Scroll Top

Joel Dudley and what happen when you let data and not hypotheses drive discovery

Episode Summary:

Harry’s guest this week is Dr. Joel Dudley from the Icahn School of Medicine at Mount Sinai, where he serves as executive vice president of precision health, associate professor of genetics and genomic sciences, and founding director of the Institute for Next Generation Healthcare. Dr. Dudley explains how his group is utilizing data to uncover health problems that can’t be detected through normal methods, as well as his groundbreaking paper on the link between Alzheimer’s disease and herpes.

Please rate and review The Harry Glorikian Show on Apple PodcastsHere’s how to do that from an iPhone, iPad, or iPod touch:

1. Open the Podcasts app on your iPhone, iPad, or Mac.

2. Navigate to The Harry Glorikian Show podcast. You can find it by searching for it or selecting it from your library. Just note that you’ll have to go to the series page which shows all the episodes, not just the page for a single episode.

3. Scroll down to find the subhead titled “Ratings & Reviews.”

4. Under one of the highlighted reviews, select “Write a Review.”

5. Next, select a star rating at the top — you have the option of choosing between one and five stars.

6. Using the text box at the top, write a title for your review. Then, in the lower text box, write your review. Your review can be up to 300 words long.

7. Once you’ve finished, select “Send” or “Save” in the top-right corner.

8. If you’ve never left a podcast review before, enter a nickname. Your nickname will be displayed next to any reviews you leave from here on out.

9. After selecting a nickname, tap OK. Your review may not be immediately visible.

That’s it! Thanks so much.

Transcript:

Harry Glorikian: Welcome to the Moneyball medicine podcast. I’m your host, Harry Glorikian. And this series is all about the data-driven transformation of the healthcare and life sciences landscape. Each episode, we dive deep through one-on-one interviews with leaders in the new cost conscious value-based healthcare economy. 

We look at the challenges and opportunities they’re facing and their predictions the years to come. 

What if we can predict the future of patients from electronic health records, diabetes, schizophrenia, and different forms of cancer. My next guest, Dr. Joel Dudley believes in the mantra. Let the data point the way. Dr. Dudley is currently associate professor of genetics and genomic sciences. And Founding Director of the Institute for next generation healthcare at the Icahn school of medicine at Mount Sinai.

In March, 2018. Dr. Dudley was named executive vice president for precision health from Mount Sinai health system. In 2017, he was awarded an endowed professorship by Mount Sinai in biomedical data science. Prior to Mount Sinai he held positions as Co-founder and Director of informatics at Numedii, Inc and Consulting Professor of systems medicine in the department of pediatrics at Stanford university school of medicine. 

His work is focused at the nexus of Omix digital health, artificial Intelligence, scientific wellness and healthcare delivery. His work has been featured in the Wall street journal, scientific American MIT technology review, CNBC and other popular media outlets. He was named in 2014 as one of the 100 most creative people in business by Fast Company magazine. 

He is the co-author of books exploring personal genomics from Oxford university, press. Dr. Dudley received a BS in microbiology from Arizona state universe. And an Ms. And PhD in biomedical informatics from Stanford university, school of medicine, 

Harry Glorikian: Dr. Dudley, welcome to the show. 

Joel Dudley: Thanks for having me here. 

Harry Glorikian: So we’ve been talking on and off on the phone for quite some time and, and it seems like you’re involved in so many different things. 

And so, you know, can you tell the audience here? So what, what do you do here at Mount Sinai?  

Joel Dudley: Yeah, it’s a good- good question. And, and, you know, it looks like we’re working in lots of different areas, but there’s actually a, uh, principle sort of approach behind this, and it’s- In one way, we’re a systems oriented people, right?

And so really we see not only the body as a system, but the disease is a system right. Of, of connections. And then, you know, the body has a system, there’s many different layers. So the one way I’ve been trying to describe it recently is using a software analogy by saying that we’re full stack translational medicine researchers. 

And what does that mean? It means that we go from, you know, the molecular to the, uh, you know, physiological to the medical, to the electronic health records, wearables environment, et cetera. And we integrate across that whole stack, just like when you’re building a software application, you know, it’s database to web server, et cetera. 

Um, but then, you know, not only do we go deep in that regard, but then we go wide across all diseases from that perspective, because we believe that the taxonomy of diseases as defined today is wrong. 

Harry Glorikian:  So this is, you know, you know, if I’m going back in time, you know, it’s, it’s sort of Leroy hood-ish, P4 medicine approach to-, except now we’re applying. 

A lot more computational capabilities. Um, sometimes I think Leroy was a little ahead of his time.

Joel Dudley: Yeah, no, I bet Lee hood is definitely a visionary, uh, who is, was early in, in sort of the trumpeting, this approach where, you know, I’m a big, big fan and friend of Lee hoods. You know, and I think, thankfully now there’s more than just one group of folks that are beating this drum. 

Harry Glorikian: Right? Right. It seems like a number of people are coming out, but, um, you seem to be having some incredibly interesting successes and I’m wondering. You know, what  you know, what do you think in your background sort of might’ve contributed to that, uh, leap?  

Joel Dudley: Sure. Um, I think there’s a couple of reasons why we’re doing things differently. 

One, you know, which I’ll talk about maybe after is that we’re positioned in a major health system, you know, and a lot of people that were early on in, in this sort of, you know, P4 medicine or however you want to describe. We’re outside of the health system and it just turns out that’s a really hard approach. 

Um, but you know, my background, two things, I think one is I actually have a chronic disease myself. I have Crohn’s disease. And that really opens your eyes to a lot of the deficiencies in the current health system. I always give an example in my talks, which is if you have Crohn’s disease, like I do. You have a gastrointestinal manifestations, but you also have skin manifestations of the disease that rashes and such. 

So if you talk to your GI doctor and your dermatologist, you end up being ping pong back and forth between them and you really find no interest in connecting those dots whatsoever. You know? So that becomes glaring that you’re sort of moving from silo to silo when there’s obvious, you know, interesting biology to be discovered by connecting those silos. 

Uh, another huge, uh, Benefit from my past is I wasn’t this unique program at Stanford, uh, in the department of biomedical informatics where most sort of bioinformatics programs today where the computer science department reaching out towards the health and bio, but more firmly rooted in computer science. 

The Stanford program was a bit unusual in that. I think back in the seventies, there are a couple of doctors that said, gee, you know, computers gotta be important in healthcare someday. I’m sure they’d be disappointed today.  If they’re still alive, I’m sure they’re disappointed now looking at where things are, but so they actually started this program from within the medical school and we actually are, are, the program is, is rooted in the medical school. 

We take med school courses like biology, so I know how to read charts and all those other sorts of things. So that was, that was really helpful.  

Harry Glorikian: So, yeah, you’re, you’re doing, you’ve been doing this Institute for next generation healthcare. How’s that going? 

Joel Dudley: Yeah, I think it’s, it’s going, you know, it’s early, we’re just a little over a year into it. 

Um, but it’s going well. It’s, it’s, it’s totally different than I think anything that’s out there on the precision medicine world I think, you know, we’ve been at the forefront of really applying, cutting edge algorithms to, for example, electronic health record data, et cetera. And what became painfully clear to me is that, you know, all of the AI or so-called big data in healthcare to date, including the stuff that we published, I’ll pick on myself. 

I say it’s like building mechanical horses to pull a carriage, you know, meaning that, you know, we’ve gotten powerful new technology in terms of deep learning, uh, et cetera. Uh, we have an old way of doing things. Um, and obviously the electronic health record data was not designed to, you know,  be input for, you know, predictive health algorithms. 

So everybody’s so focused on building this mechanical horse to pull the old carriage of health care along, and we have this compromise solution. So we switched and the Genesis for the next generation healthcare Institute was realizing we need to rebuild. We need to get rid of the carriage right , which is the healthcare system. 

And so we’ve decided now, you know, and I’ve been provocative in, in meetings saying we have all the algorithms we need for 10 years. I don’t actually believe that. And we still need new algorithms, but you know, the pace of innovation and the algorithm side is way ahead of the pace of innovation, of the data collection side of things. 

Harry Glorikian: Yeah. I mean, well, if you look at, you know, the original premise behind an electronic medical record was it was the billing system.

Joel Dudley: Yeah,

Harry Glorikian:  Because that was never designed to do anything that we keep talking about. We want to do. 

Joel Dudley: Right. 

Harry Glorikian: And, and, you know, yes, all the politicians did the reinvestment and recovery act to have everybody move to an electronic medical record system, which was at least a move to digitize everything. 

Joel Dudley: Right

Harry Glorikian: But we’re far off from having it in the format that we’d ideally like to have it in. Um, so the next generation healthcare Institute, you know, do you see it as a disruptive you know, concept? Can it be duplicated in other locations? And can you see it working outside of a academic medical center? 

Joel Dudley: Sure. And I think, you know, the premise behind our Institute was, you know, how do we, if we could start over from scratch being sort of data driven minded people, how would we redefine or rebuild the healthcare system to support this vision that we can see for a more data-driven healthcare? Um, so we’re not-

A lot of, biomedical data science and precision medicine is trying to drag 2018 medicine kicking and screaming into the future. And we have this wonderful freedom here at Mount Sinai within a well-established health system to sort of start from scratch. And if we could start from scratch, how would we rebuild the health system? 

And how would we prototype that? How would we validate it all within the walls of, uh, you know, uh, in the scale of Mount Sinai makes this possible because people who are not familiar with Mount Sinai, they acquired continuum health care, and a couple of other health systems. So we are you know, the largest healthcare provider in Manhattan, for sure. 

One of the largest, maybe the largest in the state of New York, we have a patient population that equivalent to about a, um, a small Scandinavian country. You know, so the scale we have several, you know, several million patients extremely diverse, you know, almost 40,000 physicians. So we are a big playground to test, you know, even big ideas.

So, um, so there’s a couple of core principles that, that, you know, we’ve, we did a lot of work and we felt that yeah, of course the, the future is gonna be data-driven. But I think a lot of people who are in this data-driven medicine space don’t think too hard about this sort of last mile problem or sort of, uh, well, yeah, if we get a bunch of data together, it would be great. 

And if we could measure a bunch of things on people, it’d be great. Well, of course, you know, that’s easy to see. But then how you actually build healthcare systems, um, that would engage patients and physicians and actually facilitate the practice of medicine. In addition to that, you know, data collection and do so in a way that, um, people would actively and happily engage with that system. 

So we have sort of next gen clinical programs. Uh, one that’s out in the public has the Lab100 where we’re redesigning clinical environments to not only collect data, but we also have to demonstrate that people will have a better overall health care experience. The doctors will have a better experience in this clinic and that people will stay engaged in and find it valuable. 

So I think this is what sets us apart from what a lot of our people are doing is that we’re actually, you know, building these clinics and testing things out in real life health care setting.

Harry Glorikian: So that, that, you know, begs, you know, 50 questions, but two that come to the top of my mind is, are there, have you had any successes that you can share? 

Joel Dudley: Yeah. So, uh, um, well, one is we’ve, uh, in less than a year, we designed and launched a next generation health clinic, um, that does things completely differently than the current health system. And one there’s one thing really important about this clinic Lab100 is that it’s actually open for business and seeing real patients. 

And why is that important is because designing the technology and the user experience, which while very hard. It was like 20% of the challenge. The other part of the challenge was regulatory, right? So, you know, a lot of people can conceive of a high-tech clinic, but to be able to open the doors and meet all the regulatory requirements is super hard. 

People don’t realize because when you introduce new technologies into healthcare, there’s a bunch of regulatory implications. So one thing I insisted on with Mount Sinai is that we’re going to build a high-tech clinic. It’s going to do a bunch of things differently, but from day one that we opened the doors. 

It’s not going to be a research study. It’s going to be a real clinic. That’s open to the public. So we’re forced to figure out all of these regulatory requirements, just briefly, we take blood inside the clinic. In fact, we do it very quickly. Your blood results are done by the time you’re done with your visit. 

We {inaudible} lab license, you know, and things like that to do that. So, so again, one thing to conceive and actually implement, but a whole other thing open the doors for business.  

Harry Glorikian: Well, the other thing is, is the type of people that you need to hire to design the place is more of a, I think of it more like an Apple store or an Apple graphical user interface so that when the patient is looking at something, they. 

It’s not gobbledygook. Um, so it must have required a whole different group of people.

  Joel Dudley: Yeah. Yeah. Whole different group of people. Um, and not only does it have to be aesthetically pleasing, it has to be informative on someone’s health. Right. So it has to be actionable and all sorts of quite a few bars . Um, we worked with a design firm here in New York called Cactus. 

They’re based in Brooklyn, really interesting group of folks that not only designed physical space, but also folk focused on user experience. Uh, there’s a gentleman, David Stark. Who’s an MD who really led the development of [inaudible]. But he also uh, but also has informatics training. So, you know, you have to find a doctor that understands computer science and everything that, that goes into patient care and informatics. 

So it was a really unusual group of folks that we got together and to work together, to build something like this. 

Harry Glorikian: So this sounds like, you know, basically put together the pieces of a startup with the support of a large institution behind it that had to be willing to let go a little bit.

Joel Dudley: Oh yeah. So, and, and, uh, you know, we actually had this as almost a skunkworks and actually the, the clinic, if you come to visit us as in the library. 

So we, and there’s a number of reasons for that one is because um, one clinic at Mount Sinai is really boring. So the whole concept behind Lab100 is a scalable, smart clinics that could be deployed anywhere. So we wanted to show that we could deploy one in the library and also hide out in there, uh, as a skunkworks. 

Yeah. Last place he looked for a clinic is the library. So the, um, so we, and we did operate exactly like a startup within an incumbent. And I think there’s very few places that would let us do something like this, that wasn’t under the thumb of the, hospital administrators and saying, oh, there’s this new clinic is cool, but can you help me solve 30 day readmissions or something, you know? 

 Harry Glorikian: Right, right.  So jumping to another subject. So, you know, your bachelor’s degree is in microbiology. Um, so with all the predictive analysis that EMR chart, you know, review, et cetera, do you think that we’re getting better to understanding things like either predicting sepsis or, or understanding or predicting microbial resistance in, in, in different situations where time is of the essence? 

Joel Dudley: Yeah, definitely. I think, uh, well, in both of those fronts, we’re seeing, um, some good advances on just the pure molecular diagnostic side. Sure there’s people that have been looking at ICU data streams for predicting sepsis and sepsis early warning scores, but actually on the sepsis front, I think there’s a company called Informatics that has come up with some point of care gene expression diagnostics would look really promising, uh, still, you know, it needs to be integrated into the EHR and be ordered in at the right time. 

And there are some signals you can get from that data. Um, the microbial in terms of hospital part of infections, is that what you’re sort of, sort of talking about? 

And again, that’s a whole slew of companies.

 The speed of, you know, next gen sequencing with NovaSeq and some of these new machines is finally gotten to a point where you can do very rapid turnaround of sequencing, the natural environment to look for these strands from a metogenomic perspective, right? 

Which is a big challenge to identify novel strains and species at the whole genome level. 

 Harry Glorikian: So, you know, that sort of also touches , the microbiome, which, you know, there’s, you know, that’s eventually going to jump to, you know, some of the other work that you’ve done, but, um, how do you, what do you think are the barriers for that and the technological? 

Is it computational? Is it- ?

Joel Dudley: I think, well, microbiome,  we’re almost starting over again. And what I mean by that is that, you know {inaudible} meta-genomics. It’s you know, doing whole meta-genomics of, uh, all the species in say a gut sample has now gotten to sort of a speed and price point where we can do large, uh, cohorts, people don’t often understand the difference between say 16 S sequencing, which if you’re familiar with the direct to consumer company, you buy them. 

Or a lot of academic studies have used 60 S uh, ribosomal RNA sequencing, which really tries to understand the whole, uh, you know, microbiome from really one gene, if you think about it. Um, I guess one analogy I use, it’s sort of like getting the roster for a football game and trying to predict the outcome of the game just by knowing who’s playing, you know, kind of in, and even then you might not know their, I don’t know their last name. I’m not sure. I don’t know how the analogy works there.

 But, um, you know, meta-genomics, we get fungus, we get phages. Um, the bacteria and then you can even sort of infer gene expression using some new methods from, uh metogenomic. So, so now we kind of the point being, we gotta rebuild up now this knowledge base of just trying to understand the dynamic, uh, what, what is, you know, we don’t know what a normal microbiome is. 

It’s a huge, huge issue. Right? So, um, and, and there’s really interesting work coming out. Um, I want to point out, uh, uh, research R Segal in uh, he’s not a Technion Weizmann Institute. Who’s doing some really amazing work and microbiome. Um, really for the first time we’re seeing basic questions, like do probiotics stay in your gut when you can take probiotics, it turns out by the way, they don’t, 

Harry Glorikian: They don’t It’s a great scam though. 

Joel Dudley: Well, you could argue it’s good for probiotics manufacturers because you got to keep taking the probiotics for them to have an effect. Uh, there appears to be a core, just, you know, a microbiome that’s with you from your neonatal, you know, uh, stage, but then 70% or so, you know, variability from person to person. 

So, so we’re just sort of starting over again with this high resolution metogenomic lens.  

Harry Glorikian: Yeah, no, I, I. I’ve had this, you know, hypothesis that whenever I see new parents and they say, oh my, my, my baby, my child is wonderful. He sleeps doesn’t cry, blah, blah, blah. And then whenever I’ve seen the mother develop an infection or take antibiotics within days, that kid is colicky as you can’t believe after breastfeeding. And so my hypothesis is there’s some transfer of the antibiotics and you, you alter their microbiome, right? 

Joel Dudley: Yeah. 

Harry Glorikian: And again, it’s a N of what 10?

Joel Dudley: Right. 

Harry Glorikian: But it, it just keeps happening. And so I feel that there’s something there, but. Um, and like you said, normal microbiome,  normal for who?

Joel Dudley: Yeah. And within an individual appears to be very dynamic over time as well.

Harry Glorikian: Right

Joel Dudley: So like, you know, even, you know, at what frequency would you need to sample and et cetera. But we do know that it has, significant effects on, or appears to have significant relationship with, uh, things we care about. Like, actually this Weizmann group, did a nice study on postprandial glucose response, right? 

How’s your blood sugar respond to food. Um, you know, microbiome has a clear, uh, role, it seems from their work and adjusting your blood sugar response. Right. So that does matter for diabetes risk and things like that. So, but, but it’s probably the issue again, getting to our theme of, I guess we haven’t called it out but sort of pro complexity that it’s really, it’s an ecosystem that you’re modulating. Right. So I think the belief, just like the belief that we can find magic drug targets, that cure complex diseases. There’s not going to be very unlikely to be one micro, right. That’s going to, you know, cure a disease. 

Harry Glorikian: No, I, I, I think that’s gotta be impossible. 

Um, uh, and we don’t know what we don’t know, which is sort of the interesting part of it, but, um, bringing us to we don’t know what we don’t know. You published this amazing paper in June. Um, that’s gotten tremendous attention. Um, and I remember us talking about it, you know, beforehand, and I was, you know, blown away by the direction of it, but, you know, tell me a little bit more about this, you know, this Alzheimer’s link that you’ve identified. 

And is there a downside to the hypothesis agnostic approach when using AI, AML or techniques for discovery?  

Joel Dudley: Yeah, sure. I can. So first I’ll address the study, which, uh, you mentioned, you know, we’ve, we’ve found a connection between a herpes virus, specifically HHV six and seven were among the most prominent, which are actually Rhodiola virus. 

Um, and also HSV one, but to a lesser degree. So, you know, what we set out to do there was reflected in, you know, one of my, the quotes I’m most proud of that got in the New York times, which is sort of a funny one, which is I went looking for drugs and all I found were these stupid viruses, which I think was hilarious that that was printed. 

Um, that actually is true actually. Um, because we were trying to do, I think it’s important to let people know what we’re trying to find and that, so we were working on a data set called AMP AD- Accelerating Medicines Partnership for Alzheimer’s Disease, uh, funded by the NIH. Multiple institutions are funded to sequence their brain banks. 

This is definitely one of the deepest, you know, uh, profiled, uh, Alzheimer’s cohorts ever, or several cohorts ever assembled. I mean, we got the proteomics metabolomics, epigenetics, RNA, DNA on a post-mortem brains of Alzheimer’s and, and patients, are really the largest data set ever assembled on Alzheimer’s. 

And what we were trying to do with this dataset was build these sort of multi-scale network models that we like to do to connect the DNA to the RNA, to the protein, et cetera, um, build these network models and then look at are there transitions in these sort of gene networks between the healthy brain and the Alzheimer’s brain that would reveal non-obvious drug targets?

Right? So many people are aware that we’ve targeted predominantly beta amyloid in Alzheimer’s disease to hundreds, literally hundreds of failures. And I think you’re going to see some news in the next few weeks on some new failures, uh, I believe, uh, so down that route, And then to a lesser degree, our that’s still a significant, a [inaudible] but you know that as a targeting hypothesis doesn’t seem to have worked out. So we need new targets.

 Um, so we, we felt that this network analysis could reveal targets in this transition, and we believe that we want to target networks, right? That targeting one gene. So how could we almost find a drug that would rewire the disease network in a way back to the healthy network? So what we did was we build these network models. And really, we have algorithms that look at how did the, you know, this gene and this gene seemed to be dancing partners, you know, it’s in the healthy brain. And then they stopped dancing together in the Alzheimer’s brain. We do that for all genes and we can see there’s this sort of network rewiring. 

So would that reveal drugs? So we basically ran this algorithm and, and by ran it, I mean, it takes two weeks on a supercomputer to actually do this type of analysis. And what we’re hoping to get out were, you know, network drivers and network patterns that we could map back to drugs. But as we started looking at all the genes that were mostly being rewired between the two conditions, and we looked at sort of what is the known biology around these, um, what came screaming back at us was viral biology. 

So, um, so that was the first hint that maybe this is, uh, you know, antimicrobial antiviral, but that wasn’t enough to publish this paper, um, what’s unique about this MPD data set is that we had, um, for the first time, really next gen sequencing data, uh, from brains, clinical samples, this is really important because when people asked me, why didn’t we find this until now this, you know, these viruses is really the first time we could have found it and being because, um, all previous studies or most previous studies were based on microarrays. 

Right. And if you’re not familiar with microarrays, microarrays find what they’re looking for. 

Harry Glorikian: Yes

Joel Dudley: Right. So if they weren’t looking for viral sequences. Another lot of people don’t realize too, that the software, like one popular one called GATK, which is used for, um, um, processing the next gen sequencing data by most people. It actually filters out viral sequences for the most part by default because viral sequences are often contamination. So we actually, so even someone looking at the same data, if they didn’t know how the tool worked, if they’re using it like Microsoft office, you know off the shelf, they would have not seen it either. 

So we had to go back and write our own algorithms to map the viral sequences. So, um, what we then were able to do is go back and find actual viral gene sequences in the data, but not only find them in their abundance, but put them in context of these host networks to really see, not only are the viral genes present, say more often in Alzheimer’s, but are they interacting with host genes in a way that would, you know, tell us about how they’re contributing to Alzheimer’s pathology. So its a complicated story, but, um, the, get to your second question about, um, are there any challenges, I guess I touched on it with, with, with using these approaches with the JA2K sort of stories, right? 

A lot of these algorithms have assumptions and parameters, which are very complex. Um, and you know, you have to understand those, to use these big data tools, if you will, and the data itself, you have to understand how the data is generated. Right? So a lot of people don’t, I’ve had a number of students come from computer science and find these amazing discoveries that ended up being well-known sources of, uh, measurement error. 

And the data and they’re get really excited because they find this, uh, you know, structure in the data and they’re like, all we know that the sequencer always generates this noise. Oh, was that the sequencer? We know? You know, so, you know, so if you don’t know that then, uh, the huge risk there.

Harry Glorikian: Well, it’s interesting, right? There’s a, you know, we design these things and we think that we’re doing something good by filtering something out because we think it’s going to come from [inaudible]. In reality, that could be the cause of a particular trigger or a disease or something else. So it’s uh – don’t throw the baby out with the bath water. 

Joel Dudley: And I think in another area or where we might see some, you know, discoveries along these lines is, people don’t realize in clinical imaging, uh, the Siemens, the GE machines, et cetera. There’s a huge, uh, number of steps of data reduction that happened in these machines. So these machines are capturing lots of rich and data, several steps of data reduction that are largely driven by sure, their, their sources of technical area, but largely driven by humans. Who have decided that, you know, reducing the data in this particular way 

Harry Glorikian:  Right

Joel Dudley: – Is useful because it’s what they want to see on the output. 

Harry Glorikian:  Right

Joel Dudley: My belief is that if we start unleashing deep learning on raw, you know, say MRI or CT data, we’re going to find things we didn’t know were hiding in the, in the, in the data. 

 Harry Glorikian: Yeah.  I was reading something about how you can from the MRI data. Uh, this is again, um, oh no, I forgot her name up at MIT, but he’s doing image analysis where they can see the tumor way before, like the naked eye could see the beginnings of the tumor just based on what she was looking at. I wasn’t sure if it was raw data or was just early, uh, image analysis, but very interesting on how um, even in her own breast cancer, she was like looking at the images the machine could have picked it up three years earlier, which, you know, I always tell people if we can find it early, our ability to treat goes way up 

Joel Dudley: Correct

Harry Glorikian:  Compared to finding it late. Now it’s economically not a good thing to find it early. 

Joel Dudley: Right. 

Harry Glorikian: But for the patient, it it’s a good dynamic. Thank you so much for participating in the episode. It was great to have you here and, um, hope to talk to you in the future about all the things that are that you have going on  and going forward

Joel Dudley: Yeah, thanks. It was fun. I appreciate it.  

Harry Glorikian: And that’s it for this episode? 

Hope you enjoyed the insights and discussion for more information. Please feel free to go to www.glorikiancom. Hope you join us next time until then farewell.