06:28:09
and you turn the steering wheel,
00:06:28:09 - 00:06:31:07
there's a sensor that actually then ultimately,
00:06:31:07 - 00:06:31:14
you know,
00:06:31:14 - 00:06:33:12
feeds into a linkage mechanism
00:06:33:12 - 00:06:35:10
and turns your tires right.
00:06:35:10 - 00:06:38:07
You feel like you're in control, but there's actually
00:06:38:07 - 00:06:39:07
a mechanical assist
00:06:39:07 - 00:06:41:08
that's doing all of the heavy lifting, so to speak.
00:06:41:17 - 00:06:42:16
Well, in this case,
00:06:42:16 - 00:06:45:09
the robot is technically moving the instrument,
00:06:45:09 - 00:06:47:10
even though you're holding on to it,
00:06:47:10 - 00:06:49:17
where you would normally hold on to it.
00:06:49:17 - 00:06:52:04
But the result of that is the instrument is
00:06:52:04 - 00:06:55:05
now extremely stable and
00:06:56:21 - 00:06:58:15
you can even do things like
00:06:58:15 - 00:06:58:23
there's like
00:06:58:23 - 00:07:00:00
a pedal on the floor
00:07:00:00 - 00:07:01:02
that engages the robot
00:07:01:02 - 00:07:03:00
to allow it to follow your hand.
00:07:03:00 - 00:07:04:09
If you let go of the pedal,
00:07:04:09 - 00:07:06:05
you can actually let go of the instrument.
00:07:06:05 - 00:07:07:12
Can't do that normally.
00:07:07:12 - 00:07:08:01
Like if you've got
00:07:08:01 - 00:07:08:21
an instrument down
00:07:08:21 - 00:07:09:20
like in somebody's middle
00:07:09:20 - 00:07:12:07
ear or a biopsy needle in somebody's
00:07:12:07 - 00:07:13:14
brain or whatever,
00:07:13:14 - 00:07:15:13
you can't normally let go of those
00:07:15:13 - 00:07:18:19
and have them be held steady, rock solid, steady.
00:07:19:04 - 00:07:20:13
But now you have this ability.
00:07:20:13 - 00:07:21:11
So not only do
00:07:21:11 - 00:07:22:17
you get assistance
00:07:22:17 - 00:07:24:20
while you're moving the instruments around,
00:07:24:20 - 00:07:27:08
but you also have these new capabilities
00:07:27:08 - 00:07:31:04
for kind of like a third hand that can ergonomically
00:07:31:04 - 00:07:33:13
fit into surgical environments.
00:07:34:12 - 00:07:35:20
That are very, very challenging
00:07:35:20 - 00:07:38:03
just because they're really tight.
00:07:38:03 - 00:07:38:20
You know,
00:07:39:07 - 00:07:41:11
just give me one example and then we'll move on
00:07:41:11 - 00:07:42:22
is imagine
00:07:42:22 - 00:07:43:22
There's a transsphenoidal
00:07:43:22 - 00:07:44:20
a procedure
00:07:44:20 - 00:07:48:02
where they're going up your nostril to basically do
00:07:48:21 - 00:07:50:11
different kinds of brain surgeries,
00:07:50:11 - 00:07:52:15
typically like to remove the pituitary gland
00:07:53:10 - 00:07:54:10
sometimes
00:07:54:10 - 00:07:56:12
because of the requirements of the procedure.
00:07:56:12 - 00:07:59:12
You'll have between two and four surgeons
00:07:59:21 - 00:08:01:09
all cramming their instruments,
00:08:01:09 - 00:08:04:00
plus an endoscope up into your two nostrils.
00:08:04:15 - 00:08:06:19
It's it's horrible, right?
00:08:06:19 - 00:08:07:12
Every you can
00:08:07:12 - 00:08:09:17
only so many people can lean into the table
00:08:09:18 - 00:08:11:08
and reach across like that.
00:08:11:08 - 00:08:13:03
And so those are the environments
00:08:13:03 - 00:08:17:05
where you really need better technology, better tools
00:08:17:11 - 00:08:20:13
to assist the surgeons for what they know how to do.
00:08:20:22 - 00:08:23:11
They just need that kind of leg up
00:08:23:11 - 00:08:24:18
technology wise.
00:08:24:18 - 00:08:27:00
So that's that's where our robot is going after.
00:08:27:15 - 00:08:28:11
Got it. Yeah.
00:08:28:11 - 00:08:30:12
So to speed things up a little bit,
00:08:30:12 - 00:08:33:05
I'm also, sorry, a few things now.
00:08:33:05 - 00:08:34:22
That was great. A great background.
00:08:34:22 - 00:08:36:23
But you know, to a couple other elements.
00:08:36:23 - 00:08:38:12
Number one that I picked up
00:08:38:12 - 00:08:40:08
that you don't have a disposable
00:08:40:08 - 00:08:41:21
like you're designing your
00:08:41:21 - 00:08:45:06
your platform to accommodate off the shelf
00:08:45:18 - 00:08:48:06
surgical tools, which I think is pretty unique
00:08:48:06 - 00:08:49:04
or maybe that's changed.
00:08:49:04 - 00:08:50:16
But the second thing. There is a
00:08:50:16 - 00:08:53:00
disposable in that there is a disposable.
00:08:53:00 - 00:08:55:19
adapter. Yeah, that's fair.
00:08:55:19 - 00:08:56:06
That's right.
00:08:56:06 - 00:08:56:21
But yeah,
00:08:56:21 - 00:08:59:10
the tools themselves you're partnering with
00:09:00:02 - 00:09:02:16
on market products, that's pretty
00:09:02:16 - 00:09:05:12
a great way to speed yourself to market.
00:09:05:12 - 00:09:06:10
And then the other
00:09:06:10 - 00:09:06:23
I think clever
00:09:06:23 - 00:09:09:05
part of of your business that I've read about
00:09:09:05 - 00:09:11:02
and people can read of it up on it
00:09:11:02 - 00:09:12:03
offline is that
00:09:12:03 - 00:09:12:12
you're going
00:09:12:12 - 00:09:13:11
after the subscription
00:09:13:11 - 00:09:15:20
model versus the capital expense.
00:09:15:20 - 00:09:18:05
So it allows you to get in the door easier.
00:09:18:05 - 00:09:21:13
But, but one area that I thought was interesting
00:09:21:13 - 00:09:22:09
and worth talking about
00:09:22:09 - 00:09:26:09
at a high level about your company is that one
00:09:26:09 - 00:09:27:19
there isn't one particular
00:09:27:19 - 00:09:30:10
sliver of the surgical market
00:09:30:10 - 00:09:34:06
that your robot would justify the expense but it's
00:09:34:06 - 00:09:38:13
there's multiple areas in like soft tissue surgery
00:09:38:13 - 00:09:40:03
and the combination of all of them
00:09:40:03 - 00:09:42:09
makes your business viable.
00:09:42:09 - 00:09:43:21
Can you talk a little bit about that?
00:09:43:21 - 00:09:45:12
Yeah. You know,
00:09:45:12 - 00:09:48:20
one of the big I think Achilles heels of robots in
00:09:48:20 - 00:09:50:01
this industry is so
00:09:50:01 - 00:09:52:03
many of them are optimized
00:09:52:03 - 00:09:54:05
for an individual procedure.
00:09:54:05 - 00:09:58:18
So, if you want to do a total knee replacement
00:09:58:18 - 00:10:01:01
there are robots out there for it.
00:10:01:01 - 00:10:03:18
But that robot can't be moved by another department.
00:10:03:18 - 00:10:04:08
It's not going to do
00:10:04:08 - 00:10:05:16
Lasik, it's not going to do pedicle
00:10:05:16 - 00:10:07:21
screws, it's not going to do prostatectomies
00:10:07:21 - 00:10:08:16
and things like that.
00:10:08:16 - 00:10:10:08
And so you end up with a lot of these one
00:10:10:08 - 00:10:12:03
trick ponies out there
00:10:12:03 - 00:10:15:01
that have million dollar plus price tags,
00:10:16:00 - 00:10:18:10
but very narrow utility.
00:10:18:10 - 00:10:20:18
And that's a that's a huge problem in this industry.
00:10:20:18 - 00:10:22:07
Because I'll tell you,
00:10:22:07 - 00:10:24:18
I've talked a lot of hospital administrators
00:10:24:18 - 00:10:27:07
who refer to those things as just giant
00:10:27:07 - 00:10:29:03
fancy dust collectors.
00:10:29:03 - 00:10:32:08
And so with our platform,
00:10:32:08 - 00:10:35:15
by simply changing the instrument that it's holding,
00:10:35:15 - 00:10:38:23
we can potentially move into multiple departments
00:10:38:23 - 00:10:42:04
and provide real value too to the surgeon.
00:10:42:04 - 00:10:45:19
And all that changes is what instrument we're holding.
00:10:45:19 - 00:10:47:16
And there are a lot of opportunities that way.
00:10:47:16 - 00:10:49:17
So it's a great
00:10:49:17 - 00:10:51:09
it's a great opportunity for us.
00:10:51:09 - 00:10:52:18
I've done my homework.
00:10:52:18 - 00:10:55:19
I know your your favorite book was Crossing the Chasm.
00:10:55:19 - 00:10:56:18
So here we go.
00:10:56:18 - 00:10:58:13
We're going to Cross the Chasm.
00:10:58:13 - 00:11:00:09
I have not read it myself, but
00:11:00:09 - 00:11:02:18
you know, the podcast is called Speed to Data here.
00:11:02:18 - 00:11:06:15
So I want to cover what data was most critical
00:11:06:15 - 00:11:08:13
as you were crossing the chasm.
00:11:08:13 - 00:11:09:12
And granted,
00:11:09:12 - 00:11:12:04
we're always in the chasm, so to speak, and
00:11:12:23 - 00:11:14:09
in the world we're in.
00:11:14:09 - 00:11:16:00
But we'll start with,
00:11:16:00 - 00:11:16:10
you know, right
00:11:16:10 - 00:11:18:16
when you started the company, because our audience,
00:11:18:16 - 00:11:19:11
you know, is
00:11:19:11 - 00:11:21:01
clearly at the startup world
00:11:21:01 - 00:11:21:21
where they
00:11:21:21 - 00:11:24:15
maybe are licensing technology like you did or
00:11:24:15 - 00:11:26:21
or even like a global company that sees promise.
00:11:26:21 - 00:11:28:00
And there's,
00:11:28:00 - 00:11:29:02
you know,
00:11:29:02 - 00:11:31:23
a prototype benchtop prototype that shows feasibility.
00:11:31:23 - 00:11:34:01
So there's the early stage
00:11:34:01 - 00:11:35:06
and then there's the stage
00:11:35:06 - 00:11:36:04
where you're kind of in now
00:11:36:04 - 00:11:36:21
like, you know,
00:11:36:21 - 00:11:38:00
hard core product
00:11:38:00 - 00:11:40:03
development, classic product development.
00:11:40:03 - 00:11:42:12
What data was most critical for you
00:11:42:12 - 00:11:44:04
once you got more funding
00:11:44:04 - 00:11:46:04
and then here towards the end and in the future,
00:11:46:04 - 00:11:46:21
more commonly,
00:11:48:04 - 00:11:49:00
what data is and
00:11:49:00 - 00:11:50:13
how is that data going to be leveraged?
00:11:50:13 - 00:11:51:18
So let's start,
00:11:52:07 - 00:11:54:15
you know, day or maybe month
00:11:54:15 - 00:11:57:00
six after you, you know,
00:11:57:00 - 00:11:58:16
stop doing the other work that you're doing.
00:11:58:16 - 00:12:00:23
Maybe in you're all in on on Galen,
00:12:00:23 - 00:12:04:18
what what data was most critical for you
00:12:04:18 - 00:12:08:18
as the CTO and kind of as a company?
00:12:08:18 - 00:12:10:11
To really,
00:12:10:11 - 00:12:12:23
show that this product had viability?
00:12:12:23 - 00:12:13:17
I mean, obviously,
00:12:13:17 - 00:12:15:19
the surgeons were saying, hey, we want the product.
00:12:15:19 - 00:12:18:20
And and Hopkins was happy to license it to you,
00:12:19:08 - 00:12:21:02
but what were you focused on?
00:12:21:02 - 00:12:21:20
Well, you know, it is
00:12:21:20 - 00:12:22:19
it is interesting
00:12:22:19 - 00:12:24:20
in a regulated industry like this
00:12:24:20 - 00:12:26:23
that it's very difficult to
00:12:28:08 - 00:12:31:01
obviously do a thing that smells like pre marketing.
00:12:31:20 - 00:12:33:10
But quite honestly,
00:12:33:10 - 00:12:36:09
you need to identify a clinical application
00:12:36:09 - 00:12:38:18
that has value to the surgeon
00:12:38:18 - 00:12:40:04
and by value to the surgeon.
00:12:40:04 - 00:12:42:18
I mean that they're actually going to go to bat
00:12:42:18 - 00:12:44:06
to get that paid for
00:12:46:19 - 00:12:50:13
and this is true of any industry,
00:12:51:02 - 00:12:53:16
I've taken over 40 products to market
00:12:54:04 - 00:12:55:21
and in every single case
00:12:55:21 - 00:12:58:18
the most important thing has always been
00:12:58:18 - 00:13:01:19
what does the end user actually want to do?
00:13:02:09 - 00:13:03:21
You know,
00:13:03:21 - 00:13:08:18
something that I know drives engineers just bananas is
00:13:08:18 - 00:13:10:12
when product managers just hand off
00:13:10:12 - 00:13:13:06
like a laundry list of features with no context.
00:13:13:06 - 00:13:16:20
No just tell me what they're trying to do
00:13:16:20 - 00:13:21:03
and that is certainly qualitative data
00:13:21:03 - 00:13:24:21
but it's really important and so especially for
00:13:24:21 - 00:13:28:14
for us in medical robotics,
00:13:29:03 - 00:13:30:15
we had to go into surgeries.
00:13:30:15 - 00:13:33:06
We had to see what the surgeons were doing
00:13:33:06 - 00:13:35:05
so that we could in real time
00:13:35:05 - 00:13:38:13
have them tell us, see how much of a pain this is.
00:13:38:13 - 00:13:40:04
Look at what I'm doing right here.
00:13:40:04 - 00:13:42:14
Look at the ergonomics of what I'm trying to do here.
00:13:42:14 - 00:13:46:08
And this is why this is so difficult today.
00:13:46:08 - 00:13:47:12
Right?
00:13:47:12 - 00:13:50:13
And this is why it's so difficult for this surgery
00:13:50:13 - 00:13:51:01
to be done
00:13:51:01 - 00:13:51:18
at a secondary
00:13:51:18 - 00:13:55:11
hospital, because, I have the facilities
00:13:55:11 - 00:13:57:01
and I'm one of the few people
00:13:57:01 - 00:13:59:11
that had the magic hands to get through this.
00:13:59:11 - 00:14:02:11
And yet every other surgeon
00:14:02:11 - 00:14:05:10
that maybe can't qualify to do the procedure,
00:14:05:10 - 00:14:06:23
they know all of it.
00:14:06:23 - 00:14:08:18
they're certainly trained
00:14:08:18 - 00:14:11:10
well enough to know how to do it.
00:14:11:10 - 00:14:13:12
But it's just a little bit of outreach
00:14:13:12 - 00:14:16:06
from from a dexterity or an ergonomic standpoint.
00:14:16:06 - 00:14:18:07
And so we really had to get that under our skin
00:14:18:07 - 00:14:21:08
and understand that, on the other side
00:14:21:08 - 00:14:23:11
of the same coin, I suppose,
00:14:23:11 - 00:14:25:07
is we had to then understand
00:14:26:08 - 00:14:27:11
how that
00:14:27:11 - 00:14:30:08
translated into kind of mechanical numbers.
00:14:31:10 - 00:14:33:15
You know, you take a look at something like and I'm
00:14:33:15 - 00:14:36:07
not trashing the da Vinci in any possible way here.
00:14:36:20 - 00:14:40:05
You know, it's a fantastic system for what it does,
00:14:40:05 - 00:14:43:12
but it is optimized for the kinds of motions
00:14:43:12 - 00:14:46:01
required in laparoscopic surgery, right.
00:14:46:13 - 00:14:50:13
When I'm supporting a procedure
00:14:50:13 - 00:14:52:22
exploratory, of course, because we're pre FDA
00:14:53:16 - 00:14:55:11
but when we're looking at a procedure, say,
00:14:55:11 - 00:14:59:01
for the middle ear, I'm looking at zones
00:14:59:01 - 00:15:00:19
that are less than a millimeter
00:15:00:19 - 00:15:03:12
when I'm trying to like go through like a mastoidectomy
00:15:03:12 - 00:15:06:15
I've got like sometimes a millimeter of clearance
00:15:07:21 - 00:15:10:21
that is going to be either a successful operation
00:15:10:21 - 00:15:14:08
or patient or irreparable patient injury in the case
00:15:14:08 - 00:15:15:18
of middle ear surgery. Right.
00:15:16:23 - 00:15:18:05
For prostatectomy,
00:15:18:05 - 00:15:21:18
nerve sparing technique for doing a colon resection,
00:15:21:18 - 00:15:24:02
you've got a little bit more give, right?
00:15:24:02 - 00:15:26:02
And so the guys engineering
00:15:26:02 - 00:15:28:06
something like a laparoscopic robot,
00:15:28:06 - 00:15:30:03
they don't have the requirement
00:15:30:03 - 00:15:33:15
to like have like 50 microns of precision motion
00:15:34:13 - 00:15:36:10
for their minimal increments
00:15:36:10 - 00:15:38:14
of moving the end effectors.
00:15:38:14 - 00:15:41:04
We do, right.
00:15:41:04 - 00:15:43:20
It would be over engineering for a laparoscopic
00:15:43:20 - 00:15:45:22
robot to dial in their hardware
00:15:45:22 - 00:15:47:21
to the kind of precision that we're doing
00:15:47:21 - 00:15:51:09
because it's unnecessary for laparoscopic procedures.
00:15:51:20 - 00:15:55:17
So understanding what your scale
00:15:55:17 - 00:15:58:19
is and understanding what you need to do
00:15:58:19 - 00:16:00:20
not only from a range standpoint
00:16:00:20 - 00:16:03:03
like what are my potential procedures
00:16:03:03 - 00:16:03:23
that I could be doing
00:16:03:23 - 00:16:06:07
or what are the kinds of things that this platform
00:16:06:07 - 00:16:07:21
could be doing in the future?
00:16:07:21 - 00:16:11:05
And can I reach into those with what I'm making now,
00:16:11:05 - 00:16:14:17
but also having a perspective on where
00:16:14:17 - 00:16:17:14
your current scale of motion,
00:16:18:07 - 00:16:20:19
all of those things relate to, you know, heck,
00:16:20:19 - 00:16:22:18
what's kind of screws you're going to choose
00:16:22:18 - 00:16:23:11
what kind of metal
00:16:23:11 - 00:16:24:14
are you going to choose for the chassis
00:16:24:14 - 00:16:26:02
to to provide enough stiffness?
00:16:26:02 - 00:16:27:03
All of those things
00:16:27:03 - 00:16:30:22
will come down to having that initial understanding
00:16:30:22 - 00:16:32:10
of what kind of scale
00:16:32:10 - 00:16:34:22
am I really trying to achieve here.
00:16:34:22 - 00:16:36:10
And that's critical
00:16:36:10 - 00:16:39:11
because you could really end up either
00:16:39:11 - 00:16:40:04
over engineering
00:16:40:04 - 00:16:42:00
or under engineering a product
00:16:42:00 - 00:16:44:02
and just missing the boat
00:16:44:02 - 00:16:47:20
for your domain of surgeries or procedures
00:16:47:20 - 00:16:48:20
or whatever, right?
00:16:48:20 - 00:16:49:01
Yeah.
00:16:49:01 - 00:16:49:14
No, I mean,
00:16:49:14 - 00:16:50:06
I think
00:16:50:06 - 00:16:53:20
translating market needs into product requirements.
00:16:53:20 - 00:16:55:03
I mean, you just nailed it.
00:16:55:03 - 00:16:56:08
And I totally agree.
00:16:56:08 - 00:16:59:08
You don�t want to over or under engineer your product.
00:16:59:08 - 00:17:03:21
So my my next question here is what steps
00:17:03:21 - 00:17:08:10
did you take to evaluate a university prototype,
00:17:10:09 - 00:17:11:16
to evaluate whether that prototype
00:17:11:16 - 00:17:14:12
could actually meet some of these requirements?
00:17:14:18 - 00:17:15:15
Well, fortunately,
00:17:15:15 - 00:17:18:20
a lot of legwork had been done for us.
00:17:19:11 - 00:17:20:00
But I'll tell you,
00:17:20:00 - 00:17:22:03
one of the one of the most interesting things,
00:17:22:03 - 00:17:23:00
I don't have a prop here.
00:17:23:00 - 00:17:25:09
It probably took me too long to get it.
00:17:25:09 - 00:17:27:10
But one of the interesting things that happened
00:17:27:10 - 00:17:29:04
during the early days of the development
00:17:29:04 - 00:17:33:03
for this technology was another grad student
00:17:33:03 - 00:17:36:14
had actually done kind of a joking YouTube video
00:17:36:14 - 00:17:38:11
where she was actually playing
00:17:38:11 - 00:17:41:01
the old Mattel Operation game
00:17:41:01 - 00:17:44:13
with the da Vinci that they have at Johns Hopkins.
00:17:44:13 - 00:17:47:00
And it's pretty funny to watch, right?
00:17:47:19 - 00:17:50:10
You got this giant $2 million robot and you're
00:17:50:10 - 00:17:55:01
you're plucking out the funny bone from the little
00:17:56:12 - 00:17:57:02
kids toy.
00:17:57:02 - 00:17:57:16
Right.
00:17:58:01 - 00:18:01:08
And Russ Taylor looked at that and he said, well,
00:18:01:08 - 00:18:03:12
you know, that's interesting.
00:18:03:12 - 00:18:07:01
But those pieces are gigantic compared
00:18:07:01 - 00:18:09:02
to the kinds of anatomy
00:18:09:02 - 00:18:11:03
you'd be going after with this robot.
00:18:11:03 - 00:18:13:10
You know, you'd be looking at say, for example,
00:18:14:06 - 00:18:17:07
you're sometimes doing like one potential procedure
00:18:17:07 - 00:18:19:17
is for laryngeal removing like a cyst
00:18:19:17 - 00:18:20:22
from your vocal cords.
00:18:20:22 - 00:18:23:20
You're doing that with a 25 centimeter long instrument
00:18:23:20 - 00:18:25:12
and you're trying to grasp on tissue
00:18:25:12 - 00:18:27:17
by bringing it down into the person's mouth
00:18:27:17 - 00:18:29:17
and grabbing on this one millimeter,
00:18:29:17 - 00:18:31:15
maybe one and a half millimeter polyp.
00:18:31:15 - 00:18:33:15
That's enough to really mess up your voice.
00:18:33:15 - 00:18:35:00
And you're trying to grab on that,
00:18:35:00 - 00:18:37:09
you know, make an incision, all that sort of thing.
00:18:37:09 - 00:18:40:06
So how can you support
00:18:40:06 - 00:18:43:00
those hand motions to make that more precise?
00:18:43:13 - 00:18:45:20
And so instead of the Mattel Monopoly[3] game,
00:18:45:20 - 00:18:48:02
what they ended up developing
00:18:48:02 - 00:18:52:14
was this little piece of sheet metal with a spiral
00:18:52:14 - 00:18:55:20
that was a millimeter the channel, and the spiral was
00:18:55:20 - 00:18:56:17
a millimeter
00:18:57:22 - 00:18:58:22
gap.
00:18:58:22 - 00:19:02:06
And the and the game basically was to take
00:19:02:06 - 00:19:04:01
a laryngeal instrument,
00:19:04:01 - 00:19:05:23
which is about 25 centimeters long.
00:19:05:23 - 00:19:07:21
And you put the tip into the channel
00:19:07:21 - 00:19:09:18
and you had to run it through the channel
00:19:09:18 - 00:19:11:14
without ever touching the sides.
00:19:11:14 - 00:19:14:11
Now, that is very analogous to the kinds of hand
00:19:14:11 - 00:19:17:11
motions that you would do in these narrow
00:19:17:11 - 00:19:18:23
corridor procedures
00:19:18:23 - 00:19:21:09
that you're doing of upper head, neck.
00:19:22:01 - 00:19:24:09
And so just being able to do that
00:19:25:03 - 00:19:26:19
was actually a really big deal.
00:19:26:19 - 00:19:27:02
In fact,
00:19:27:02 - 00:19:28:01
there was a study
00:19:28:01 - 00:19:29:18
that ended up getting published
00:19:29:18 - 00:19:31:18
showing just great results
00:19:31:18 - 00:19:33:09
just by being able to do that.
00:19:33:09 - 00:19:37:02
And we ended up taking our first commercial prototype
00:19:37:02 - 00:19:40:01
of the robot to a otolaryngology
00:19:41:00 - 00:19:42:16
educational conference.
00:19:42:16 - 00:19:44:13
And we had over 70
00:19:45:10 - 00:19:49:10
ENT Surgeon Department heads come into it
00:19:49:10 - 00:19:52:19
by invitation only demo.
00:19:52:19 - 00:19:55:04
And we actually had them play the game
00:19:56:15 - 00:19:59:10
and we had people their jaws are on the floor.
00:19:59:10 - 00:20:04:04
They're like, Holy cow, we only had one person
00:20:04:04 - 00:20:07:22
that could actually beat the robot freehand
00:20:07:22 - 00:20:11:06
and they couldn't do it repeatedly.
00:20:11:06 - 00:20:13:11
So just so they had one good shot, right?
00:20:13:11 - 00:20:15:18
They did one good thing.
00:20:15:18 - 00:20:18:02
But that was but that was great because what
00:20:18:02 - 00:20:21:13
it was what that really helped us be able to see
00:20:21:13 - 00:20:24:22
was that even the less nerdy people in the company
00:20:24:22 - 00:20:27:23
could go look at what we're helping surgeons
00:20:27:23 - 00:20:30:23
do that was something that we could really point at.
00:20:30:23 - 00:20:33:16
And then by using, camera vision and things
00:20:33:16 - 00:20:36:05
like that, we're able to actually measure
00:20:36:05 - 00:20:38:20
motion of the tool inside the spiral.
00:20:38:20 - 00:20:40:00
And that's how we were able to
00:20:40:00 - 00:20:42:12
then eventually quantify
00:20:42:12 - 00:20:44:13
how much we were stabilizing the instrument.
00:20:45:06 - 00:20:46:07
But, you know,
00:20:46:07 - 00:20:50:03
just being able to as funny as it sounds,
00:20:50:03 - 00:20:54:15
being able to build the equivalent of a kid's game
00:20:55:14 - 00:20:56:06
and actually
00:20:56:06 - 00:20:59:12
use that as a mechanism for,
00:20:59:12 - 00:21:03:15
validating the function of the tool with surgeons.
00:21:03:15 - 00:21:04:13
It just clicked.
00:21:04:13 - 00:21:06:21
It just every one of the surgeons was like,
00:21:06:21 - 00:21:09:11
holy cow, this, you know, not only will
00:21:09:11 - 00:21:10:11
this make my life easier,
00:21:10:11 - 00:21:11:07
this will make it easier
00:21:11:07 - 00:21:12:15
for me to train residents
00:21:12:15 - 00:21:13:19
because they'll spend
00:21:13:19 - 00:21:16:22
less time focusing on all of the annoying techniques
00:21:16:22 - 00:21:17:21
for mitigating hand
00:21:17:21 - 00:21:19:08
tremor and blah, blah, blah, blah, blah, blah.
00:21:19:08 - 00:21:22:11
And they're going to focus on clinical effectiveness.
00:21:22:11 - 00:21:25:01
How do I do the surgery correctly? Right.
00:21:25:01 - 00:21:25:19
And so it was
00:21:25:19 - 00:21:27:22
it was huge for us to be able to
00:21:27:22 - 00:21:30:10
to be able to make that connection.
00:21:30:10 - 00:21:30:22
We still do.
00:21:30:22 - 00:21:31:09
Yeah.
00:21:31:09 - 00:21:32:06
It's such a great story
00:21:32:06 - 00:21:35:11
that I mean, it's not rocket science here
00:21:35:11 - 00:21:36:02
to the extent
00:21:36:02 - 00:21:40:04
you can use the benchtop prototype that you inherit
00:21:40:04 - 00:21:43:12
or even mock up yourself and,
00:21:43:12 - 00:21:46:20
have your end users use that.
00:21:46:20 - 00:21:48:23
And there's either published data
00:21:48:23 - 00:21:51:18
or if you can use computer vision, as you said, to,
00:21:51:18 - 00:21:52:11
to show
00:21:52:11 - 00:21:54:08
technically that you're going to meet
00:21:54:08 - 00:21:56:16
a few of those key requirements,
00:21:56:16 - 00:21:59:08
then, you have that data and you're ready
00:21:59:08 - 00:22:00:23
to talk to investors
00:22:00:23 - 00:22:03:06
and say there's a market need.
00:22:03:06 - 00:22:06:05
You have some technical feasibility here, you're ready
00:22:06:05 - 00:22:09:13
to form for a real company and actually proceed
00:22:09:13 - 00:22:10:15
with product development,
00:22:10:15 - 00:22:13:22
which leads me deep into the chasm now
00:22:15:04 - 00:22:16:17
now that we're we're in it.
00:22:16:17 - 00:22:18:10
You've been doing it for a few years.
00:22:18:10 - 00:22:20:19
And if I may
00:22:20:19 - 00:22:23:07
ask the question again, so our audience is
00:22:23:07 - 00:22:24:09
is developing these
00:22:24:09 - 00:22:26:12
these medical devices just like you are.
00:22:26:12 - 00:22:29:06
You know what data has maybe to phrase it,
00:22:29:06 - 00:22:29:23
you know what
00:22:29:23 - 00:22:30:14
data has been
00:22:30:14 - 00:22:34:14
the hardest to capture or most critical
00:22:34:14 - 00:22:35:19
for you
00:22:35:19 - 00:22:38:15
to get through the woods in in product development?
00:22:38:19 - 00:22:41:14
Well, I'll give you another leather example.[4][5]
00:22:41:14 - 00:22:42:19
Just a quick story.
00:22:42:19 - 00:22:45:02
So we took the benchtop version
00:22:45:02 - 00:22:46:20
that Johns Hopkins had developed
00:22:46:20 - 00:22:49:21
and we tried to turn it into kind of like the first
00:22:49:21 - 00:22:51:10
prettier looking prototype.
00:22:51:10 - 00:22:53:16
And I think you actually saw that.
00:22:53:16 - 00:22:56:18
And so the the main guts of this
00:22:56:18 - 00:22:59:19
technology is a Delta parallel robot.
00:23:00:09 - 00:23:01:03
And
00:23:02:21 - 00:23:06:09
the original version is actually relatively small
00:23:06:09 - 00:23:08:14
and so to get it into the surgical field,
00:23:08:14 - 00:23:12:02
another student at the school is as one of their own.
00:23:12:02 - 00:23:13:20
I think it was a masters project.
00:23:13:20 - 00:23:15:17
They'd actually built this mobile base
00:23:15:17 - 00:23:17:07
that would actually hold
00:23:17:07 - 00:23:20:02
the Delta robot and move it into position.
00:23:20:02 - 00:23:21:05
So we took what we knew,
00:23:21:05 - 00:23:22:18
we took what had already been done,
00:23:22:18 - 00:23:23:22
and we just made it look,
00:23:23:22 - 00:23:26:14
we made it a little bit bigger
00:23:26:14 - 00:23:29:03
and we made it look a little prettier.
00:23:29:03 - 00:23:31:14
And we took it to this conference and
00:23:31:14 - 00:23:34:13
ergonomically it was completely unusable.
00:23:34:13 - 00:23:36:09
And the reason for that is
00:23:36:09 - 00:23:38:10
and we'd completely forgotten about this
00:23:39:13 - 00:23:40:16
otolaryngologists
00:23:40:16 - 00:23:44:07
so rhinologists, laryngologist, otologists
00:23:44:07 - 00:23:46:10
they're using surgical microscopes
00:23:46:10 - 00:23:49:10
that are like right here and the robot is supposed to
00:23:49:10 - 00:23:52:05
then also come into the basically the same area
00:23:52:22 - 00:23:55:12
and it's supposed to ergonomically fit
00:23:55:12 - 00:23:56:13
into the environment
00:23:56:13 - 00:23:59:15
that you already have well, that Delta,
00:23:59:15 - 00:24:02:12
which we ended up calling the flying trashcan,
00:24:02:12 - 00:24:04:12
was banging into the microscope.
00:24:04:12 - 00:24:06:08
We completely forgotten
00:24:06:08 - 00:24:10:11
about this critical environment that was in the O.R.
00:24:10:11 - 00:24:10:20
We're like,
00:24:10:20 - 00:24:12:11
well, is it taking up too much floor space?
00:24:12:11 - 00:24:14:10
I mean, we had all these KPIs
00:24:14:10 - 00:24:16:03
that were important,
00:24:16:03 - 00:24:19:00
but we completely forgotten about the air space
00:24:19:00 - 00:24:22:15
being occupied by the meat and potatoes of the robot.
00:24:22:15 - 00:24:25:11
And so this this really goes back to.
00:24:25:11 - 00:24:27:08
Yeah, I mean, we
00:24:27:08 - 00:24:28:23
I'm not sure that that really
00:24:28:23 - 00:24:31:19
was going to end up being the final product anyway,
00:24:32:20 - 00:24:34:04
but it was
00:24:34:04 - 00:24:37:01
a bit of a, oh, gosh, what are we going to do now?
00:24:37:23 - 00:24:39:19
And you need those kind of failures.
00:24:39:19 - 00:24:41:22
You know, honestly,
00:24:41:22 - 00:24:45:04
I get nervous talking to potential users
00:24:45:04 - 00:24:46:18
and potential customers and things like that
00:24:46:18 - 00:24:48:22
and not having at least somebody
00:24:48:22 - 00:24:50:04
tell me that I'm an idiot
00:24:50:04 - 00:24:53:10
and this is not going to work and you're a moron.
00:24:53:10 - 00:24:54:15
And this is why, right?
00:24:54:15 - 00:24:56:17
If I can't find somebody like that, I get stuck.
00:24:56:17 - 00:24:58:08
I start getting fidgety because I'm like,
00:24:58:08 - 00:25:00:11
I don't like it when people just like
00:25:00:11 - 00:25:02:02
they're either being polite
00:25:02:02 - 00:25:04:00
and just, like, going, Oh, yeah, that'll be beautiful.
00:25:04:00 - 00:25:06:17
I'd love to see your robot.
00:25:07:16 - 00:25:10:12
You got to find somebody that goes, you screwed
00:25:10:12 - 00:25:12:06
this up dude, and here's why.
00:25:12:06 - 00:25:13:18
And so,
00:25:13:18 - 00:25:16:10
even though it was a little discouraging
00:25:16:10 - 00:25:17:15
we pivoted really fast
00:25:17:15 - 00:25:19:10
and we came up with a newer design
00:25:19:10 - 00:25:20:13
that actually
00:25:20:13 - 00:25:23:12
meets more needs than the original intention anyway.
00:25:23:12 - 00:25:25:13
So it was a great discovery
00:25:26:13 - 00:25:28:19
but you need to have some failures.
00:25:28:19 - 00:25:29:08
You got it
00:25:29:08 - 00:25:32:17
if you're not getting some kind of negative feedback.
00:25:32:17 - 00:25:33:14
And I mean,
00:25:33:14 - 00:25:36:02
and I mean somebody actually like getting in your face
00:25:36:02 - 00:25:37:14
and giving you four letter words, telling you
00:25:37:14 - 00:25:39:18
how badly you screwed up with your prototype.
00:25:39:18 - 00:25:41:13
If you can't find at least one person
00:25:41:13 - 00:25:42:18
that does that, you're
00:25:42:18 - 00:25:44:17
probably not even trying hard enough
00:25:44:17 - 00:25:46:22
with what you're developing in the first place.
00:25:47:08 - 00:25:49:04
And you just keep...
00:25:49:04 - 00:25:53:14
Yeah, I know. So can you give our audience an idea
00:25:53:14 - 00:25:56:17
for like how frequently you would poll,
00:25:56:17 - 00:26:00:01
market leaders for this kind of feedback.
00:26:00:01 - 00:26:01:00
It's just like quarterly.
00:26:01:00 - 00:26:02:01
You would kind of try
00:26:02:01 - 00:26:03:11
to get in front of stakeholders
00:26:03:11 - 00:26:06:14
and go out beyond Hopkins, at what clip
00:26:06:14 - 00:26:07:20
were you getting this content?
00:26:07:20 - 00:26:09:23
Constantly.
00:26:09:23 - 00:26:12:21
Bruce and I are doing at least
00:26:12:21 - 00:26:16:04
one really in-depth demo with a KOL
00:26:17:07 - 00:26:19:16
each month, that's a minimum.
00:26:19:16 - 00:26:21:23
In the beginning, it was probably
00:26:21:23 - 00:26:25:00
at least once a week, and we were and in some cases
00:26:25:00 - 00:26:25:21
we're going to well,
00:26:25:21 - 00:26:28:11
the first year Bruce and I were on a plane
00:26:28:11 - 00:26:32:12
all the time, so we were in every major hospital
00:26:32:12 - 00:26:33:11
showing them, look,
00:26:33:11 - 00:26:35:23
this is the technology we're developing
00:26:35:23 - 00:26:37:14
and we're trying to figure it out.
00:26:37:14 - 00:26:38:23
You know, are we going in the right place?
00:26:38:23 - 00:26:39:21
And, you know,
00:26:39:21 - 00:26:43:19
how would this solve your problems?
00:26:43:19 - 00:26:46:06
We were getting in front of as many people as possible
00:26:46:06 - 00:26:49:04
and just, showing them technology. Right.
00:26:49:04 - 00:26:50:15
Because we didn't
00:26:50:15 - 00:26:51:07
at that point,
00:26:51:07 - 00:26:53:05
I didn't even have a rollout strategy for a product.
00:26:53:05 - 00:26:56:05
It was really still raw technology,
00:26:57:00 - 00:26:59:04
but we were validating it constantly.
00:26:59:04 - 00:27:00:19
Just making sure that we're going in
00:27:00:19 - 00:27:01:16
the right direction
00:27:01:16 - 00:27:03:18
was is this going to solve your problem?
00:27:04:01 - 00:27:06:08
So we hear you loud and clear. You want
00:27:07:13 - 00:27:08:09
rapid and
00:27:08:09 - 00:27:09:17
constant customer feedback
00:27:09:17 - 00:27:10:16
and you want people to tell you
00:27:10:16 - 00:27:12:03
you're doing the wrong things.
00:27:12:03 - 00:27:15:06
If you have to build a prototype with papier mach�,
00:27:15:06 - 00:27:16:12
do it.
00:27:16:12 - 00:27:18:11
You know, don't don't wait for,
00:27:18:11 - 00:27:20:01
you know, all the fancy actuators
00:27:20:01 - 00:27:20:20
and you know,
00:27:20:20 - 00:27:22:14
for the motors to come in from wherever
00:27:22:14 - 00:27:24:21
you're sourcing them from and that kind of stuff
00:27:24:21 - 00:27:26:07
do whatever it takes.
00:27:26:07 - 00:27:26:19
In fact, there
00:27:26:19 - 00:27:30:02
literally were paper mach� mockups done
00:27:30:02 - 00:27:33:12
before the original Mark Zero was produced at
00:27:33:12 - 00:27:34:05
Johns Hopkins.
00:27:34:05 - 00:27:36:04
So I'm not even kidding
00:27:36:04 - 00:27:39:02
when I say papier mach�, whatever it takes
00:27:39:18 - 00:27:41:23
get a mock up made and get feedback.
00:27:41:23 - 00:27:44:05
So you're on your way to market.
00:27:44:05 - 00:27:47:13
And if I were, if I recall, that's de novo, right?
00:27:47:13 - 00:27:49:19
So and then talk a little bit about
00:27:51:06 - 00:27:52:02
using on the
00:27:52:02 - 00:27:55:02
market tools with your consumable.
00:27:55:02 - 00:27:56:14
What does that look like?
00:27:56:14 - 00:27:58:14
You know, that was a big deal.
00:27:58:14 - 00:27:59:20
We've got enough challenges
00:27:59:20 - 00:28:01:02
with the technology that we're building.
00:28:01:02 - 00:28:03:13
We didn't want to become instrument manufacturers.
00:28:03:13 - 00:28:05:10
You know, there's Teleflex and Integra.
00:28:05:10 - 00:28:06:15
There's plenty of folks out there
00:28:06:15 - 00:28:08:09
that make perfectly fine.
00:28:08:09 - 00:28:09:20
You know, Johann Grasper
00:28:09:20 - 00:28:12:05
and scissors and forceps and all this kind of stuff.
00:28:13:14 - 00:28:14:23
Who wants to go into that business?
00:28:14:23 - 00:28:15:12
Right.
00:28:16:10 - 00:28:19:02
And so what was really important
00:28:19:02 - 00:28:21:00
was that we could meet the surgeons
00:28:21:00 - 00:28:22:14
where they are with the instruments
00:28:22:14 - 00:28:23:22
they already know how to use.
00:28:23:22 - 00:28:26:16
And so that was to me, that's a huge
00:28:26:16 - 00:28:28:01
for at least for us,
00:28:28:01 - 00:28:30:08
it ends up being a huge piece of the puzzle
00:28:30:08 - 00:28:31:22
in terms of customer adoption
00:28:31:22 - 00:28:33:00
and customer acceptance,
00:28:33:00 - 00:28:34:15
because we're not asking them
00:28:34:15 - 00:28:35:22
to do a procedure differently.
00:28:35:22 - 00:28:37:11
We're not even asking them necessarily
00:28:37:11 - 00:28:40:00
to use a different instrument
00:28:40:00 - 00:28:42:16
because of FDA regulatory.
00:28:42:16 - 00:28:45:13
You know, we have to specify exactly
00:28:45:13 - 00:28:47:01
what instrument we're holding
00:28:47:01 - 00:28:48:14
like down to the part number
00:28:48:14 - 00:28:49:18
so right now,
00:28:49:18 - 00:28:51:06
I'm only going to be supporting one
00:28:51:06 - 00:28:52:15
brand of instrument,
00:28:52:15 - 00:28:54:23
but over time, we could be supporting,
00:28:54:23 - 00:28:57:01
you know, any arbitrary number of instruments.
00:28:57:01 - 00:28:59:14
And it's just it's a surreal process over time.
00:29:00:17 - 00:29:02:14
But that's huge for us
00:29:02:14 - 00:29:05:15
because it means that we don't have to spend a lot of
00:29:05:15 - 00:29:07:18
we don't have to burn a lot of calories
00:29:07:18 - 00:29:10:20
trying to figure out how to reinvent the instrument
00:29:10:20 - 00:29:12:00
that's already out there.
00:29:12:00 - 00:29:13:03
They're class one instruments.
00:29:13:03 - 00:29:15:14
They're already cleared. It's great.
00:29:15:14 - 00:29:16:21
And so by supporting them
00:29:16:21 - 00:29:19:18
and by giving them more capability,
00:29:20:12 - 00:29:22:07
you know, my favorite analogy is always,
00:29:22:07 - 00:29:23:21
chocolate stands on its own.
00:29:23:21 - 00:29:25:00
Peanut butter stands on its own.
00:29:25:00 - 00:29:27:00
You put them together and it's my favorite candy
00:29:27:00 - 00:29:29:19
and to me, this technology
00:29:29:19 - 00:29:32:06
is that kind of peanut butter cup solution
00:29:32:21 - 00:29:36:02
because we're taking two standalone technologies
00:29:36:02 - 00:29:37:03
and we're bringing them together
00:29:37:03 - 00:29:39:08
and creating a new solution
00:29:39:08 - 00:29:41:01
that does not exist on its own.
00:29:41:01 - 00:29:42:12
And I think that's great
00:29:42:12 - 00:29:44:01
if you can do something like that.
00:29:44:01 - 00:29:45:11
I think it's a huge win.
00:29:45:11 - 00:29:46:01
Let's talk a little bit
00:29:46:01 - 00:29:47:22
more about testing with the users.
00:29:47:22 - 00:29:51:15
so there's this hybrid between market research and,
00:29:51:15 - 00:29:53:22
you know, papier mach� and,
00:29:55:03 - 00:29:57:12
kind of like more formative evaluations.
00:29:57:12 - 00:30:01:08
And so I noticed that your your product has a GUI
00:30:01:08 - 00:30:03:03
a separate screen.
00:30:03:03 - 00:30:03:22
Can you talk a little bit
00:30:03:22 - 00:30:06:19
about how the surgeon interacts with the screen,
00:30:06:19 - 00:30:09:10
how you determined kind of what goes on the screen?
00:30:09:22 - 00:30:12:12
And I assume that it's different for each procedure.
00:30:12:23 - 00:30:16:10
Yeah, we obviously, the name of the game
00:30:16:10 - 00:30:18:10
is to keep your user interface
00:30:18:10 - 00:30:20:19
not looking like DOS or Windows 3.1.
00:30:20:19 - 00:30:22:06
So you don't just like cram
00:30:22:06 - 00:30:25:02
every icon under the screen that you possibly can
00:30:25:02 - 00:30:27:15
so it needed to be very, very clean.
00:30:27:15 - 00:30:30:01
It's a touch screen interface
00:30:30:01 - 00:30:33:15
and it's as adaptive as it possibly can be.
00:30:33:15 - 00:30:35:21
So if you're doing a laryngeal procedure,
00:30:35:21 - 00:30:37:06
you're only going to be presented
00:30:37:06 - 00:30:38:23
with laryngeal instruments.
00:30:38:23 - 00:30:40:11
And then eventually as we add
00:30:40:11 - 00:30:44:08
other surgical disciplines by adding other instruments
00:30:44:23 - 00:30:47:14
that are used in those other disciplines,
00:30:48:01 - 00:30:51:14
you would log in and say, I'm doing neurosurgery
00:30:51:14 - 00:30:52:00
and here's
00:30:52:00 - 00:30:54:07
my selection of supported neurosurgery instruments.
00:30:55:02 - 00:30:58:19
And so that keeps the clutter down significantly
00:30:59:20 - 00:31:03:04
when you're actually doing a procedure or,
00:31:03:04 - 00:31:05:21
testing the robot or doing a trial or whatever
00:31:07:19 - 00:31:10:14
we try again, we want to keep the information
00:31:10:14 - 00:31:12:00
that's on the screen
00:31:12:00 - 00:31:15:18
usable and actionable, but not like just,
00:31:15:18 - 00:31:18:17
you know, a ton of telemetry and numbers,
00:31:18:17 - 00:31:21:06
you know, zipping along like The Matrix,
00:31:21:06 - 00:31:24:09
but have things be very, very simple and
00:31:24:10 - 00:31:26:04
and easy to see.
00:31:26:04 - 00:31:28:22
So when you're moving the robot around we�ll show you
00:31:28:22 - 00:31:31:21
where the instrument is within the surgical workflow.
00:31:32:06 - 00:31:33:20
Yeah, for the most part,
00:31:33:20 - 00:31:37:01
you probably don't even need that information.
00:31:37:01 - 00:31:38:17
But sometimes if the
00:31:38:17 - 00:31:40:12
if the robot is is draped,
00:31:40:12 - 00:31:41:06
you might not know
00:31:41:06 - 00:31:43:23
that you're at the edge of your work space,
00:31:43:23 - 00:31:44:16
that sort of thing.
00:31:44:16 - 00:31:46:23
So it is kind of it can occasionally be useful
00:31:46:23 - 00:31:49:07
to look over the screen and go, oh, I'm over,
00:31:49:14 - 00:31:51:15
but maybe I need to shift my robot six
00:31:51:15 - 00:31:53:01
inches to the left or something like that
00:31:53:01 - 00:31:55:14
because I'm just I'm in a weird space ergonomically.
00:31:55:22 - 00:31:59:03
So we've got that kind of feedback, but
00:32:00:05 - 00:32:01:10
we just want to keep it as
00:32:01:10 - 00:32:04:01
clean as possible
00:32:05:10 - 00:32:06:15
and
00:32:07:12 - 00:32:10:11
you know, just avoid it.
00:32:12:00 - 00:32:13:11
How do I usually put this?
00:32:13:11 - 00:32:16:02
I think one of the frustrations from a lot of surgeons
00:32:16:02 - 00:32:17:20
is that a lot of surgical robots
00:32:17:20 - 00:32:20:03
kind of become the centerpiece of the O.R.
00:32:20:16 - 00:32:23:02
instead of the surgeon
00:32:24:01 - 00:32:25:10
and so we want to be
00:32:25:10 - 00:32:26:11
a piece of equipment
00:32:26:11 - 00:32:30:02
that's there to serve the surgeon, not turn it into
00:32:30:02 - 00:32:32:03
the surgeon is there to serve the robot,
00:32:32:03 - 00:32:33:05
you know what I mean?
00:32:33:14 - 00:32:36:23
And so it's all about what can we provide
00:32:36:23 - 00:32:41:08
that is easy to digest and is not distracting to it.
00:32:41:08 - 00:32:44:09
To go along the terminology of being simple, like
00:32:44:09 - 00:32:47:12
the surgeon can always override the robot.
00:32:47:12 - 00:32:49:15
Is that fair? OK, absolutely. Yeah.
00:32:50:05 - 00:32:50:14
Yeah.
00:32:50:14 - 00:32:52:22
Like there's gain controls and things like that.
00:32:53:10 - 00:32:56:03
So, you know, you can slow down
00:32:56:03 - 00:32:57:10
the robot's response
00:32:57:10 - 00:32:59:20
if you're in a really, really tight situation.
00:32:59:20 - 00:33:01:08
You're just like, I just don't
00:33:01:08 - 00:33:03:08
I don't want the instrument to move
00:33:03:10 - 00:33:06:11
as fast as it could potentially move in,
00:33:06:11 - 00:33:08:06
and in free space,
00:33:08:06 - 00:33:10:18
I want the robot to slow me down sometimes.
00:33:10:18 - 00:33:11:09
That's actually
00:33:11:09 - 00:33:14:11
what the surgeon wants and so are there's slider bars
00:33:14:11 - 00:33:15:08
that they can jump to that
00:33:15:08 - 00:33:16:23
just say, you know, drop the
00:33:16:23 - 00:33:18:17
top level of the gain down.
00:33:19:20 - 00:33:22:01
And it's, you know,
00:33:22:01 - 00:33:23:10
that's something that they pick up
00:33:23:10 - 00:33:26:13
during the training process, which is a very short
00:33:27:06 - 00:33:27:18
process.
00:33:27:18 - 00:33:29:18
In fact, we've got another
00:33:29:18 - 00:33:32:18
big usability study coming up in a few weeks.
00:33:33:07 - 00:33:34:11
And
00:33:35:09 - 00:33:37:11
this round of surgeons, like the last time
00:33:37:11 - 00:33:38:13
they get about 2 hours
00:33:38:13 - 00:33:40:13
worth of training for the user interface.
00:33:40:13 - 00:33:44:09
And then they're there in a cadaver doing work.
00:33:44:09 - 00:33:46:06
And they seem to be picking it up
00:33:46:06 - 00:33:48:18
really, really quickly.
00:33:48:18 - 00:33:50:23
But yeah, it's important to
00:33:52:00 - 00:33:53:23
especially with something like a surgical robot,
00:33:53:23 - 00:33:56:10
you need to make sure that there's enough
00:33:56:10 - 00:33:58:08
you want to keep things simple.
00:33:58:08 - 00:33:59:08
But at the same time,
00:33:59:08 - 00:34:01:09
you want to make sure that the power users
00:34:01:09 - 00:34:03:03
or somebody who's had their hands on the robot
00:34:03:03 - 00:34:04:19
for a few weeks and they start to go,
00:34:04:19 - 00:34:06:20
you know, I like the default,
00:34:06:20 - 00:34:07:22
but I just want to have it
00:34:07:22 - 00:34:10:04
feel a little bit, whatever in my hand.
00:34:10:04 - 00:34:10:19
I want it
00:34:10:19 - 00:34:12:09
I want to give me a little bit more resistance
00:34:12:09 - 00:34:14:02
or I want it to be a little bit looser.
00:34:14:02 - 00:34:17:23
Let them make those changes as long as you can do so.
00:34:19:01 - 00:34:21:04
Obviously in a safe and efficacious manner,
00:34:22:06 - 00:34:24:12
depending on however your hardware works.
00:34:24:12 - 00:34:26:11
But the settings are important,
00:34:26:11 - 00:34:30:08
but you can't just like blast them in the faces of
00:34:30:08 - 00:34:32:04
of the user either. Got it.
00:34:32:04 - 00:34:32:09
All right.
00:34:32:09 - 00:34:35:18
So we've covered who Galen Robotics is.
00:34:35:18 - 00:34:38:06
We've covered what data is critical.
00:34:38:06 - 00:34:39:05
And it's clear,
00:34:39:05 - 00:34:40:11
you know, you need to market need
00:34:40:11 - 00:34:42:09
you need to test within users.
00:34:42:09 - 00:34:44:10
And to the extent that
00:34:44:10 - 00:34:46:16
you have technical demonstration already
00:34:46:16 - 00:34:47:07
out of the box
00:34:47:07 - 00:34:49:06
with the technology or licensing that,
00:34:49:06 - 00:34:51:10
that certainly helps give you the confidence
00:34:51:10 - 00:34:53:21
and you transition into formal product development.
00:34:54:19 - 00:34:56:20
And so the next section here is
00:34:57:10 - 00:34:57:21
just kind of
00:34:57:21 - 00:35:00:06
classic product development questions for you.
00:35:01:05 - 00:35:02:15
So the first question,
00:35:02:15 - 00:35:04:04
a lot of these startup companies
00:35:04:04 - 00:35:05:19
that we talk with are,
00:35:05:19 - 00:35:08:20
in the exploratory phase, like a phase zero
00:35:08:20 - 00:35:10:19
and at what point did
00:35:10:19 - 00:35:13:20
you know that Galen was ready to go from phase zero
00:35:14:08 - 00:35:16:14
to formal product development?
00:35:17:01 - 00:35:19:02
I mean, because you're building
00:35:19:02 - 00:35:20:19
this airplane as you're flying it.
00:35:20:19 - 00:35:24:04
When did you lock, product requirements?
00:35:24:04 - 00:35:26:03
What were you looking for?
00:35:26:13 - 00:35:29:05
You know, when we came back from that that big
00:35:29:05 - 00:35:30:16
otolaryngology conference,
00:35:31:19 - 00:35:33:07
we clearly knew that that we
00:35:33:07 - 00:35:35:08
had some work to do on the requirements,
00:35:35:08 - 00:35:37:02
although I guess maybe the requirements
00:35:37:02 - 00:35:39:04
themselves were relatively stable,
00:35:39:04 - 00:35:41:21
but it was the engineering output response
00:35:41:21 - 00:35:42:23
to those requirements
00:35:42:23 - 00:35:46:08
that really needed to be changed.
00:35:46:08 - 00:35:47:23
But but we knew
00:35:47:23 - 00:35:50:04
we had some we still had some work to do.
00:35:51:02 - 00:35:53:12
I think when we had our Mark 2 prototype
00:35:53:12 - 00:35:56:16
and we started putting it in front of people's
00:35:56:16 - 00:35:58:18
you know, putting in front of people again, and doing,
00:35:58:18 - 00:36:01:04
you know, phantoms and just different studies.
00:36:02:21 - 00:36:05:05
That was when the feedback was coming back
00:36:05:05 - 00:36:06:12
and it was like, oh,
00:36:06:12 - 00:36:08:00
I love how this does this
00:36:08:00 - 00:36:10:04
and if you could support this or
00:36:10:04 - 00:36:12:08
if the four sensor was a little bit more durable,
00:36:12:08 - 00:36:14:00
I could put this bigger instrument.
00:36:14:00 - 00:36:16:18
Suddenly the kind of feedback was come
00:36:16:18 - 00:36:18:09
to us in the context of
00:36:19:14 - 00:36:23:05
I believe in what you've done as a base model here.
00:36:23:05 - 00:36:23:13
And so
00:36:23:13 - 00:36:26:06
I'm now able to start thinking a little bit beyond
00:36:26:06 - 00:36:28:05
what you've what you've given me
00:36:28:05 - 00:36:29:20
and that's when we
00:36:29:20 - 00:36:33:03
I would say anecdotally knew that we were we were
00:36:33:03 - 00:36:36:22
we were ready to like start moving more aggressively
00:36:37:16 - 00:36:39:23
into commercialization.
00:36:39:23 - 00:36:41:12
Now, how do you quantify that?
00:36:41:12 - 00:36:44:01
I mean, that's that's pretty tricky,
00:36:44:01 - 00:36:46:18
that comes down to the whole
00:36:46:18 - 00:36:48:03
talking to customers and
00:36:48:03 - 00:36:49:15
and making sure that those requirements
00:36:49:15 - 00:36:51:12
are really reflections of
00:36:51:12 - 00:36:53:20
what is it the user wants to do.
00:36:53:20 - 00:36:55:08
And that's important.
00:36:55:08 - 00:36:59:02
You've got to and it takes you know, we've done those
00:37:00:04 - 00:37:03:04
we do those we've actually done requirement
00:37:03:04 - 00:37:06:19
reviews company wide where we've had
00:37:07:09 - 00:37:09:00
almost every single person in the company
00:37:09:00 - 00:37:10:12
sitting down as we would walk
00:37:10:12 - 00:37:12:15
through one requirement after another
00:37:12:15 - 00:37:13:04
and actually
00:37:13:04 - 00:37:16:15
even like sometimes hash out semantics going
00:37:16:15 - 00:37:17:16
yeah but is this
00:37:17:16 - 00:37:20:07
is this requirement really describing what it is that
00:37:20:07 - 00:37:21:07
we're trying to achieve
00:37:22:14 - 00:37:25:23
and you know, it takes
00:37:25:23 - 00:37:27:07
more than one person in the room
00:37:27:07 - 00:37:28:04
to do that sort of thing.
00:37:28:04 - 00:37:31:22
So we made sure that that,
00:37:31:22 - 00:37:35:23
everybody had at least visibility into that process.
00:37:35:23 - 00:37:37:15
So it was tough.
00:37:37:15 - 00:37:38:18
It was tough.
00:37:38:18 - 00:37:39:19
It sounds like there was a little bit
00:37:39:19 - 00:37:41:18
of like a gut feeling that,
00:37:41:18 - 00:37:43:18
you're on to something and,
00:37:43:18 - 00:37:44:05
you read it,
00:37:44:05 - 00:37:46:00
you're ready to lock requirements
00:37:46:00 - 00:37:49:20
and you've got this combination of agreement
00:37:49:20 - 00:37:51:04
from the market.
00:37:51:04 - 00:37:53:13
Plus now you have a team
00:37:53:13 - 00:37:54:07
and you sort of
00:37:54:07 - 00:37:55:14
it sounds like you crowd source
00:37:55:14 - 00:37:57:20
reviewed your requirements. I think that's great.
00:37:58:07 - 00:38:00:09
You know, rather than just having the engineers,
00:38:00:16 - 00:38:02:00
put requirements together
00:38:02:00 - 00:38:03:19
and then just proceed with product
00:38:03:19 - 00:38:05:13
with design control development,
00:38:05:13 - 00:38:08:05
you get to this infinite loop of making design changes
00:38:08:05 - 00:38:09:10
you have to document
00:38:09:10 - 00:38:10:21
and then you find,
00:38:10:21 - 00:38:13:04
user needs come out of left field and then your
00:38:13:04 - 00:38:14:16
you have to go back to phase zero.
00:38:14:16 - 00:38:17:09
So I like how you at least focused on the end user
00:38:17:09 - 00:38:19:22
and then crowd sourced, your requirements.
00:38:19:22 - 00:38:21:19
So we're pretty technical company here.
00:38:21:19 - 00:38:24:00
I'm going to ask a couple software
00:38:24:00 - 00:38:25:16
related questions here.
00:38:25:16 - 00:38:27:22
Dave, I remember on one of the other podcasts
00:38:28:09 - 00:38:31:16
you said that surgical robots are close to 80%
00:38:31:16 - 00:38:32:21
software anyways.
00:38:32:21 - 00:38:36:20
So can you talk a little bit about the software
00:38:37:04 - 00:38:40:01
and what you licensed from Hopkins
00:38:40:01 - 00:38:42:18
and clearly you know this is a public forum here.
00:38:42:18 - 00:38:44:08
We don't need to get into the details of the control
00:38:44:08 - 00:38:45:11
algorithms and things like that.
00:38:45:11 - 00:38:46:12
But you know other
00:38:46:12 - 00:38:48:03
startup companies are inheriting
00:38:48:03 - 00:38:50:13
SOUP �software of unknown pedigree�. Right.
00:38:50:21 - 00:38:53:08
So how has this software evolved
00:38:53:16 - 00:38:55:23
from when you first started to where we are today?
00:38:56:10 - 00:38:59:18
As as nicely as I can say this, the software
00:38:59:18 - 00:39:01:02
that was handed over to us from
00:39:01:02 - 00:39:03:07
the university was for demo purposes only.
00:39:04:16 - 00:39:06:15
There's no way that it was
00:39:06:15 - 00:39:09:02
going to just roll straight into commercialization.
00:39:10:07 - 00:39:11:18
It definitely demonstrated to us
00:39:11:18 - 00:39:13:21
that things like the algorithms were understood
00:39:13:21 - 00:39:15:04
you know, there's math
00:39:15:04 - 00:39:16:01
with kinematics
00:39:16:01 - 00:39:18:10
and things like that that can be really tricky.
00:39:18:10 - 00:39:21:09
And so the software as it exists
00:39:21:09 - 00:39:23:04
certainly provided proof
00:39:23:04 - 00:39:26:21
that those algorithms had been sufficiently vetted.
00:39:26:21 - 00:39:29:12
And so there were guts in there that were good.
00:39:30:15 - 00:39:32:19
But, you know,
00:39:32:19 - 00:39:36:04
grad students and all of that working
00:39:36:04 - 00:39:39:09
on different class projects and things
00:39:39:09 - 00:39:40:02
like that.
00:39:40:02 - 00:39:41:23
They're not having these big architecture
00:39:41:23 - 00:39:44:18
meetings deciding how to like modularly produce
00:39:45:18 - 00:39:49:00
software for for a gizmo
00:39:49:06 - 00:39:51:21
that�s being developed inside of a university.
00:39:52:06 - 00:39:55:22
So it was it was spaghetti
00:39:56:09 - 00:40:00:14
soup, spaghetti, and it was it was rough.
00:40:00:14 - 00:40:01:19
And so
00:40:01:19 - 00:40:03:04
we had the challenge
00:40:03:04 - 00:40:05:09
of not only moving development forward,
00:40:05:09 - 00:40:08:03
but also re-architecting it live.
00:40:08:16 - 00:40:11:09
And so our Director of Software Engineering,
00:40:11:17 - 00:40:14:05
I think, did a just an incredible job.
00:40:14:23 - 00:40:18:15
In fact, we've been able to add in some additional,
00:40:18:15 - 00:40:21:00
future features for demo purposes
00:40:21:12 - 00:40:25:01
that he was able to drop in within days
00:40:25:11 - 00:40:29:10
because the code has become so much more robust
00:40:29:10 - 00:40:30:20
and modularized.
00:40:31:19 - 00:40:32:19
And that was critical.
00:40:32:19 - 00:40:37:02
But you know, my warning to anybody
00:40:37:11 - 00:40:39:07
that is going to license code
00:40:39:07 - 00:40:41:06
is the code is going to show you that
00:40:41:06 - 00:40:43:09
technologically it's feasible
00:40:43:09 - 00:40:45:22
but do not expect that you're just going to like slap
00:40:45:22 - 00:40:48:17
some new icons on the code and then take it to market.
00:40:48:17 - 00:40:51:08
It is not ready for that.
00:40:51:08 - 00:40:55:09
And as good as you might hope it is
00:40:55:09 - 00:40:57:09
and you might interview some post-grad,
00:40:57:09 - 00:40:58:18
it seems like just the savant
00:40:58:18 - 00:41:00:13
for coding or something like that.
00:41:00:13 - 00:41:03:05
You just I think it's a bad idea
00:41:03:05 - 00:41:06:12
to even assume that, the code looks great,
00:41:06:12 - 00:41:08:03
Well, no, it doesn't.
00:41:08:03 - 00:41:10:19
You need to formally go through it
00:41:11:05 - 00:41:15:22
on your side of the fence and you need to treat it
00:41:15:22 - 00:41:18:10
as a prototype and you need to,
00:41:18:10 - 00:41:20:05
you need to work it out and make it your own
00:41:22:18 - 00:41:23:16
see what you got there.
00:41:23:16 - 00:41:23:21
Yeah.
00:41:23:21 - 00:41:26:15
So it's a six degrees of freedom robot
00:41:27:04 - 00:41:29:07
and looking at the patents and whatnot
00:41:29:07 - 00:41:33:19
it�s a series of transducers and sensors
00:41:33:19 - 00:41:34:15
in an array
00:41:34:15 - 00:41:36:07
and you collect all this data
00:41:36:07 - 00:41:38:12
and it helps guide the user.
00:41:38:12 - 00:41:41:14
But can you talk maybe,
00:41:41:14 - 00:41:43:11
philosophically on the control
00:41:43:11 - 00:41:44:02
algorithm, like,
00:41:44:02 - 00:41:47:13
for example, how does the robot differentiate between,
00:41:48:00 - 00:41:50:15
a hand tremor that's not intentional
00:41:50:15 - 00:41:54:15
and an intended movement from the from the surgeon?
00:41:55:05 - 00:41:56:18
Or does it just filter out,
00:41:56:18 - 00:41:58:04
you know, micro hand movement?
00:41:58:04 - 00:42:00:14
And so it turns out from the literature
00:42:01:15 - 00:42:03:23
that naturally occurring hand tremor
00:42:03:23 - 00:42:07:07
actually has a very specific cycle to it.
00:42:07:07 - 00:42:09:09
It's actually around 50 to 60 hertz
00:42:11:08 - 00:42:14:18
weirdly, as that number is as a coincidence.
00:42:14:18 - 00:42:19:04
But so certainly the first approach is really just,
00:42:19:04 - 00:42:22:19
filter out that that that that jitter
00:42:22:19 - 00:42:24:01
because you just don't need it. Right.
00:42:25:09 - 00:42:25:22
I'm going to
00:42:25:22 - 00:42:26:12
move from point
00:42:26:12 - 00:42:28:19
A to point B, I don't need all this fuzz.
00:42:29:17 - 00:42:30:20
So that turned out to actually
00:42:30:20 - 00:42:33:02
be a relatively simple solution.
00:42:33:12 - 00:42:36:15
But then yeah, then it gets into
00:42:37:02 - 00:42:37:14
the control
00:42:37:14 - 00:42:40:14
algorithms, gets way more complex from there.
00:42:40:14 - 00:42:42:21
Because you still are asking,
00:42:43:16 - 00:42:46:06
the kinematics to follow your hand
00:42:46:17 - 00:42:49:02
and you wanted to do it naturally. Right?
00:42:49:02 - 00:42:52:04
And so you don't want to feel the, the ground grind and pop
00:42:52:04 - 00:42:53:18
from each of those individual motors
00:42:53:18 - 00:42:56:08
as it's trying to decide to go into the next position
00:42:56:08 - 00:42:57:15
as it's following,
00:42:57:15 - 00:42:58:06
you know, whatever
00:42:58:06 - 00:43:00:12
vector travel your hand is going through.
00:43:01:13 - 00:43:03:20
So that's where it gets rough.
00:43:03:20 - 00:43:06:22
And in fact, we have a person who's a Ph.D.
00:43:06:22 - 00:43:11:08
in hand controls and that's all he does
00:43:11:08 - 00:43:15:06
is just tuning the feel of,
00:43:15:06 - 00:43:17:23
of moving the instrument along while you're holding it
00:43:18:16 - 00:43:21:13
so that it just it feels as natural.
00:43:21:13 - 00:43:25:11
And I often use the when it's not feeling good,
00:43:25:19 - 00:43:29:00
especially in the early days, I describe it as crunchy
00:43:29:18 - 00:43:31:17
as you were actually trying to move the instrument.
00:43:31:17 - 00:43:32:20
You could almost feel
00:43:32:20 - 00:43:33:21
almost felt like you were like
00:43:33:21 - 00:43:36:07
dragging it across a gravel road.
00:43:36:07 - 00:43:38:19
And that's, we could get into why that is,
00:43:38:19 - 00:43:40:12
but that's you know, just that's it.
00:43:40:12 - 00:43:43:06
That's a response factor of all the individual motors
00:43:43:06 - 00:43:44:03
trying to like
00:43:44:03 - 00:43:46:11
move into the next position kinematically
00:43:47:05 - 00:43:48:15
as you're traveling along.
00:43:48:15 - 00:43:52:09
And so getting that to feel natural was rough.
00:43:52:09 - 00:43:54:11
And it's an ongoing process, honestly.
00:43:54:22 - 00:43:58:13
It will we'll always be tweaking and tuning that
00:43:58:13 - 00:44:00:07
you know, based on different
00:44:01:22 - 00:44:03:13
applications.
00:44:03:13 - 00:44:04:02
The tuning
00:44:04:02 - 00:44:07:08
might change depending on the procedure type.
00:44:07:08 - 00:44:09:17
So the tuning is currently fine
00:44:09:17 - 00:44:13:02
for our laryngeal indication, which is our first
00:44:13:19 - 00:44:15:02
go to market.
00:44:15:02 - 00:44:18:03
It might be different when you're doing neurosurgery
00:44:18:03 - 00:44:19:02
I don't know.
00:44:19:16 - 00:44:21:05
So that's that's something we have to explore.
00:44:21:05 - 00:44:22:04
But that
00:44:22:04 - 00:44:24:07
That's actually where it gets rough.
00:44:24:07 - 00:44:25:18
Canceling out hand tremor
00:44:25:18 - 00:44:26:23
turns out to actually be pretty easy.
00:44:26:23 - 00:44:29:09
So I'm just curious when you get into V & V,
00:44:30:02 - 00:44:32:00
there's the essential performance requirements,
00:44:32:00 - 00:44:34:17
and I would imagine your essential performance is,
00:44:35:00 - 00:44:36:15
you know, reduced
00:44:36:15 - 00:44:40:04
hand tremor aid surgeon positioning by X,
00:44:40:13 - 00:44:43:23
either dimension or percent, whatever.
00:44:44:10 - 00:44:45:05
Are you, I'm curious,
00:44:45:05 - 00:44:49:05
are you developing a robot to test your robot
00:44:49:05 - 00:44:51:02
or are you using
00:44:51:02 - 00:44:53:04
how are you actually verifying that.
00:44:54:02 - 00:44:58:14
A lot of it is is hand feel because if it doesn't feel
00:44:58:14 - 00:45:00:16
good to the surgeon,
00:45:00:16 - 00:45:02:21
it really doesn't matter what your numbers are.
00:45:02:21 - 00:45:04:15
It still has to feel good
00:45:04:15 - 00:45:06:19
because this is a cooperatively controlled robot.
00:45:07:18 - 00:45:09:04
If it was fully autonomous, I
00:45:09:04 - 00:45:10:18
could probably get away with something
00:45:10:18 - 00:45:12:22
a little bit more automated.
00:45:12:22 - 00:45:14:05
But in a lot of ways,
00:45:14:05 - 00:45:15:21
we have to take it through the processes
00:45:15:21 - 00:45:16:18
with a human being.
00:45:16:18 - 00:45:18:12
At the controls stick
00:45:18:12 - 00:45:21:10
because because there is that qualitative factor, it's
00:45:21:10 - 00:45:24:06
got to feel right
00:45:24:06 - 00:45:26:06
or it doesn't matter what the math says.
00:45:26:06 - 00:45:27:16
So that was a challenge for us.
00:45:27:16 - 00:45:28:22
And that doesn't necessarily apply
00:45:28:22 - 00:45:30:07
to all surgical robots.
00:45:30:07 - 00:45:32:20
I'm going to ask a second part of that question
00:45:32:20 - 00:45:35:17
as we enter the final round of our podcast here.
00:45:36:04 - 00:45:38:14
We we use a lightning round now.
00:45:38:14 - 00:45:39:04
I'm going to call it,
00:45:39:04 - 00:45:40:16
questions from Key Techers.
00:45:40:16 - 00:45:42:01
So to get on the market,
00:45:42:01 - 00:45:44:05
medical devices need to be safe and effective.
00:45:44:05 - 00:45:47:01
So you just described a way to measure
00:45:47:13 - 00:45:49:17
efficacy, right? It's got to feel right.
00:45:49:17 - 00:45:55:01
But how do you justify safety when you have this tool
00:45:55:01 - 00:45:59:12
on that's mounted on something that could very easily
00:45:59:12 - 00:46:01:04
move very fast
00:46:01:04 - 00:46:02:18
in a direction you don't want it to go?
00:46:02:18 - 00:46:05:08
So there are philosophically, how are you
00:46:05:08 - 00:46:06:23
justifying the safety here.
00:46:06:23 - 00:46:08:06
That is part of your tuning
00:46:08:06 - 00:46:09:05
is that you've got to
00:46:09:05 - 00:46:11:04
you have to understand what those bounds are
00:46:11:04 - 00:46:16:13
for normal intended motions like a surgeon that is,
00:46:17:04 - 00:46:18:02
got an instrument
00:46:18:02 - 00:46:19:07
buried down into somebody's
00:46:19:07 - 00:46:21:11
larynx isn't going to do this.
00:46:21:11 - 00:46:23:04
It's never going to happen.
00:46:23:04 - 00:46:24:15
And so you've got to find some way
00:46:24:15 - 00:46:25:13
of characterizing
00:46:25:13 - 00:46:26:10
what that unwanted
00:46:26:10 - 00:46:27:17
motion would look like
00:46:27:17 - 00:46:28:09
and make sure
00:46:28:09 - 00:46:30:07
that you are basically restricting
00:46:30:07 - 00:46:35:05
those motions within normal bounds for that procedure.
00:46:35:05 - 00:46:38:07
So you do have to put numbers to things like that.
00:46:38:07 - 00:46:41:00
But yeah, there's another great
00:46:41:00 - 00:46:42:15
example is when you're doing
00:46:42:15 - 00:46:44:14
something like a mastoidectomy,
00:46:44:14 - 00:46:48:16
you�re drilling with 80,000 RPM, glorified dremel tool.
00:46:49:05 - 00:46:50:00
And
00:46:50:23 - 00:46:54:03
the mastoid is not solid bone.
00:46:54:03 - 00:46:56:04
It looks like a it looks like an English muffin.
00:46:56:04 - 00:46:58:11
There's like nooks and crannies all over the place.
00:46:58:11 - 00:46:59:18
And one of the biggest risks
00:46:59:18 - 00:47:02:00
that you have with a handheld drill
00:47:02:00 - 00:47:03:00
is that you open up
00:47:03:00 - 00:47:04:08
one of these nooks and crannies
00:47:04:08 - 00:47:07:06
and your drill can actually, like, dive in.
00:47:07:06 - 00:47:10:03
And inside of that little pocket
00:47:10:03 - 00:47:12:04
could be one of two facial nerves.
00:47:12:04 - 00:47:13:09
It could be a blood vessel.
00:47:13:09 - 00:47:15:18
There's a lot of real dangerous stuff inside
00:47:15:18 - 00:47:19:04
the mastoid just waiting for you to to expose it.
00:47:19:16 - 00:47:22:20
And so, you know understanding
00:47:22:20 - 00:47:24:07
and this isn't something that we currently do
00:47:24:07 - 00:47:28:11
because we're not supporting a drill yet commercially.
00:47:28:11 - 00:47:29:20
This is something that we're doing
00:47:29:20 - 00:47:32:04
as research for this technology.
00:47:32:04 - 00:47:35:01
But understanding what a normal drilling
00:47:35:01 - 00:47:36:17
motion would look like versus
00:47:36:17 - 00:47:39:04
the thing jerking out of your hand,
00:47:39:04 - 00:47:40:22
that's something that you can actually
00:47:40:22 - 00:47:42:17
you don't even need machine learning for that.
00:47:42:17 - 00:47:43:06
That is something
00:47:43:06 - 00:47:46:06
you can actually write an algorithm to and go where
00:47:46:23 - 00:47:49:19
if the surgeon says, you know, turn on anti jerk,
00:47:51:08 - 00:47:53:20
don't let the tool jerk out of my hand.
00:47:53:20 - 00:47:56:15
And that is a motion that you can characterize.
00:47:56:15 - 00:47:59:04
And it's a motion that is unwanted. Right?
00:47:59:04 - 00:48:02:01
It is very easy to get any otologist
00:48:02:01 - 00:48:03:18
just to sit down and go
00:48:03:18 - 00:48:05:22
I don't want a tool jerking out of my hand.
00:48:05:22 - 00:48:06:18
OK, well, what is it?
00:48:06:18 - 00:48:08:05
What is what is jerking out of your hand
00:48:08:05 - 00:48:08:23
look like
00:48:08:23 - 00:48:10:05
actually characterizing
00:48:10:05 - 00:48:11:03
that mathematically
00:48:11:03 - 00:48:13:19
and squelching that potential motion out
00:48:14:13 - 00:48:15:22
from the robot's repertoire.
00:48:15:22 - 00:48:19:12
That's where we add safety to manual instruments
00:48:19:12 - 00:48:22:09
that cannot be achieved under normal means.
00:48:22:22 - 00:48:24:14
So yeah, it sounds like your your
00:48:24:14 - 00:48:26:06
your first launch is implementing
00:48:26:06 - 00:48:28:15
these safety features that don't exist with humans.
00:48:29:07 - 00:48:32:16
So that's a great sort of baseline platform
00:48:32:16 - 00:48:34:11
launch and then, you know,
00:48:34:11 - 00:48:36:17
let's talk a little bit more about the future
00:48:36:17 - 00:48:38:00
with autonomy.
00:48:38:00 - 00:48:39:17[6]
And you need to collect the data
00:48:39:17 - 00:48:41:13
from each procedure eventually and use it.
00:48:41:13 - 00:48:44:08
But what you have the motion.
00:48:44:19 - 00:48:45:13
I guess talk a little bit
00:48:45:13 - 00:48:47:11
more about what data you're collecting
00:48:47:11 - 00:48:49:00
or potentially going to collect with this product
00:48:49:00 - 00:48:50:11
when it's on the market
00:48:50:11 - 00:48:53:01
and maybe an idea of what what where
00:48:53:01 - 00:48:53:18
the first couple
00:48:53:18 - 00:48:56:06
layers of autonomy you'd want to go to market with.
00:48:56:19 - 00:48:59:06
I'll give you a little I'll give you a hint.
00:48:59:06 - 00:49:01:22
If you look at something that has been done by
00:49:01:22 - 00:49:04:07
Dr. Greg Hager in the computer science department
00:49:04:07 - 00:49:05:14
at Johns Hopkins
00:49:05:14 - 00:49:07:18
since 2006, he's been working on something
00:49:07:18 - 00:49:09:13
called the language of surgery.
00:49:09:13 - 00:49:12:18
And one of the examples of his research involves
00:49:12:18 - 00:49:16:04
basically having a evaluation algorithm
00:49:16:04 - 00:49:19:01
tied to the hand controls of a da Vinci.
00:49:19:08 - 00:49:22:08
And it can almost instantly tell you
00:49:22:15 - 00:49:26:13
if how experienced the user is that has their hands
00:49:26:13 - 00:49:27:23
in the control system
00:49:27:23 - 00:49:30:00
that can be applied all over the place.
00:49:30:07 - 00:49:32:08
And so one of the things that we want to do
00:49:32:08 - 00:49:35:17
is as we're collecting all the motion data from
00:49:35:17 - 00:49:37:06
the surgeon that's driving
00:49:38:22 - 00:49:39:23
the robot
00:49:39:23 - 00:49:42:23
is actually looking at the efficiency of hand motions
00:49:43:04 - 00:49:44:18
and how can we help support
00:49:44:18 - 00:49:47:04
what they ultimately want to do
00:49:47:04 - 00:49:49:07
even without, even with like, you know,
00:49:49:07 - 00:49:51:05
tremor control and things like that,
00:49:51:05 - 00:49:52:03
there's still going to be
00:49:52:03 - 00:49:54:04
an inefficiency of of motion.
00:49:54:10 - 00:49:55:18
Are there ways
00:49:55:18 - 00:49:57:21
that we can actually tighten that envelope
00:49:57:21 - 00:49:59:06
and make the instrument
00:49:59:06 - 00:50:02:15
do what we can tell the surgeon is really trying to do
00:50:03:02 - 00:50:06:00
and help them dial in the instruments
00:50:06:05 - 00:50:08:06
so that they're just going straight for,
00:50:08:06 - 00:50:10:07
the disease tissue or whatever it is.
00:50:10:16 - 00:50:14:06
And we can actually glean a lot of that
00:50:14:06 - 00:50:15:18
from just the motion data.
00:50:15:18 - 00:50:18:15
So that, to me is is really cool.
00:50:19:00 - 00:50:21:15
Because that's that's that's fruit that can be done
00:50:22:10 - 00:50:25:17
algorithmically because, machine learning in surgical
00:50:25:17 - 00:50:28:04
robots is not something that's going to go together
00:50:29:07 - 00:50:30:05
in the foreseeable
00:50:30:05 - 00:50:33:17
future because there is no clearance path with the FDA
00:50:34:02 - 00:50:37:17
for basically having machine learning guide a
00:50:38:14 - 00:50:41:03
a robotic system that is holding a weapon
00:50:41:13 - 00:50:42:18
because that's really what
00:50:42:18 - 00:50:44:19
now they're looking at it right now. Right.
00:50:44:19 - 00:50:46:22
And so there's so much that we can do
00:50:47:01 - 00:50:49:14
just by looking at the motion data and saying,
00:50:49:14 - 00:50:52:10
how can we dial it in and really ultimately do
00:50:52:18 - 00:50:55:20
what the surgeons are trying to do in the first place
00:50:55:20 - 00:50:59:14
and actually be able to tighten that motion envelope.
00:50:59:14 - 00:51:00:21
And so I think that's pretty exciting.
00:51:00:21 - 00:51:03:23
But second to that is
00:51:04:12 - 00:51:07:19
we have been exploring ways of adding
00:51:07:19 - 00:51:10:20
even though we're using existing surgical instruments
00:51:10:20 - 00:51:13:04
without modifying the instrument,
00:51:13:04 - 00:51:16:19
how can we effectively add sensor data to the use of
00:51:16:19 - 00:51:17:21
the instrument itself?
00:51:17:21 - 00:51:20:21
Like, for example, if we were holding a drill
00:51:21:20 - 00:51:23:06
with the robot,
00:51:23:06 - 00:51:25:03
is there a way that we can actually tell
00:51:25:03 - 00:51:26:05
how much pressure
00:51:26:05 - 00:51:27:07
is being applied
00:51:27:07 - 00:51:29:23
between patient tissue and the tip of the drill?
00:51:29:23 - 00:51:31:08
And what kinds of things
00:51:31:08 - 00:51:33:15
can we improve for the surgeon by that?
00:51:33:15 - 00:51:34:18
Like, for example,
00:51:34:18 - 00:51:37:01
can we make sure that if you're like in the middle ear
00:51:37:01 - 00:51:39:08
and you're doing something called a stapedectomy,
00:51:39:08 - 00:51:42:00
can I make sure that you can't push the drill so hard
00:51:42:00 - 00:51:43:21
that you actually break the stapes
00:51:43:21 - 00:51:45:01
or the stapes foot plate,
00:51:45:01 - 00:51:47:06
which is a really bad thing to have happen?
00:51:48:05 - 00:51:51:02
Or or if I'm drilling those nooks and crannies
00:51:51:02 - 00:51:52:01
I was talking about
00:51:52:01 - 00:51:54:08
and I've got pressure from the drill,
00:51:54:08 - 00:51:56:11
and then suddenly I'm in outer space
00:51:56:11 - 00:51:57:12
and there's no pressure.
00:51:57:12 - 00:51:59:19
There's no resistance to the tip of the drill.
00:51:59:19 - 00:52:02:08
Can I slow down the response of the drill
00:52:02:08 - 00:52:04:11
just to make sure that it doesn't plunge?
00:52:04:11 - 00:52:07:16
And these are all things that we can do with
00:52:08:01 - 00:52:11:09
with by adding maybe a little bit of additional sensor
00:52:11:09 - 00:52:12:17
that might go into like a
00:52:12:17 - 00:52:15:02
the the instrument adapter itself.
00:52:15:02 - 00:52:17:03
So so the instrument adapter
00:52:17:03 - 00:52:18:18
might have additional sensors on it
00:52:18:18 - 00:52:21:11
that just gives us better information
00:52:21:11 - 00:52:23:19
about how the tool is being used and
00:52:24:19 - 00:52:26:10
what kind of tissue response
00:52:26:10 - 00:52:27:23
we're getting from the tool.
00:52:27:23 - 00:52:30:16
And again, these are things that you cannot do
00:52:31:07 - 00:52:34:20
in any meaningful way with a manually held instrument.
00:52:35:02 - 00:52:38:18
So the cool thing about this whole industry is that
00:52:38:18 - 00:52:41:22
robots are actually giving you access to information
00:52:41:22 - 00:52:44:03
about how surgeries are being done.
00:52:44:03 - 00:52:45:16
That has been anecdotal
00:52:45:16 - 00:52:48:17
for literally thousands of years.
00:52:48:17 - 00:52:51:03
And now suddenly we can actually get
00:52:51:03 - 00:52:54:04
we can put numbers to hand motions,
00:52:54:04 - 00:52:56:13
we can put numbers to, you know,
00:52:57:18 - 00:52:59:20
tissue contact and all those sorts of things.
00:52:59:20 - 00:53:00:16
That is that
00:53:00:16 - 00:53:02:06
even if you could have collected it
00:53:02:06 - 00:53:03:18
now ten years ago or something like that,
00:53:03:18 - 00:53:04:16
it would have been useless
00:53:04:16 - 00:53:07:04
because you're still holding the instrument manually.
00:53:07:04 - 00:53:09:08
So what are you going to do? Right.
00:53:09:08 - 00:53:13:00
And so, that's to us, that's really exciting
00:53:13:00 - 00:53:15:05
because that also opens the door to
00:53:15:14 - 00:53:18:11
what if we were talking to a nerve root monitor?
00:53:18:22 - 00:53:20:11
Well, the way that we're currently works
00:53:20:11 - 00:53:22:11
is you hear a beep and you stop your hand.
00:53:22:11 - 00:53:25:12
While we were talking directly to the nerve monitor,
00:53:25:12 - 00:53:29:06
we could get a signal from it and we could stop you
00:53:29:06 - 00:53:31:22
in a way that the human hand cannot match.
00:53:31:22 - 00:53:34:08
And so there's all kinds of safety capabilities
00:53:34:08 - 00:53:36:02
that we could do
00:53:36:02 - 00:53:38:09
just by being more connected in the O.R.
00:53:38:09 - 00:53:39:22
And to me, that's what
00:53:39:22 - 00:53:42:05
that's what I find really exciting because
00:53:42:05 - 00:53:43:10
I'm not a roboticist.
00:53:43:10 - 00:53:45:03
We've got expert roboticists here.
00:53:45:03 - 00:53:46:04
That's not my thing.
00:53:46:04 - 00:53:50:01
My thing is what kinds of applications can we provide
00:53:50:01 - 00:53:51:23
using this robot as a platform?
00:53:51:23 - 00:53:53:23
And that's really where we take this in the future.
00:53:54:12 - 00:53:56:18
And that's that's cool to me. I love that stuff.
00:53:57:18 - 00:53:58:00
Yeah.
00:53:58:00 - 00:53:58:19
I hadn�t appreciate it until
00:53:58:19 - 00:53:59:15
right here until the end
00:53:59:15 - 00:54:01:11
where you keep talking about safety.
00:54:01:11 - 00:54:03:08
And when you think about devices
00:54:03:08 - 00:54:04:17
yeah, they say safety and efficacy,
00:54:04:17 - 00:54:07:08
but kind of the engineer in us always kind of jumps
00:54:07:08 - 00:54:09:10
to, well, how effective is it?
00:54:09:10 - 00:54:12:14
And you kind of, downplay safety in a way.
00:54:12:14 - 00:54:13:20
Just want to check that box.
00:54:13:20 - 00:54:17:06
But your platform in surgical robotics in general,
00:54:17:06 - 00:54:20:16
it sounds like safety almost could be more important
00:54:21:02 - 00:54:23:17
in what you're preventing than the efficacy.
00:54:24:13 - 00:54:25:16
Oh, heck, yeah.
00:54:25:16 - 00:54:28:09
You know, one of one of Russ Taylor's,
00:54:28:09 - 00:54:31:09
famous phrases is the biggest risk of surgical robots
00:54:31:17 - 00:54:33:13
is that it does exactly the right thing,
00:54:33:13 - 00:54:35:07
but in the wrong place.
00:54:35:07 - 00:54:38:16
And we think of that all the time.
00:54:38:16 - 00:54:39:23
You know, in our code, it's
00:54:39:23 - 00:54:43:09
just it's baked in our DNA nine lines out of ten
00:54:43:20 - 00:54:46:20
in the in the software code are about double
00:54:46:20 - 00:54:48:20
checking what you're about to do and and safety
00:54:48:20 - 00:54:50:03
and all those sorts of things.
00:54:50:03 - 00:54:53:01
And so, yeah, sometimes we almost take it for granted
00:54:53:01 - 00:54:56:03
because it's such an integral part of everything
00:54:56:03 - 00:54:57:18
that you tell the robot to do.
00:54:57:18 - 00:55:00:15
You got to make sure that it's doing it safely.
00:55:00:23 - 00:55:02:23
And that's got to be your baseline
00:55:03:05 - 00:55:05:10
and then build on top of that.
00:55:05:10 - 00:55:06:12
But yeah, absolutely.
00:55:06:12 - 00:55:07:13
The safety part
00:55:07:13 - 00:55:11:19
and the opportunities for making surgery safer,
00:55:11:19 - 00:55:14:08
although man, that's a difficult indication to prove
00:55:14:19 - 00:55:17:15
from a marketing clearance standpoint.
00:55:17:20 - 00:55:19:20
But that's got to be your guiding light,
00:55:19:20 - 00:55:20:19
I think, because that's
00:55:20:19 - 00:55:22:17
where the opportunities really are.
00:55:22:17 - 00:55:24:08
Two more questions for you.
00:55:24:08 - 00:55:25:19
And this one's a little bit of a curveball,
00:55:25:19 - 00:55:26:08
but I like this.
00:55:26:08 - 00:55:29:02
When we polled the audience, the Key Techers here
00:55:29:05 - 00:55:32:00
are there certain actions, surgical actions
00:55:32:15 - 00:55:36:10
that a human is dramatically better than a robot
00:55:36:10 - 00:55:37:15
at doing?
00:55:37:21 - 00:55:40:21
As you see it now, or humans
00:55:41:12 - 00:55:43:23
losing and they've already lost?
00:55:43:23 - 00:55:46:10
Well, I think one is
00:55:47:02 - 00:55:48:02
where that where the robot
00:55:48:02 - 00:55:50:20
I think is better is actually stabilizing instruments.
00:55:50:20 - 00:55:51:11
I think
00:55:51:11 - 00:55:53:13
I think a robot is better capable of holding
00:55:53:13 - 00:55:56:01
something steady and never getting tired
00:55:56:10 - 00:55:57:20
that a human being is.
00:55:57:20 - 00:55:59:08
But here's the thing.
00:55:59:08 - 00:56:00:07
Human beings are better
00:56:00:07 - 00:56:02:03
at getting the lay of the land.
00:56:02:03 - 00:56:04:16
You know, you can look inside a
00:56:04:16 - 00:56:05:14
part of the anatomy
00:56:05:14 - 00:56:07:17
and you can figure out heads or tails
00:56:07:17 - 00:56:11:09
as a human being, way better than any algorithm
00:56:11:09 - 00:56:12:11
is going to be able to do,
00:56:12:11 - 00:56:14:08
at least in the foreseeable future.
00:56:14:08 - 00:56:16:17
And so to me, that's why cooperative,
00:56:16:21 - 00:56:18:07
cooperative robotics
00:56:18:07 - 00:56:18:20
is such a it's
00:56:18:20 - 00:56:20:08
such a great opportunity
00:56:20:08 - 00:56:22:16
because you get to bring together
00:56:22:16 - 00:56:23:19
the best of both worlds.
00:56:23:19 - 00:56:26:21
Human beings are still the best motivated
00:56:26:21 - 00:56:28:11
decision makers
00:56:28:11 - 00:56:30:22
they can look at something and they can motivate.
00:56:30:22 - 00:56:32:17
You know, they're motivated to go,
00:56:32:17 - 00:56:33:18
OK, this is what I'm going to do.
00:56:33:18 - 00:56:35:00
OK, I've opened the patient up.
00:56:35:00 - 00:56:36:07
Oh, my God, they're a mess.
00:56:36:07 - 00:56:38:10
And actually make a game plan
00:56:38:10 - 00:56:39:12
and then apply it.
00:56:39:12 - 00:56:43:01
But the robot can actually make it to me
00:56:43:09 - 00:56:47:18
easier for them to do what they are trained to do.
00:56:47:23 - 00:56:49:21
And that's the best of both worlds.
00:56:49:21 - 00:56:50:15
I totally agree.
00:56:50:15 - 00:56:53:00
So last question for you.
00:56:53:00 - 00:56:54:07[7]
So you've been developing this robot
00:56:54:07 - 00:56:55:23
for six years.
00:56:55:23 - 00:56:58:23
You started as two guys in a garage, sounds like.
00:56:59:18 - 00:57:02:00
And here you are today, thriving.
00:57:02:00 - 00:57:03:15
So if you could go back
00:57:03:15 - 00:57:05:09
and do one thing differently
00:57:05:09 - 00:57:07:15
from a product development perspective,
00:57:08:15 - 00:57:11:13
what do you think that would be and why?
00:57:11:13 - 00:57:12:23
And again, our audience is,
00:57:12:23 - 00:57:16:06
entrepreneurs at global companies and,
00:57:16:15 - 00:57:18:20
folks like yourselves trying to get to market.
00:57:18:20 - 00:57:22:03
I wish I had more copies of our earlier prototypes.
00:57:23:03 - 00:57:25:01
I feel like even with the Mark 1
00:57:25:01 - 00:57:27:05
and the flying trashcan and all of that kind of stuff,
00:57:27:05 - 00:57:27:18
I feel like
00:57:27:18 - 00:57:29:06
we could have learned more
00:57:29:06 - 00:57:31:09
if we had more copies of that robot
00:57:31:09 - 00:57:32:12
that we could have.
00:57:32:12 - 00:57:34:17
I've got five different universities
00:57:34:17 - 00:57:36:15
that I work with on a regular basis
00:57:36:15 - 00:57:39:22
that are using our platform or building tools
00:57:40:06 - 00:57:41:12
or software modules
00:57:41:12 - 00:57:43:15
that are anticipating the use of our platform.
00:57:44:02 - 00:57:48:03
And if I could just send them a robot to have,
00:57:49:03 - 00:57:50:22
I think that would be such a
00:57:50:22 - 00:57:52:21
they would have been such a huge win.
00:57:52:21 - 00:57:55:20
And it's tough because especially,
00:57:55:20 - 00:57:57:04
in the money raising environment,
00:57:57:04 - 00:57:58:15
unless you're willing to go to some,
00:57:58:15 - 00:57:59:20
you know, $10 billion
00:57:59:20 - 00:58:01:07
fund and bring in a hundred million
00:58:01:07 - 00:58:03:22
and get deluded, you know, beyond belief,
00:58:04:21 - 00:58:06:23
it's not always easy to afford,
00:58:06:23 - 00:58:09:12
a half dozen prototypes of something
00:58:09:12 - 00:58:12:06
that you're not sure if it's really the final version.
00:58:12:06 - 00:58:12:20
Right.
00:58:13:19 - 00:58:16:16
But but if I had a magic wand
00:58:17:00 - 00:58:19:19
and I could make that decision
00:58:20:04 - 00:58:22:19
without consequences, consequences being dilution,
00:58:23:21 - 00:58:28:01
I would have absolutely have had more copies of each
00:58:28:01 - 00:58:30:11
of our rounds of prototypes than we were then
00:58:31:00 - 00:58:33:07
than what we actually ended up building.
00:58:33:07 - 00:58:34:10
I just think it's more valuable.
00:58:34:10 - 00:58:36:04
It's way more valuable.
00:58:36:04 - 00:58:36:10
You know,
00:58:36:10 - 00:58:38:03
it sounds like you wanted
00:58:38:03 - 00:58:40:10
why did you want more prototypes?
00:58:40:16 - 00:58:42:17
Just more user feedback early.
00:58:43:03 - 00:58:45:18
More user feedback.
00:58:46:18 - 00:58:47:23
You know, and then sometimes
00:58:47:23 - 00:58:50:03
maybe something screwy happens and you just like,
00:58:50:03 - 00:58:51:03
oh, man, if I could just, like,
00:58:51:03 - 00:58:53:01
take this robot half apart,
00:58:53:01 - 00:58:53:16
I could, like,
00:58:53:16 - 00:58:55:11
learn so much and I could figure out, oh,
00:58:55:11 - 00:58:57:12
why did you give that weird response? Right.
00:58:57:12 - 00:59:00:12
But if you've only got one or two of them,
00:59:00:12 - 00:59:02:00
you take the thing apart
00:59:02:00 - 00:59:03:17
so that you can play with it on the bench.
00:59:03:17 - 00:59:04:16
You suddenly have taken
00:59:04:16 - 00:59:06:08
that prototype out of commission
00:59:06:08 - 00:59:07:13
and now the software guys
00:59:07:13 - 00:59:09:11
maybe don't have a platform to develop it on
00:59:09:11 - 00:59:11:15
or you can't take it over to the university
00:59:11:15 - 00:59:13:10
or you can't take it to a cadaver lab.
00:59:13:10 - 00:59:15:01
You've got limited resources, right?
00:59:15:01 - 00:59:16:13
There's only there's only so
00:59:16:13 - 00:59:18:15
especially with something like a surgical robot,
00:59:18:15 - 00:59:20:14
you've only got so many resources
00:59:20:14 - 00:59:21:23
and yet you've got multiple.
00:59:21:23 - 00:59:22:23
Well, in our case,
00:59:22:23 - 00:59:24:12
we had surgeons from,
00:59:24:12 - 00:59:26:18
a half dozen disciplines going,
00:59:26:18 - 00:59:27:18
I've got a cadaver.
00:59:27:18 - 00:59:29:10
And if you can just bring the robot over,
00:59:31:05 - 00:59:32:15
we'll do a cadaver lab with you.
00:59:32:15 - 00:59:34:22
And it's like, oh, my God, yeah.
00:59:34:22 - 00:59:37:19
Just the opportunities create themselves.
00:59:37:19 - 00:59:39:23
If you've got multiple copies of the your prototype.
00:59:40:09 - 00:59:40:17
Yeah.
00:59:40:17 - 00:59:41:23
Did you consider, like,
00:59:41:23 - 00:59:44:07
making an analog a cheaper version?
00:59:44:07 - 00:59:44:23
I'm thinking
00:59:44:23 - 00:59:46:10
like I don't want to say
00:59:46:10 - 00:59:48:10
duct tape, but like, something that,
00:59:49:02 - 00:59:50:09
kind of mimicked it
00:59:50:09 - 00:59:52:06
enough to get you the user feedback
00:59:52:06 - 00:59:53:13
that you're looking for,
00:59:53:13 - 00:59:55:02
or is it just way too complicated
00:59:55:02 - 00:59:56:18
even consider that the other way to do it
00:59:56:18 - 00:59:58:14
cheaply is do a subsystem.
00:59:58:14 - 01:00:00:11
It would have been tough in our case
01:00:00:11 - 01:00:02:00
because the precision levels
01:00:02:00 - 01:00:04:07
were actually creating new opportunities.
01:00:04:07 - 01:00:05:18
There are minimally invasive
01:00:05:18 - 01:00:06:03
and narrow
01:00:06:03 - 01:00:07:09
corridor procedures
01:00:07:09 - 01:00:10:01
where just no other robot could touch that thing.
01:00:10:14 - 01:00:13:01
And the surgeons were like, well,
01:00:13:01 - 01:00:15:08
I think your technology can
01:00:15:08 - 01:00:17:16
but I need to try it myself.
01:00:17:16 - 01:00:22:21
And it was kind of one of those tricky things
01:00:22:21 - 01:00:25:21
where it's like what we're trying to evaluate
01:00:25:21 - 01:00:28:19
really required a fully working platform.
01:00:28:19 - 01:00:31:06
So we also did build simulators.
01:00:31:06 - 01:00:31:21
So I mean,
01:00:31:21 - 01:00:33:18
there are things that we
01:00:33:18 - 01:00:35:20
in fact do through software simulation.
01:00:35:20 - 01:00:38:03
We actually simulate a transsphenoidal approaches
01:00:38:03 - 01:00:39:17
and things like that completely
01:00:39:17 - 01:00:41:03
incidentally through Blender
01:00:43:06 - 01:00:43:23
we actually
01:00:43:23 - 01:00:47:20
built an animation model with rigging in Blender,
01:00:47:20 - 01:00:50:00
and we're actually able to validate
01:00:50:00 - 01:00:51:22
things like a transsphenoidal approach
01:00:51:22 - 01:00:53:18
and just make sure the geometry
01:00:53:18 - 01:00:57:03
was sufficient to get in and out and all of that.
01:00:57:03 - 01:00:57:20
So I mean, there
01:00:57:20 - 01:01:00:02
certainly were things that we did
01:01:00:02 - 01:01:02:13
do through through software simulation but
01:01:03:02 - 01:01:05:18
yeah, I looking back and if I had extra
01:01:05:18 - 01:01:07:20
versions of the Mark 1 and the Mark 2,
01:01:07:20 - 01:01:09:20
I would have been a happy camper. Yeah.
01:01:09:20 - 01:01:11:23
So there was obviously funding limitations
01:01:12:13 - 01:01:14:01
but I guess another question,
01:01:14:01 - 01:01:15:02
if you could go back.
01:01:15:02 - 01:01:15:06
Yeah.
01:01:15:06 - 01:01:15:11
I guess
01:01:15:11 - 01:01:16:14
what was the most challenging
01:01:16:14 - 01:01:18:12
part of what you're doing?
01:01:18:12 - 01:01:20:04
In some cases,
01:01:20:04 - 01:01:21:20
it was actually taking some of those things
01:01:21:20 - 01:01:24:07
that we were talking about earlier, those qualitative,
01:01:25:01 - 01:01:27:04
you know, kind of how does it feel
01:01:27:10 - 01:01:30:09
and actually turning that into numbers.
01:01:30:09 - 01:01:31:03
I mean, at some point
01:01:31:03 - 01:01:32:12
you're dealing with kinematics, right?
01:01:32:12 - 01:01:34:14
You've got to turn it into math somehow,
01:01:34:14 - 01:01:35:08
even though it
01:01:35:08 - 01:01:37:03
even though, like I said, it's got to feel right
01:01:37:03 - 01:01:39:06
or it doesn't matter what the math says.
01:01:39:06 - 01:01:40:19
But the flip side of that is
01:01:40:19 - 01:01:42:12
you've got to take what feels right
01:01:42:12 - 01:01:44:22
and you've got to be able to express that as math.
01:01:44:22 - 01:01:47:08
And that was tricky.
01:01:47:08 - 01:01:48:22
We've we've got some hardcore
01:01:48:22 - 01:01:52:18
math nerds here and we had some scary whiteboards with
01:01:54:15 - 01:01:55:10
all kinds of
01:01:55:10 - 01:01:57:07
crazy algorithms being worked out of there.
01:01:57:07 - 01:01:58:06
And,
01:01:58:06 - 01:02:01:13
that stuff can get really hairy at times.
01:02:02:00 - 01:02:06:14
And sometimes we would almost
01:02:07:06 - 01:02:09:06
there were a couple of times when we almost spent
01:02:09:06 - 01:02:11:10
nothing but
01:02:11:10 - 01:02:14:06
a full week just sitting there vetting math
01:02:16:00 - 01:02:17:18
because because ultimately that is
01:02:17:18 - 01:02:19:01
where your safety comes from.
01:02:19:01 - 01:02:21:19
I mean, if your algorithms bad
01:02:21:19 - 01:02:22:11
especially for these
01:02:22:11 - 01:02:25:14
some of these weird kinematic equations,
01:02:26:19 - 01:02:28:12
there could be a, you know,
01:02:28:12 - 01:02:31:18
a hidden motion that could actually cause an injury.
01:02:32:02 - 01:02:34:05
And so it's you got it.
01:02:34:05 - 01:02:36:05
That is rough to do it is it
01:02:36:05 - 01:02:38:20
it is a hair pulling situation at times.
01:02:39:17 - 01:02:42:14
So you recommend like creating kind of a baseline
01:02:42:15 - 01:02:46:18
algorithm and, create this sort of inventor analytics.
01:02:46:18 - 01:02:49:04
Yeah. Through analytics.
01:02:50:10 - 01:02:50:20
Yeah.
01:02:50:20 - 01:02:51:12
You got to you got
01:02:51:12 - 01:02:53:05
to you got to prove your math, prove your math.
01:02:53:05 - 01:02:56:07
Maybe the next episode sometime in the future.
01:02:56:07 - 01:02:57:23
So hey,
01:02:57:23 - 01:02:59:12
thank you so much for your time today.
01:02:59:12 - 01:03:00:20
And obviously,
01:03:00:20 - 01:03:01:19
best of luck, you know,
01:03:01:19 - 01:03:03:01
getting to market here
01:03:03:01 - 01:03:04:19
in the next couple of quarters or so
01:03:04:19 - 01:03:08:13
and yeah, go Baltimore.
01:03:08:13 - 01:03:10:00
So thanks, everybody.
Is this a real phrase? "Kick the fairs"
It doesn't look like it (according to google) but I think that's what he said.
Do you think he meant Mattel ?
Is this a real phrase?
I don't think so - I would just change it to "I'll give you another example"
Background noise here. Sounds like Andy is repeating himself
Lots of backend noise here