Ophthalmology: Mass Eye & Ear • Rishi Singh
Selective Senses
What if some of medicine’s biggest opportunities lie not in inventing new treatments, but in questioning what we already accept as standard practice?
Dr. Rishi Singh, Chair of Ophthalmology at Mass General Brigham and Professor at Harvard Medical School, joins Nic to explore the evolution of ophthalmology through AI, imaging, oculomics, precision medicine, and remote care.
They examine how AI can democratize expertise, how the eye can serve as a window into systemic disease, and why longitudinal data could replace “one-size-fits-all” treatment.
The deeper lesson: innovation happens when technology changes not only what we can do, but what we consider possible… and what we stop doing simply because it has always been done.
-
Nic (00:00)
Rishi, welcome up to UnNatural Selection.
Rishi Singh (00:03)
Thank you for having me, Nick.
Nic (00:04)
This is truly a pleasure and an honor to have you here. you you have a a long and distinguished career in and publication record. So I'm really looking forward to first of all learn more about ophthalmology. I'm not an expert in the field, I know very little about it, so I'm coming at it from a very external standpoint. But before we get going, could you please let us know what need or impact drives your work?
Rishi Singh (00:28)
Well, I think Nick, the need and impact that really drives my work is why I got into the field of ophthalmology in the first place. I I remember back when I started twenty plus years ago, researching in a laboratory, actually back in undergraduate and then in medical school, realizing that there was very little therapeutic options for patients with eye diseases. many of the things were applied from other areas, but really there wasn't any any specific therapeutic options for
The most common causes of blindness, macular degeneration, diabetic retinopathy, uveitis. And so while we had accomplished a lot maybe in the cataract side of things, where that is the global leader of blindness, in the industrialized nation, these areas that were that I just mentioned to you were not really attacked with newer therapeutic options. So it was a field in which I knew had a huge impact from from the standpoint of possibly innovation, but also realizing from welfare. When you talk to a patient who has vision loss, you hear about the fact that.
It's everything to them. It's sometimes worse than a cancer diagnosis, sometimes worse than surviving a heart attack, and sometimes even worse than losing a limb. So realizing that that is such an impact on quality of life was really my calling to the field and really bringing new therapeutic options to patients is something that I have really focused on in all of my career and insights as far as figuring out who should receive therapeutic options and maybe who.
might be more benefited by these options. Really tailoring therapy has been something that I've worked on for the past twenty-five years.
Nic (01:54)
That's fascinating. And you mentioned then there that, when you you one of the things that drove you to this was the the shortage of innovation in the field. can you tell us a little bit about what was driving that shortage of innovation?
Rishi Singh (02:06)
Well, I think it's just science in general. I mean, I think that when you look at the field as a as a totality, obviously basic science has really propelled us in many ways, the chemical understanding of some of these diseases. But also because of the ability to capture data in these patient populations was challenging at best. many of them weren't necessarily showing up. Vision loss is not something that somebody necessarily complains of right away. It's usually after the most advanced stages step in.
So it was the lack of that. Also technology in regards to imaging. imaging has been a huge boon to the insights around ophthalmology. There's a technology called optical coherence tomography. It's applied in heart, it's applied in dermatology, but in ophthalmology really has where it's been a huge impact. It was developed here right at MIT, actually, next door in Boston. but that has led to a lot of insights as far as disease pathogenesis and prognosis over time.
And hence allow us to image the eye in a much more different fashion and pick up disease more readily than we ever had in the past.
Nic (03:10)
Interesting. And I think you said right there that one of the reasons is because people typically don't complain about eye problems until the whatever is causing symptoms is fairly advanced, right? Is that is that the case that I actually hear that right?
Rishi Singh (03:26)
Yeah, and that's the case because I think Nick, unfortunately, we didn't have maybe the ability to screen large populations. I think also when you talk about some of these conditions, aging, it's something that I think people take for granted that they say, Well, I'm aging and therefore my vision should decline. I have diabetes and my vision should decline. those are not common situations, or I have a family history of blindness in my family and therefore my vision should decline. Those are sort of acceptable.
norms for the population, at least back 20 years ago. Now I think a lot of the population has been become educated because of the community that we've created around this, the therapeutic options available. They hear about their friends coming to us and realizing they can see finally after therapeutic options are delivered. These are all very, very important things to change the norm. And I'm sure it's the same in cancer or in other fields where people sort of accepted the fact that once you have a disease, maybe it's heart disease, maybe it's cancer,
That was sort of the next step of of whatever that progression was. But we can change the norms and we certainly have changed norms and ophthalmology in the past few years.
Nic (04:33)
Yeah, my background. So as we mentioned before we started recording, I run the International Consortium on Newborn Sequencing. So there the idea is how do we bring genomics at the to the point of care during during the, you know, birth, right after newborns come to life and come to this earth. And the idea there is how do we use genomics in a way that we can use it for screening, identifying disease, but then potentially preventing things in the long run. And so on the genetic front, very much to what you're saying, it's you know.
Decades ago, if you had a genetic disease, it was almost a life sentence of which you just had to accept. And then, you know, a few decades, maybe a, you know, two decades ago, we could put a name to a disease and maybe do something about the diagnosis and maybe some treatment. And now we're getting to the phase where, you know, with CRISPR and other technologies, we can actually potentially alter the course of a disease. So a genetic disease isn't necessarily a life sentence. Now it's something that you can potentially control. and it's changing the dialogue.
In
the the narrative around some of these rare diseases. It sounds like on the ophthalmology front as well, people just accepted it, which I actually I can identify with because I only started wearing glasses about two years ago. fuzzy vision for a long time, then I realized I had astigmatism, and so they gave me glasses for correcting that. but yeah, very much I had that mentality, which is like, well, at some point I'm just gonna lose my vision, or not entirely, but I'm gonna start losing.
some of my vision as I get older. so for all I know I could have had something far more serious that I overlooked and ignored because I thought, well, at my age of course I should lose my vision.
Rishi Singh (06:13)
Mm-hmm. Yeah, it's the same in cataracts in third world countries, right? As you as people get older, you know, those who are breadwinners then became dependent on their younger family members. And so I remember when I was a a a a resident going out to India and operating on patients with cataracts, one of the things they would tell me is that they became so dependent on their family members for their livelihood, their meals, their money that they might be earning. And they felt burdensome, actually.
And that was one of the most largest, I think, pushes for me and doing what I'm doing is I remember those people telling me how burdensome they had become to others. And with a simple cataract surgery or some sort of procedure that can be done, you change your livelihood altogether. Now that they're working age adults again, or working age individuals, they can actually do things. They can contribute meaningfully to the population. that's really where the welfare and ophthalmology can change everything.
Nic (07:10)
Are there are there genes that code for certain eye diseases that you can actually screen for before and you talked, you know, this part of the conversation has been about detecting things almost when it's too late. Are there preventive measures or early screening measures that you can do before people start picking up symptoms?
Rishi Singh (07:28)
There's some. I think they're isolated. there may be some her hereditary retinal degeneration sort of diseases that we know that run in families or have there. But even within those genetic issues, they're not a full penetrance. And so just because you have a genine mutation or gene deletion of some sort doesn't confer that you'll have the disease itself. It may mean a higher risk of progressing, which might mean improved screening mechanisms, but that still hasn't really been borne out.
I think the biggest thing that people can do, and and this is something that all of us have been become more accustomed to, is really seeing a eye care provider, and there's many of them out there, optometrists, ophthalmologists, people like me who are highly specialized with the field of ophthalmology, retina specialist. Each of us has sort of an ability to look at the eye from the front to the back and look at various areas of the eye to look for diseases.
And I that a knowledge base has become far more acceptable and commonplace for many of our individuals in those fields. So any of us can screen for eye disease at this given stage, thankfully, through some of the technological advancements like called coma measurement devices, which we had about 30, 40 years ago. And now we have the OCT device I mentioned to you now. those are sort of really huge advancements in our field that have led to that that process of screening.
Nic (08:45)
And if I recall my biology class as well, the r the the retina is a direct connection to our central nervous system, right? It's actually part of the brain, an extension of it. the do I
Rishi Singh (08:54)
Yeah.
Nic (08:55)
have that right?
Rishi Singh (08:56)
Yeah, I we consider the eye as far as brain. you can see
Nic (08:59)
Yeah.
Rishi Singh (09:00)
strokes, you can see cardiovascular disease, you can see cancers. all of those things can appear within your eye. And so that can be one of the first signs. We we're learning more, in fact, the past couple years about diseases like Parkinson's and Alzheimer's. We're actually measuring eye tissue to look at those portions of the eye that which may be indicative of Alzheimer's or progression.
And even MS, MS neurologists right now do OCTs of a specific layer of the retina to look at OCT progression in those patient populations who are under therapies for MS related diseases.
Nic (09:38)
That's fascinating. I that's that's exactly where I was going with that question. So we're it's it's possible to screen or use imaging from the retina to identify non eye related diseases, neurological disease, systemic diseases. You're you're finding patterns that are actually start indicating other possible diseases that or throughout the body.
Rishi Singh (10:02)
Absolutely, Nick. And and you know, I think some of the work that's really been driven has been AI related. And one of the most exam best examples of that has been Google's work with Verlilly. And they've done some work just on simplistically looking at even things like gender. they can actually determine gender by looking at your eyes.
Nic (10:19)
Really?
Rishi Singh (10:19)
but they also looked at vascular changes in those eyes and been able to pick up hypertension and early signs of diabetes. There's been huge data sets published by them on.
some of the work they've done. Others have done the same work. We now have AI screening in clinics. So we can do that as part of a primary care eye visit. now and a primary care visit, I should say, with your medical doctor. We're deploying cameras within our system to actually pick up diabetes without using an AI based camera system.
Nic (10:48)
That's fascinating. And how how receptive is the field to AI? Obviously, whenever a disruption like this comes around, it's it's disruptive to society in general, but then some people across the adoption curve, some will be the early adopters and the innovators that'll just feverishly want that technology. And then you get your like late adopters and laggards that are gonna be much more resistant to that. And so I can see some people across the spectrum thinking in terms of like, wow, yes, give me access to this.
Whereas others might think like they might feel a little threatened now when AI is starting to review scans or records or any number of data that it can potentially make analyses and determinations faster or a greater scale than people. So is AI generally transforming the field at scale or is it still in small pockets?
Rishi Singh (11:41)
No, it's absolutely gonna be transformative at scale. I think the acceptance and the education is something that all of us struggle with. I I you know, it used to be, Nick, and I'm not aging myself, but back in the day when I was a parochial school kid, they used to say, Well, your next language you need to learn is Spanish and Chinese, because the entire world economy is gonna be either speaking Spanish or Chinese. The language now, Nick, is absolutely AI.
And so one of the the opportunities for me as the chair of ophthalmology at Mass General Brigham and Mass INEAE is to educate and to make our department fully AI enabled. what does that mean? It means that everyone sees value in AI and learns about the language of AI and finds ways of incorporating it into their daily work. there's some really simple examples. There are more higher technical examples of this, but it can really be leveraged for all groups. And I think that the
field is realizing that A we can't do it alone. so we need the bandwidth that AI can offer. B obviously it speeds up care delivery. And so hence that's an important part of what we're doing. And C it really reduces the minutiae the things that we always used to spend hours or technically difficult things to do. Is there an opportunity to do that? And the last piece I'll say about that is it also allows for equity. You know one of the the
The first patients I saw when I came here was a s a Spanish speaking patient who was diabetic who came to her eye doctor or I'm sorry, her to her primary care doctor for 10 plus years. And she faithfully came and her Spanish speaking physician was able to take care of her diabetes. But what they didn't a well weren't able to communicate to each other was about her need for eye screening and eye appointments. And obviously it's another appointment for someone to come in. Well, he happened to have one of those cameras.
In a community-based setting. He screened her on the camera. The camera has AI in it. She didn't have to do anything because it's an automated robotic camera that takes photographs of the eye. You just have to sit there and it does it for you. And technically, skill is not really a needed for this camera. And it determined she had the worst form of diabetic eye disease, which required laser treatments. And so we were able to save her vision. and she would have suffered catastrophic vision loss if she didn't come to see us.
But it was through AI and through the use of this camera and deploying this camera and having the screening that she's able to continue to see today. And I just saw her in clinic actually just last week and she's doing phenomenally well with great vision. And we've been able to save her vision. But in the normal circumstances, these are the people that went blind. These are people that caused economic burden to their families, to their to their communities. think about the population we can create of people if we help them this way. That'd be amazing.
Nic (14:27)
that's extraordinary. And yeah, yeah, it's I I would have imagined that with the extraordinary amount of data that ophthalmology deals with, that AI would be transformative. And where where on the scale of disruption and innovation, would you compare ophthalmology with other parts of the healthcare ecosystem? Do you think that it's on the leading edge of adopting AI and modern technologies or is it surprisingly behind where where where would you categorize
it?
Rishi Singh (14:55)
Yeah.
Well, I I mean I I guess progressively I I always think we're behind a little bit, but I I actually have to compare it to other fields of medicine. I think we're a little bit ahead of most other fields of medicine, to be honest with you. I think we're probably close to radiology. I mean, if you look at the algorithms that were approved, the first D FDA approved AI algorithm was an ophthalmology algorith algorithm,
Nic (15:13)
Mm-hmm.
Rishi Singh (15:14)
actually. So we we led the field and obviously using it for very, very good use case scenarios.
the radiologists have been really great about working alongside with this because many of their protocols and what they're sort of doing in multi-disease state algorithmic development is pioneering. And we also are doing the same in ophthalmology. So trying to develop some of the standards and some of the FDA approvals and the regulatory guidelines that go along with that because a lot of it is based upon those those issues alone.
have been sort of figured out by both ophthalmology and radiology. So I would say we're on the bleeding edge of it, although you always want progress to occur quicker. I think we are on the bleeding edge of much of what is gonna occur right now.
Nic (16:00)
And yeah, it's w what are the kind of things that excite you the most that you think we're like on the verge of potentially solving or curing or treating that the average person is not aware of?
Rishi Singh (16:15)
Well, I I think that the idea of really developing precision related diagnostics, meaning that, you know, all patients are different and we have to learn who's gonna progress and who's not gonna progress, who benefits from a therapeutic option versus a patient who needs to be progressed to another therapeutic option is something that's really within our horizon. right now we're on the cusp of of of of as many medications come out, first and foremost, we apply them plain vanilla to a lot of patients, right? We we give them out.
The the protocol says every six months or every three months or every month. And you have to follow what's on the label. And that's what determines the approval process through our clinical trial. What we're realizing is that there's some people that are super responders, some people that are medium responders, and some people that are poor responders through a lot of different studies that are done. So can we call this information together to understand better what the natural history is of these patients? Can we personalize the treatment approach for those patients? So they don't end up
you know, spinning their wheels for six months or a year before getting a newer therapeutic option. That's the type of thing that we're really trying to advance in the field. And I think that that's some some work that AI really will help us with. And the second that is probably least touched right now, but is happening is the field of robotics. Much of our our surgical approaches have all been manual. It takes a very, very heav fine, proficient surgeon to do some of this work. We're talking about
microns of tissue being removed from people. That's how thin the tissue is that we work with. So if you think about that on a precision scale, there obviously is the human factors of which we all suffer from. Could robotics and could AI help us with this? And the answer is absolutely yes and equivocally it is starting to have that process. It's probably many years down the road from now, but it's something that could be really valuable to our field and really make care delivery
that much more equitable and available to most people, rather than having to go to specialized centers and sort of do that work in a certain area. So that's gonna be a really great game changer for us, but it's probably many years down the road.
Nic (18:22)
Hmm. And you you mentioned there kind of stratifying patients to see which therapeutics would work better for some versus others. earlier in the conversation you mentioned that genetic markers are spotty, you know, so it's whether they're, you know, how effective they're gonna they're gonna be, the the penetrators, and whether a patient will actually get a disease or not. What are the kind of markers that are you're using for ophthalmology to try to do this kind of precision?
therapeutic
approach. Is it kind of pharmacogenomic base or is it looking at other types of markers?
Rishi Singh (18:53)
Yeah, so there's a a terminology called oculomics, which is really where we're looking at biomarkers within the tissues and the images that we take. And those oculomics can then indicate us to a better or worse prognosis as far as outcome. AI has enabled measurements of some of these very small biomarkers within the retina, for example. We can measure the size of a a certain layer that's missing from the eye.
That then leads to precision as far as whether you're going to be a sub-threshold or super therapeutic responder to a medication as an example. We've learned that recently from a clinical trial. the same is happening of oculomics in the sense of it being the window to the brain, as I mentioned to you earlier. You know, neurologists currently use our technology for MS-related follow-up. And that's an area we can look at because if you put somebody through a serial MRI every time, obviously it's it's very time consuming, it's expensive.
But think about the fact that you can put them through an OCT or optical coherence tomography test, measure the layer of the retina, which is readily available, and determine if they're having progression or or or worsening of their MS-related disease. This is again where oculomics and sort of determination of titration of medications can be very valuable. a quick example I can also give you is some a drug called Plaquinil. Plaquinil has been available to rheumatologists for years.
one of the things that we have seen is that plaquinol can be overdosed and there can be toxicity. It's through ophthalmology and rheumatology working together that we figured out there's optimal doses of plaquenyl. It's not just of plain vanilla, one size fits all. And we've actually challenged the paradigm. And I think many patients are seeing value and benefit to reducing their dose of plaquenyl and not having ocular toxicity or vision loss and being able to follow up over long periods of time.
So that's really been some of the areas in which I think again, working together across the field has been very valuable to pro to improve patient prognosis and outcome.
Nic (20:56)
Yeah,
so what do you think are the biggest obstacles to in barriers to innovation and ophthalmology? Is it the science? Is it regulation, reimbursement, clinical adoption, implementation, economics? Where do you see like the biggest kind of friction happening in being able to develop these therapeutics, diagnostics and so on in the field?
Rishi Singh (21:20)
Well, some of it is is certainly driven by the data sets we have. you know, they've been curated, there's larger data sets every day. I mean, the data is definitely there, but one of the challenges is just managing and storing this data data as it comes out and then figuring out how we share it across platforms. Certainly that's probably one of the largest challenges. The second is that, you know, as we all look at the US healthcare system, and this is not just universal for what we're talking about here, but
In a fee for service model, there is less motivation to do things that would be of outcome benefit. And I think
Nic (21:54)
Mm-hmm.
Rishi Singh (21:54)
that we've seen some movement in the healthcare field, which I'm very much in favor of of improving outcomes, showing the at best outcome for patients, that just not it's we give patient therapeutic option A and we see the benefit potentially, you know, six months or a year down the road. Is there a way of accelerating these outcomes? And then is there a way of again moving the healthcare system from a fee for service model to an outcome based model?
That would be a value. And I think that that's where this field can be very transformative at helping both payers and also physicians get to what they ultimately want, which is a great outcome, right? We all want our patients to see, we all want our patients to do better. But is there a way for both of us to win at this game where we maximize the therapeutic options, but also minimize and and prevent those disastrous outcomes and improve the outcomes for all patients involved?
So there's probably those biggest challenges, I would say, from both the fee-for-service model and managing large data. I think the regulatory pathways actually become far more reasonable now than ever before. There is precedence set than ever before on what it takes to get to that level. The FDA has been and other areas of the or agency have really really helpful as far as outlining what is necessary. They've done run cross-disciplinary discussions with both technology advancers as well as researchers and scientists.
I think that that's been sort of looked at already and figured out. It's these other pieces that need to be solved over time.
Nic (23:18)
Hmm. Yeah, yeah, the reimbursement model in our healthcare system is leaves a lot to the imagination, especially when it comes to prevention, right? like you know, like you said, it's it the the you know, when you're if you're looking for outcomes, that's obviously a much better way than if you're just treating disease. And genomics definitely suffers from that. And you know, you have to come up with economic models to try to justify getting reimbursed for especially on the preventative side.
I could totally see how the economic argument could be made for society. If you keep prevent people from losing their site, then that obviously benefits society because now you don't lose workers and they don't become a burden to their families and their communities and so on. Is there an economic argument to be made to the payers as far as like if I do tr if I prevent this, it'll save you money? Or
Rishi Singh (24:12)
Yeah, great, great commentary, Nick. And absolutely there is. a patient who has vision loss is and legally blind, let's call a legally blind patient to the actual healthcare system is three times more expensive to the healthcare system for insurance than a patient that who is cited. And so that's been a well established kind of num no numerical value we've seen across different studies, across different disease states and ophthalmology.
And that's something the payers can really dig into. I think the challenge is that in looking at the preventative care model, sometimes it's not always there. they I think they've changed their tune. I'm not gonna, this is not a a time to necessarily hurt our our payer colleagues in this conversation. Actually, I think they've come around to this conversation. The the whole screening discussion we just had about AI that was incredibly enabled by a CMS code that came out that had a reimbursement model associated with it that.
places like ours are using and others are using to kind of pay for the industru of what has to happen in a screening based program. and it seems sufficient to to run the program and to certainly pick out and weed out blindness in our patient population. So that's an example of where it worked really well. so so there are bright spots happening. I think that they're just taking time to develop. And once they develop fully, I think we're gonna see a huge benefit from those.
Nic (25:35)
I
totally agree. And you know, I think innovating on the economics, we we talked a little bit before, you said some of the limitations in this field would be things like the access to data, storage of data, and being able to exchange that data across centers and so on. What are the what what other areas do you think as you see technology evolving these days and LLM technology, AI technology's really revolutionized so much of life sciences, healthcare, all the knowledge-based
fields. But if you were to wave a magic wand, or maybe even point certain young entrepreneurs out there that are eager to try to come up with new and exciting technologies, wh what do you think are the kind of solutions that aren't available yet that you think would drive material impact in this field that you would
Rishi Singh (26:25)
Yeah.
Nic (26:25)
and again, you could just envision and say like if somebody could solve X, Y, and Z, what what would those things look like?
Rishi Singh (26:33)
Yeah, I think, you know, developing tools that are remote-based, maybe even smartphone based for some of the activities would be a huge benefit. You know, could you do some of these tests at home and monitor some of these patients from remote areas that you would have to come to see an ophthalmologist? You know, one of the things that I'm always impressed and perplexed by is that some of the patients who came in, just like this morning, I saw a patient. He had driven in from New Hampshire to see us down in Boston and left his house at five thirty for an eight thirty.
So, so you know, that that tells you, you know, patients are coming from far distances to see you. And and yes, we are experts at what we can do, but is there an opportunity to do some of this stuff remotely? I think that that's the opportunity. I think, you know, there's obviously areas in which we're learning more about how we can deploy technology. So one of the examples is, you know, our emergency department at Mass Sign Ear is one of the busiest emergency departments in the in the country for ophthalmology.
And certainly it's because we get a catchment area from many of the states in and around Boston in Massachusetts that that send our pay their patients there with highly complex situations. These are patients that showed up, for example, at an emergency department, saw an emergency doctor, and the doctor said, you know what, you've got an eye problem that I can't take care of. So go down to Boston for your next evaluation. And we have a 24-7 emergency room that works there. Can we deploy technology to some of those places so we can do some remote screening for them?
So that it prevents the patient from having to show up urgently and determine why they're there in the first place. We just conducted a study that found that 59% of the visits we saw at this emergency room really weren't ophthalmic urgent, meaning that they didn't require urgency for evaluation. Certainly they were meaningful to take care of, but could they be done 12 to 24 hours later in a in an office set-based setting versus an ED setting? That's a huge value proposition for everyone
Nic (28:23)
Mm-hmm.
Rishi Singh (28:24)
involved, including
you know, the systems of of healthcare. So these are some areas which I think that we're gonna learn more about how we can democratize this. And that's the kind of use of I used the word equity before, but I think the other word to use would be democratize this. Allow for people to have the same ability to have ec expertise and have certainly linkage to an expert like myself and others who can do this sort of evaluation remotely is going to hugely transform the field. And it's something that we're all working towards
Nic (28:54)
I could totally relate to that. On the genomic front, there's always a shortage of genetic counselors and and experts. and so some of the thinking here is can we you kind of apply an 80-20 rule, right? Is there like a 20% that you can equip primary care clinicians to be able to deal with? And when it goes beyond that, then obviously refer them to a specialist. And I think you can even do more now, especially with modern technology and how you apply that to empower.
primary care clinicians to be able to address some of these things more in the office before you send them off to a specialist. It sounds like ophthalmology is right up that alley.
Rishi Singh (29:30)
No, absolutely. And I think that, you know, we look for care delivery models to become more diverse. We're not immune to the fact that we might not have these things in our office. We're willing to work with partners, colleagues, other clinical providers to help these situations be delivered to their offices.
Nic (29:49)
So if we if we were to think about where the field is going, and obviously like you know, when I talk to people about AI, and I'm very, very deep in that field right now, AI and life sciences and healthcare, it's almost impossible to think about where this technology is gonna be next week, let alone a year from now. But but that said, as you've seen the trajectory of ophthalmology and how it's evolving, then my guess is that
that pace of innovation is increasing because it's basically increasing everywhere. So the acceleration is going up. But if you could project forward, say 10 years or so, not guessing where AI is, but the field of ophthalmology generally, what what are the kind of implications you think technology's driving in this field and what that means for society or for people, especially from a lay perspective, just average person thinking about what their
Rishi Singh (30:40)
Sure.
Nic (30:40)
you
know, if if glaucoma or something runs in their family, how might they be thinking differently in terms of when they hit forty, fifty, sixty years old?
Rishi Singh (30:49)
Yeah. Well, the good news about my field is a lot of it is longitudinal nature and a lot of it is imaging based. And so image interpretation, automation of image interpretation, automation of longitudinal assessments is very important. you know, I I've I've said this before and I think it's true of most of medicine. Most of what we do in medicine actually hasn't been proven. somebody told us, maybe it was a colleague, maybe it was our mentor, maybe it was our attending when we were junior residents or whatever it was.
this is the way we should do something. And the the fact or the reality of that is some price in between what is really truly a fact and what is reality, right?
Nic (31:27)
Mm-hmm.
Rishi Singh (31:28)
And I I actually ask this question a lot of of my faculty sometimes is, you know, we obviously are very well healed in the sciences. You know, what is really evidence based in what we're doing and what is truly just a pass down from generations or a thought around the field? Now, you know, there's always challenges with that because community standards sometimes base themselves in that. And I'll give you a quick example of that.
Antibiotics following a a surgical procedure. Sounds standard, sounds routine, sounds like something you'd want to have.
Nic (31:55)
Mm-hmm.
Rishi Singh (31:56)
Does it actually prevent infection? Does it actually change the outcome of what we're seeing? are there studies that have validated where doing these sort of interventions matter? The the reality is that truly there's been no studies around this. And somebody told us a long time ago, every patient knee-jerk after a a procedure should get an antibiotic. And
Lo and behold, we've done some of these procedures now 40, 50,000, 60,000 times without any antibiotics at all and seen no rates of infection that are higher than others because it's an immune privilege kind of space we're dealing with. So there's there's, I think, components of this that I think are valuable to see, where, you know, if you look at like some of the care acceleration that we're doing, is really in like data interpretation, data analyses, but also best practice sharing.
And kind of really validating what is out there and what is really, you know, potentially those things I talked about, which are passed downs versus real. And that's where I'm excited about the feel because I think once we see that, it makes it that much easier for anyone to really do a good job of taking care of some of these patients.
Nic (33:00)
It's so true. I spend also a lot of time in the maternal health space and there there's so many practices that we think are best best practice and yet they've been passed down for generations. often anecdotally, sometimes via studies, but when you actually look at the the details of the study you realize there are some significant flaws.
And how they came
Rishi Singh (33:25)
Yeah.
Nic (33:25)
to those conclusions and yet we still take them as being almost like law that you should do these things.
Rishi Singh (33:31)
Yeah. And and just to add to that, you know, today I was listening to a talk of an individual talking about you know, blindness rates in sub-Saharan Africa. It turns out that young babies that are born prematurely are really at risk. And, you know, the challenge is obviously the the typical screening trained individuals to do all this sort of work. but you're also dealing with the knowledge deficit. And here's the big one, which I I didn't pick up on until I heard this talk today.
you don't even know the gestational age of the baby that's that's present there, which is a huge biomarker thing to follow, right? We know
Nic (34:05)
Mm-hmm.
Rishi Singh (34:06)
the size of the baby when they're born. We know some of their other, you know, sort of comorbidities when they're born. But if you don't know the gestational age, which is something that seems so basic to us as as in the healthcare industry within the industrialized nations, that we can't figure that out in areas where healthcare is sparse or or others, this is where AI can really be helpful. Maybe it's
Cranial facial recognition and determining, you know, phi physical development and sort of what the gestational age is. maybe it's something about their hand prints or where their fingerprints develop over time. Is there other ways we can determine gestational age to help us? Because we know gestational age is so important and many different diseases, including eye diseases. And this is just an example where we can look at this and sort of really be insightful.
As far as how we're looking at some of these activities and technologies.
Nic (35:00)
absolutely. You don't even have to go very far to see that problem because I know that in some marginalized communities, especially people that speak different languages as a first language, a lot of times mothers pregnant women will show up at the ER when it's time to deliver and they haven't had any care. And so at that point it's like, you know, good luck trying to guess the gestational age. So you don't even have to travel very far outside of Massachusetts actually to see that problem.
Rishi Singh (35:26)
Yeah.
Yep. No, I agreed. And and wouldn't it be great to figure this out for the population? I mean, that's where this all could lead to.
Nic (35:34)
absolutely. Well, and especially because we're living longer, right? And so in theory, issues with eyes and ears, what we'll start seeing more of that as people live beyond seventy six, seventy eight, eighty and beyond, we're only gonna start encountering more age related decline in other diseases that are gonna start becoming more prevalent, th I I would imagine.
Rishi Singh (35:57)
Absolutely.
Nic (35:58)
So so Rishi, this has been an incredible conversation, really enlightening for me for someone that doesn't know very much about ophthalmology. you're you're you're young, you're at the prime of your career and clearly nowhere close to being done, and you've accomplished so much already. But I always like to finish with this question just to kind of think about what you know, what goals or or what would be a best case scenario for you. But if you could picture yourself sometime in the future reflecting on your career.
current and maybe things that you're still working towards, what do you think a best case scenario would look like where you feel like you've achieved the fullest impact that you were capable of? Something that would really make you feel fulfilled for what you would have accomplished throughout your career.
Rishi Singh (36:40)
Yeah, Nick, thank you first and foremost for the opportunity to come talk to you and your audience today. I I would say to you that it goes toward two things for me as an individual. the first is that, you know, I've always marked that I I would like to make a contribution to the field, and I think I have through my time frame here, and I will continue to do so. And contributions take many different ways. And I I say this in all kindness to my colleagues who probably get Nobel Prizes and sort of advanced fields in that way.
They do phenomenal work and I I've changed the practice of of medicine and some in in my realm of things and what I have found from an insight perspective. The other way I've I've actually impacted the field and I continue to impact the field, and this is the one that I'm gonna hopefully leave behind to others, is a legacy of individuals that have trained with me that I've been able to inspire, they're able to lead, whether they're trainees and fellows that I spend time with.
through my practice who say that they're my disciples or or people that worked with me. I I don't like the word disciple per se, but I I think it's all kind that they refer to me as their their primary teacher. I I think that that I try to leave on the impact side of it that they are asking questions. And that to me is a a bevy of people that are asked more questions about the the whys, the hows, the can we do betters?
that's that leaving that legacy of people that potentially have that ability to do that work is kind of what's what's gonna really define our field. And that's what I'm most excited by is if I can leave a legacy of those communities behind me in this field of ophthalmology and maybe cross pollinate into other fields because of it, what wouldn't the world be a better place because of it? And that that's really what I'm excited about.
Nic (38:20)
yeah wise words and you know after this past hour of talking to you in my own small way I consider myself your student as well. So with that, Rishi,
Rishi Singh (38:29)
Yes.
Nic (38:30)
thank you so much for your time. It really has been a pleasure. I hope to see you at the World Medical Innovation Center and thank you for being on UnNatural Selection.
Rishi Singh (38:38)
Thank you for having me, Nick.
