Rare diseases present daunting challenges for researchers and clinicians. With limited data and resources available, developing effective treatments can feel like trying to solve a complex puzzle with missing pieces. However, Shweta Maniar, Director of Global Healthcare & Life Sciences Industry Strategy at Google Cloud, believes there is hope - and it comes in the form of Generative AI. Listeners will come away with a deeper understanding of how this transformative technology can crack the code of rare diseases and unlock a future of better outcomes for patients in critical need. Copy
Shweta Maniar, Director of Global Healthcare & Life Sciences Industry Strategy, Google Cloud
Megan Antonelli, Chief Executive Officer, HealthIMPACT Live
Intro 0:00: Welcome to Digital Health talks. Each week, we meet with the healthcare leaders making an immeasurable difference in equity, access and quality. Hear about what tech is worth investing in and what isn't. As we focus on the innovations that deliver. Join Megan Antonelli, Janae sharp and Shahid Shah for a weekly no BS deep dive on what's really making an impact in healthcare.
Megan Antonelli 0:31: Hi everyone. This is Megan Antonelli, Welcome to Health Impact, live digital health talks. I'm thrilled to introduce our guest. Shweta manyar, Director of Global Health Care and life science industry strategy at Google Cloud, is a visionary leader in applying cutting edge technology to health care challenges. I had the pleasure of meeting her back in 2016 when we were just launching, you know, node health looking at digital evidence. And her career has spanned, you know, a number of years, and I'm so excited to have her here. Her expertise in leveraging generative AI for Rare Disease Research has positioned her at the forefront of a potential revolution in medical science with a background spanning digital health, pharmaceuticals and biotech. Schweitzer brings a unique perspective on how innovative technologies can transform patient care. Today, we'll explore how generative AI is, cracking the code of rare diseases and opening new frontiers in treatment development. Hi, Shweta, how are you good?
Shweta Maniar 1:29: Thank you for having me here today. Really appreciate it.
Megan Antonelli 1:32: I'm so glad to see you and just so excited to hear about you know, what you've been up to lately. So share with our audience a little bit about your background, and you know how you've ended up at Google?
Shweta Maniar 1:45: Yes. So as you know, you and I were interacting and working together during my time at Genentech, where we had overlapping interests on how do we bring how do we look at technology and innovation in the healthcare and life science realm. And so it's that same is, it's the same passion or the same focus, except now I have the opportunity to do this on the other side, which is, now I'm I'm on the other side of the table, looking at, how do you think about various technologies and then applying them to the healthcare and life science sector. And so it's actually really, really satisfying, because often I'm sitting in these conversations where part of the credibility or the rapport building is, well, I was on the receiving end of these type of conversations just a few years ago. So it's wonderful to see not only thinking about, what are the challenges that the healthcare ecosystem has, but how do you apply the technology? And so that was so, so enticing, and so I've been at Google for for about six years now, doing, doing entirely focused in the in the healthcare and life science industry,
Megan Antonelli 3:04: wow. First of all, that time flies, huh?
Shweta Maniar 3:08: No kidding, I can't believe it's good to be reconnected.
Megan Antonelli 3:11: Yes, it is. And that's, you know, it, is it? It's amazing. And also, the technology has changed quite a bit, right? And we are what we are capable of now, you know, we talk about how slow healthcare changes, but you know, at this point the technology is changing so fast, it's, it's, you know, hard for healthcare to keep up, still, but it is, it's doing, I think, a better job than the first time we we met. But I'd love to hear how, you know, obviously, coming from Genentech and coming from that life sciences side of things, and then being at Google and seeing the power of that technology, how has that, you know, sort of brought you to this research and work around rare diseases?
Shweta Maniar 3:53: So I love, I love when I get asked this question, because, and for the use of AI in the in the life science sector, right? Rare Disease, we can, we can bring it specifically to rare disease. But it didn't happen overnight. It was, it really was a gradual process, right? That that stemmed from my personal experiences, both, you know, personal meaning, you know, from truly personal as well as professional experiences. But then also, like the growing recognition that there is a role of technology and AI in healthcare and life sciences. If you think about the ecosystem as a whole, really think about, you know, technology actually has, or is part of that ecosystem to support the industry as a whole. So of course, 10 years ago, I think we saw this, and this was already happening, but it really is now, you know, this role that we have here, this is what I'm doing here today. Didn't exist right when we were in school or when, you know, when we had our first jobs. And so earlier in my career, I had the opportunity to work closely with. With patients who are suffering from a variety of rare diseases. While at Genentech, I worked across different different parts of the portfolio, but it was heavily focused in the rare disease space, and so being able to see the value and some of the struggles that individual patients had firsthand because of, you know, different diagnoses or the lack of different options, really left a lasting impact on me, and so I really became acutely aware that there's got to be something more that we can do. It's not just from a science, but how do you support rare disease overall, right from a patient perspective, from the coming the community perspective, from being able to access technology. So there's that piece. And then at the same time, I was always interested right in, how do we think about technology and innovation? And so AI, you know, could potentially revolutionize the way we approach rare diseases, and I think we're starting to see a lot of that happen bits and pieces and very carefully. And so the ability that AI has to analyze a lot of data to uncover these hidden, hidden patterns generate different hypotheses. I think really does offer a glimmer of hope in fields where data is limited, and so that's now about the technology and AI overall when it comes to generative AI, think you mean that even is like an added bonus, right? The potential to be able to think of some of the most pressing challenges in rare disease from synthetic data simulating biological processes so many ways that you could accelerate some of the drug discovery processes or the diagnosis processes, and time is of the essence, often with these rare diseases. So I think that there's a lot of potential here, but it really wasn't something that, you know, I woke up one day and said, this is, you know, I want to be at the crossroads of healthcare and AI, or healthcare and, like science. And it really is, I think, combination of experiences and time as well as some of the experiences that I had personally as well.
Megan Antonelli 7:19: Yeah, well, it's, it's really amazing with rare disease that, you know, and I remember from working on it and that there's to get the dollars and to get the concentration and to get the research funding behind it, is challenging, and the work that goes into that is challenging. So is it? Is there something, you know? Is it the capabilities or even the cost of generative AI that makes it, you know, better positioned to get rid of some of those barriers, what you know, what is helping sort of what makes it the powerful tool that it is to kind of refocus attention on rare diseases.
Shweta Maniar 8:02: So so let's talk about this as like the traditional AI that we all know, right? It's excellent at pattern recognition, identifying trends, looking at existing data, making predictions based on those data, as classifying or categorizing information, that's fantastic for tests like looking at triaging for diseases or looking at images, maybe identifying patterns with symptoms, but not fully understanding what you know, it's more around understanding what is rather than what it could be. And I think that's where generative AI can play a slightly different role. On the could be part, on the other hand, it's, it's now the realm of creation. I think that's the key difference. When you think about traditional AI and generative AI, it learns it's learning from existing it's learning from existing data, and then it uses that knowledge to actually create new content or new endpoints. And so thinking about this, it's a difference between a student who memorizes facts and then a student who could write a compelling essay based on those facts, right? And if you think that's actually two very different skill sets and and frankly, very you know the latter can be very, very valuable, because now the student can write a variety of essays, right, because they have the core facts. So when you think about it, in rare disease, you can have generative AI can help with synthetic data. Because what is the biggest problem in rare disease? It's actually the scarcity of data right to be able to then identify those patterns and and so synthetic data is a big piece of rare diseases, being able to simulate some of those biological processes that I had mentioned. Earlier, right creating virtual models. So being able to simulate is a big piece redesigning or designing novel molecules that you couldn't because of just the amount of time it takes or just the right scenarios or the right conditions. So being able to design novel novel treatments or novel molecules, and then you can also, with generative AI, theoretically, you can explore hypothetical scenarios, so being able to simulate a variety of rare diseases and thinking about, how do you test different intervention strategies without, you know, and this is completely in a virtual environment. So these are a lot of unique ways and capabilities that generative AI can actually play a role that would be different than traditional AI. It will take some time, and I think that we have, you know, this is going to be done in a much more measured way, but I'm talking about the potential of how generative AI, generative AI, can affect the rare disease space because of that data, or the lack of data challenge that exists,
Megan Antonelli 11:03: Wow, I've never really thought about it at that level. But that is, you know, in terms of just that scarcity of data and what it can what the potential of being able to create that that is, is huge. So tell me, we're all about impact and health impact, right? So what about, you know, are there case studies? Have you seen this work, you know, in places? And tell us a little bit about what you know, what you're seeing out there?
Shweta Maniar 11:30: Yes, you know, I there's a variety of use cases that we're starting to see in, in in the rare disease space. I think that those that are looking at the use of generative AI, it is around data augmentation, because of the limited patient population, there are organizations and companies that are starting to look at the small data sets that they have that traditionally would hinder some of their research and development, and then starting to just explore initially, how can you maybe use some of that and, you know, create some synthetic data that closely resembles some of that real patient data that they have to then maybe increase The diversity and the size of those data sets, and then train some models, some machine learning models. And we're very early stages as because this is, again, this is an area that is very new for everyone, in terms of, how do we actually do this in a measured and methodical way, where you're keeping the human in the loop in the entire process, I'm starting to see organizations that have a rare disease portfolio, or focus on rare disease really think about, how do they augment their existing data? We're also seeing many companies leverage in silico drug discovery as a piece of the use of generative AI, again, as I mentioned, accelerating the processes by designing these novel molecules with whatever desired property that they want. And now, if you can do it in silico, now you can predict the interactions of the biological targets. And so there are several, several organizations, some that I can't mention right now that are looking at is silico drug discovery as a piece of how they're looking at applying generative AI as part of their processes. Both companies as well as academic research centers and university centers are very much spending their time and some of the earlier stages of what in silico actually means, and when it comes to the discovery processes, right?
Megan Antonelli 13:47: So, I mean, it almost seems like there's applications at every, you know, at every phase of development, right? I mean, at the discovery stage, at the uncovering the disease state all the way through.
Shweta Maniar 13:59: And one other thing I would also mention is we didn't even talk about the kind of the other end of that spectrum, which is the patient stratification part, right about, like the truly personalized medicine aspect of it. By looking at the again, the complex patient data, which is limited because it's rare disease relative to other diseases. But then you can take that, you know, that full understanding of that individual right there, for genomic information, their medical images, their their records right trying to under understand everything so and then the use of the use of this generative AI could potentially help organizations identify different subgroups, you know, identify patients with maybe similar disease characteristics, right how they potentially would respond to different therapies. So now, then you're on the other side where you could actually, one day, one day. You could create more targeted or personalized approaches to how patients are diagnosed, how their what their clinical trial treatments are, or how you know what clinical trials they do or do not go into because of being able to stratify patients a little bit more with generative AI. So I just wanted to give one more example. I'm kind of on the other end of the other end of the spectrum, right?
Megan Antonelli 15:23: Well, it, you know, that's it is an important piece of that, you know, and I that makes me then think about kind of as you get to this, right? So rare diseases have always been sort of an ethical dilemma, right? You know people, you know, they tend to be championed by patients or communities, or, yes, there's some pharma companies and life life science companies that you know only address rare diseases. But it is a tough financial angle has been traditionally to get to, because you can spend, you still spend as much time as you do discovering and researching, and then there isn't as big info population to take that medicine. And then the population is small and often not able to afford them to do it. So it's a tough, you know, it's a tough financial and ethical problem, and now we layer artificial intelligence on it. But in some ways, I mean, when I'm listening to you, you know, when we talk about, you know, the dangers of AI and the risks and the concerns. But in here. I mean, the application is very positive. I mean, it's, it's, it's, it's going into a territory where, you know, there wasn't enough research, there weren't enough resources. And here, you know, here we're able to find it so and I know, in your role, you you know, talk and and sort of manage the regulatory guardrails of that. So talk a little bit about that broadly in terms of, you know, sort of attitudes, and the sort of attitude towards this adoption and applications with rare diseases,
Shweta Maniar 16:50: yeah, and so and so, just to clarify, right? So it's not my responsibility to think about right to to be responsible for, you know, how our regulatory bodies make their decisions. But certainly they are looking at all these regulatory bodies like the FDA, are looking at various technology providers. We have such a powerful technology. How do they create, how do they inform themselves so that they can create the right guardrails and, you know, create a point of view, so that AI generative, AI can be used potentially in the near future in a way that will be accepted. One thing before, you know, before I kind of go into that side, is be able to share a little, you know, I was giving, I was giving a keynote in Washington, DC, around, around the similar topic very recently. And a question that I had received is, and it's a question that I think we all get in this space pretty frequently as well. What if the data is biased? You know, we're a regulated industry. This has to be, and I always emphasize right, that it's not the output is not going to be biased, unless the data that you put in is going is biased to begin with, right? And so when you think about generative AI models, they have to be trained on diverse and representative data to avoid bias. And the only people that can do that are going to be the people who are putting data into these systems. Right? Not the, not only the tech providers, but it is also right, the health care, the life science company, the startup, whoever it might be, need to think about what you're putting in there so that you can avoid bias on the on that's coming out on the outside, that I think that and data privacy are going to be part of the part of how we think about like, I guess we call it right, responsible, AI. The other piece of it that both the regulators as well as the healthcare and life science ecosystem are looking at is being clear. So there's a transparency you need to be able to talk about what's under the hood all the time, because I can't ask you to trust me. You know that you're a doctor, and here's, you know, 100 images, and you need to tell these five patients from these five images that they have this disease, and don't worry about the other 95 without understanding what's under the hood. Why did I triage these five for you to take a look at? And said no to these other 95 so transparency is going to be key in how we've developed these models. Being able to educate not just the tech experts or your IT folks, but then also educating that researcher or that physician about how how this technology is working and how it's going to be used, before we started recording you and adding a little bit about this. And it's not that AI is coming for. Your job, but AI will come for people who don't try to understand how this is going to materially update or change the future of of their work, right? And so I think there is this misconception, and I know I'm going a little bit further outside of your question, but there's this misconception of, how do you think about, you know, I don't need to worry about AI, because it's, you know, I'm, you know, I'm a I'm a nurse. I don't need to worry about it, or I'm a researcher. I don't need to think about this. It's going to be part of all of what we do right moving forward. So it's not just the tech providers responsibility on having that responsible, AI, having, you know, thinking about the bias of the data, thinking about the data privacy and the security. But also, how do we actually explain it transparently to the ecosystem, to the users of these different technologies in a way that they can trust it, so that, yes, you say, I'm going to focus on these five images and not these 95 for this reason, I understand why this was, you know, this was the output. So that's, I think, a big piece of how, how this is going to be more widely adopted, is the transparency and understanding what's under the hood is going to be paramount.
Megan Antonelli 21:20: Yeah, absolutely. And it is around where, you know, people have their reservations around it, and their concerns, particularly around jobs and taking jobs. I mean, just listening to you, I can think of so many jobs that would be created by this because of the opportunities that are created. And of course, it applies across, you know, much more broadly across rare disease. You know, beyond rare diseases to the life science in general, one thing I've found, you know, and I love your perspective on this, you know, so Pharma is a pretty technology forward industry, you know, however they've grappled with digital transformation across the ecosystem. So this isn't necessarily generative AI or rare disease, but I'd love to see your thoughts on how that you know how that is happening, because this is, this is pretty advanced tech you are talking about, you know, literally the forefront of this. And yet, we often talk about how they're not quite, you know, they're not quite there so, but they clearly must be, because this is your job. So I'd love that, that you know your thoughts on that. And you know, also, even just as, as a woman, we I also work in a, we do a women in generative AI group, and it's actually not healthcare specific, but you know, it's been a lot about upskilling. And what does it mean, and how is, you know, how is it going to impact so I love your perspective, both as a woman, as a woman in pharma and as a woman in tech, as has that converging, and how that adoption rate seems to be going? Because certainly, listening to you, it seems to be going pretty well.
Shweta Maniar 23:03: Yes, optimistic, right? I think it's it all. We are all responsible and income, you know, it's incumbent upon all of us, right? So certainly, I am optimistic about how we look at technology in the application in healthcare and life science, you know, and I, I have such a soft spot for this, for what you're asking, it's because that's how we originally met, is right when we're working and thinking, working in a group on the Genentech Roche side, thinking about digital transformation, and working with you on, you know, some of The initiatives that you were on, helping Genentech to participate and understand how other organizations are also digitally transforming, particularly in pharma. And this is not a rare disease comment, right? The challenges in this industry have been, right? We all, we as a pharma company, say, right, you know, we have, you know, we're a regulated industry. It's very, very challenging. There's so many data silos that exist, many by design and many because it's required thinking it's if you can't get things wrong. So it's a very risk averse place. And then when it comes to AI, right, there could be, and I wouldn't say this is for everyone, but there could be even a perceived skill gap. Right as things are changing so rapidly, how you hire has to also be updated, right for what kind of skills you're looking for, where you can have people who can look at emerging technologies. So one of the things, and I love this, the idea of women in tech, women in STEM. Now we're talking about women in AI and generative AI. It is so important to have a variety, not just women, but people of all different backgrounds, because of the perspectives that they provide. Because, again, as I mentioned earlier, it's, you know, garbage in garbage. When it comes to your data and what your outputs are, and that's how bias can be formed. And so having a variety of perspectives and people working in the generative AI field is really important to be able to mitigate some of the biases to think about how collaboration and creativity is going to be done if generative AI is partially about creating new things. You're going to need diverse teams to think differently and to think about what are some of the unique ways that we could solve different problems and all of that. And I think that there's the reality of addressing some gender specific, gender specific needs. You know, if you're going to talk about some of the angst around maternal health, it certainly will help to have at least some perspective from those who have actually gone through some of those, from some of those experiences. And then, I think it's also just important generally, because we need to have more representation today, because that's how we're going to inspire the future generations. Because, frankly, I hope that we don't have to have this conversation in the near future, because it shouldn't even be part like, it's just part of how, how we all work, right? We don't need to think that there has to be a woman, or there has to be somebody of color, or, you know, this shouldn't be part of the part, part of the dialog as we move forward, but we only used to
Megan Antonelli 26:26: that. Or maybe we'll need to Ai do it for us. But
Shweta Maniar 26:29: exactly right, exactly, exactly. So anyhow, I think it's not just a matter of, you know, it's not just a matter of equality, but really it is. These are how the AI systems are going to be fair and innovative and actually benefit everyone, right?
Megan Antonelli 26:48: And in some ways it's a false assumption, I think that they, I mean, yes, the body of data that they could be trained on is inherently biased to and that exists, but the efforts can be made to train it on data that isn't biased, and that, that is what has to be done to get us to a place of, you know, where the where it's not garbage in and garbage out. Ultimately, well, that, and I think that it's optimistic, but I think I It sounds also a little realistic. It doesn't sound overly optimistic. But here, tell me a little bit about as you look five years, 10 years. I mean, I feel like technology is moving so fast right now that 10 years is it feels like, whoa, you know. I mean, needless to say, everything that's going on in the world, it's a little. 10 years is a lot. So let's go with five. When you look ahead five years in terms of where you see, you know, applications of generative AI really impacting rare disease. What does that future state look to you? Look like to you?
Shweta Maniar 27:54: Okay, so for rare disease, so I think it's, it's so exciting because of everything that I was mentioning earlier, right? Like you can model, think synthetic data, there's in silico, and so I actually do think that there is a reality in the not too distant future where we're going to say that this rare disease was cured, or a patient was taken care of because of a medicine that was personalized for them that was created with the help of, if not entirely, from the use of generative AI. So synthetic data in silico research, I also see, because of the way generative AI is created, has been created, we're going to see a lot of repurposing of existing drugs, because you've got all this data, and you can elicit these patterns that you couldn't see before. So if you can repurpose what you already have, you are instantly and dramatically shortening the time. You know, we all talk about that 10 to 20 year time frame for getting a new drug to market. You're dramatically decreasing that and so, and then I would also say, also, sorry. One more thing also is around generative AI is going to play a key role in in new biomarkers and the diagnostics part, right? So much more it's, it's going to be much more proactive in the rare disease space than it is going to be reactive once you have, once you have the physical symptoms. And do you think that gender of AI is going to play a huge role on these biomarkers and the diagnostics, where people will be able to lead much healthier lives, because you're able to curb those diseases that normally are quite debilitating, and you can only start to try to treat it once you have the physical manifestations.
Megan Antonelli 29:41 : Yeah, well, the future is certainly bright. From that perspective. I think it is. It's a sort of the possibilities are really endless. It does so I when I think about this, and I think we've had Susannah Fox on the show, and she recently came out with her book rebel health. And in that, of course, she talks about the different, you know. Places in the you know healthcare journey that people are and their personas. And you know you have the you know, when you think about the patient and what they go through, having having a rare disease. And I know that you are living in this from the technology and the life science perspective, where is, you know, where do you think, just from a patient perspective, they fit into this story in terms of how they can, then, you know, inspire or get, you know, get pharma and technology, you know, working behind them when, if they do have a rare disease.
Shweta Maniar 30:38: So from a patient perspective, I think it's going to be so important that to have an understanding of your own disease, right? I think we what has shifted right is we used to be at a point where patients just listened to whatever their doctor said and then went home and took that medicine or followed the regimen, and now we have much more empowered patients, right rare disease or otherwise, but particularly those in a rare disease, because you as an individual, or at least your immediate caregivers, are needing to really understand what that you know your disease, what are your options? And so it's so important, as you look at kind of thinking about tech from the technology perspective of what you know, what is the potential? How do you think about rare disease and the use of AI, thinking about patients and having a curiosity mindset, thinking about what are the common misconceptions about what AI could do, and how do I think about where I can start to spend some time to understand about what is the research that's being done and being much more vocal about where where you can spend some time. I think one of my biggest pieces of advice is particularly in rare disease. Your biggest outside of your immediate care team, your biggest advocacy, advocacy is going to be those patient advocacy groups where information is starting to cascade around how different technologies and different innovation is being used. And so my advice is really at the very highest level, it's really around being curious about understanding that technology is going to play a role in how diseases are going to be treated and cured, and taking the time to understand that because each patient is going to be very critical in in this kind in this workflow of how, how the research that's being done today to help the patients of tomorrow,
Megan Antonelli 32:53: that's really powerful advice. And it really, you know, I think of it as, you know, people talk about the have so much fear or hesitation around it. And, you know, you think back, and even I think Susanna Fox sort of frames this is that before the internet, you know, it was very difficult to find this and generative. AI is only you know, sort of that continued expend exponential, you know, power of that. You know, sharing of information and putting that power of the data behind it. So, you know, it is amazing, and the potential is, is there. But I know you have a big job to do. You have a lot of work to do, so unfortunately, we are out of time, and I could talk to you all day. It's so great to catch up. But thank you so much for joining us. And you know, sharing, sharing your insights around this.
Shweta Maniar 33:42: Thank you very much for having me. It's been such a pleasure to spend some time together and have a conversation around a topic that we both are very excited about.
Megan Antonelli 33:51: Yeah. Well, thanks again, Shweta, and to our audience, thanks for listening. Shweta, how can our audience get in touch with you if they have questions? Of course, they can go through health impact, but are you on LinkedIn? What's the best you know? What's the best way get get in touch with you?
Shweta Maniar 34:07: Yes, connecting with me on LinkedIn is going to be the fastest and the most the easiest way you can search me up under srita. Many are, I'm the only one there, and would love to have a dialog around the use of generative AI in the rare disease space, it is such a powerful tool, and it's really going to happen with a collaboration with a variety of industries. So I am would love to stay connected with those who'd like to have a further conversation. So thanks again for having me today.
Megan Antonelli 34:38: Absolutely. Thanks so much for being here. Shweta, many our insights offer a glimpse into the future where rare diseases are no longer unsolvable puzzles. By harnessing the power of generative AI, we stand on the brink of a new era in medical research and patient care. Thank you, Shweta, for joining us and your expertise and your vision. Thank you audience for being here today. I. And that's Megan Antonelli signing off from Health Impact digital health talks.
Thank you 35:05 : Thank you for joining us for this week's health impacts digital health talk. Don't miss another podcast. Subscribe@digitalhealthtalks.com and to join us at our next face to face event. Visit Health Impact live.com you