July 14, 2025 • 20 mins
Interested in being a guest? Email us at admin@evankirstel.com
The race to develop life-saving medications has always been constrained by time, cost, and the challenge of gathering enough patient data. But what if we could create AI-powered replicas of trial participants that accurately predict how they would respond to treatment?
Chief Executive Officer Steve Herne from Unlearn.AI joins us to reveal how digital twins technology is revolutionizing clinical trials by creating virtual patient forecasts that run alongside actual participants. This breakthrough approach enables researchers to compare observed versus predicted outcomes, dramatically reducing statistical noise and allowing true treatment signals to emerge more clearly. The results are nothing short of remarkable—Alzheimer's trials with 33% smaller sample sizes while boosting statistical power by 10%, Parkinson's disease studies completed nearly 10 months faster than traditional methods, and rare disease research accelerated by months rather than years.
We dive deep into how regulatory bodies like the FDA are adapting to this AI revolution, with new frameworks ensuring these approaches meet rigorous scientific standards. Steve shares fascinating insights into how pharmaceutical companies are transitioning from reluctance to enthusiasm, increasingly reaching out to Unlearnat the earliest stages of protocol development. The vision? Making digital twins standard practice across all therapeutic areas, accelerating the "miracle of medicine" to patients worldwide.
Whether you're a healthcare professional, technology enthusiast, or simply curious about how AI is transforming medicine, this episode offers a fascinating glimpse into a future where digital innovation and clinical research converge to create better outcomes for patients. Subscribe now and join the conversation about technology's profound impact on healthcare's most challenging problems.
More at https://linktr.ee/EvanKirstel
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Hey everybody, fascinating discussion and
important discussion today, aswe talk about AI-powered
replicas, or digital twins, thatare revolutionizing how we run
clinical trials, and joined by atrue insider and thought leader
in the space from unlearnaiSteve, how are you?
Speaker 2 (00:20):
I'm doing very well.
Thank you for having me today.
Speaker 1 (00:23):
Well, thanks for being here.
I'm doing very well.
Thank you for having me today.
Well, thanks for being here.
Yes, Evan and Irma here fromAvira Health Really fascinated.
Maybe we'll start from thebeginning.
What's your background and howdid you land at unlearnai?
Speaker 2 (00:37):
Sure, no, absolutely.
So I'm coming up around 30years in the clinical trial
space.
I started my life out ofacademia, getting into clinical
trials on the operations sideand then spent a lot of my
career in the commercialmarketing and then for the last
10-15 years I've been spendingmore of my time on the clinical
technology side in supportingclinical trials and the kind of
(01:00):
advantages that technology canreally do to shape our clinical
trials going forward.
And that really fast forwardsto my journey over the last year
here at Unlearn AI, wheretechnology AI meets clinical
trials.
So for me it's an excitinghopefully near end of my career
where I can really takeeverything I've learned over the
last three decades in clinicaltrials and see if we can really
(01:23):
start to make a difference usingartificial intelligence and
technology out there in themarket.
Speaker 3 (01:29):
Fantastic.
So today we will be diving intothe world of digital twins and
for our listeners, who may notbe familiar or new to the term,
what exactly is a digital twin,and particularly in the context
of medicine?
Speaker 2 (01:48):
is a digital twin and particularly in the context of
medicine Sure, no, absolutely,to make it nice and simple and
very easy for the listeners, adigital twin is basically an
AI-generated forecast of apatient's clinical outcomes,
predominantly in our world,obviously in a clinical trial.
So that really helps usforecast how that patient will
progress with their diseasealong the life cycle of the
clinical trial.
(02:08):
So we're spending a lot oftechnology behind the scenes, a
lot of algorithmic activity toreally forecast that
longitudinal activity of thatindividual.
So you go into that clinicaltrial and your twin comes in
with you.
So we create that digital twinthat runs along with you
throughout the clinical trial.
Speaker 1 (02:29):
Amazing, and this has moved from the sort of lab into
the real world.
How are you actually applyingthese digital twins at Unlearn
to medical research and what arethe different use cases?
Speaker 2 (02:41):
Sure, and what are the different use cases?
Sure, so again, I think mostdigital twins we're really using
to enable us to compareobserved outcomes versus
predicted outcomes.
So in clinical terms, that letsus really look at the
variability, helping to reducethe noise so that we can really
see true signals in the data bylowering that variance.
(03:03):
That allows us to really startto really see true magnified
signals that people can thenmake much more effective what we
call, in clinical triallanguage, go-no-go decisions.
So people want to shorten thelength of a clinical trial as
quickly but as effectively asthey can.
So the quicker we get to thosego-no-go decisions ie, is the
(03:24):
therapeutic working well?
Are we seeing the response welike to those go-no-go decisions
?
Ie, is the therapeutic workingwell?
Are we seeing the response welike to see, then the happier
everybody is.
And we do that obviously byreducing sample size, which
basically means we're bringingthe number of patients on the
trial or the number ofvolunteers that are in the
clinical trial.
We bring that down because wecan obviously run a digital
trial sorry, a digital twinalong with a patient at the same
(03:45):
time.
So smaller sample sizes equalsfaster enrollment and therefore
we're boosting what we call thepower of the study, reducing the
error and making that trialmuch more effective and much
quicker.
Speaker 3 (03:57):
Wow, wow, that's right.
There are already greatbenefits of using digital twins
compared to just the traditionalclinical trial methods.
Can you elaborate a little bitmore?
Do you have advantage in termsof geographical area you can
cover or any other benefits tousing this approach versus
(04:17):
traditional clinical trials?
Speaker 2 (04:19):
So I think again, main main benefit again just for
everybody listening is reducing.
You know endpoints with fewerpatients and you know obviously
shorter durations.
That is definitely equals, youknow, digital twin period.
But from a geographyperspective, across the board
we're pretty agnostic in bothgeography and in therapeutic
area.
(04:39):
We're trying to, as we workwith companies and customers
right at the beginning of theclinical trial development
program, in analyzing theprotocols and understanding
different ways to work with them, we're obviously trying to
maximize how they can run thattrial as effective as possible.
So we have lots and lots andlots of data out there that we
obviously learn these modelsthrough so that we can really
(05:02):
help predict the future forthese trials.
So in particular areas we mightsay go into this geography or
look at this type of population.
We also might say examine thisendpoint along with this
endpoint because you will seemaximum benefit that we've seen
maybe as we've supported othertrials or done other work with
other customers.
So again, it's case by case oneach situation, but the main
(05:25):
main thrust obviously we'rereducing sample size and we're
decreasing clinical timelines.
That's amazing.
Speaker 1 (05:32):
And we all know the cost and time associated with
drug development.
It's hard to believe drugs getto market, given the challenges.
Can you walk us through howthey help reduce time and cost,
and can you quantify that at all?
Speaker 2 (05:48):
Sure, so we've spent a lot of time in the evolution
of Unlearn in what we call theneurodegenerative disease space,
or sometimes known as thecentral nervous system, cns.
So evolution over the lastseven, eight years in the
company we've spent a lot oftime predominantly in
Alzheimer's disease.
So again, this is public and onour website and publicly
available data, so I'm notsharing anything too
(06:10):
confidential today.
But as we managed anAlzheimer's drug with a customer
, we managed to reduce thesample size of up to 33% in that
particular trial and we boostedthe power over 10%.
So we went from just under 80%to over 90% in power and brought
the trial in at least fourmonths earlier than the original
(06:32):
plan date.
So that was a phase two trial.
These are slightly smaller thanthe larger phase threes but
seeing that type of activity ona phase two is very exciting.
To obviously get us into thatno-go decision for phase three
and some of the rare diseases orsome of the more rare CNS
profiles like ALS, we did see arecent trial that we ran with
(06:53):
one of our customers where wereduced our sample size by 18%
and cut those timelines byaround three to four months.
And the last one that we justactually not long ago finished
actually was in Parkinson'sdisease and again we actually
increased our sample size thereby about 33% and increased the
power by about 65%, cuttingabout nine, 10 months off the
(07:15):
entire journey on that one too.
So again, obviously the largertrials that we get into, the
larger cuts that we can make intime and also, hopefully, the
stronger boost we can make inpower.
But some of these things forphase two are very, very
important for customers and very, very lucrative for them to
move forward.
Speaker 1 (07:32):
Exciting.
Speaker 3 (07:47):
These are great advancements, and also from
regulators like the FDA,specifically to your approach
and maybe kind of in general tothe growing use of AI in
medicine and clinical trials.
Speaker 2 (08:00):
Sure, and I think it's evolving for all of us and
every day, you know, sometimestongue in cheek every day we see
, is a school day in the worldof clinical trials, meeting AI.
I think from the customer side,very much engagement.
Most large pharma companies atthe moment are exploring
artificial intelligence in someshape or form to really help
(08:20):
decrease these timelines.
Many, many of them are underbig, big pressure from their
stakeholders to make sure thatthey can bring some of these
drugs to market way quicker thanoriginal plans and
methodologies of trying to dothat obviously are engaging AI
and tech companies to try andhelp catalyze that.
At the same time, they want tomake sure that the companies
that they choose to partner withare very credible and that they
(08:42):
have what we would call in ourindustry, validated solutions.
And that's where it kind oflinks into your question with
regards to the kind of FDA andEMA some of the large governing
bodies worldwide out there thatsupport clinical trial
development.
They have to make sure thatwe're doing things proper and
correct.
So we went through, obviously,creditations with the EMA and
(09:03):
FDA to get alignment andagreement that we're very valid
in what we do and very crediblethat we're very valid in what we
do and very credible.
And actually pretty recently theFDA has published a seven-step
framework for AI in drugdevelopment, setting very clear
expectations on how AI companiesshould be working with pharma
companies to make sure that theyprovide the right things.
(09:26):
So it's a structured approachthat really emphasizes on
transparency, risk assessmentand continuous evaluation and
integration of AI into the drugdevelopment spectrum.
And if that gets a tick box orgets a green light, then you
know we all go off to the racesand apply AI in that clinical
trial.
So we've produced a few whitepapers and some case studies
(09:47):
recently on how you know we'reinteracting with FDA also how
we're managing to help validateand be credible through that
process.
And we go to a lot of meetingswith our customers at the FDA so
that when we're having thoseintense discussions about the
clinical trial impacts of AI,we're their hand on, we're right
beside the customer havingthose discussions.
(10:08):
So a lot of things are changingand we will continue to see
change.
If we look back in the mirror,over the last year, four years,
we've really came on leaps andbounds and I'm sure the next
year four years in this spaceare going to continue to really
accelerate dramatically.
Speaker 1 (10:22):
Yeah, exciting times.
I mean zooming out big picture.
You do have a lot of differentstakeholders, obviously clinical
development ecosystem, theresearchers that you mentioned,
payers, patients.
How do you think about thatbalancing act with all of those
different interested andinvolved parties?
Speaker 2 (10:41):
No, absolutely, and I think the luxurious position we
have here at Unlearn is we doget a chance to interact with
most of that ecosystem in someshape or form.
If I take it big picture andthen come to probably a day in
the life of Unlearn, obviouslyin big picture I spend as the
CEO of Unlearn, with fellow CEOsor fellow C-suite members of
the team on the pharma side inlooking at how we can reduce
(11:03):
that risk and cost and how wecan help with our trial
timelines so that ultimately,those stakeholders are getting a
chance to really accelerate keydecisions with confidence.
At the same time, we areobviously supporting that
regulatory world that I justspoke about.
So we're interfacing with thoseregulatory teams to make sure
that if there's areas of theirbusiness that we need to fine
(11:24):
tune or get back on track forwith regards to how AI is being
interpreted with the regulators,we're supporting that with the
team across the board.
And that sometimes encapsulatesmany, many projects that are in
flight time.
But probably in a day in thelife of Unlearn we spend a lot
of our time with what we callour clinical development team or
the research and developmentteam, who are really, you know,
(11:46):
looking at the intimacy of thatprotocol, wondering where, how
patients are performing, howthings are just running in the
health and just the life cycleof that study in its entirety.
So as we transgress through that, we do peak and trough with
different of the more specialistroles in clinical development
and in this occasion we spend alot of time with what we call
(12:07):
biostatisticians, so people thatare looking at the statistical
outcomes and the statisticalevidence that's prevalent in
that clinical trial, becausedata is how ultimately we are
making decisions on how thingsgo forward.
So we have to make sure thatthat data is well addressed and
obviously analyzed.
So we spend an intimate amountof time with the data management
(12:28):
and biostats teams and ninetimes out of ten it's been seen
as an extension to their team,but sometimes we run solo on
certain projects and then bringthat back in-house.
So a lot of ecosystem we touchwhich makes the job very, very
exciting and at different partsof the clinical trial process
obviously many, many others areinvolved.
Speaker 3 (12:47):
Wow thanks for that great overview of kind of behind
the scenes.
How much is involved in gettingall this data to provide useful
insights and actionableinsights, expected or exciting
use cases that you had found inapplying Digital Twin's approach
to real-world drug developmentor clinical trials?
Speaker 2 (13:21):
Again, I think it's a little cliched, but probably
the most exciting thing is theactual adoption curve that we
are seeing.
So you know, as we look at thebeginning of our journey here at
Unlearn, you know probably AImight have been seen as that
spooky word or you know not surewhat that really means and
what's happening to my data,what's going to happen to this
(13:42):
trial overall and what I'vedefinitely seen, and especially
in my own time here at Unlearnthe adoption and the
understanding.
I think the industry is reallymoving, which is very, very
exciting for us.
Obviously, we had a much moreearlier onsite or early
excitement about the adoption ofAI in clinical trials.
Hence the reason we started theorganization and the company.
(14:02):
But that's not necessarily.
You know, the pharmaceuticalindustry, without sounding
terribly derogative, you knoware a little bit more of a
lagger industry.
They like to make sure othershave tried and tested and seen
before they move in.
Sure others have tried andtested and seen before they move
in.
So I'm very excited to see thatwe spend a lot of our time in
the top 30, 50 pharma surgicalcompanies in the world who are
leading drug development.
And then, you know, what's evenmore exciting for me is now
(14:23):
that they're picking up thephone and calling us and saying
we're just about to start trialX and we know that we need
Unlearn with us in thissituation because we are really
focused in on getting this doneproperly, excited, but at the
same time we're trying to reduceour numbers and we're trying to
bring in our timelines and weknow that unlearn will be able
to help.
So, having seen that adoptioncurve, I think is exciting and
(14:46):
of course, like most of us whocame into the clinical trial
business or come from some formof medical training and some
description, we're very excitedabout how we can get that
miracle of medicine to thepatient as quickly as we can.
So most of us every day aregetting out of bed to make sure
that you know that Alzheimer'spatient, that oncology patient,
that you know thatcardiovascular patient is
(15:07):
getting the help and the needthat they get.
And the quicker we manage toget that medicine to them, the
quicker that hopefully we givethem a much better life going
forward.
So again, you know, from a muchmore evangelical concept or big
picture, here too I do thinkwe're making a big difference in
patients lives going forward,which definitely keeps us very
motivated here on them oh, it'swonderful to hear, um, some
(15:30):
would say the drug industry, thepharmaceutical industry is sort
of stuck in the Excelspreadsheet era of drug
development.
Speaker 1 (15:37):
You're clearly on the cutting edge, the leading edge,
but what has to happen toreally let AI and data science
take the wheel in clinicaltrials?
Beyond working with unlearn,what else does the industry need
to do to transform itself?
Speaker 2 (15:53):
So I think actually you know the industry need to do
to transform itself.
So I think actually you knowthe industry will transform
itself.
I've seen the evolution of youknow things in the in 10, 15
years ago where we did what wecall edc electronic data capture
.
So we moved from paper toelectronic.
Everybody thought we were crazyas an industry and everybody
was worried about it.
And now here we are, 15, 20years later.
I don't really think there'smany, if at all any, clinical
(16:15):
trials really done on paper outthere, unless something very
urgent and very quick has to bedone.
So I think AI will become thenext part of that.
It will become, hopefully in mykids' generation or in the
youth of today going forward asthey take clinical trials to the
next stage, that AI will justbe part of that protocol.
We'll be part of that clinicaldevelopment.
(16:36):
I think we're just gettingthrough, you know, some of the
cycles where we can see a drugmove from phase two into phase
three, into post-marketing, andwhere was AI influencing in that
?
So people can get a feel forthe benefit and the excitement
of it.
But we don't have, you know,hundreds and hundreds of cases
of it and we had an industrythat likes to see that.
You know we there are hundredsand hundreds of cases so that we
(16:58):
feel secure about there.
So I think you know thatadoption curve that I was
speaking about being excited, Ithink overall is going to change
.
We are moving out of excelspreadsheet into much more
trusting, you know, algorithmicbased activity, um, especially
the AI world, and there are many, many ways of technology
touching everybody in our lifetoday, but also in the clinical
(17:21):
trial world.
So we're a small part of thatAI ecosystem.
You know we've got way back atthe beginning as we look at the
cataloging of some of thesecompounds that you think you
would like to target.
So we call it target profilingand we're using AI at the
beginning to help us profilethose targets much more
sophisticated, much quicker, sothat we don't pick a loss leader
or we don't pick out of thefive drugs, we pick two that you
(17:43):
know we had to go back and findthe other three.
So there's many, many, manysteps of clinical development
and there's definitely now many,many areas where AI has been
very prevalent.
Speaker 1 (17:54):
Incredible.
Speaker 3 (18:00):
Oh, you're giving us already a little bit of a
preview.
Maybe future, your future focus.
So, as we wrap up here, I wantto ask you about, kind of like,
over the next few years maybelooking forward five years what
are um evidence you're going topursue something new and
exciting or are you going tojust try to drill down on, on
what you've already shown as asvery beneficial and, uh,
(18:23):
producing huge results?
Like what's what's, in the nextfew years on on your radar?
Speaker 2 (18:29):
so definitely picking up on your last part, in the
next few years I'd like toobviously continue to build our
you know our profile out there.
As you know, an excellentcompany in you know drug
development with the power of aithat can support that going
forward.
Um, my big, big vision or oneof my personal ambitions for
unlearn is that we will becomeexceptionally agnostic to any
(18:52):
therapeutic area in any clinicaltrial out there and basically
every single time a cliniciangoes to start drafting a
protocol that when they go tomove their tools over in their
what do, I must need list.
Without question.
Unlearn is right at the top ofthat must need list that we will
then just naturally become partof every single clinical
(19:13):
protocol out there as we goforward.
So we are not at that stage atthe moment.
As I said, we've been journeyingvery much in the CNS space.
We're moving now recently intooncology and into inflammation
and into some of thecardiometabolic disorders out
there too.
So we are definitely taking ourstatewide approach to that.
We're a small company so wehave to be careful that we don't
(19:35):
overcommit to things we candeliver.
But eventually, hopefully, aswe grow ourself, as we grow the
industry's knowledge and aseverybody embraces technology
and AI and clinical trials, thenUnlearn becomes the center part
of that protocol and that willbe very exciting for not just
the industry but definitely forme personally and the team here
on there.
Speaker 1 (19:57):
And for all of us.
We are really rooting for yoursuccess and the mission and
vision is amazing.
Godspeed, and onwards andupwards.
Thanks for joining and sharinga peek behind the curtain.
Speaker 2 (20:06):
Thank you very much for having me.
Thanks again.
Speaker 1 (20:08):
And thanks for listening.
Speaker 3 (20:09):
Thank you everyone for watching and listening and
check out our new.
Speaker 1 (20:13):
TV show at techimpacttv now on Bloomberg
and Fox Business, and thanks forsharing everyone.
Take care.
Thanks, Steve.
Hey everybody, fascinating discussion and
important discussion today, aswe talk about AI-powered
replicas, or digital twins, thatare revolutionizing how we run
clinical trials, and joined by atrue insider and thought leader
in the space from unlearnaiSteve, how are you?
Speaker 2 (00:20):
I'm doing very well.
Thank you for having me today.
Speaker 1 (00:23):
Well, thanks for being here.
I'm doing very well.
Thank you for having me today.
Well, thanks for being here.
Yes, Evan and Irma here fromAvira Health Really fascinated.
Maybe we'll start from thebeginning.
What's your background and howdid you land at unlearnai?
Speaker 2 (00:37):
Sure, no, absolutely.
So I'm coming up around 30years in the clinical trial
space.
I started my life out ofacademia, getting into clinical
trials on the operations sideand then spent a lot of my
career in the commercialmarketing and then for the last
10-15 years I've been spendingmore of my time on the clinical
technology side in supportingclinical trials and the kind of
(01:00):
advantages that technology canreally do to shape our clinical
trials going forward.
And that really fast forwardsto my journey over the last year
here at Unlearn AI, wheretechnology AI meets clinical
trials.
So for me it's an excitinghopefully near end of my career
where I can really takeeverything I've learned over the
last three decades in clinicaltrials and see if we can really
(01:23):
start to make a difference usingartificial intelligence and
technology out there in themarket.
Speaker 3 (01:29):
Fantastic.
So today we will be diving intothe world of digital twins and
for our listeners, who may notbe familiar or new to the term,
what exactly is a digital twin,and particularly in the context
of medicine?
Speaker 2 (01:48):
is a digital twin and particularly in the context of
medicine Sure, no, absolutely,to make it nice and simple and
very easy for the listeners, adigital twin is basically an
AI-generated forecast of apatient's clinical outcomes,
predominantly in our world,obviously in a clinical trial.
So that really helps usforecast how that patient will
progress with their diseasealong the life cycle of the
clinical trial.
(02:08):
So we're spending a lot oftechnology behind the scenes, a
lot of algorithmic activity toreally forecast that
longitudinal activity of thatindividual.
So you go into that clinicaltrial and your twin comes in
with you.
So we create that digital twinthat runs along with you
throughout the clinical trial.
Speaker 1 (02:29):
Amazing, and this has moved from the sort of lab into
the real world.
How are you actually applyingthese digital twins at Unlearn
to medical research and what arethe different use cases?
Speaker 2 (02:41):
Sure, and what are the different use cases?
Sure, so again, I think mostdigital twins we're really using
to enable us to compareobserved outcomes versus
predicted outcomes.
So in clinical terms, that letsus really look at the
variability, helping to reducethe noise so that we can really
see true signals in the data bylowering that variance.
(03:03):
That allows us to really startto really see true magnified
signals that people can thenmake much more effective what we
call, in clinical triallanguage, go-no-go decisions.
So people want to shorten thelength of a clinical trial as
quickly but as effectively asthey can.
So the quicker we get to thosego-no-go decisions ie, is the
(03:24):
therapeutic working well?
Are we seeing the response welike to those go-no-go decisions
?
Ie, is the therapeutic workingwell?
Are we seeing the response welike to see, then the happier
everybody is.
And we do that obviously byreducing sample size, which
basically means we're bringingthe number of patients on the
trial or the number ofvolunteers that are in the
clinical trial.
We bring that down because wecan obviously run a digital
trial sorry, a digital twinalong with a patient at the same
(03:45):
time.
So smaller sample sizes equalsfaster enrollment and therefore
we're boosting what we call thepower of the study, reducing the
error and making that trialmuch more effective and much
quicker.
Speaker 3 (03:57):
Wow, wow, that's right.
There are already greatbenefits of using digital twins
compared to just the traditionalclinical trial methods.
Can you elaborate a little bitmore?
Do you have advantage in termsof geographical area you can
cover or any other benefits tousing this approach versus
(04:17):
traditional clinical trials?
Speaker 2 (04:19):
So I think again, main main benefit again just for
everybody listening is reducing.
You know endpoints with fewerpatients and you know obviously
shorter durations.
That is definitely equals, youknow, digital twin period.
But from a geographyperspective, across the board
we're pretty agnostic in bothgeography and in therapeutic
area.
(04:39):
We're trying to, as we workwith companies and customers
right at the beginning of theclinical trial development
program, in analyzing theprotocols and understanding
different ways to work with them, we're obviously trying to
maximize how they can run thattrial as effective as possible.
So we have lots and lots andlots of data out there that we
obviously learn these modelsthrough so that we can really
(05:02):
help predict the future forthese trials.
So in particular areas we mightsay go into this geography or
look at this type of population.
We also might say examine thisendpoint along with this
endpoint because you will seemaximum benefit that we've seen
maybe as we've supported othertrials or done other work with
other customers.
So again, it's case by case oneach situation, but the main
(05:25):
main thrust obviously we'rereducing sample size and we're
decreasing clinical timelines.
That's amazing.
Speaker 1 (05:32):
And we all know the cost and time associated with
drug development.
It's hard to believe drugs getto market, given the challenges.
Can you walk us through howthey help reduce time and cost,
and can you quantify that at all?
Speaker 2 (05:48):
Sure, so we've spent a lot of time in the evolution
of Unlearn in what we call theneurodegenerative disease space,
or sometimes known as thecentral nervous system, cns.
So evolution over the lastseven, eight years in the
company we've spent a lot oftime predominantly in
Alzheimer's disease.
So again, this is public and onour website and publicly
available data, so I'm notsharing anything too
(06:10):
confidential today.
But as we managed anAlzheimer's drug with a customer
, we managed to reduce thesample size of up to 33% in that
particular trial and we boostedthe power over 10%.
So we went from just under 80%to over 90% in power and brought
the trial in at least fourmonths earlier than the original
(06:32):
plan date.
So that was a phase two trial.
These are slightly smaller thanthe larger phase threes but
seeing that type of activity ona phase two is very exciting.
To obviously get us into thatno-go decision for phase three
and some of the rare diseases orsome of the more rare CNS
profiles like ALS, we did see arecent trial that we ran with
(06:53):
one of our customers where wereduced our sample size by 18%
and cut those timelines byaround three to four months.
And the last one that we justactually not long ago finished
actually was in Parkinson'sdisease and again we actually
increased our sample size thereby about 33% and increased the
power by about 65%, cuttingabout nine, 10 months off the
(07:15):
entire journey on that one too.
So again, obviously the largertrials that we get into, the
larger cuts that we can make intime and also, hopefully, the
stronger boost we can make inpower.
But some of these things forphase two are very, very
important for customers and very, very lucrative for them to
move forward.
Speaker 1 (07:32):
Exciting.
Speaker 3 (07:47):
These are great advancements, and also from
regulators like the FDA,specifically to your approach
and maybe kind of in general tothe growing use of AI in
medicine and clinical trials.
Speaker 2 (08:00):
Sure, and I think it's evolving for all of us and
every day, you know, sometimestongue in cheek every day we see
, is a school day in the worldof clinical trials, meeting AI.
I think from the customer side,very much engagement.
Most large pharma companies atthe moment are exploring
artificial intelligence in someshape or form to really help
(08:20):
decrease these timelines.
Many, many of them are underbig, big pressure from their
stakeholders to make sure thatthey can bring some of these
drugs to market way quicker thanoriginal plans and
methodologies of trying to dothat obviously are engaging AI
and tech companies to try andhelp catalyze that.
At the same time, they want tomake sure that the companies
that they choose to partner withare very credible and that they
(08:42):
have what we would call in ourindustry, validated solutions.
And that's where it kind oflinks into your question with
regards to the kind of FDA andEMA some of the large governing
bodies worldwide out there thatsupport clinical trial
development.
They have to make sure thatwe're doing things proper and
correct.
So we went through, obviously,creditations with the EMA and
(09:03):
FDA to get alignment andagreement that we're very valid
in what we do and very crediblethat we're very valid in what we
do and very credible.
And actually pretty recently theFDA has published a seven-step
framework for AI in drugdevelopment, setting very clear
expectations on how AI companiesshould be working with pharma
companies to make sure that theyprovide the right things.
(09:26):
So it's a structured approachthat really emphasizes on
transparency, risk assessmentand continuous evaluation and
integration of AI into the drugdevelopment spectrum.
And if that gets a tick box orgets a green light, then you
know we all go off to the racesand apply AI in that clinical
trial.
So we've produced a few whitepapers and some case studies
(09:47):
recently on how you know we'reinteracting with FDA also how
we're managing to help validateand be credible through that
process.
And we go to a lot of meetingswith our customers at the FDA so
that when we're having thoseintense discussions about the
clinical trial impacts of AI,we're their hand on, we're right
beside the customer havingthose discussions.
(10:08):
So a lot of things are changingand we will continue to see
change.
If we look back in the mirror,over the last year, four years,
we've really came on leaps andbounds and I'm sure the next
year four years in this spaceare going to continue to really
accelerate dramatically.
Speaker 1 (10:22):
Yeah, exciting times.
I mean zooming out big picture.
You do have a lot of differentstakeholders, obviously clinical
development ecosystem, theresearchers that you mentioned,
payers, patients.
How do you think about thatbalancing act with all of those
different interested andinvolved parties?
Speaker 2 (10:41):
No, absolutely, and I think the luxurious position we
have here at Unlearn is we doget a chance to interact with
most of that ecosystem in someshape or form.
If I take it big picture andthen come to probably a day in
the life of Unlearn, obviouslyin big picture I spend as the
CEO of Unlearn, with fellow CEOsor fellow C-suite members of
the team on the pharma side inlooking at how we can reduce
(11:03):
that risk and cost and how wecan help with our trial
timelines so that ultimately,those stakeholders are getting a
chance to really accelerate keydecisions with confidence.
At the same time, we areobviously supporting that
regulatory world that I justspoke about.
So we're interfacing with thoseregulatory teams to make sure
that if there's areas of theirbusiness that we need to fine
(11:24):
tune or get back on track forwith regards to how AI is being
interpreted with the regulators,we're supporting that with the
team across the board.
And that sometimes encapsulatesmany, many projects that are in
flight time.
But probably in a day in thelife of Unlearn we spend a lot
of our time with what we callour clinical development team or
the research and developmentteam, who are really, you know,
(11:46):
looking at the intimacy of thatprotocol, wondering where, how
patients are performing, howthings are just running in the
health and just the life cycleof that study in its entirety.
So as we transgress through that, we do peak and trough with
different of the more specialistroles in clinical development
and in this occasion we spend alot of time with what we call
(12:07):
biostatisticians, so people thatare looking at the statistical
outcomes and the statisticalevidence that's prevalent in
that clinical trial, becausedata is how ultimately we are
making decisions on how thingsgo forward.
So we have to make sure thatthat data is well addressed and
obviously analyzed.
So we spend an intimate amountof time with the data management
(12:28):
and biostats teams and ninetimes out of ten it's been seen
as an extension to their team,but sometimes we run solo on
certain projects and then bringthat back in-house.
So a lot of ecosystem we touchwhich makes the job very, very
exciting and at different partsof the clinical trial process
obviously many, many others areinvolved.
Speaker 3 (12:47):
Wow thanks for that great overview of kind of behind
the scenes.
How much is involved in gettingall this data to provide useful
insights and actionableinsights, expected or exciting
use cases that you had found inapplying Digital Twin's approach
to real-world drug developmentor clinical trials?
Speaker 2 (13:21):
Again, I think it's a little cliched, but probably
the most exciting thing is theactual adoption curve that we
are seeing.
So you know, as we look at thebeginning of our journey here at
Unlearn, you know probably AImight have been seen as that
spooky word or you know not surewhat that really means and
what's happening to my data,what's going to happen to this
(13:42):
trial overall and what I'vedefinitely seen, and especially
in my own time here at Unlearnthe adoption and the
understanding.
I think the industry is reallymoving, which is very, very
exciting for us.
Obviously, we had a much moreearlier onsite or early
excitement about the adoption ofAI in clinical trials.
Hence the reason we started theorganization and the company.
(14:02):
But that's not necessarily.
You know, the pharmaceuticalindustry, without sounding
terribly derogative, you knoware a little bit more of a
lagger industry.
They like to make sure othershave tried and tested and seen
before they move in.
Sure others have tried andtested and seen before they move
in.
So I'm very excited to see thatwe spend a lot of our time in
the top 30, 50 pharma surgicalcompanies in the world who are
leading drug development.
And then, you know, what's evenmore exciting for me is now
(14:23):
that they're picking up thephone and calling us and saying
we're just about to start trialX and we know that we need
Unlearn with us in thissituation because we are really
focused in on getting this doneproperly, excited, but at the
same time we're trying to reduceour numbers and we're trying to
bring in our timelines and weknow that unlearn will be able
to help.
So, having seen that adoptioncurve, I think is exciting and
(14:46):
of course, like most of us whocame into the clinical trial
business or come from some formof medical training and some
description, we're very excitedabout how we can get that
miracle of medicine to thepatient as quickly as we can.
So most of us every day aregetting out of bed to make sure
that you know that Alzheimer'spatient, that oncology patient,
that you know thatcardiovascular patient is
(15:07):
getting the help and the needthat they get.
And the quicker we manage toget that medicine to them, the
quicker that hopefully we givethem a much better life going
forward.
So again, you know, from a muchmore evangelical concept or big
picture, here too I do thinkwe're making a big difference in
patients lives going forward,which definitely keeps us very
motivated here on them oh, it'swonderful to hear, um, some
(15:30):
would say the drug industry, thepharmaceutical industry is sort
of stuck in the Excelspreadsheet era of drug
development.
Speaker 1 (15:37):
You're clearly on the cutting edge, the leading edge,
but what has to happen toreally let AI and data science
take the wheel in clinicaltrials?
Beyond working with unlearn,what else does the industry need
to do to transform itself?
Speaker 2 (15:53):
So I think actually you know the industry need to do
to transform itself.
So I think actually you knowthe industry will transform
itself.
I've seen the evolution of youknow things in the in 10, 15
years ago where we did what wecall edc electronic data capture
.
So we moved from paper toelectronic.
Everybody thought we were crazyas an industry and everybody
was worried about it.
And now here we are, 15, 20years later.
I don't really think there'smany, if at all any, clinical
(16:15):
trials really done on paper outthere, unless something very
urgent and very quick has to bedone.
So I think AI will become thenext part of that.
It will become, hopefully in mykids' generation or in the
youth of today going forward asthey take clinical trials to the
next stage, that AI will justbe part of that protocol.
We'll be part of that clinicaldevelopment.
(16:36):
I think we're just gettingthrough, you know, some of the
cycles where we can see a drugmove from phase two into phase
three, into post-marketing, andwhere was AI influencing in that
?
So people can get a feel forthe benefit and the excitement
of it.
But we don't have, you know,hundreds and hundreds of cases
of it and we had an industrythat likes to see that.
You know we there are hundredsand hundreds of cases so that we
(16:58):
feel secure about there.
So I think you know thatadoption curve that I was
speaking about being excited, Ithink overall is going to change
.
We are moving out of excelspreadsheet into much more
trusting, you know, algorithmicbased activity, um, especially
the AI world, and there are many, many ways of technology
touching everybody in our lifetoday, but also in the clinical
(17:21):
trial world.
So we're a small part of thatAI ecosystem.
You know we've got way back atthe beginning as we look at the
cataloging of some of thesecompounds that you think you
would like to target.
So we call it target profilingand we're using AI at the
beginning to help us profilethose targets much more
sophisticated, much quicker, sothat we don't pick a loss leader
or we don't pick out of thefive drugs, we pick two that you
(17:43):
know we had to go back and findthe other three.
So there's many, many, manysteps of clinical development
and there's definitely now many,many areas where AI has been
very prevalent.
Speaker 1 (17:54):
Incredible.
Speaker 3 (18:00):
Oh, you're giving us already a little bit of a
preview.
Maybe future, your future focus.
So, as we wrap up here, I wantto ask you about, kind of like,
over the next few years maybelooking forward five years what
are um evidence you're going topursue something new and
exciting or are you going tojust try to drill down on, on
what you've already shown as asvery beneficial and, uh,
(18:23):
producing huge results?
Like what's what's, in the nextfew years on on your radar?
Speaker 2 (18:29):
so definitely picking up on your last part, in the
next few years I'd like toobviously continue to build our
you know our profile out there.
As you know, an excellentcompany in you know drug
development with the power of aithat can support that going
forward.
Um, my big, big vision or oneof my personal ambitions for
unlearn is that we will becomeexceptionally agnostic to any
(18:52):
therapeutic area in any clinicaltrial out there and basically
every single time a cliniciangoes to start drafting a
protocol that when they go tomove their tools over in their
what do, I must need list.
Without question.
Unlearn is right at the top ofthat must need list that we will
then just naturally become partof every single clinical
(19:13):
protocol out there as we goforward.
So we are not at that stage atthe moment.
As I said, we've been journeyingvery much in the CNS space.
We're moving now recently intooncology and into inflammation
and into some of thecardiometabolic disorders out
there too.
So we are definitely taking ourstatewide approach to that.
We're a small company so wehave to be careful that we don't
(19:35):
overcommit to things we candeliver.
But eventually, hopefully, aswe grow ourself, as we grow the
industry's knowledge and aseverybody embraces technology
and AI and clinical trials, thenUnlearn becomes the center part
of that protocol and that willbe very exciting for not just
the industry but definitely forme personally and the team here
on there.
Speaker 1 (19:57):
And for all of us.
We are really rooting for yoursuccess and the mission and
vision is amazing.
Godspeed, and onwards andupwards.
Thanks for joining and sharinga peek behind the curtain.
Speaker 2 (20:06):
Thank you very much for having me.
Thanks again.
Speaker 1 (20:08):
And thanks for listening.
Speaker 3 (20:09):
Thank you everyone for watching and listening and
check out our new.
Speaker 1 (20:13):
TV show at techimpacttv now on Bloomberg
and Fox Business, and thanks forsharing everyone.
Take care.
Thanks, Steve.
Popular Podcasts
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.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.