October 11, 2024 • 35 mins
What if the key to detecting cancer earlier—and saving thousands of lives—lies in AI? Host Ed McNamara explores how Kheiron Medical is making that a reality. Joined by Peter Kecskemethy, CEO and co-founder, and Sarah Kerruish, Chief Strategy Officer, they delve into the groundbreaking technology behind Mia® – Kheiron’s AI-powered tool that is revolutionizing breast cancer detection by detecting what the human eye might miss.
Named one of the biggest medical breakthroughs of 2023, Mia® is proving its potential in real-world clinical settings, boosting breast cancer detection rates and offering hope to thousands of women. Peter and Sarah share the challenges, triumphs, and future of AI in healthcare, offering a glimpse into how technology is reshaping cancer diagnostics and changing lives.
Listen to the full episode below or wherever you get your podcasts.
Featuring: Peter Kecskemethy, CEO and co-founder of Kheiron Medical, and Sarah Kerruish, Chief Strategy Officer.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Peter (00:00):
On a on a day to day basis, the whole development cycle of AI
is very, very complicated becauseevery time you need to figure
out, okay, is it good enough?
Is it good enough?
What element do I need to improve on it?
But even that test of whetherit's good enough is hard.
Ed (00:13):
In the world of technology, heroes are everywhere.
They're overcoming disruption, deliveringsustainable outcomes, and fearlessly
forging the future to solve what's next.
Join me.
Ed McNamara, as we meet the peopleand businesses driving change in
our constantly disrupted world.
This is Innovation Heroes, apodcast brought to you by SHI.
(00:39):
Think about the everyday ways AI haschanged our lives, from turning speech
into text to managing our daily schedules.
But in today's episode, we're going tosee how AI is being used to save lives.
I'm joined by Peter Kecskamaty.
The CEO and co founder ofChiron Medical and Sarah Karush,
their Chief Strategy Officer.
(00:59):
They're here to share their journeyof how Chiron Medical's talented deep
learning team developed an AI technologyto improve breast cancer screenings, so
that patients can experience improvedhealth outcomes through earlier detection.
We'll dive into the technology, thereal world impact it's having, and what
the future holds for AI in medicine.
(01:21):
Peter and Sarah, welcometo Innovation Heroes.
Sarah (01:23):
Thank you so much, Ed.
Peter (01:24):
Very nice.
Ed (01:27):
And thanks for being here.
Peter, can you share a bit aboutyour background and what inspired
you to co found Chiron Medical?
Peter (01:36):
Yes.
And thanks again for having us.
Um, I had a pretty longjourney towards this.
Um, actually, um, my backgroundis technical AI, uh, software
engineering, computer science, uh,high performance computing and stats,
um, and, and population genetics.
(01:57):
But I think the most important, uh,Part of the background for this is
my, my mom's a radiologist and, uh,I actually spent most of my childhood
sort of like after school hours, etcetera, uh, in radiology departments.
And I would say where Chiron comesfrom is the recognition that AI and
(02:20):
generally computer science and AIare extremely powerful in producing
tools for, For doctors and also torecognition that in health care, it's
not just the patients who suffer, buta lot of times the doctors as well.
You often see the doctor having extremelyunsuitable software that they need to use.
(02:42):
Uh, at their work sort of all day andthey get out of the hospital and they
have the most sophisticated, uh, um,software and algorithms recommending, um,
various movies or restaurants to go to.
So I think what we need to doand what Karen comes from is.
bringing the bestpossible tools to doctors.
(03:04):
And I do believe that AI is one of themost important piece of technologies to,
uh, to produce that tool, those tools.
And it, since doctors are helping patientsand, and are absolutely on the front
lines of saving lives this way, I thinkone of the, just one of the best ways.
Uh, innovation, especiallytech, uh, software engineering
can, uh, can, can save lives.
Ed (03:27):
And Peter, we're one question in, and I'm going, you know, somewhat off script
already, but it just, it just brought tomind in terms of like all of the trainings
and the years of practice that doctorshave and radiologists, I mean, how much
of that is computer training and how bigof a gap could there be from knowledge of,
you know, perfecting their craft in termsof treating patients the way that they
do versus the technology that they use?
(03:49):
So is there one?
Peter (03:51):
Um, I think it depends on the on the field for sure.
Um, and as I see, radiology has alwaysbeen on the forefront of adopting
really somewhat esoteric but extremelyfrontline technologies like if you
think about the X ray or and thencomputer tomography and now MRI, which
(04:13):
is extremely high level of technology.
Um, then, then digitization in thehospital, a lot of, uh, that was
driven in, uh, by, by radiologists.
And I do believe AI is one of the, um,like radio is also one of the most natural
areas to, to adopt, uh, AI, especiallyimage, image recognition systems.
Ed (04:37):
Excellent.
And Sarah, you've had an amazingcareer as a journalist and a filmmaker.
You know, what brought you to ChironMedical and, and how do you define
your role as chief strategy officer?
Sarah (04:49):
That's a good question.
So, um, I started out in media, asyou said, in filmmaking and I was
invited to go and make a film ofa company called General Magic in
Silicon Valley many, many years ago.
And that was my introductionto the world of tech.
But I very quickly became interested inhow you use technology and software and
hardware to bring important ideas to life.
And I got, I was very privileged towork with some incredible people,
(05:12):
some of the big innovators of our age.
And, um, you know, from them, I learneda tremendous amount and thought this
is something that I want to, Spend mylife doing and in particular I'm very
interested in how technology can beapplied in healthcare and in the field
of cancer And you know your earlierpoint Doctors when they when you have
a scan when you have a suspicious massor there's something wrong when you
(05:32):
have a scan You sort of expect thatthe doctor will look at that and go.
Oh, this is a very specific kind ofcancer It's a two millimeter something
coma Um, but actually, cancer detectionsare really difficult tasks, even
for the most experienced doctors.
And that's where AI canreally add a lot of value.
So when I was introduced to Chiron,and you know, this incredible team
of machine learning world experts,and Peter, who really understood the
(05:54):
pain points of doctors, it seemed tome that that was the right formula.
for creating a technology like the onewe've created called NEAR to really
help doctors with that very difficulttask of, of detecting cancer earlier.
And as it relates to my, as it relatesto your question about Chief Strategy
Officer, the way I see that is really,you know, Peter leads on strategy,
but I'm sort of his right arm.
(06:16):
Um, and, and we look together forareas of the business where we can
have the most meaningful growth.
Um, And depending on the stage of thebusiness, that can be different stages,
but it's things like for us creatingacademic partnerships, working to
develop relationships with organizationslike the NHS or big partners in the
US, so that we can test and developthe technology and make sure that it's
working for, um, you know, the womenthat we are hoping to, um, to help.
Ed (06:42):
And Sarah, I think that, you know, in terms of developing partnerships, I was
reading and I'm consulting my notes now.
Um, your company has experts and engineersthat your partner with that were from
Oxford, Cambridge, um, McGill, uh, KTHUniversity, uh, sorry, KTH University
College, London Imperial College, uh,from a recruiting perspective, I'd
(07:05):
imagine that makes a pretty attractiveplace for people on the cutting edge
to, uh, to want to come to work.
Is that fair?
Sarah (07:12):
Yeah.
And I think that, um, you know, we'vedefinitely attracted incredible talent.
Uh, and the most special thing though,about the people that we work with is not
the illustrious universities they wentto, although they did well, the incredible
science that they develop or the AIthat they, you know, develop and test.
It's really the, the mission.
And, you know, we have a teamthat's so dedicated to this
(07:32):
mission of finding cancer earlier.
And, you know, as you, as youprobably know, you know, if
you can't, if you detect cancerearlier in our case, breast cancer.
That can mean an entirelydifferent prognosis.
So you're not looking at necessarilyradiation or mastectomies.
You're potentially looking justat lumpectomies and maybe, you
know, maybe some radiation,but possibly not chemotherapy.
So, you know, we're hearing now fromthe women whose cancers were picked up
(07:55):
earlier by Mia, which is our technology.
And it's just the greatest privilegeof our lives to know that there are
women, many women now walking outin the world whose cancer was picked
up earlier by our technology and whohave a better prognosis as a result.
Ed (08:08):
Yeah, I mean, maybe it's time to talk about, um, the, that technology.
Um, you know, Chiron's, uh, medicalAI technology, MIA, um, as you
said, Sarah, it's quite innovative.
Um, Peter, maybe, could you explainwhat sets your AI driven diagnostics
apart in the fight against cancer?
Peter (08:25):
Yeah, so first of all, um, our, um, Number one flagship, uh,
product or AI service is it's calledMia is for breast cancer screening.
Uh, very specifically, um, it isbased on, um, image processing or
AI, uh, the AI, uh, the arm or the.
(08:47):
Um, the part of AI or, uh, that, uh,the type of AI that is, uh, working with
images, image inputs, uh, so basicallythe radiologists are producing the images
and, uh, or the, uh, the hospitals areproducing the images that radiologists
look at, the AI also looks at it.
Um, and pretty much attemptsto make the same decision.
(09:08):
Um, however, the AI hasvery different capabilities.
So the AI is infinitely diligent.
It looks at every single pixelby the humans don't do that.
Um, but the AI doesn't have thesame strategic overview or a
contextual background understandingthat the trained doctor does.
So what we optimized for is twofold.
(09:28):
Um, uh, one.
for the AI to be useful.
So really being ascomplimentary as possible.
So additive, try to do bring allthe capabilities that that is hard
for humans, but not try to replicatethe capability that humans have.
So we're not trying to imitate howdoctors make the decisions, but,
but bring in extra informationthat is, for instance, not.
(09:50):
visible to the human eye or sort ofnot accessible to the brain, the visual
cortex of how humans process images.
Uh, so, so that's, that's how wecombine the two capabilities and focus
on that is one of the ways that Ithink our approach is very different.
And, and that is through Um, uh, both interms of what the AI looks at information
(10:11):
it processes and also how it collaborateswith the doctors and how it collaborates
with the doctors is through variousworkflows and it's a different in the U.
S.
than, than, than in Europe.
Um, so that's one aspect.
And the other aspect is for, forthe AI to be working and, and, you
know, like I'm sure everyone thinksand they develop a piece of AI,
they, they believe it, it works.
(10:32):
But actually.
Assessing whether AI works ornot is extremely complicated.
Uh, it requires statistical thinkingand, and it's a completely new paradigm.
Like we never had a kind of tool.
Humanity never had a tool like thisbefore, when it's hard to assess
whether, whether the tool works or not.
Like if you have a piece of hardware,like a knife, you could figure
(10:53):
out whether that cuts or not.
Uh, you have a piece of software,you know what it's supposed to do
and you test it and then, and then,you know, whether it works or not
for AI, it's extremely hard because.
It is strained on that set of images,for instance, a set of inputs.
And then it's assessed on similar,but different, but fundamentally
different inputs and that's when, if,if it can perform well on, on those
(11:16):
unseen cases, that's when it works.
But what if the case is too unseen?
Like it, it sees it not like a.
breast mammogram, but like a chest xray, then of course it shouldn't work.
But, but what if it, um, it is amammogram, uh, for, for breast cancer
screening, but it's not for like alady from Europe, but a lady from
(11:37):
Southeast Asia, like it, you wouldexpect it still should work, right?
So that exact level of specialism andgeneralizability needs to be gotten right.
As it happens, it's fundamentallyimpossible to prove, extremely hard
to test, and takes a lot of effort.
And, uh, but it, but that kind oftesting has to be a fundamental
(12:00):
part of how AI is being developed.
So you can imagine that it's, on a dayto day basis, the whole development
cycle of AI is very, very complicated,because every time you need to
figure out, okay, is it good enough?
Is it good enough?
What element do I need to improve on it?
But even that test of whetherit's good enough is hard.
So, so that is a very involvedtechnical and, and, uh,
(12:21):
statistical and scientific task.
Um, and, and taking that seriously, thattask to a level that I haven't seen from
others is what I think also sets us apart.
So basically being useful, being, be,uh, bringing AI capability beyond human
capability and, uh, making sure it works.
Ed (12:42):
And I'm just putting myself in the shoes of the patient for a second
to put myself in the shoes of anybodythat has to make a phone call to
a patient that could be, you know,either the best or worst phone call
that they might ever receive, right?
I mean, you really don'twant to be chasing.
I'll say chasing ghosts.
That's that's a layman'sterms or something.
But you you really wantto be sure of the data.
(13:04):
And it's just such anamazingly sensitive topic.
And, you know, Sarah, the waythat you described the company
before, it just sounds like.
Pretty much everyone at Chiron is going tohave that kind of sensitivity because of,
um, at the end of the day, we're talkingtechnology, but there's real, um, you
know, life changing conversations that arehappening as a result of the technology.
Is that fair?
Sarah (13:26):
Yeah, and I can tell you about one of those conversations.
I think the first thing that'simportant to note is that Mia never
makes a decision in isolation.
So Mia will make a recommendationas to whether a woman should be
recalled for further testing or not.
But ultimately, it's the doctorsthat decide, and that's very much our
worldview, that Mia and AI should becomplementary to doctors decision making.
(13:47):
So that's the first thing.
And then in terms of the actualconversation itself, um, we have
one wonderful patient who, um, wasinvolved in a study we did in Scotland.
And when she was told about her verysmall mass, uh, in, in her breast,
that was in fact a cancer post biopsy,they knew that it was a cancer.
She said, Oh, she said,I don't have the big C.
(14:10):
I have a little C.
And so that conversation is never an easyconversation, but we want more of those
little c conversations And you know,my my my ultimate dream is that um, you
know in five ten years That when somebodycalls you to say, hey, you know, we've
done further testing, we think you've gotcancer, or we know that you have cancer.
I want people to have that feelingof like, oh, that's a bummer, but
(14:32):
it's not the, am I going to die?
And I think, you know, there'sbeen so much development, so much,
um, advance in therapeutics, uh,but not so much in diagnostics.
And I think that's where AI isplaying a really critical role.
And a combination of those two thingsis really going to transform cancer and
the very meaning of the word, I think.
Ed (14:51):
And Sarah, earlier you mentioned, you know, early detection and
the impact in early detection.
How has Kiran's technology improvedthe accuracy and efficiency of
breast cancer screening in realworld health care settings?
Sarah (15:07):
Well, that's been the part that's so unbelievably exciting because
we, we started, we did our firstprospective real world study in Hungary.
And we saw extraordinary results there.
So, you know, up to 15 percent ofincreased, uh, cancer detection.
And that was, you know, that was sogroundbreaking that it was reported on
the front page of the New York Times.
(15:27):
You know, one of the first real worldexamples of AI actually being used
in clinical practice and having,you know, extraordinarily good
results beyond our expectations.
Those results we've reproduced in theUK, they were published recently in
the BBC and again made global headlinesbecause we're just beginning to see that.
you know, our technology and technologieslike ours really will have huge,
(15:48):
make huge differences in the accuracyand the, um, and the ability of
doctors to detect cancer earlier.
Um, you know, and, and in many casesyears earlier, and we're projecting that
in breast cancer screening, 20 percentof cancers are missed, uh, routinely
because it's such a difficult task.
It's not like doctorsaren't doing their job.
It's just very difficult tosee and detect breast cancer.
(16:10):
Um, and we, we think, you know, basedon the stats we're seeing now, that MEER
will halve that rate of missed cancers.
And, and what that translatesinto is 2, 000 women a year in
the UK, 40, 000 in, in the U.
S.
These are, these are big numbers and verymeaningful impact that we hope to have.
Ed (16:27):
And I guess from a, from a technology perspective, Um, Peter, it's been just
under 20 years since you studied AIand cybermetrics at the University of
Reading and just under nine since youwere co founder of Chiron Medical.
You know, what's changed most about theAI technology since you first started?
You know, not, not just, you know,your, your development of it, but
(16:47):
the actual, like, capabilitiesin, let's say, nine years since,
since you founded the company.
Peter (16:53):
Yeah, I think Actually, the last, I would say 11 years is the big deep
learning AI revolution, um, when it hasbeen shown and proven that AI can be
extremely accurate, um, especially thenew novel, uh, AI algorithms, these deep
(17:15):
learning convolutional neural networks,uh, as they, as they call it a bit.
The GPU, so graphical processingunits, uh, that are normally used
for gaming, but now are used for,uh, for AI, um, the combination of
that about 11 years ago, and a proventheir proof for image processing.
Um, is I think the hallmark of, ofthe starting, uh, of, of this process
(17:39):
of AI actually proving itself.
Um, interestingly enough, the technologyor sort of the, uh, the methods that
we use now were already in the sixties.
Uh, but, uh, we're not used becauseof computing power wasn't there.
So, so the development and areal honing of the technology
was not possible at a time.
But now, uh, with, uh, with good computingcapability, that, that became a reality.
(18:03):
And since that, I think, uh, you know,like even 2016, so basically three, four
years after the event of, of the successof deep learning, uh, we've seen that,
uh, it, it, it is becoming better at tasksvery specific, well-defined tasks, it can
get better than, than, than humans can.
(18:23):
Um, but most important thingis, uh, that we do see that
it, it, it acts differently.
And I think.
Seeing that, uh, both the strengths andthe weaknesses, uh, of it is, and, and
the more and more precise understandingof that over the years is what I
think is propelling the field forward.
So I do believe that right now,it's not actually what algorithm
(18:48):
do we use is really the question.
The question is how do we use it?
Like what input do we give thealgorithms in terms of not just
for training, but on every day.
Um, uh, data, like do we want topre process what we give it or
do we not want to pre process?
I think people have experiencewith that with chat GPT.
If you have a different prompt,you get a very different answer.
Uh, but also understanding the limitationsof, okay, which part do I believe?
(19:13):
And which part do I not believe for,uh, for the AI, um, And, and I think
what is becoming more and more clearis that in critical use cases like
medicine or security or some, someother aspects, uh, when my life and
death questions are, are really, reallyasked, um, uh, how do we like assessing
(19:36):
the reliability of the software?
So, so basically monitoring.
Um, uh, better, better, I can, shouldstill trust the AI or has it learned
something that I, I did not intendit to, to learn, I think those are
the new questions, so it's not reallythe algorithm itself, but everything
around it, like how do we integratethat into our, into everyday usage
(19:56):
for real world usage and for medical,I think that's extremely critical.
That's where the, that's where thefront of the field is sort of like
the front lines, um, Um, to make surethat the software works when we bring
it to a new hospital for breast cancerscreening specifically, and make sure
that it works in the second month andthe third month and the sixth month
(20:18):
and like the second year, not just onthe first week when we installed it.
Um, and I think that intense monitoringand the capabilities, the, the
scientific technical capabilitiesaround that is where I see.
Uh, that the most important nextsteps, uh, need to be taken.
And that's, that's exactlywhere we put our, we're putting
our efforts in right now.
Ed (20:40):
Excellent.
I mean, it's, it seems pretty clear.
If you've, I'm pretty confidentabout the, the technology.
So Sarah, I kind of wanted to turnto you for a second since, since
joining Chiron, like what are someof the key initiatives or strategies?
You've implemented that havehelped shape the company's
direction since you've joined.
Sarah (20:57):
Well, we've talked a little bit about the importance of the
clinical trials and the studies andthe academic partners and the KOLs.
But one of the things I'm most proudof, actually, that we've done is
really brought the patient voice in.
Throughout everything we've done, wehave a patient advisory board, incredible
women and one man who, um, advisesfrom everything about what we call
the product, how we invite women toscreening, how we think about describing
(21:21):
AI to women, uh, in the context oftheir screening and the insights and
the joy, the pure joy, actually, ofjust getting together with these women.
Um, and their excitementabout what we're doing.
I think that's been, for me,one of the greatest privileges.
And then I would say the otherthing is, um, it's been really
humbling, um, the amount of coverage,press coverage, we've gotten.
(21:45):
And I think that's something that, youknow, Um, you know, obviously the clinical
trial results are very, very exciting,and we have a whole bunch more coming
out now, and I can't remember actuallyever being so excited professionally to
share this news with the world when wecan, um, but it's been very gratifying
that this has been picked up, that peopleare starting to understand the importance
of AI and technology in cancer and what,what this kind of technology can do.
(22:08):
So I think, um, the fact that the wordis spreading and that the pickup, the
media pickup has been all over theworld has been very, very encouraging.
And I think lays a foundation for what wereally want to do in our ambition, which
is to make sure that breast screening isimproved for women all over the world.
And one of the things that we haven'ttalked about a lot, which is really
(22:28):
important in AI is eliminating bias.
And in healthcare, that meansmaking sure that Mia works for
every woman, doesn't matter whereshe comes from or her ethnicity.
And, you know, one of the studies thatI can sort of hint at is a big study
that proves how Mia works well for womenfrom different ethnicities and doesn't
increase health healthcare inequality.
(22:48):
So, you know, I, I can't claimany credit for any of that work.
But it's just this amazing teamof scientists and AI specialists.
Um, the results, it's just, It'sunbelievably gratifying, and
no matter what happens next,I shall be proud of this work,
you know, till the end of my days.
Ed (23:06):
That's fantastic.
I mean, I'm just thinking, like, thepatient and the doctor, they just
want to get to the right answer.
I mean, are most patients evenaware of the technology that
got them to the right answer?
And, because to them, there's alreadybigger issues to be concerned with, but
are they even aware of how they got there?
Yeah.
(23:26):
Well, that's a really good question.
Generally not, you know, when you haveany kind of tests, the doctor doesn't
say, Hey, we use this kind of CATscan or this kind of software or the
processing software or any of that.
And in fact, that's the women Iwas telling you about earlier.
That was, that was some of their feedback.
They're like, we don't need toknow that AI has read my scan.
As long as the doctor's makingthe decision, um, you know, we
don't need to know the details.
And I thought that was actually avery interesting insight for us.
(23:49):
Um, the, the, the, the only occasionreally when women explicitly know is
during the studies, and obviously thenthey're informed that Mia is going to
be reading their scan in addition todoctor's and just, you know, giving
them the option to, to opt out if theydon't want to participate in the study.
So in that case, you know, there's fulltransparency and a lot of information
given to women about, you know, what'sgoing to happen regarding their scan.
(24:09):
And then, you know, when they do getthe results, if Mia has picked up a
cancer, they have some context andthe doctor will explain then in that
case that Mia flagged something thatperhaps they hadn't seen originally.
Um, and that's what they want,you know, that, and at that point
then they would do a biopsy.
And I'm just, I'm, I'm just love theenthusiasm that with which you, you
speak of it because like, you know,and everybody at, at Chiron knows, you
(24:31):
know, and I, I can tell that that's,you know, what's, what's important
to you guys that they're getting, youknow, the answers that, that, that
they need to get one way or another.
Um, which kind of brings me to likefuture goals and expansion, you know,
looking ahead, you know, what are thenext big goals for, for Chiron Medical?
Are there any plans to expand?
Expand, uh, me or any otherAI solution to other types of
(24:51):
cancer screenings or diagnostics.
Peter (24:55):
Um, our, our main focus is, uh, more and more becoming the United States.
Um, we had some great results in the UKand, and, and in Europe, and thankfully
we can, um, count the number of lives thatwe impacted, but specifically in the UK.
Kind of at an, at an impasse.
(25:15):
Um, and the UK is forming its strategyUK and National Health Services,
um, performing their, their, um,approaches on how to assess ai.
And that's unfortunately preventing,uh, uh, timely, timely adoption.
Uh, but we are seeing, uh, interest,uh, and, and quite a bit of momentum
in, in, in the us uh, Bitcoin.
(25:37):
Speak about those very, very openly yet.
Um, but, uh, I think the in terms ofthe impact that we can have as I think
we can impact about 2, 000 women andtheir, and their families who, who would
probably have a late cancer detectionwithout the software and with the
(25:57):
software, uh, most, most likely that wecan, we can find those cancers earlier.
And in the U.
S.
that number is actually probably over40, 000, uh, if we extrapolate some
of the results that we previously had.
So if you think about.
40, 000 women and their families, um,impacted across the U S every year.
That's, that's a prettysignificant number.
(26:18):
That's very exciting to work towards.
Um, and, uh, and I think that's just thecurrent state of the, of the technology.
Uh, if we, uh, improve the performanceeven further and, uh, the integrations
and, and, and beyond, then, you know,that could be significantly improved,
uh, even beyond, so if you think about.
(26:39):
Can, can you have a more exciting pieceof software technology that you just bring
into a hospital, install it, and prettymuch from next day, next week, you, you
start, um, um, helping doctors save lives.
You can actually measure it and, uh, Um,it's just such an exciting and scalable
thing to do, uh, in terms of human impact.
(27:02):
I can't imagine any, uh, uh, any,any more exciting thing to work on.
Ed (27:09):
And I can't help myself.
And you could totally pass on thisquestion if you want, but we're in.
You know, full scale electionseason here in the U.
S.
Is there any any like non technical nonmedical just hurdles that need to need
need to be reached, you know, in termsof regulation or anything like that?
Does it?
Does it?
You know, does any one geography lenditself to earlier or more more thorough
(27:35):
adoption of this technology somehow?
Sarah (27:40):
I think it requires, I think it's a really interesting question.
And what we've seen is that there needsto be a new level of understanding
among our politicians about both theopportunity and the risks with AI.
And right now you hear actuallyquite a lot about about the risks
and the dangers, which are real.
Um, but less about the opportunities andhow we as quickly as possible leverage the
power of AI and what Peter was referringto earlier in the, in the NHS is that,
(28:03):
you know, we've developed the technologyproven that it works for women all over
the UK, but the policy hasn't caught up.
So what I would say to, um, MadamPresident, Kamala Harris, is to make that
a core part of her administration andto, um, and to, or whoever comes to the
White House, uh, to make it a core partof the administration and making sure
(28:24):
that they have task forces that reallyunderstand the opportunities, but also
how you can Accelerate the adoption of thetechnologies like chirons because we're
solving really, really important problems,not just in terms of cancer detection,
but saving physicians a lot of time.
For example, in the UK, you know, we'reseeing that, um, we can save in it.
(28:45):
The context is that in the UK to whattwo doctors read every mammogram,
but with me, we can save that secondreader, something like 80 percent of
the, of the time required to read.
So there's a lot of economic health,economic savings, doctors, time savings.
Inpatient livesaving.
So I think, you know, getting those peopleinto office, into political office that
(29:06):
understand AI, the opportunities of riskand how to deploy and monitor it safely.
And then, you know, the benefitswill be much more rapidly felt
by everybody in wherever it is,the US, UK, Canada, and hopefully
sometime soon, the rest of the world.
Ed (29:23):
Innovation heroes is proudly produced by S.
H.
I.
A leading global I.
T.
Solutions provider dedicated to makinglife simpler for business and I.
T.
Professionals for over 30 years.
We've bridged the gap between digitalstrategy and execution, delivering
cutting edge technology and servicesto thousands of customers worldwide.
Whether you're navigating the cloudor Strengthening cybersecurity
(29:45):
or transforming your workplace.
SHI is here to help.
To learn more, visit hi.com/solve.
What's next?
That's s hi.com/solve.
What's next To embrace the Futurewith stress-free, scalable solutions
that you and your people will love?
Yeah, I think you, you vary.
(30:07):
eloquently said.
What we we talk about on innovationhere is a lot is that lawmakers tend
to, uh, not exactly early adopterswhen it comes to new technology or
the regulations that come with it.
So I was just kind of curious asto as to what what you were facing.
But that was the perfect answer,which kind of brings me to, um, just
the advice for aspiring innovators.
(30:29):
Um, What advice would, would each ofyou give, Peter, we'll start with you.
What advice would you give toinnovators who are looking to leverage
technology to tackle complex healthchallenges like cancer or, or, or, or
any, any others that they may face?
Peter (30:45):
Yeah.
I think the list is extremely long on whatneeds to be, uh, What one needs to get
right in order for innovation to be tobe brought, especially into healthcare.
Um, you know, there arelots of generalities.
A lot of methodology has been developed,uh, over the last decade for innovation,
(31:06):
like, um, lean, um, uh, methodologieslike lean startup and, and, um, uh,
being able to have strong opinions,weakly held, uh, having an experimental
mindset, learning as fast as possible.
staying, um, scrappy and lean and,and, and, um, uh, uh, sort of, uh,
(31:29):
conservative on, on, on, on, on expenses.
I was thinking the current, current,uh, um, economic, uh, Uh, environment
is very, very important, but I think,I think maybe the biggest advice above
it all is make sure that when they'reinnovating, we actually like our, our
(31:49):
market and, and those we innovatingfor, and, and we solve a problem.
Um, I, I see a lot, um, that apiece of exciting technology is.
Someone's excited about it and then theyare trying to just bring in for whatever
problem they can, they can think of.
But rather than focusing on solvinga problem that someone has, um, a
(32:12):
genuine problem and really makingthat, uh, making the problem go away.
It's more about obsession on,on the piece of technology.
And I think for AI, it's evenmore of a, of a, even high, higher
frequency than anything else.
So, um, yeah.
So I think focusing on, on, on problems,genuine problems to be, to be solved
that are worth solving, I think wouldprobably be the number one thing
(32:33):
that gets people, people to success.
Um, so it would be my biggestrecommendation probably.
Ed (32:39):
Sarah, how about you?
Sarah (32:40):
Yeah, I'm just going to add to that.
I think, you know, solve meaningfulproblems because the kinds of, um,
work that we've been doing at Curran isreally difficult and really demanding.
And we've had lots of sleeplessnights and scary moments, but, um,
it's Ultimately incredibly rewarding.
So I would say work on abig, important problem.
That's meaningful to you as Peter said,but then I would say, just never give up.
(33:01):
I think I've been lucky to work withsome amazing innovators, including the
original Apple team and the differencebetween those people who succeed and
those who fail are the people whojust keep going and maybe it doesn't
happen in this current iteration.
Maybe it happens in the nextiteration or maybe even the next one.
But if it's something you reallycare deeply about, that sort of keeps
you You know, makes you want to keeptrying and keep at it, just keep
(33:24):
going, because ultimately at the endof the day, that's what distinguishes
those who succeed from those who fail.
Ed (33:29):
And just listening to both of you talk in terms of collaboration,
you just talk about it seamlessly,whether it's within your own team
or whether it's with literally theteams that are out on the field on the
front lines using your technologies.
I mean, that collaboration's just got tobe there and you speak so easily to it.
So it's just wanted tocompliment you both.
It sounds like it sounds likea great environment, both for.
(33:50):
You know, the, the work that youdo, but also, you know, it, it
is a workplace, you know, andit, it sounds, it sounds awesome.
So kudos to you both for that.
Sarah (33:58):
It is.
Ed (33:59):
Um, I just wanted to say, you know, thank you so much for being with me here
today, uh, for listeners interested inSarah, maybe you could say for listeners
interested in, in learning more ormaybe want to reach out to you directly
or find out more about the company.
Uh, what's the best place to find you.
Sarah (34:17):
Yeah, well, you can, um, our website and, um, send email
inquiries via a website or anyone's,you know, feel free to reach out to
me directly, I'm sarah at chironmed.
com.
So always happy to speak to people whoare interested in the field or who just
want to talk about startups or an idea.
As Hillary Clinton said, it takesa village and we've been very lucky
to work with some of the greats.
And as you said, we have our own teamthat's amazing, but it's also nice to
(34:41):
support other people with their ideas.
Ed (34:42):
Excellent.
Um, so in closing, just wanted to saytoday's discussion has highlighted the
incredible potential of AI to changethe face of healthcare as we know it.
It's truly amazing to think that AIcan be leveraged, not just to detect
cancer earlier, but To make it amanageable disease, the little C,
as you said earlier, Sarah, um, andas we move forward, Cari Medical's
(35:05):
work inspires us to think about howwe can use technology to solve the
most pressing issues in health care.
A big thank you to Peter andSarah for joining me today.
Until next time, keep innovating.
On a on a day to day basis, the whole development cycle of AI
is very, very complicated becauseevery time you need to figure
out, okay, is it good enough?
Is it good enough?
What element do I need to improve on it?
But even that test of whetherit's good enough is hard.
Ed (00:13):
In the world of technology, heroes are everywhere.
They're overcoming disruption, deliveringsustainable outcomes, and fearlessly
forging the future to solve what's next.
Join me.
Ed McNamara, as we meet the peopleand businesses driving change in
our constantly disrupted world.
This is Innovation Heroes, apodcast brought to you by SHI.
(00:39):
Think about the everyday ways AI haschanged our lives, from turning speech
into text to managing our daily schedules.
But in today's episode, we're going tosee how AI is being used to save lives.
I'm joined by Peter Kecskamaty.
The CEO and co founder ofChiron Medical and Sarah Karush,
their Chief Strategy Officer.
(00:59):
They're here to share their journeyof how Chiron Medical's talented deep
learning team developed an AI technologyto improve breast cancer screenings, so
that patients can experience improvedhealth outcomes through earlier detection.
We'll dive into the technology, thereal world impact it's having, and what
the future holds for AI in medicine.
(01:21):
Peter and Sarah, welcometo Innovation Heroes.
Sarah (01:23):
Thank you so much, Ed.
Peter (01:24):
Very nice.
Ed (01:27):
And thanks for being here.
Peter, can you share a bit aboutyour background and what inspired
you to co found Chiron Medical?
Peter (01:36):
Yes.
And thanks again for having us.
Um, I had a pretty longjourney towards this.
Um, actually, um, my backgroundis technical AI, uh, software
engineering, computer science, uh,high performance computing and stats,
um, and, and population genetics.
(01:57):
But I think the most important, uh,Part of the background for this is
my, my mom's a radiologist and, uh,I actually spent most of my childhood
sort of like after school hours, etcetera, uh, in radiology departments.
And I would say where Chiron comesfrom is the recognition that AI and
(02:20):
generally computer science and AIare extremely powerful in producing
tools for, For doctors and also torecognition that in health care, it's
not just the patients who suffer, buta lot of times the doctors as well.
You often see the doctor having extremelyunsuitable software that they need to use.
(02:42):
Uh, at their work sort of all day andthey get out of the hospital and they
have the most sophisticated, uh, um,software and algorithms recommending, um,
various movies or restaurants to go to.
So I think what we need to doand what Karen comes from is.
bringing the bestpossible tools to doctors.
(03:04):
And I do believe that AI is one of themost important piece of technologies to,
uh, to produce that tool, those tools.
And it, since doctors are helping patientsand, and are absolutely on the front
lines of saving lives this way, I thinkone of the, just one of the best ways.
Uh, innovation, especiallytech, uh, software engineering
can, uh, can, can save lives.
Ed (03:27):
And Peter, we're one question in, and I'm going, you know, somewhat off script
already, but it just, it just brought tomind in terms of like all of the trainings
and the years of practice that doctorshave and radiologists, I mean, how much
of that is computer training and how bigof a gap could there be from knowledge of,
you know, perfecting their craft in termsof treating patients the way that they
do versus the technology that they use?
(03:49):
So is there one?
Peter (03:51):
Um, I think it depends on the on the field for sure.
Um, and as I see, radiology has alwaysbeen on the forefront of adopting
really somewhat esoteric but extremelyfrontline technologies like if you
think about the X ray or and thencomputer tomography and now MRI, which
(04:13):
is extremely high level of technology.
Um, then, then digitization in thehospital, a lot of, uh, that was
driven in, uh, by, by radiologists.
And I do believe AI is one of the, um,like radio is also one of the most natural
areas to, to adopt, uh, AI, especiallyimage, image recognition systems.
Ed (04:37):
Excellent.
And Sarah, you've had an amazingcareer as a journalist and a filmmaker.
You know, what brought you to ChironMedical and, and how do you define
your role as chief strategy officer?
Sarah (04:49):
That's a good question.
So, um, I started out in media, asyou said, in filmmaking and I was
invited to go and make a film ofa company called General Magic in
Silicon Valley many, many years ago.
And that was my introductionto the world of tech.
But I very quickly became interested inhow you use technology and software and
hardware to bring important ideas to life.
And I got, I was very privileged towork with some incredible people,
(05:12):
some of the big innovators of our age.
And, um, you know, from them, I learneda tremendous amount and thought this
is something that I want to, Spend mylife doing and in particular I'm very
interested in how technology can beapplied in healthcare and in the field
of cancer And you know your earlierpoint Doctors when they when you have
a scan when you have a suspicious massor there's something wrong when you
(05:32):
have a scan You sort of expect thatthe doctor will look at that and go.
Oh, this is a very specific kind ofcancer It's a two millimeter something
coma Um, but actually, cancer detectionsare really difficult tasks, even
for the most experienced doctors.
And that's where AI canreally add a lot of value.
So when I was introduced to Chiron,and you know, this incredible team
of machine learning world experts,and Peter, who really understood the
(05:54):
pain points of doctors, it seemed tome that that was the right formula.
for creating a technology like the onewe've created called NEAR to really
help doctors with that very difficulttask of, of detecting cancer earlier.
And as it relates to my, as it relatesto your question about Chief Strategy
Officer, the way I see that is really,you know, Peter leads on strategy,
but I'm sort of his right arm.
(06:16):
Um, and, and we look together forareas of the business where we can
have the most meaningful growth.
Um, And depending on the stage of thebusiness, that can be different stages,
but it's things like for us creatingacademic partnerships, working to
develop relationships with organizationslike the NHS or big partners in the
US, so that we can test and developthe technology and make sure that it's
working for, um, you know, the womenthat we are hoping to, um, to help.
Ed (06:42):
And Sarah, I think that, you know, in terms of developing partnerships, I was
reading and I'm consulting my notes now.
Um, your company has experts and engineersthat your partner with that were from
Oxford, Cambridge, um, McGill, uh, KTHUniversity, uh, sorry, KTH University
College, London Imperial College, uh,from a recruiting perspective, I'd
(07:05):
imagine that makes a pretty attractiveplace for people on the cutting edge
to, uh, to want to come to work.
Is that fair?
Sarah (07:12):
Yeah.
And I think that, um, you know, we'vedefinitely attracted incredible talent.
Uh, and the most special thing though,about the people that we work with is not
the illustrious universities they wentto, although they did well, the incredible
science that they develop or the AIthat they, you know, develop and test.
It's really the, the mission.
And, you know, we have a teamthat's so dedicated to this
(07:32):
mission of finding cancer earlier.
And, you know, as you, as youprobably know, you know, if
you can't, if you detect cancerearlier in our case, breast cancer.
That can mean an entirelydifferent prognosis.
So you're not looking at necessarilyradiation or mastectomies.
You're potentially looking justat lumpectomies and maybe, you
know, maybe some radiation,but possibly not chemotherapy.
So, you know, we're hearing now fromthe women whose cancers were picked up
(07:55):
earlier by Mia, which is our technology.
And it's just the greatest privilegeof our lives to know that there are
women, many women now walking outin the world whose cancer was picked
up earlier by our technology and whohave a better prognosis as a result.
Ed (08:08):
Yeah, I mean, maybe it's time to talk about, um, the, that technology.
Um, you know, Chiron's, uh, medicalAI technology, MIA, um, as you
said, Sarah, it's quite innovative.
Um, Peter, maybe, could you explainwhat sets your AI driven diagnostics
apart in the fight against cancer?
Peter (08:25):
Yeah, so first of all, um, our, um, Number one flagship, uh,
product or AI service is it's calledMia is for breast cancer screening.
Uh, very specifically, um, it isbased on, um, image processing or
AI, uh, the AI, uh, the arm or the.
(08:47):
Um, the part of AI or, uh, that, uh,the type of AI that is, uh, working with
images, image inputs, uh, so basicallythe radiologists are producing the images
and, uh, or the, uh, the hospitals areproducing the images that radiologists
look at, the AI also looks at it.
Um, and pretty much attemptsto make the same decision.
(09:08):
Um, however, the AI hasvery different capabilities.
So the AI is infinitely diligent.
It looks at every single pixelby the humans don't do that.
Um, but the AI doesn't have thesame strategic overview or a
contextual background understandingthat the trained doctor does.
So what we optimized for is twofold.
(09:28):
Um, uh, one.
for the AI to be useful.
So really being ascomplimentary as possible.
So additive, try to do bring allthe capabilities that that is hard
for humans, but not try to replicatethe capability that humans have.
So we're not trying to imitate howdoctors make the decisions, but,
but bring in extra informationthat is, for instance, not.
(09:50):
visible to the human eye or sort ofnot accessible to the brain, the visual
cortex of how humans process images.
Uh, so, so that's, that's how wecombine the two capabilities and focus
on that is one of the ways that Ithink our approach is very different.
And, and that is through Um, uh, both interms of what the AI looks at information
(10:11):
it processes and also how it collaborateswith the doctors and how it collaborates
with the doctors is through variousworkflows and it's a different in the U.
S.
than, than, than in Europe.
Um, so that's one aspect.
And the other aspect is for, forthe AI to be working and, and, you
know, like I'm sure everyone thinksand they develop a piece of AI,
they, they believe it, it works.
(10:32):
But actually.
Assessing whether AI works ornot is extremely complicated.
Uh, it requires statistical thinkingand, and it's a completely new paradigm.
Like we never had a kind of tool.
Humanity never had a tool like thisbefore, when it's hard to assess
whether, whether the tool works or not.
Like if you have a piece of hardware,like a knife, you could figure
(10:53):
out whether that cuts or not.
Uh, you have a piece of software,you know what it's supposed to do
and you test it and then, and then,you know, whether it works or not
for AI, it's extremely hard because.
It is strained on that set of images,for instance, a set of inputs.
And then it's assessed on similar,but different, but fundamentally
different inputs and that's when, if,if it can perform well on, on those
(11:16):
unseen cases, that's when it works.
But what if the case is too unseen?
Like it, it sees it not like a.
breast mammogram, but like a chest xray, then of course it shouldn't work.
But, but what if it, um, it is amammogram, uh, for, for breast cancer
screening, but it's not for like alady from Europe, but a lady from
(11:37):
Southeast Asia, like it, you wouldexpect it still should work, right?
So that exact level of specialism andgeneralizability needs to be gotten right.
As it happens, it's fundamentallyimpossible to prove, extremely hard
to test, and takes a lot of effort.
And, uh, but it, but that kind oftesting has to be a fundamental
(12:00):
part of how AI is being developed.
So you can imagine that it's, on a dayto day basis, the whole development
cycle of AI is very, very complicated,because every time you need to
figure out, okay, is it good enough?
Is it good enough?
What element do I need to improve on it?
But even that test of whetherit's good enough is hard.
So, so that is a very involvedtechnical and, and, uh,
(12:21):
statistical and scientific task.
Um, and, and taking that seriously, thattask to a level that I haven't seen from
others is what I think also sets us apart.
So basically being useful, being, be,uh, bringing AI capability beyond human
capability and, uh, making sure it works.
Ed (12:42):
And I'm just putting myself in the shoes of the patient for a second
to put myself in the shoes of anybodythat has to make a phone call to
a patient that could be, you know,either the best or worst phone call
that they might ever receive, right?
I mean, you really don'twant to be chasing.
I'll say chasing ghosts.
That's that's a layman'sterms or something.
But you you really wantto be sure of the data.
(13:04):
And it's just such anamazingly sensitive topic.
And, you know, Sarah, the waythat you described the company
before, it just sounds like.
Pretty much everyone at Chiron is going tohave that kind of sensitivity because of,
um, at the end of the day, we're talkingtechnology, but there's real, um, you
know, life changing conversations that arehappening as a result of the technology.
Is that fair?
Sarah (13:26):
Yeah, and I can tell you about one of those conversations.
I think the first thing that'simportant to note is that Mia never
makes a decision in isolation.
So Mia will make a recommendationas to whether a woman should be
recalled for further testing or not.
But ultimately, it's the doctorsthat decide, and that's very much our
worldview, that Mia and AI should becomplementary to doctors decision making.
(13:47):
So that's the first thing.
And then in terms of the actualconversation itself, um, we have
one wonderful patient who, um, wasinvolved in a study we did in Scotland.
And when she was told about her verysmall mass, uh, in, in her breast,
that was in fact a cancer post biopsy,they knew that it was a cancer.
She said, Oh, she said,I don't have the big C.
(14:10):
I have a little C.
And so that conversation is never an easyconversation, but we want more of those
little c conversations And you know,my my my ultimate dream is that um, you
know in five ten years That when somebodycalls you to say, hey, you know, we've
done further testing, we think you've gotcancer, or we know that you have cancer.
I want people to have that feelingof like, oh, that's a bummer, but
(14:32):
it's not the, am I going to die?
And I think, you know, there'sbeen so much development, so much,
um, advance in therapeutics, uh,but not so much in diagnostics.
And I think that's where AI isplaying a really critical role.
And a combination of those two thingsis really going to transform cancer and
the very meaning of the word, I think.
Ed (14:51):
And Sarah, earlier you mentioned, you know, early detection and
the impact in early detection.
How has Kiran's technology improvedthe accuracy and efficiency of
breast cancer screening in realworld health care settings?
Sarah (15:07):
Well, that's been the part that's so unbelievably exciting because
we, we started, we did our firstprospective real world study in Hungary.
And we saw extraordinary results there.
So, you know, up to 15 percent ofincreased, uh, cancer detection.
And that was, you know, that was sogroundbreaking that it was reported on
the front page of the New York Times.
(15:27):
You know, one of the first real worldexamples of AI actually being used
in clinical practice and having,you know, extraordinarily good
results beyond our expectations.
Those results we've reproduced in theUK, they were published recently in
the BBC and again made global headlinesbecause we're just beginning to see that.
you know, our technology and technologieslike ours really will have huge,
(15:48):
make huge differences in the accuracyand the, um, and the ability of
doctors to detect cancer earlier.
Um, you know, and, and in many casesyears earlier, and we're projecting that
in breast cancer screening, 20 percentof cancers are missed, uh, routinely
because it's such a difficult task.
It's not like doctorsaren't doing their job.
It's just very difficult tosee and detect breast cancer.
(16:10):
Um, and we, we think, you know, basedon the stats we're seeing now, that MEER
will halve that rate of missed cancers.
And, and what that translatesinto is 2, 000 women a year in
the UK, 40, 000 in, in the U.
S.
These are, these are big numbers and verymeaningful impact that we hope to have.
Ed (16:27):
And I guess from a, from a technology perspective, Um, Peter, it's been just
under 20 years since you studied AIand cybermetrics at the University of
Reading and just under nine since youwere co founder of Chiron Medical.
You know, what's changed most about theAI technology since you first started?
You know, not, not just, you know,your, your development of it, but
(16:47):
the actual, like, capabilitiesin, let's say, nine years since,
since you founded the company.
Peter (16:53):
Yeah, I think Actually, the last, I would say 11 years is the big deep
learning AI revolution, um, when it hasbeen shown and proven that AI can be
extremely accurate, um, especially thenew novel, uh, AI algorithms, these deep
(17:15):
learning convolutional neural networks,uh, as they, as they call it a bit.
The GPU, so graphical processingunits, uh, that are normally used
for gaming, but now are used for,uh, for AI, um, the combination of
that about 11 years ago, and a proventheir proof for image processing.
Um, is I think the hallmark of, ofthe starting, uh, of, of this process
(17:39):
of AI actually proving itself.
Um, interestingly enough, the technologyor sort of the, uh, the methods that
we use now were already in the sixties.
Uh, but, uh, we're not used becauseof computing power wasn't there.
So, so the development and areal honing of the technology
was not possible at a time.
But now, uh, with, uh, with good computingcapability, that, that became a reality.
(18:03):
And since that, I think, uh, you know,like even 2016, so basically three, four
years after the event of, of the successof deep learning, uh, we've seen that,
uh, it, it, it is becoming better at tasksvery specific, well-defined tasks, it can
get better than, than, than humans can.
(18:23):
Um, but most important thingis, uh, that we do see that
it, it, it acts differently.
And I think.
Seeing that, uh, both the strengths andthe weaknesses, uh, of it is, and, and
the more and more precise understandingof that over the years is what I
think is propelling the field forward.
So I do believe that right now,it's not actually what algorithm
(18:48):
do we use is really the question.
The question is how do we use it?
Like what input do we give thealgorithms in terms of not just
for training, but on every day.
Um, uh, data, like do we want topre process what we give it or
do we not want to pre process?
I think people have experiencewith that with chat GPT.
If you have a different prompt,you get a very different answer.
Uh, but also understanding the limitationsof, okay, which part do I believe?
(19:13):
And which part do I not believe for,uh, for the AI, um, And, and I think
what is becoming more and more clearis that in critical use cases like
medicine or security or some, someother aspects, uh, when my life and
death questions are, are really, reallyasked, um, uh, how do we like assessing
(19:36):
the reliability of the software?
So, so basically monitoring.
Um, uh, better, better, I can, shouldstill trust the AI or has it learned
something that I, I did not intendit to, to learn, I think those are
the new questions, so it's not reallythe algorithm itself, but everything
around it, like how do we integratethat into our, into everyday usage
(19:56):
for real world usage and for medical,I think that's extremely critical.
That's where the, that's where thefront of the field is sort of like
the front lines, um, Um, to make surethat the software works when we bring
it to a new hospital for breast cancerscreening specifically, and make sure
that it works in the second month andthe third month and the sixth month
(20:18):
and like the second year, not just onthe first week when we installed it.
Um, and I think that intense monitoringand the capabilities, the, the
scientific technical capabilitiesaround that is where I see.
Uh, that the most important nextsteps, uh, need to be taken.
And that's, that's exactlywhere we put our, we're putting
our efforts in right now.
Ed (20:40):
Excellent.
I mean, it's, it seems pretty clear.
If you've, I'm pretty confidentabout the, the technology.
So Sarah, I kind of wanted to turnto you for a second since, since
joining Chiron, like what are someof the key initiatives or strategies?
You've implemented that havehelped shape the company's
direction since you've joined.
Sarah (20:57):
Well, we've talked a little bit about the importance of the
clinical trials and the studies andthe academic partners and the KOLs.
But one of the things I'm most proudof, actually, that we've done is
really brought the patient voice in.
Throughout everything we've done, wehave a patient advisory board, incredible
women and one man who, um, advisesfrom everything about what we call
the product, how we invite women toscreening, how we think about describing
(21:21):
AI to women, uh, in the context oftheir screening and the insights and
the joy, the pure joy, actually, ofjust getting together with these women.
Um, and their excitementabout what we're doing.
I think that's been, for me,one of the greatest privileges.
And then I would say the otherthing is, um, it's been really
humbling, um, the amount of coverage,press coverage, we've gotten.
(21:45):
And I think that's something that, youknow, Um, you know, obviously the clinical
trial results are very, very exciting,and we have a whole bunch more coming
out now, and I can't remember actuallyever being so excited professionally to
share this news with the world when wecan, um, but it's been very gratifying
that this has been picked up, that peopleare starting to understand the importance
of AI and technology in cancer and what,what this kind of technology can do.
(22:08):
So I think, um, the fact that the wordis spreading and that the pickup, the
media pickup has been all over theworld has been very, very encouraging.
And I think lays a foundation for what wereally want to do in our ambition, which
is to make sure that breast screening isimproved for women all over the world.
And one of the things that we haven'ttalked about a lot, which is really
(22:28):
important in AI is eliminating bias.
And in healthcare, that meansmaking sure that Mia works for
every woman, doesn't matter whereshe comes from or her ethnicity.
And, you know, one of the studies thatI can sort of hint at is a big study
that proves how Mia works well for womenfrom different ethnicities and doesn't
increase health healthcare inequality.
(22:48):
So, you know, I, I can't claimany credit for any of that work.
But it's just this amazing teamof scientists and AI specialists.
Um, the results, it's just, It'sunbelievably gratifying, and
no matter what happens next,I shall be proud of this work,
you know, till the end of my days.
Ed (23:06):
That's fantastic.
I mean, I'm just thinking, like, thepatient and the doctor, they just
want to get to the right answer.
I mean, are most patients evenaware of the technology that
got them to the right answer?
And, because to them, there's alreadybigger issues to be concerned with, but
are they even aware of how they got there?
Yeah.
(23:26):
Well, that's a really good question.
Generally not, you know, when you haveany kind of tests, the doctor doesn't
say, Hey, we use this kind of CATscan or this kind of software or the
processing software or any of that.
And in fact, that's the women Iwas telling you about earlier.
That was, that was some of their feedback.
They're like, we don't need toknow that AI has read my scan.
As long as the doctor's makingthe decision, um, you know, we
don't need to know the details.
And I thought that was actually avery interesting insight for us.
(23:49):
Um, the, the, the, the only occasionreally when women explicitly know is
during the studies, and obviously thenthey're informed that Mia is going to
be reading their scan in addition todoctor's and just, you know, giving
them the option to, to opt out if theydon't want to participate in the study.
So in that case, you know, there's fulltransparency and a lot of information
given to women about, you know, what'sgoing to happen regarding their scan.
(24:09):
And then, you know, when they do getthe results, if Mia has picked up a
cancer, they have some context andthe doctor will explain then in that
case that Mia flagged something thatperhaps they hadn't seen originally.
Um, and that's what they want,you know, that, and at that point
then they would do a biopsy.
And I'm just, I'm, I'm just love theenthusiasm that with which you, you
speak of it because like, you know,and everybody at, at Chiron knows, you
(24:31):
know, and I, I can tell that that's,you know, what's, what's important
to you guys that they're getting, youknow, the answers that, that, that
they need to get one way or another.
Um, which kind of brings me to likefuture goals and expansion, you know,
looking ahead, you know, what are thenext big goals for, for Chiron Medical?
Are there any plans to expand?
Expand, uh, me or any otherAI solution to other types of
(24:51):
cancer screenings or diagnostics.
Peter (24:55):
Um, our, our main focus is, uh, more and more becoming the United States.
Um, we had some great results in the UKand, and, and in Europe, and thankfully
we can, um, count the number of lives thatwe impacted, but specifically in the UK.
Kind of at an, at an impasse.
(25:15):
Um, and the UK is forming its strategyUK and National Health Services,
um, performing their, their, um,approaches on how to assess ai.
And that's unfortunately preventing,uh, uh, timely, timely adoption.
Uh, but we are seeing, uh, interest,uh, and, and quite a bit of momentum
in, in, in the us uh, Bitcoin.
(25:37):
Speak about those very, very openly yet.
Um, but, uh, I think the in terms ofthe impact that we can have as I think
we can impact about 2, 000 women andtheir, and their families who, who would
probably have a late cancer detectionwithout the software and with the
(25:57):
software, uh, most, most likely that wecan, we can find those cancers earlier.
And in the U.
S.
that number is actually probably over40, 000, uh, if we extrapolate some
of the results that we previously had.
So if you think about.
40, 000 women and their families, um,impacted across the U S every year.
That's, that's a prettysignificant number.
(26:18):
That's very exciting to work towards.
Um, and, uh, and I think that's just thecurrent state of the, of the technology.
Uh, if we, uh, improve the performanceeven further and, uh, the integrations
and, and, and beyond, then, you know,that could be significantly improved,
uh, even beyond, so if you think about.
(26:39):
Can, can you have a more exciting pieceof software technology that you just bring
into a hospital, install it, and prettymuch from next day, next week, you, you
start, um, um, helping doctors save lives.
You can actually measure it and, uh, Um,it's just such an exciting and scalable
thing to do, uh, in terms of human impact.
(27:02):
I can't imagine any, uh, uh, any,any more exciting thing to work on.
Ed (27:09):
And I can't help myself.
And you could totally pass on thisquestion if you want, but we're in.
You know, full scale electionseason here in the U.
S.
Is there any any like non technical nonmedical just hurdles that need to need
need to be reached, you know, in termsof regulation or anything like that?
Does it?
Does it?
You know, does any one geography lenditself to earlier or more more thorough
(27:35):
adoption of this technology somehow?
Sarah (27:40):
I think it requires, I think it's a really interesting question.
And what we've seen is that there needsto be a new level of understanding
among our politicians about both theopportunity and the risks with AI.
And right now you hear actuallyquite a lot about about the risks
and the dangers, which are real.
Um, but less about the opportunities andhow we as quickly as possible leverage the
power of AI and what Peter was referringto earlier in the, in the NHS is that,
(28:03):
you know, we've developed the technologyproven that it works for women all over
the UK, but the policy hasn't caught up.
So what I would say to, um, MadamPresident, Kamala Harris, is to make that
a core part of her administration andto, um, and to, or whoever comes to the
White House, uh, to make it a core partof the administration and making sure
(28:24):
that they have task forces that reallyunderstand the opportunities, but also
how you can Accelerate the adoption of thetechnologies like chirons because we're
solving really, really important problems,not just in terms of cancer detection,
but saving physicians a lot of time.
For example, in the UK, you know, we'reseeing that, um, we can save in it.
(28:45):
The context is that in the UK to whattwo doctors read every mammogram,
but with me, we can save that secondreader, something like 80 percent of
the, of the time required to read.
So there's a lot of economic health,economic savings, doctors, time savings.
Inpatient livesaving.
So I think, you know, getting those peopleinto office, into political office that
(29:06):
understand AI, the opportunities of riskand how to deploy and monitor it safely.
And then, you know, the benefitswill be much more rapidly felt
by everybody in wherever it is,the US, UK, Canada, and hopefully
sometime soon, the rest of the world.
Ed (29:23):
Innovation heroes is proudly produced by S.
H.
I.
A leading global I.
T.
Solutions provider dedicated to makinglife simpler for business and I.
T.
Professionals for over 30 years.
We've bridged the gap between digitalstrategy and execution, delivering
cutting edge technology and servicesto thousands of customers worldwide.
Whether you're navigating the cloudor Strengthening cybersecurity
(29:45):
or transforming your workplace.
SHI is here to help.
To learn more, visit hi.com/solve.
What's next?
That's s hi.com/solve.
What's next To embrace the Futurewith stress-free, scalable solutions
that you and your people will love?
Yeah, I think you, you vary.
(30:07):
eloquently said.
What we we talk about on innovationhere is a lot is that lawmakers tend
to, uh, not exactly early adopterswhen it comes to new technology or
the regulations that come with it.
So I was just kind of curious asto as to what what you were facing.
But that was the perfect answer,which kind of brings me to, um, just
the advice for aspiring innovators.
(30:29):
Um, What advice would, would each ofyou give, Peter, we'll start with you.
What advice would you give toinnovators who are looking to leverage
technology to tackle complex healthchallenges like cancer or, or, or, or
any, any others that they may face?
Peter (30:45):
Yeah.
I think the list is extremely long on whatneeds to be, uh, What one needs to get
right in order for innovation to be tobe brought, especially into healthcare.
Um, you know, there arelots of generalities.
A lot of methodology has been developed,uh, over the last decade for innovation,
(31:06):
like, um, lean, um, uh, methodologieslike lean startup and, and, um, uh,
being able to have strong opinions,weakly held, uh, having an experimental
mindset, learning as fast as possible.
staying, um, scrappy and lean and,and, and, um, uh, uh, sort of, uh,
(31:29):
conservative on, on, on, on, on expenses.
I was thinking the current, current,uh, um, economic, uh, Uh, environment
is very, very important, but I think,I think maybe the biggest advice above
it all is make sure that when they'reinnovating, we actually like our, our
(31:49):
market and, and those we innovatingfor, and, and we solve a problem.
Um, I, I see a lot, um, that apiece of exciting technology is.
Someone's excited about it and then theyare trying to just bring in for whatever
problem they can, they can think of.
But rather than focusing on solvinga problem that someone has, um, a
(32:12):
genuine problem and really makingthat, uh, making the problem go away.
It's more about obsession on,on the piece of technology.
And I think for AI, it's evenmore of a, of a, even high, higher
frequency than anything else.
So, um, yeah.
So I think focusing on, on, on problems,genuine problems to be, to be solved
that are worth solving, I think wouldprobably be the number one thing
(32:33):
that gets people, people to success.
Um, so it would be my biggestrecommendation probably.
Ed (32:39):
Sarah, how about you?
Sarah (32:40):
Yeah, I'm just going to add to that.
I think, you know, solve meaningfulproblems because the kinds of, um,
work that we've been doing at Curran isreally difficult and really demanding.
And we've had lots of sleeplessnights and scary moments, but, um,
it's Ultimately incredibly rewarding.
So I would say work on abig, important problem.
That's meaningful to you as Peter said,but then I would say, just never give up.
(33:01):
I think I've been lucky to work withsome amazing innovators, including the
original Apple team and the differencebetween those people who succeed and
those who fail are the people whojust keep going and maybe it doesn't
happen in this current iteration.
Maybe it happens in the nextiteration or maybe even the next one.
But if it's something you reallycare deeply about, that sort of keeps
you You know, makes you want to keeptrying and keep at it, just keep
(33:24):
going, because ultimately at the endof the day, that's what distinguishes
those who succeed from those who fail.
Ed (33:29):
And just listening to both of you talk in terms of collaboration,
you just talk about it seamlessly,whether it's within your own team
or whether it's with literally theteams that are out on the field on the
front lines using your technologies.
I mean, that collaboration's just got tobe there and you speak so easily to it.
So it's just wanted tocompliment you both.
It sounds like it sounds likea great environment, both for.
(33:50):
You know, the, the work that youdo, but also, you know, it, it
is a workplace, you know, andit, it sounds, it sounds awesome.
So kudos to you both for that.
Sarah (33:58):
It is.
Ed (33:59):
Um, I just wanted to say, you know, thank you so much for being with me here
today, uh, for listeners interested inSarah, maybe you could say for listeners
interested in, in learning more ormaybe want to reach out to you directly
or find out more about the company.
Uh, what's the best place to find you.
Sarah (34:17):
Yeah, well, you can, um, our website and, um, send email
inquiries via a website or anyone's,you know, feel free to reach out to
me directly, I'm sarah at chironmed.
com.
So always happy to speak to people whoare interested in the field or who just
want to talk about startups or an idea.
As Hillary Clinton said, it takesa village and we've been very lucky
to work with some of the greats.
And as you said, we have our own teamthat's amazing, but it's also nice to
(34:41):
support other people with their ideas.
Ed (34:42):
Excellent.
Um, so in closing, just wanted to saytoday's discussion has highlighted the
incredible potential of AI to changethe face of healthcare as we know it.
It's truly amazing to think that AIcan be leveraged, not just to detect
cancer earlier, but To make it amanageable disease, the little C,
as you said earlier, Sarah, um, andas we move forward, Cari Medical's
(35:05):
work inspires us to think about howwe can use technology to solve the
most pressing issues in health care.
A big thank you to Peter andSarah for joining me today.
Until next time, keep innovating.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.