“When a phone can recognize my face, why can’t software recognize a polyp?”- Dr. Sravanthi Parasa (Interview)
“When a phone can recognize my face, why can’t software recognize a polyp?” – Interview with Dr. Sravanthi Parasa
Every so often an interview reminds me yet again that the future is coming at us faster — much faster. I’m still wrapping my head around what gastroenterologist and AI researcher Dr. Sravanthi Parasa told me: there would be a “1,000 algorithms” in 3 years in GI.
This is big implications for the business of gastroenterology. The AI will indeed operate and guide us like a self driving car. In such a world, how would your role evolve as the endoscopist? What would happen to all the PE valuations that are based on GI manual productivity? How would the workflow change when you combine liquid biopsy and AI? Would your practice prefer hiring an AI vs. an average endoscopist? What if the insurance mandates the use of AI?
Welcome to GI 2.0.
Here’s the interview.
◘ Dr. Sravanthi Parasa’s journey: At the intersection of computer science, medicine and gastroenterology
◘ “I thought, when my cell phone can recognize my face, why can’t a computer software recognize a polyp?”
◘ “My goal is to understand what technology is out there, so that you can bring that technology into medicine and gastroenterology so we can make our lives and patient lives better.”
◘ “I don’t know of any gastroenterologist yet who does only AI as their real job. I wish I was there.”
◘ “Artificial intelligence in gastroenterology if I have to break it down in four or five things, what are those four or five things?”
◘ “Calculation of ADR needs at least 30+ staff hours to calculate for one physician, with NLP (Natural Language Processing), that comes down to a few minutes”
◘ “What is the status of AI in GI? Where is the field right now, and where is the field going to be in the next one year?”
◘ “What advice would you have for 3 categories of gastroenterologists: People who are experienced in the field, the mid career people and those who are just about to enter GI?”
The Transcribed Interview:
Praveen Suthrum: Dr. Sravanthi Parasa, welcome to the Scope Forward show. I’m so glad that we’re chatting about this very interesting topic, artificial intelligence in gastroenterology.
Dr. Sravanthi Parasa: Oh, thank you so much for having me..
Praveen Suthrum: I want to lay out to the audience about your background talk about it before we start. So Dr. Sravanthi Parasa is a gastroenterologist and clinical researcher at the Swedish Medical Center in Seattle. Her research is at the intersection of epidemiology, biostatistics, and machine learning. She’s passionate about advancing clinical care through the meaningful application of artificial intelligence. She serves on several IEEE and engineering and computer science conferences, and program committees apart from GI societies. By partnering with and advising or serving on advisory boards of several world- renowned institutes, she’s published papers in the areas of high fidelity, risk prediction models, application of computer vision, and natural language processing in the medical space. That’s a fantastic background. And I’m sure our audience will get much out of this conversation. So how did all this happen? Where did you begin? And how did you end up being at this intersection of computer science, and medicine and gastroenterology?
Dr. Sravanthi Parasa: Okay, I think the first thing that happened was med school. And then I did my Master’s in Epidemiology and Biostatistics at University of Washington. And that’s where I was introduced to the concepts of clinical research, you know, large data sets and how we link databases and try to basically turn the data to find meaningful insights. And that went on, and this was in 2007. And then I did my fellowship and all that stuff. And then, right when I graduated, that was at the cusp, when the human, the human error and the AI era were intersecting. And then that was the inflection point, where you had AI technologies from, computer science standpoint, data standpoint, and also from the computing standpoint. And GI was the perfect use case, right? Because that’s when people started using cell phones for facial recognition and all this stuff. And I thought, when the cell phone can recognize my face, why can’t a computer software recognize a polyp, they’re not as diverse as the human face. So that’s when I started my journey. And then over the years, talk to a lot of computer scientists learn from them. I’m not a computer scientist, but my goal is to understand what technology is out there so that we can bring that technology into medicine and gastroenterology so we can make our lives and patients lives better.
Praveen Suthrum: So, most doctors that I come across, are, in fact afraid about technology, and they keep a certain distance, how come you did the opposite?
Dr. Sravanthi Parasa: Well, if technology is helping me, I want to use that.
Praveen Suthrum: Okay. And also, there’s an inherent fear of Math, but again, like, here you are embracing data. You’re a biostatistician. How did that happen?
Dr. Sravanthi Parasa: I always kind of liked mathematics, not that I know anything more now than 12th standard. We are at this point where technology will be infused and you know, thrown at us what a classic example that I give for physicians is, you know, when EHR came, right, we probably didn’t pay enough attention. It was just thrown at us a ton of EHR systems. And now we feel the burnout in the AI space also, unless the domain experts to whom this is going to be applied to or not involved, then we cannot come up with meaningful use cases meaning the right question which would be relevant for the physician as well as the patient. Because we see our patients everyday we know what are the pain points, we know what questions could be better answered with technology. So having physician involvement is very, very important. And now once you realize that, that’s the case. I mean, the rest is like you’re inspired and you’re motivated to learn whatever needs to be learned to get there.
Praveen Suthrum: AI in GI. Is this your day job? Do you have a clinical practice in the day and you know, this is like Superman and Clark Kent or you know, take your superhero?
Dr. Sravanthi Parasa: I don’t know of any gastroenterologist yet who does early AI as a real job. So, I wish I was there. But now I am fully clinical person seeing patients scoping, like anybody else four days a week, and then on my day off, I work on AI related problems. And that’s where my passion comes up.
Praveen Suthrum: What’s been the most exciting discovery for you in AI in gastroenterology so far?
Dr. Sravanthi Parasa: I think what we see on the market is just a tip of the iceberg. These are commercialized products that most people are familiar with. But just like any other research, there’s a lot of a lot of research happening in the AI space within gastroenterology. And commonly what you will see in the journals and more around computer vision, meaning what an endoscopist sees and what it’s recognizing and so forth. Similar to what we do from a diagnostic standpoint as a gastroenterologist, but there are several others. There’s clinical workflow, new ways of thinking about how to report quality. And then NLP space, you have literature- based stuff. There’s a time coming live, it’s just so hard to keep up with all that.
Praveen Suthrum: Artificial intelligence and gastroenterology have to break it down into four or five things. What are those four of five things?
Dr. Sravanthi Parasa: Most people are familiar with computer vision, meaning AI system being trained to recognize lesions, because radiology took off that’s been there for a while. And most of the current applications that we see whether it’s pathology of the MALDI are all computer vision, so that’s one bucket. The second bucket is a what we call prediction models, prognostication. So, in the past, when we were doing, like maybe the Framingham Heart Study, right, the one of the pivotal studies. What they did was collect patients’ data over the years, and then found out, okay, these are the risk factors that could be associated with a bad outcome. But now you have several signatures from the patient, whether it’s the health records, the social demographics, social data and social determinants of health, the genomics, lab data. And then you have the pathology, and then endoscopy. And you know, you’re combining all that information. And the regular statistical models cannot provide insights with that richness of data and granularity of data. That’s where machine learning comes in. And that’s where you see these prediction models. That’s the second bucket. The third bucket is what we call natural language processing, right? Trying to understand the jargon of our clinical paragraphs that we dictate or type- in for a patient and trying to extract meaningful information and make it relevant to the patient. A classic example within endoscopy is ADR calculation of ADR, and it’s at least like 30 Plus staff hours to calculate for one physician. With NLP that kind of comes down to a few minutes or something like that. And then the last one, I know really, the buckets keep going on. But the other ones are, like speech recognition. Where do you have the ambient clinical intelligence, what we call where, let’s say we are in telemedicine chat, and all I’m doing is just talking to you and the intelligence system can actually find the relevant pieces of information and transcribe the note for you for you to sign. So, you’re not dictating or even typing after you’re done with the patient. It’s happening right then. And this is in a very structured format, where you can pull that information again, and use it for your prediction model. So, these are the big buckets that we are looking at and a combination of these can be used for different applications.
Praveen Suthrum: What is the status of AI in GI across these four buckets? Where is the field right now? And where is the field going to be in the next one year? I’m interested in the immediate future.
Dr. Sravanthi Parasa: So immediate future will be we already have the first FDA approved computer vision algorithm for polyp detection. So that’s happening. A bunch of other different companies also working on validation of their algorithms for FDA approval. So that that is coming up. If you say in the immediate next one year, you will have different players in the market for a similar use case meaning polyp detection or characterization, those kinds of things. In terms of the computer vision and clinical trial recruitment, I think a lot of companies are also using computer vision algorithms to be deployed into your endoscopy documentation software, so you can identify patients for a specific clinical trial. So those will come up as well. In terms of ambient clinical intelligence, it’s already there in the market. And then in terms of adenoma detection rate and these kind of NLP related metrics, I think in the next one year that should be available as well for commercial use.
Praveen Suthrum: Okay. So, when you say this, I’m assuming all of this is being actively tested right now probably finished the testing gate, and is waiting for FDA or the others.
Dr. Sravanthi Parasa: They are in various stages of trials. And as with anything else, people want to get their product out as soon as possible.
Praveen Suthrum: And that’s, that’s amazing, you know, computer vision for polyp detection. That’s been doing the rounds for a few years now. And that’s been seeing those green boxes in so many conferences. And now my question to you is, are we talking like five companies, or 10 Companies 15, 20? Like, you know, what is the number that you’re sensing?
Dr. Sravanthi Parasa: I think I can say, I mean, it’ll be definitely more than five or six companies. But the issue here is not which company is doing. I think in the future, it just becomes so easy for just a clinician to develop those algorithms specific for your patient population. So that’s where AI in general is moving. So, I wouldn’t be surprised at the end of maybe three years, you would have 1000 algorithms floating around.
Praveen Suthrum: And what would these 1000 algorithms do?
Dr. Sravanthi Parasa: 1000 algorithms are 1000 ways of developing a polyp detection model, or whatever the use case is. A lot of times what happens is developing the model itself is not the hard part, it’s how well it’s validated. And that the quality of data that is collected and what different things that is the model being trained on meaning the computer is it’s a polyp from one angle and a different size polyp, just like how human eyes be trained on different types of polyps. So that would be the differentiating factors between all these algorithms and how generalizable those algorithms are.
Praveen Suthrum: So, we’re actually talking about classification of polyps. So which algorithm is able to most specifically classify a certain polyp, like, so it gets more and more sophisticated? Is that what you mean?
Dr. Sravanthi Parasa: No, what I’m trying to say is just, for example, just say it’s just polyp detection model, right? So, I may start collecting data tomorrow, and I may be doing 1000 colonoscopies a year, have 1000 videos available, I annotate them and I build a model. My colleague, somewhere in Boston will do the same thing. And somebody else in India is doing the same thing. Africa is doing the same thing. And then you can cross validate it within your own data set, or you know, prospectively validate it right. Now you have, let’s say, already five algorithms between me and my colleagues, they have deadlocked. Now, how do you say x algorithm is better than y? Right? So that’s where the market will be flooded with all these algorithms. And it will become really hard for a clinician to understand which one to use. A lot of times, obviously, you have to go through the FDA approval process. I think FDA actually just put out their guidance for software as medical device today, November 4. But that is just in terms of the number of algorithms for one use case. But the question that you’re asking is, how many use cases can we apply this to? It could be polyp detection, or classification meaning telling us is that a hyperplastic polyp versus tubular adenoma. And then dysplasia detection may be Barrett’s detection, and then gastric cancer detection. And then you know, this can go on, the size of your pancreatic cyst. How big is it and all that stuff?
Praveen Suthrum: Exciting. But I’m curious, why didn’t you go to private practice gastroenterology? I’m sure many of the GI fellows that you met are probably doing that.
Dr. Sravanthi Parasa: I don’t know. I like my job. So I just did.
Praveen Suthrum: Okay, no. So, the reason that question popped up in my head is because what you’re doing is very, very exciting. If you have the choice to let’s say going to private practice GI and doing regular colonoscopies morning tonight, would you do that? Or given what you know, right now, where the field is going? What would you feel like if you were presented with that kind of an option?
Dr. Sravanthi Parasa: I like to take a time break and then understand what’s happening in the world because you know, you have only limited energy and how you use it. So, if I’m working on same type of problem every day, probably I won’t have a challenge.
Praveen Suthrum: If you were in private practice gastroenterology. And we were so busy, you know, in the endoscopy room and in your consultation room for the right reasons with your patients and all that. And you didn’t have time to keep track of AI in gastroenterology and you wouldn’t bother about it. Now given what you know as Sravanthi. And with all the research and everything, would you see this as a risk for me?
Dr. Sravanthi Parasa: AI is just like any other AI in gastroenterology that space is just like any other research space, right? It’s telling you okay, you there’s a new device for barrette’s treatment, or there is a new way of how you do EMRs. So even in your practice, you will have to keep up with that information. The same thing will happen within AI space also, it’s no different. It’s almost integrated into one of the hot topics where people will learn and will be expected to know about what’s happening in the field. In fact, the next generation of medical students or gastroenterology fellows will probably be even trained on some of the basics for AI so that they can understand the field better so that they can move it forward.
Praveen Suthrum: I was quite amused to learn that when some endoscopists are testing the AI the kind of competing with the AI in order to detect polyps, can you shed some light about this?
Dr. Sravanthi Parasa: The way I look at AI in, in medicine at this point, maybe it will change in 20- 30 years, we don’t have a general artificial intelligence yet. It is designed for very narrow tasks. I looked at AI as something that will augment or improve the way I’m working. So a similar analogy would be let’s say you have a smart car, in the sense of not a self- driving car, but a regular car with Lane Assist Device, which is like level one autonomous on driving scale or level two at the most, right. So, when you are driving the car, you’re just driving the car. And when the Lena says devices, oh, there’s the car coming up, you’re not competing with that image. They’re saying that, hey, you know I detected the car first, you’re just going with the flow or thing like that detected, I’m just going to stay in my lane. So that is exactly how I look at AI when there was a bounding box falling on the mucosa of the colon. We just look at it and say, hey, yeah, it’s, it is a polyp. And I recognize it’s a polyp, let’s resect it. Sometimes it could be false positive, it was just flashed for a fraction of a second. And it could be just a bubble, similar to a human eye. So, I’m not competing trying to find that bounding box faster than myself. So that’s how I would look at it.
Praveen Suthrum: So maybe a slight difference between the self- driving car, and you know, the detection of the polyp in that the reason that maybe I don’t tend to compete with the self- driving car, because it’s not going after my bread and butter. Here, you know, this green box is perhaps going after my bed in particular. Now I’m the boss, you know, I know how to detect the polyp. And what I do currently is still the gold standard. And now you’re coming here and telling me that the software is able to do what I probably trained for maybe 15/ 20/ 25 years.
Dr. Sravanthi Parasa: No, I don’t think AI will come and remove the polyp. The decision making will still be in the hands of the physician. Let’s say I have a cardiac problem and AI devices running my echo and telling to the cardiologist, hey, insulins ejection fraction is 50%. I’m not going to trust the AI to give me that decision. I’m still going to talk to my physician, let them make the decision. If it if they think that the AI system will help them make a decision. I don’t have any problem of them using it. But the decision making and how you manage the patient and what happens in endoscopy room. You know, I still think the endoscopist is the boss.
Praveen Suthrum: But here’s the thing. All the artificial intelligence researchers, this is what they all say. But as a user, you know, when I do not use maps anymore, I want to ask you, you know, when was the last time you tried to remember a street and I just rely on a device to drive wherever I need to go. I trust AI so much that I will get in to anybody’s car literally, I don’t know who this person is. I don’t know what this car is. I don’t know anything about it. But I’m willing to trust it enough to get into a stranger’s car to pick me up and drop me off and someplace. I trust very soon I’ll be trusting an AI lead device to even maybe fly me from one place to the other. I trust one typing like so I noticed personally that I’ve become lazy in typing, I just let the spelling mistakes happen. And I know it will figure it out, and it does. So, when all this is happening, don’t you think we’ll reach a day where whatever we’re doing at the basic level, we’ll get past this so much? You know, frankly, I don’t think it’s a threat. But I’m interested in your, your view, we’ll get past this baseline, where yes, like, it is what we are used to doing that we don’t have to do it anymore. And we’ll rely on the technology so much that we are willing to give the keys of the car to the somebody else.
Dr. Sravanthi Parasa: So, the two points here. One is, when we’re thinking about medicine, the gold standard still be the physician because the license is on us, right? So the malpractice, the whole thing. So you will, as a physician, continue to learn and excel in what you’re supposed to do. So that’s the number one basic thing. The second point is, once you have reached that stage that you know your accuracy is 99.9%, or whatever the gold standard ground truth is right? Then your question is, I’ll ask you a reverse question here. So when you are using the Maps, let’s say you get into the car, you put it on the GPS, you’re letting the GPS talk to you and you’re driving, you’re using that time for something else, you’re using that time for something more productive. The same thing, when you’re typing something, right? You have typing it up, you trust the algorithm enough that it is predicting the next word, or grammatically correcting you and so forth. Instead, now you’re focusing on the content, how do I present this content to somebody? Now, the same analogy might happen in medicine, you’ve already reached a certain level of expertise, when you know exactly what our polyp looks or how our dysplastic lesion looks. Now you can focus on something else that’s happening in the endoscopy room. Maybe you see a new lesion, and you don’t know what it is, maybe you can use AI to tell you, Okay, this looks like a new endocrine tumor, can you pull up images, which look like neuroendocrine tumor. So, at that time, you are taking the time biopsies and stuff going back. So, there’ll be different ways as to how we, as humans will adapt and use that technology to make our lives better. I would never think that we will go back to just moving forward.
Praveen Suthrum: Let’s do the analogy a little bit more. I know plenty of people who write very poorly compared to an AI. And very similarly, again, like in this show, the scope forward show, we pretty much talk about everything and anything without hesitating, so I’m going to go to a point where people think this in their heads, maybe they never been up. It may be entirely possible that AI may be better than at least some endoscopists, it may detect very clearly, more adenomas than a below average endoscopy assumption. What happens to them, let’s say somebody who has until now, very quickly survived, maybe being an average or below average endoscopy. So what happens to that category of the field?
Dr. Sravanthi Parasa: Let’s, let’s forget about AI for a second. And let’s say you have an endoscopist, who’s performing below your threshold. Now, the standard we do these now is that you provide education to the person and the chances that they will improve the most would be highest, right, the delta will be much higher. Now with AI, what we’re doing is are the assistant like augmented device, what’s happening is there will be some basic level of standardization because you cannot ignore that there is a bounding box it helps them look at things better. Now, of course, the technique itself as poor are the bowel prep is poor there’s nothing, we can do about it. But at least for the visual inspection, assuming that they’re doing the same thing that anybody else is doing. There is some standardization of the procedure itself.
Praveen Suthrum: Okay. I was actually looking for a broader answer on how the world of GI might settle if AI becomes an integral part of the team. Let’s say a GI leader tomorrow has a choice. I can recruit an average endoscopist or I can, you know, give this tool to some who are more open and embracing and then put them on wheels. I may prefer the AI versus human being? Is that scenario possible? I know we don’t like to say. And we always like to say that AI cannot be licensed, but the laws and the rules are changing, so that you change. From a reimbursement standpoint, if an insurance company comes out and says that, look, I trust this AI enough that if you use it, we’ll reimburse you tax. If you want to drive manually, that’s fine, then we reimburse you why? Like saying, what would be the scenario?
Dr. Sravanthi Parasa: So the scenarios could be multiple. One is we as gastroenterologist, just embrace it. And it’s not imposed on us understanding that there is some value or some standardization across all the patients that we sculpt, right? That’s one way of thinking about it. Now, will the CMS pay or who’s going to pay? We don’t know. The second thing is the comparison of AI, like what we call augmented intelligence, plus the human itself, versus the same a human by themselves. So, if that delta is significant at a population level, the insurance companies will start working on that piece because they want quality that should be some value-based care, that’s where healthcare is moving. In the future. If you’re no longer thinking about repeat colonoscopies every two years, or five years, or 10 years, or whatever was happening, you need a quality exam. And did that happen? And because this is a way that you can standardize it, that could be a realistic option that might be available in the future. And a classic example for that, which we already see in healthcare, is, if I exercise every day, and I connect my Fitbit or whatever device to a third-party company that gives that information over to my insurance. I have a significant reduction in my pay my co- pay for insurance. So, that’s already happening, right? There is some value in you know, what data you generate, and how insurance companies are gauging how well you’re doing. The same thing will happen on the physician. And as well, how you will get reimbursed, maybe your quality reimbursement will change, it will no longer be whether you are telling the patient to come back for surveillance in five years, because maybe that’s not the real metric. Maybe the real metric is are you doing a good job visualizing the mucosa and removing all the polyps. So, a lot of new ways of how we can use this technology will come up in the future. And I’m pretty sure everybody is already eyeing on that space.
Praveen Suthrum: So, in this in this scenario, do you think that gastroenterologist and endoscopist will gravitate towards more advanced procedures like so therefore, the gravity shift like so? Is it more advanced procedures? Or is it more volume with the aid of technology? What would happen?
Dr. Sravanthi Parasa: The question about doing just screening procedures versus more therapeutic procedures is not totally reliant on the AI technology itself. But it’s more because we have other options, non- invasive options for screening, right. So there are more advancements, we no longer just doing FOBT, which is just seeing if there’s some blood in the stool. We have moved on to fit and then you have additional, like DNA base tests, and now we have DNA based blood tests as well. So those might be what people would like to deploy. For screening standpoint, nobody wants to spend $5,000 for a screening test unless there is some value in terms of, you know, therapeutic. So, I think in general, the field might move more towards therapeutic because it will be driven by the non- invasive technologies per se, not AI. And for some of those non- invasive technologies to work, you’re going to need artificial intelligence, because there’s a lot of computation that needs to go in for finding that specific signal.
Praveen Suthrum: Can you reimagine future workflow with digital biology, let’s say we’ve arrived at, you know, at the point of time, where liquid biopsy is a reality, and it’s just not one type of cancer, it’s just not colon cancer, but 15 different types of cancers with a blood sample. Assuming that that day becomes a reality at some point. And AI is all pervasive. So it’s there everywhere, and all the three or four different categories that you talked about in the beginning, they’re all a reality, and they’re at the point where, you know, as a researcher that you would like them to be in that scenario, can you reimagine the workflow of GI and endoscopy? What might happen on a typical day in private practice?
Dr. Sravanthi Parasa: I think that is almost like the future. Right, the future could be anything. Assuming that there is a widely available non-invasive test like a blood test or a simple stool test that is pretty accurate at detecting precancerous lesions, because that’s the bread and butter for us, right? It taking even like five millimetre adenomas with reasonable accuracy, then that would be a screening test. Once that becomes positive, that’s when you will start doing more invasive procedures like colonoscopy and so forth. In terms of how the workflow will change, you can probably like from a private practice standpoint, you can accurately predict which patient might not show on your clinic schedule or your endoscopy schedule. And you can probably overbook that slot. So that way you’re, you know, you’re more efficient in your workflow.
The second thing is, let’s say you’re running late in endoscopy, because you have a complicated patient or a complicated procedure, you can send real time alerts to your patient as to when they need to arrive, and not waiting in your pre op for an hour. And then, when you’re doing your procedure, most of the documentation is happening while the procedure is being done. And optical biopsy is working in real time, meaning you know exactly that this is a five-millimetre tubular adenoma versus a hypoplastic polyp and the you know, you just resect and discard the polyp, and then you’re just telling the patient come back in five years, or, you know, whatever the surveillance need to be. And the patient will probably get a report, not of the overall quality of the physician that performed it, but how was the quality for that particular patient at that particular time. So, saying that, okay, this physician has visualized 85% of the mucosa of the colon, and your surveillance is five years because we found a five millimetre to bladder, no muscle, so forth. So that the whole loop is connected, and it is closed in one kind of an interaction with the patient rather than me going back and forth multiple times.
Praveen Suthrum: Sravanthi that is an amazing, amazing layout of what the landscape looks like. Wanted to ask you, what advice would you have for three categories of gastroenterologist. One, people who are very experienced in the field, and for probably 25- 30 years, have been doing endoscopy in the field, or the other the mid- career people and the final category are people who are just about to enter GI. As a researcher, given what you know, given what you’re seeing, like so you have, you’re seeing what is going to happen in the future, and you have access to whatever is coming out. So, given what you know, what advice would you give for three categories of GIs?
Dr. Sravanthi Parasa: So, the first very easy category would be the just people who are entering the field, right, because that those are totally moldable and malleable. I think, learning a little more about the basics of AI, whether it’s computer vision, what does this machine learning mean, you know how to understand because that’s how the data will come out in the future. We already seen, like a lot of publications within GIE, and some of the major GI journals focus completely on AI. And in radiology, 70% of their research is on AI. So that’s how it will, the field will transform. So, understanding the basics will become very important. And knowing to lead the field is very important than taking ownership of the domain as gastroenterologists becomes important.
Now, the mid- career people, there’s nothing, anybody can switch to do what they want at any point of time, right? So mid- career, if you’re already committed to a particular lifestyle, you don’t have to do this the rest of your life, there’s no, there’s no reason somebody has to do that. But from a more practical standpoint, understanding where the field is going, just like what we’re doing today, is very important, so that you can plan things or your finances according to that, and also being early adopters of this technology will help you understand, you know how to make it better, and how to kind of use it in clinical practice and maybe even guide the industry as to what use cases might be very relevant. So that’s, that’s an opportunity right there.
Now, if you are a very experienced gastroenterologist, really not very keen on changing tracks. In the sense let’s say you’re a basic science researcher or a gastroenterologist in private practice, have scope for like 20- 30 years. Just understanding like what I said, if you need to know what’s latest and greatest happening in the gold space, or you know, colon cancer space, how would you approach literature the same thing you would approach AI based technologies as well. So that way you’re well informed, to know when to embrace a particular technology. And when I say AI, it’s just not computer vision, right? It could be telemedicine, digital technologies, you know, interact with your patients also will change in the future.
Praveen Suthrum: Sravanthi thank you so much for sharing your perspective with us today. I found many, many of the points very, very insightful, and it is quite clear to me on the field is going and I’m sure people who are listening of watching this would also feel the same now but thanks so much for joining us.
Dr. Sravanthi Parasa: Thanks for the opportunity, Praveen.
By Praveen Suthrum, President & Co-Founder, NextServices.