September 17, 2025 • 23 mins
An interview on oculomics (using retinal imaging and AI to assess systemic disease) with Tien Wong, MD, PhD, Chair Professor and Founding Head of Tsinghua Medicine at Tsinghua University, Beijing, China.
Listen
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
David Klonoff (00:13):
Welcome to Diabetes Technology Report.
I'm Dr David Klonoff from MillsPeninsula Medical Center in San
Mateo, California.
I'm here with my co-host and wehave a very special guest from
very far away from California.
David will introduce him.
David Kerr (00:31):
Hello everyone.
It's great to be with you,david Klonoff, today, and, of
course, with our very, veryspecial guest, dr Chen Wong,
who's speaking to us fromBeijing.
We are so fortunate because youare a guru when it comes to the
use of technology for thedetection of diabetic
(00:51):
retinopathy, so I'm very excited.
But what we'd like to ask ourguests first of all is how did
you end up where you are with aninterest in technology and eye
disease and diabetes?
What was that journey for you,dr Wong?
Tien Wong (01:07):
Okay, thank you very much.
I'm so pleased to be here.
My journey really began as anophthalmologist and I was doing
my PhD in the US at JohnsHopkins, and I was also doing
field work and a postdoctoralfellowship at the University of
Wisconsin Medicine, and at thattime diabetic retinopathy was
(01:32):
just beginning to be recognizedas a major cause of blindness in
the US and elsewhere in theworld.
Working with Ronald and BarbaraKlein from the landmark
Wisconsin Epidemiological Studyof Diabetic Retinopathy, I went
into the field whereby lookingat not just the epidemiology,
but how do we prevent thedisease, how do we screen for
(01:54):
the disease and therefore leadto better treatment.
So a lot of my early researchwas focused on that.
I therefore stumbled onto whatI'll say is quite an important
part of managing diabeticretinopathy as well as other
diabetes complications which Ican get into subsequently.
(02:15):
But at that time we werebeginning to move from
film-based photography todigital photography, photography
to digital photography, andtherefore we you know the whole
group had to be very involved inunderstanding what does digital
photography mean?
How do you standardize thetechnology?
What does a high qualitydigital image offer you?
(02:41):
And then we started workingwith computer scientists, and
that was really before the eraof AI and deep learning.
Right, you know very basicpixels and understanding digital
retinal photographs.
So I would say it's reallyabout 20 plus years ago almost
(03:01):
30 years when we delved intodigital diabetic retinopathy
screening and I think that's thebasis of our current technology
today globally.
David Kerr (03:12):
So what are you up to at the moment?
Presumably everything isartificial intelligence, deep
learning.
Tien Wong (03:20):
So, at the moment, we have taken, I would say, a
longish journey towards usingthe retinal images that we now
take for granted.
It's all digital, right and it'staken from smaller and more
portable retinal cameras, andthat's for primarily two
(03:41):
purposes.
One is, of course, to detectearly stages of diabetic
retinopathy and therefore torefer them, when appropriate, to
the eye specialist for them tohave treatment, usually at eye
centers or eye clinics.
So that's a fundamental part ofdiabetic retinopathy screening,
(04:08):
part of diabetic retinopathyscreening.
Secondly, we're using theseretinal images to look at a
really even newer area in thisfield of using the retinal
photographs digital, which canallow us to look and interrogate
, for example, the blood vesselsin the not just in the eye but
the rest of the body, as well asthe nerve fiber layers.
That's seen in the eye, but therest of the body as well as the
nerve fiber layers that's seenin the eye but again
representative of nervous andneural systems in the rest of
(04:32):
the body, and this whole fieldof using the eye as what I would
say a window or a channel tounderstand the systemic
complications of diabetes which,as everyone knows, you know,
diabetes affects the heart, thebrain, the kidneys and so on and
so forth.
So that allows us to use asingle technology, in other
(04:57):
words a digital retinalphotograph, to not only
understand eye diseases but also, probably, the systemic
complications of diabetes, andthat's a very exciting new field
.
David Klonoff (05:10):
Tian, can you comment on the idea of oculomics
and how it will be used toevaluate and treat people with?
Tien Wong (05:18):
diabetes.
So oculomics is again the termcoined, whereby we are using the
eye images and the informationin the eye, so the omics from
the eye or the ocular structure,and using it to interrogate the
different systemiccomplications related to
(05:40):
diabetes, and even in peoplewithout diabetes.
But let's concentrate ondiabetes.
So, for example, there has beena very strong biological and
historical link between whathappens in the eye and what
happens in the kidneys in peoplewith diabetes.
Almost concurrently, peoplethat have diabetic retinopathy,
(06:02):
but particularly the severestages, would also have kidney
damage, diabetic nephropathy.
So using a single screening, Iwould say biomarker, in other
words, information from theretinal images, might offer you
simultaneous understanding ofdamages seen in the eye as well
(06:25):
as the rest of the body.
So this is the fundamentalconcepts of oculomics.
So no different fromunderstanding, for example, a
gene, whether it affects boththe eye and the heart and the
kidneys.
So what we call genomics orprotein, what we call proteomics
, and this is now what we calloculomics.
Now, the potential of oculomicsis number one.
(06:47):
The retinal photograph isdigital, it is non-invasive.
The technology to capture thisinformation happens in seconds
and could be, in the future,widely available in what I would
say easily accessible settings,For example, in the optometrist
(07:07):
shop.
There's been discussion that itcould be even in a pharmacy, it
could be in the supermarket andit could be in your
neighborhood post office.
So what we are potentiallydoing is to take a single
technology that could be cheaplyand widely available towards
(07:31):
what we really need to do, whichis to have not just early
screening but detection of thedifferent stages of complication
that diabetes affects across.
You know, US as well as theworld.
David Klonoff (07:44):
Ken, in the US, a problem that we've seen is that
even with screening, peopleoften don't have follow-through
to see the eye doctor for thetreatment.
Are you working on some type ofa follow-through program so
it's not just here's yourdiagnosis, but getting people in
for treatment?
Tien Wong (08:03):
And this is the new, exciting era of large language
models.
As most of us in the world noware aware, the large language
models offers what we call verytailored, very specific
instructions that can be used tohelp patients understand their
(08:25):
problem, understand the severityof their condition, as well as
have very targetedrecommendations for them to
either have follow-up, forexample, with the eye doctors or
with their kidney doctors, orto have better control of their
sugars and blood pressure.
And therefore, in one of ourrecent papers, we have combined
(08:48):
what we call the retinal imageswith a large language model
component, so that they extractthe information from the patient
, as well as the retinal images,and give a targeted, specific
advice that might hopefullyprompt the patient to take
better control of their ownhealth situation.
(09:10):
Now, this is different fromwhat was previously, prior to
the era of large language model,prior to large chat, gpts and
so forth, whereby people wouldget a pamphlet, which will be
good advice, but it will be verygeneric advice and it doesn't
really, I would say, tailor to aspecific individual or patient.
(09:31):
And we think that in the era ofAI, digital imaging combined
with large language models, weshould be better able to tackle
this part, which is thatpatients get information they
don't know what to do with it.
They're not motivated to doanything with it and hopefully
this large language model offersthat patient assistance.
David Kerr (09:54):
Tian, this is absolutely fascinating something
that we're very interested in.
My world at Sutter Health herein California, we're also
contemplating whether you cancombine more frequent retinal
imaging with, say, otherwearables such as continuous
glucose monitors or bloodpressure monitoring, and so you
(10:16):
can actually have, from aperspective of time you can see,
changes in the retina whichreflect changes in glucose.
Are you involved in work inthat area as well?
Tien Wong (10:28):
I think a lot of groups are working combining
multiple modalities, andcertainly we have some work with
people on the wearables field,as well as those using what
we'll call mobile technology,and information that feeds this
(11:03):
mobile technology will likelycome from multiple places, right
From the standard doctor'snotes to the laboratory records,
continuous glucose monitoringand, of course, you know, the
periodic visits for us will bethe optometrist and for some
people would be the kidneyspecialist and I think,
ultimately, an individualpatient will, at the point of
(11:29):
diagnosis, have a digital recordof information that's collected
over time from various sources,no different from, you know,
from our own digital history.
I mean, we have our own Facebookhistory, our Instagram history
I likely see a little bit ofthis our diabetes history.
What we are not seeing yet,it's a holistic platform whether
(11:55):
you can call it an app or adevice that is able to extract
these very diverse sources ofdata and to make sense of these
data.
But I think that will likelycome soon.
It's really what we call a newfield, where people say it's
convergent science.
Right, you know, we havescience that is very siloed,
(12:20):
kind of developing on their own,and what we need is a platform
that brings together data andtechnology, ai you know large
language models into somethingsimpler.
You know we need a little bitlike a Steve Jobs iPhone moment
(12:41):
right, whereby we're able tohave that single companion, that
buddy that helps us with ourhealth over time.
David Kerr (12:51):
Essentially and I think that that's something that
you know, we wish to see forour patients and for our
providers- so, just following onfrom that, on a more negative
aspect of that, if you take 100people and you try and look at
the back of the eye withoutdilating the pupils a proportion
of them you just don't get verygood images.
(13:13):
Is that a problem that's alwaysgoing to be there, or do you
think the technology is gettingbetter and better, that the
ungradable or unviewable imagesis going to be zero cameras?
They?
Tien Wong (13:27):
were poor quality images and then newer cameras
have tracking devices, therewill likely be some sort of
automatic kind of artifactremoval.
(13:48):
There will be likely somegenerative parts of the images
that will be synthetic but willfollow the pattern of what the
large language models ispredicting generative AI.
So you don't need all theinformation on everything right
(14:08):
I think you need substantialinformation and it will be able
to produce an image that is asclose as possible to the real
image.
So I think that technologywhereby the new generative AI
algorithms allow will herald anera whereby most of the images
(14:30):
will be gradable and readableand interpretable for everyone.
David Klonoff (14:35):
Tian, one of the most fascinating areas of AI is
agent hospital.
Could you talk about that, whatyou're doing and what the
hospital is all about?
Tien Wong (14:46):
We are starting to think about how do we build
something that is for healthcareusing AI, and there are really,
I would say, three approaches,right?
The first approach is thetraditional hospital, the
traditional clinics.
All of us are very familiarwith it.
They are not nice places tovisit.
They are usually crowded, thereare usually lots of gaps in the
(15:11):
care.
We spend three hours in theclinic and we see our physician
for maybe five minutes, right,so there are many challenges
that happens in the clinic andwe see our physician for maybe
five minutes, right, so thereare many challenges that happens
in the traditional hospital.
So one way, of course, is totransform this hospital by
putting in more applications,having back-end AI tools, using
(15:32):
some front-end AI tools where,before the patients see the
doctors, they have an interview,so-called with an AI chatbot,
right, so those are being done.
What I would say traditionalhospital plus small AI.
Now, is that doable?
Is that going to be successful?
I think there's many, manyexperiments going on everywhere
(15:54):
and we have yet to see majortransformation.
In fact, there's probably a lotof mental fatigue, there's
probably increase in costs,there's uncertainty from
patients and physicians and oneapp leads to another and there's
very little convergence.
So what we are trying to do isto say let's build it from
scratch, right?
(16:14):
Almost a hospital that is builtdigitally from beginning, and
that's the concept of the AIagent hospital, and therefore
that hospital should have, inone sense, co-development by
physicians and engineers, anidea whereby the financial model
(16:35):
or the sustainability of themodel is not dependent on how
many patients come in, how manyprocedures are done, how many
tests are done, but on a modelin which we are able to
integrate the mostcost-effective digital platforms
and artificial intelligencewith as few touch points by the
(16:58):
physicians as possible.
In fact, we envision that atsome stage, half, maybe even
more, of these patients will notneed to interact with a
physical healthcare provider,right?
So you need to start from thatkind of basis, and that's where
we are not calling it an AIhospital, but an AI agent
(17:19):
hospital, whereby the agent isthe primary coordinator of care
within the hospital andtherefore, as I said, it's
envisioned that 50% probably donot ever need to see that
physical human healthcareprovider or physician for that
matter and therefore we hope tohave targeted improvement in
(17:43):
efficiency, of course, indiagnosis and safety.
That's the basis of thetraining the agent, but also
things that we don't really havea good handle of, you know, the
throughput of patients, theshortness of the waiting times
in the hospital and, ultimately,whether they even need to come
(18:04):
to the hospital, because we canimagine that the AI agent
hospital a lot of those carethat's provided will be via
mobile devices, will be viatelemedicine and in their homes,
right.
So that requires a completethinking.
The AI agent hospital reallyneeds a concept that we do not
(18:29):
yet know, but we are trying tobuild.
David Klonoff (18:33):
Tian.
One topic that we've talkedabout is how do you get patients
to trust the AI If they're notseeing a doctor.
The patient has to really findthat it's trustworthy.
Tien Wong (18:44):
Yeah, I think trust is a very key component of the
entire AI ecosystem.
I don't think we pay sufficientattention to that.
We usually pay number one, thetechnology development.
Right, how robust?
How big is the data set?
That's training the algorithm.
We looked at another aspect,which is the clinical outcomes.
(19:08):
Right, the clinical outcomes.
Is it efficacious?
Is it doing what it's supposedto do?
Is it picking up the disease?
Is it missing cases?
But in between is what I wouldsay a very important area, which
is on this entire system oftrust, and you can say it is
behavioral, it is psychological,it is experiential and it
(19:35):
involves multiple stakeholders.
Now, how do we define trust?
You can say that wetraditionally have put a trust
in the patient-doctorrelationship.
Now you are adding a thirdparty involved in this
patient-doctor, so you need tobuild a trust in this
three-party relationship.
(19:55):
It's not that easy, right?
I mean, we know who we like,who we can interact with.
Do we want someone else in theroom?
Right?
And I think that, therefore,that trust needs to be from a
multi-faceted approach.
Number one patient needs totrust.
And what does the patient trustmean?
The doctor needs to trustbecause you know there's another
(20:18):
person in the room.
So there's also a healthcareprovider trust relationship, and
then the system needs to trust,because the system needs to
know that we're building asystem whereby it's not just a
two-way interaction between thepatient and the healthcare
professional, but another partin the system, and the system's
trust could involve manydifferent things that we are not
(20:41):
able to see.
So I think that a whole idea oftrust is something that we need
.
Other groups of stakeholdersinvolved in this AI relationship
, and currently I see two majorstakeholders the computer
scientists, the engineers andthe physicians who want to use
the technology.
We need that third group in theroom.
David Kerr (21:03):
Tian.
I have one final question, justto bring it back to the retinal
imaging.
I was brought up that theretina is the gateway to the
soul.
If you move beyond diabetes andthe complications, what other
diseases do you think are goingto be detectable at the back of
the eye at a very early stage?
Tien Wong (21:24):
going forward, I think that one of the most
promising aspects beyonddiabetes, is using the eye as
really a marker of aging andbrain health.
I think these are the two mostimportant things we are now
interested in healthy aging.
(21:45):
You can say it's longevity orit is better quality of life as
we age, and I think the retinaoffers a lot of promise for that
.
For example, in one of thealgorithms we have developed,
what we call a retinal-phenolage.
It's really an age of theperson based on the changes or
(22:11):
the health or the damage of theretina.
I will have to tell you thatwhen I did that algorithm, my
retinal age is five years olderthan my chronological age, which
is not a good sign, essentially, right.
So I'm kind of older than myown timeline should be by
chronological age.
I've seen people that haveretinal age that is five or ten
(22:34):
years younger than theirchronological age.
So, in other words, thechronological age it's just a
time, it's a number, whereas theretinophenol age is a
biological marker of our bodyessentially, and I think that
that's a very interesting andvery exciting era as we move
(22:55):
into.
As I said, the interest of manyolder people of our entire
world is interested in how do wemaintain healthy aging and
longevity in this very complexworld that we live in?
David Klonoff (23:13):
Ken, thank you very much for discussing AI,
retinal health, oculomics andsomething I had not heard before
, which is the health age fromthe retina.
I'm going to read about that.
Thank you for joining us.
We hope to work with you in thefuture.
And now I'm going to saygoodbye from Diabetes Technology
(23:33):
Report.
We're available on Spotify andat the Apple Store and at the
Diabetes Technology Societywebsite.
We look forward to catching upwith you at our next Diabetes
Technology Report.
So for now, goodbye everybody.
Goodbye.
Tien Wong (23:51):
Thank you very much.
Welcome to Diabetes Technology Report.
I'm Dr David Klonoff from MillsPeninsula Medical Center in San
Mateo, California.
I'm here with my co-host and wehave a very special guest from
very far away from California.
David will introduce him.
David Kerr (00:31):
Hello everyone.
It's great to be with you,david Klonoff, today, and, of
course, with our very, veryspecial guest, dr Chen Wong,
who's speaking to us fromBeijing.
We are so fortunate because youare a guru when it comes to the
use of technology for thedetection of diabetic
(00:51):
retinopathy, so I'm very excited.
But what we'd like to ask ourguests first of all is how did
you end up where you are with aninterest in technology and eye
disease and diabetes?
What was that journey for you,dr Wong?
Tien Wong (01:07):
Okay, thank you very much.
I'm so pleased to be here.
My journey really began as anophthalmologist and I was doing
my PhD in the US at JohnsHopkins, and I was also doing
field work and a postdoctoralfellowship at the University of
Wisconsin Medicine, and at thattime diabetic retinopathy was
(01:32):
just beginning to be recognizedas a major cause of blindness in
the US and elsewhere in theworld.
Working with Ronald and BarbaraKlein from the landmark
Wisconsin Epidemiological Studyof Diabetic Retinopathy, I went
into the field whereby lookingat not just the epidemiology,
but how do we prevent thedisease, how do we screen for
(01:54):
the disease and therefore leadto better treatment.
So a lot of my early researchwas focused on that.
I therefore stumbled onto whatI'll say is quite an important
part of managing diabeticretinopathy as well as other
diabetes complications which Ican get into subsequently.
(02:15):
But at that time we werebeginning to move from
film-based photography todigital photography, photography
to digital photography, andtherefore we you know the whole
group had to be very involved inunderstanding what does digital
photography mean?
How do you standardize thetechnology?
What does a high qualitydigital image offer you?
(02:41):
And then we started workingwith computer scientists, and
that was really before the eraof AI and deep learning.
Right, you know very basicpixels and understanding digital
retinal photographs.
So I would say it's reallyabout 20 plus years ago almost
(03:01):
30 years when we delved intodigital diabetic retinopathy
screening and I think that's thebasis of our current technology
today globally.
David Kerr (03:12):
So what are you up to at the moment?
Presumably everything isartificial intelligence, deep
learning.
Tien Wong (03:20):
So, at the moment, we have taken, I would say, a
longish journey towards usingthe retinal images that we now
take for granted.
It's all digital, right and it'staken from smaller and more
portable retinal cameras, andthat's for primarily two
(03:41):
purposes.
One is, of course, to detectearly stages of diabetic
retinopathy and therefore torefer them, when appropriate, to
the eye specialist for them tohave treatment, usually at eye
centers or eye clinics.
So that's a fundamental part ofdiabetic retinopathy screening,
(04:08):
part of diabetic retinopathyscreening.
Secondly, we're using theseretinal images to look at a
really even newer area in thisfield of using the retinal
photographs digital, which canallow us to look and interrogate
, for example, the blood vesselsin the not just in the eye but
the rest of the body, as well asthe nerve fiber layers.
That's seen in the eye, but therest of the body as well as the
nerve fiber layers that's seenin the eye but again
representative of nervous andneural systems in the rest of
(04:32):
the body, and this whole fieldof using the eye as what I would
say a window or a channel tounderstand the systemic
complications of diabetes which,as everyone knows, you know,
diabetes affects the heart, thebrain, the kidneys and so on and
so forth.
So that allows us to use asingle technology, in other
(04:57):
words a digital retinalphotograph, to not only
understand eye diseases but also, probably, the systemic
complications of diabetes, andthat's a very exciting new field
.
David Klonoff (05:10):
Tian, can you comment on the idea of oculomics
and how it will be used toevaluate and treat people with?
Tien Wong (05:18):
diabetes.
So oculomics is again the termcoined, whereby we are using the
eye images and the informationin the eye, so the omics from
the eye or the ocular structure,and using it to interrogate the
different systemiccomplications related to
(05:40):
diabetes, and even in peoplewithout diabetes.
But let's concentrate ondiabetes.
So, for example, there has beena very strong biological and
historical link between whathappens in the eye and what
happens in the kidneys in peoplewith diabetes.
Almost concurrently, peoplethat have diabetic retinopathy,
(06:02):
but particularly the severestages, would also have kidney
damage, diabetic nephropathy.
So using a single screening, Iwould say biomarker, in other
words, information from theretinal images, might offer you
simultaneous understanding ofdamages seen in the eye as well
(06:25):
as the rest of the body.
So this is the fundamentalconcepts of oculomics.
So no different fromunderstanding, for example, a
gene, whether it affects boththe eye and the heart and the
kidneys.
So what we call genomics orprotein, what we call proteomics
, and this is now what we calloculomics.
Now, the potential of oculomicsis number one.
(06:47):
The retinal photograph isdigital, it is non-invasive.
The technology to capture thisinformation happens in seconds
and could be, in the future,widely available in what I would
say easily accessible settings,For example, in the optometrist
(07:07):
shop.
There's been discussion that itcould be even in a pharmacy, it
could be in the supermarket andit could be in your
neighborhood post office.
So what we are potentiallydoing is to take a single
technology that could be cheaplyand widely available towards
(07:31):
what we really need to do, whichis to have not just early
screening but detection of thedifferent stages of complication
that diabetes affects across.
You know, US as well as theworld.
David Klonoff (07:44):
Ken, in the US, a problem that we've seen is that
even with screening, peopleoften don't have follow-through
to see the eye doctor for thetreatment.
Are you working on some type ofa follow-through program so
it's not just here's yourdiagnosis, but getting people in
for treatment?
Tien Wong (08:03):
And this is the new, exciting era of large language
models.
As most of us in the world noware aware, the large language
models offers what we call verytailored, very specific
instructions that can be used tohelp patients understand their
(08:25):
problem, understand the severityof their condition, as well as
have very targetedrecommendations for them to
either have follow-up, forexample, with the eye doctors or
with their kidney doctors, orto have better control of their
sugars and blood pressure.
And therefore, in one of ourrecent papers, we have combined
(08:48):
what we call the retinal imageswith a large language model
component, so that they extractthe information from the patient
, as well as the retinal images,and give a targeted, specific
advice that might hopefullyprompt the patient to take
better control of their ownhealth situation.
(09:10):
Now, this is different fromwhat was previously, prior to
the era of large language model,prior to large chat, gpts and
so forth, whereby people wouldget a pamphlet, which will be
good advice, but it will be verygeneric advice and it doesn't
really, I would say, tailor to aspecific individual or patient.
(09:31):
And we think that in the era ofAI, digital imaging combined
with large language models, weshould be better able to tackle
this part, which is thatpatients get information they
don't know what to do with it.
They're not motivated to doanything with it and hopefully
this large language model offersthat patient assistance.
David Kerr (09:54):
Tian, this is absolutely fascinating something
that we're very interested in.
My world at Sutter Health herein California, we're also
contemplating whether you cancombine more frequent retinal
imaging with, say, otherwearables such as continuous
glucose monitors or bloodpressure monitoring, and so you
(10:16):
can actually have, from aperspective of time you can see,
changes in the retina whichreflect changes in glucose.
Are you involved in work inthat area as well?
Tien Wong (10:28):
I think a lot of groups are working combining
multiple modalities, andcertainly we have some work with
people on the wearables field,as well as those using what
we'll call mobile technology,and information that feeds this
(11:03):
mobile technology will likelycome from multiple places, right
From the standard doctor'snotes to the laboratory records,
continuous glucose monitoringand, of course, you know, the
periodic visits for us will bethe optometrist and for some
people would be the kidneyspecialist and I think,
ultimately, an individualpatient will, at the point of
(11:29):
diagnosis, have a digital recordof information that's collected
over time from various sources,no different from, you know,
from our own digital history.
I mean, we have our own Facebookhistory, our Instagram history
I likely see a little bit ofthis our diabetes history.
What we are not seeing yet,it's a holistic platform whether
(11:55):
you can call it an app or adevice that is able to extract
these very diverse sources ofdata and to make sense of these
data.
But I think that will likelycome soon.
It's really what we call a newfield, where people say it's
convergent science.
Right, you know, we havescience that is very siloed,
(12:20):
kind of developing on their own,and what we need is a platform
that brings together data andtechnology, ai you know large
language models into somethingsimpler.
You know we need a little bitlike a Steve Jobs iPhone moment
(12:41):
right, whereby we're able tohave that single companion, that
buddy that helps us with ourhealth over time.
David Kerr (12:51):
Essentially and I think that that's something that
you know, we wish to see forour patients and for our
providers- so, just following onfrom that, on a more negative
aspect of that, if you take 100people and you try and look at
the back of the eye withoutdilating the pupils a proportion
of them you just don't get verygood images.
(13:13):
Is that a problem that's alwaysgoing to be there, or do you
think the technology is gettingbetter and better, that the
ungradable or unviewable imagesis going to be zero cameras?
They?
Tien Wong (13:27):
were poor quality images and then newer cameras
have tracking devices, therewill likely be some sort of
automatic kind of artifactremoval.
(13:48):
There will be likely somegenerative parts of the images
that will be synthetic but willfollow the pattern of what the
large language models ispredicting generative AI.
So you don't need all theinformation on everything right
(14:08):
I think you need substantialinformation and it will be able
to produce an image that is asclose as possible to the real
image.
So I think that technologywhereby the new generative AI
algorithms allow will herald anera whereby most of the images
(14:30):
will be gradable and readableand interpretable for everyone.
David Klonoff (14:35):
Tian, one of the most fascinating areas of AI is
agent hospital.
Could you talk about that, whatyou're doing and what the
hospital is all about?
Tien Wong (14:46):
We are starting to think about how do we build
something that is for healthcareusing AI, and there are really,
I would say, three approaches,right?
The first approach is thetraditional hospital, the
traditional clinics.
All of us are very familiarwith it.
They are not nice places tovisit.
They are usually crowded, thereare usually lots of gaps in the
(15:11):
care.
We spend three hours in theclinic and we see our physician
for maybe five minutes, right,so there are many challenges
that happens in the clinic andwe see our physician for maybe
five minutes, right, so thereare many challenges that happens
in the traditional hospital.
So one way, of course, is totransform this hospital by
putting in more applications,having back-end AI tools, using
(15:32):
some front-end AI tools where,before the patients see the
doctors, they have an interview,so-called with an AI chatbot,
right, so those are being done.
What I would say traditionalhospital plus small AI.
Now, is that doable?
Is that going to be successful?
I think there's many, manyexperiments going on everywhere
(15:54):
and we have yet to see majortransformation.
In fact, there's probably a lotof mental fatigue, there's
probably increase in costs,there's uncertainty from
patients and physicians and oneapp leads to another and there's
very little convergence.
So what we are trying to do isto say let's build it from
scratch, right?
(16:14):
Almost a hospital that is builtdigitally from beginning, and
that's the concept of the AIagent hospital, and therefore
that hospital should have, inone sense, co-development by
physicians and engineers, anidea whereby the financial model
(16:35):
or the sustainability of themodel is not dependent on how
many patients come in, how manyprocedures are done, how many
tests are done, but on a modelin which we are able to
integrate the mostcost-effective digital platforms
and artificial intelligencewith as few touch points by the
(16:58):
physicians as possible.
In fact, we envision that atsome stage, half, maybe even
more, of these patients will notneed to interact with a
physical healthcare provider,right?
So you need to start from thatkind of basis, and that's where
we are not calling it an AIhospital, but an AI agent
(17:19):
hospital, whereby the agent isthe primary coordinator of care
within the hospital andtherefore, as I said, it's
envisioned that 50% probably donot ever need to see that
physical human healthcareprovider or physician for that
matter and therefore we hope tohave targeted improvement in
(17:43):
efficiency, of course, indiagnosis and safety.
That's the basis of thetraining the agent, but also
things that we don't really havea good handle of, you know, the
throughput of patients, theshortness of the waiting times
in the hospital and, ultimately,whether they even need to come
(18:04):
to the hospital, because we canimagine that the AI agent
hospital a lot of those carethat's provided will be via
mobile devices, will be viatelemedicine and in their homes,
right.
So that requires a completethinking.
The AI agent hospital reallyneeds a concept that we do not
(18:29):
yet know, but we are trying tobuild.
David Klonoff (18:33):
Tian.
One topic that we've talkedabout is how do you get patients
to trust the AI If they're notseeing a doctor.
The patient has to really findthat it's trustworthy.
Tien Wong (18:44):
Yeah, I think trust is a very key component of the
entire AI ecosystem.
I don't think we pay sufficientattention to that.
We usually pay number one, thetechnology development.
Right, how robust?
How big is the data set?
That's training the algorithm.
We looked at another aspect,which is the clinical outcomes.
(19:08):
Right, the clinical outcomes.
Is it efficacious?
Is it doing what it's supposedto do?
Is it picking up the disease?
Is it missing cases?
But in between is what I wouldsay a very important area, which
is on this entire system oftrust, and you can say it is
behavioral, it is psychological,it is experiential and it
(19:35):
involves multiple stakeholders.
Now, how do we define trust?
You can say that wetraditionally have put a trust
in the patient-doctorrelationship.
Now you are adding a thirdparty involved in this
patient-doctor, so you need tobuild a trust in this
three-party relationship.
(19:55):
It's not that easy, right?
I mean, we know who we like,who we can interact with.
Do we want someone else in theroom?
Right?
And I think that, therefore,that trust needs to be from a
multi-faceted approach.
Number one patient needs totrust.
And what does the patient trustmean?
The doctor needs to trustbecause you know there's another
(20:18):
person in the room.
So there's also a healthcareprovider trust relationship, and
then the system needs to trust,because the system needs to
know that we're building asystem whereby it's not just a
two-way interaction between thepatient and the healthcare
professional, but another partin the system, and the system's
trust could involve manydifferent things that we are not
(20:41):
able to see.
So I think that a whole idea oftrust is something that we need
.
Other groups of stakeholdersinvolved in this AI relationship
, and currently I see two majorstakeholders the computer
scientists, the engineers andthe physicians who want to use
the technology.
We need that third group in theroom.
David Kerr (21:03):
Tian.
I have one final question, justto bring it back to the retinal
imaging.
I was brought up that theretina is the gateway to the
soul.
If you move beyond diabetes andthe complications, what other
diseases do you think are goingto be detectable at the back of
the eye at a very early stage?
Tien Wong (21:24):
going forward, I think that one of the most
promising aspects beyonddiabetes, is using the eye as
really a marker of aging andbrain health.
I think these are the two mostimportant things we are now
interested in healthy aging.
(21:45):
You can say it's longevity orit is better quality of life as
we age, and I think the retinaoffers a lot of promise for that
.
For example, in one of thealgorithms we have developed,
what we call a retinal-phenolage.
It's really an age of theperson based on the changes or
(22:11):
the health or the damage of theretina.
I will have to tell you thatwhen I did that algorithm, my
retinal age is five years olderthan my chronological age, which
is not a good sign, essentially, right.
So I'm kind of older than myown timeline should be by
chronological age.
I've seen people that haveretinal age that is five or ten
(22:34):
years younger than theirchronological age.
So, in other words, thechronological age it's just a
time, it's a number, whereas theretinophenol age is a
biological marker of our bodyessentially, and I think that
that's a very interesting andvery exciting era as we move
(22:55):
into.
As I said, the interest of manyolder people of our entire
world is interested in how do wemaintain healthy aging and
longevity in this very complexworld that we live in?
David Klonoff (23:13):
Ken, thank you very much for discussing AI,
retinal health, oculomics andsomething I had not heard before
, which is the health age fromthe retina.
I'm going to read about that.
Thank you for joining us.
We hope to work with you in thefuture.
And now I'm going to saygoodbye from Diabetes Technology
(23:33):
Report.
We're available on Spotify andat the Apple Store and at the
Diabetes Technology Societywebsite.
We look forward to catching upwith you at our next Diabetes
Technology Report.
So for now, goodbye everybody.
Goodbye.
Tien Wong (23:51):
Thank you very much.
Popular Podcasts
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.
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.
Betrayal Season 5
Saskia Inwood woke up one morning, knowing her life would never be the same. The night before, she learned the unimaginable – that the husband she knew in the light of day was a different person after dark. This season unpacks Saskia’s discovery of her husband’s secret life and her fight to bring him to justice. Along the way, we expose a crime that is just coming to light. This is also a story about the myth of the “perfect victim:” who gets believed, who gets doubted, and why. We follow Saskia as she works to reclaim her body, her voice, and her life. If you would like to reach out to the Betrayal Team, email us at betrayalpod@gmail.com. Follow us on Instagram @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.