Bio-sciences have made huge strides recently, and the pace is only accelerating. Much of this is being driven by data. Now, a new generation of Artificial Intelligence is pushing the boundaries even further.
Hey everybody, this is Chris Brandt here with another FUTR podcast.
We have with us Luca Parisi, Director of Clinical Analytics and Data Science at Citeline, a Bio-Pharma R&D Intelligence company dedicated to taking ideas from inception to delivery, using a lot of data a bit of artificial intelligence. So let's hear how they navigate this complex process to bring important new discoveries to market.
Welcome Luca
Citeline: https://www.citeline.com/
Click Here to Subscribe:
FUTR.tv focuses on startups, innovation, culture and the business of emerging tech with weekly podcasts featuring Chris Brandt and Sandesh Patel talking with Industry leaders and deep thinkers.
Occasionally we share links to products we use. As an Amazon Associate we earn from qualifying purchases on Amazon.
Bio-sciences have made huge strides recently, and the pace is only accelerating. Much of this is being driven by data. Now, a new generation of Artificial Intelligence is pushing the boundaries even further.
Hey everybody, this is Chris Brandt here with another FUTR podcast.
We have with us Luca Parisi, Director of Clinical Analytics and Data Science at Citeline, a Bio-Pharma R&D Intelligence company dedicated to taking ideas from inception to delivery, using a lot of data a bit of artificial intelligence. So let's hear how they navigate this complex process to bring important new discoveries to market.
Welcome Luca
Citeline: https://www.citeline.com/
Click Here to Subscribe:
FUTR.tv focuses on startups, innovation, culture and the business of emerging tech with weekly podcasts featuring Chris Brandt and Sandesh Patel talking with Industry leaders and deep thinkers.
Occasionally we share links to products we use. As an Amazon Associate we earn from qualifying purchases on Amazon.
The biosciences have made huge strides recently, and the pace is only accelerating, and much of this is being driven by data. Now a new generation of artificial intelligence is pushing the boundaries even further. Hey everybody, this is Chris Brandt here with another future podcast. We have with us Luca Parisi, Director of Clinical Analytics and Data Science at Sightline, a biopharma R& D intelligence company dedicated to taking ideas from inception to delivery. So let's hear how they navigate this complex process to bring important new discoveries to market. Welcome, Luca.
Luca Parisi:Thank you so much, Chris, for having me today. So I lead a team of data scientists focused on leveraging advanced analytics techniques and algorithms to develop and deliver industry leading clinical analytics solutions that enhance and optimize clinical trial planning and execution strategies, therefore positively impacting millions of patients lives.
Chris Brandt:That's a mouthful. That's a mouthful of we're gonna have to break that down a little bit because that because I think what's interesting about about it is you guys do actually a lot of stuff. And that's why that's why it's it's it's it's a big, big, it's a lot of big words here. But yeah, I mean, you guys literally take the idea of a for drug from the sort of the pathology of it to all the way to Manufacturing it and distributing it, right? So that's that's you don't you help companies take it from the inception, you know, taking from the pathology all the way to delivering. A drug for your customers.
Luca Parisi:That's right. In fact, we provide the comprehensive real time are in the intelligence services, uh, to really collect data from a global landscape of clinical trials with our products, trial trial, for example, as well as organizations and sites that actually run the trials and investigators that conduct them in the side trough and, and also to inform on the drug development pipeline. We did our. product called Pharma Projects that has got data on over 100, 000 drugs. So you can imagine the wealth of data that we provide and the insights that are coming out of it that are ultimately speeding up the process that it takes to manufacture a drug and ultimately deliver it to the patients that need them the most.
Chris Brandt:You know that that's a whole lot of stuff that you guys offer. Can you sort of talk through the process of developing a drug or a pharmaceutical? You know it I think it's a really complex and Expensive and time consuming process Can you walk us through those stages and how that all comes together?
Luca Parisi:So, you may have heard this many times, but, um, in the life sciences industry, it takes about 10 to 12 years to take a drug to the market. Wow. And it costs on average 1. 3 billion. That's crazy. That's really like a lot of time and costs, but most importantly, a very complex process that starts from scientific breakthroughs or discoveries, uh, when it comes to identifying the target of interest, uh, which is effectively, um, a particular molecule that then can target a particular disease to be able to treat it or, or to be able to slow down the progression of a disease and ultimately identifying how, uh, it's, it's. It is acting on the human body in terms of, for example, the specific sites of the cells and or proteins that it has to bind with so that it can effectively have its intended purpose to treat a disease. And after that. target identification. Um, there are some validation studies conducted to be able to guarantee the effectiveness. So the efficacy of that molecule. And then, um, there effectively from you can imagine millions and millions of molecules. We now reduce that number down to a few thousands. And, and effectively all of these molecules, thousands of molecules are being optimized and improved in terms of their biochemical structure. to ultimately yield the best response for a patient. And effectively you have, um, all of these combinations of molecules that ultimately have to yield the best efficacy as well as, for example, improve absorption, solubility and developability of the drug within the human body. And after that, we move on to the next phase of the clinical trial development, which is the preclinical phase. And that is where animal studies are conducted and
Chris Brandt:basically you identify, you know, a disease and you come up with an idea for like what kind of a molecule will interact with, you know, that disease. Right. And, and you're pulling this from a database of millions and millions of molecules. Where are you getting all, where are you sourcing all that information? How do you, how do you create that database? Where does that come from?
Luca Parisi:We ingest, uh, data from several clinical trials databases and public registries that effectively report, uh, drugs information alongside trial related outcomes. And then we are able to depict all that history from, uh, you know, preclinical through to phase one to through phase four studies and really trace the history of that molecule as it goes through the various phases of the clinical development. And really that's important to be able to understand the validity and the robustness of that molecule to treat a disease.
Chris Brandt:You're taking, um, a need here in terms of like we need to identify an effective molecule that will work against this disease and you've already got millions of these molecules that have been tested in some capacity, um, in other studies for other types of things. So you not only are you like saying, like, here's a molecule. molecule that can be built that will work to target this, but we also have history on it. We know it's been studied. It's maybe even made it through some pre, you know, preclinical trials, things like that. So you have some history on this too. So not only are you like helping to find things, just the realm of possibilities that could work in this situation, you're also bringing to it, a lot of information about that, that piece, right? To shortcut some of the processes here, right?
Luca Parisi:Exactly, Chris. So, you know, there are two complementary aspects of the drug development. One is around the scientific suitability of a molecule to treat a drug. The other one is around how likely it is that it's going to succeed to make it through the various stages of the clinical development. And that's, that is what is most important because you could have a candidate molecule that scientifically could be suitable to target a particular disease. But then, you know, if that or similar molecules did not make it through the clinical trial development, they are not never going to impact the lives of. patients that need them the most. So what we are trying to do is to strike a balance between the two and really understand suitability of the biochemical component alongside the predicted likelihood that is going to succeed to be marketed and ultimately ship to the patients that need them the most. So
Chris Brandt:how do you, how do you, um, you know, you said you, you, you take it down from like a million possible molecules that could, you know, target this, um, and you, you, you narrow it down to just a handful that, you know, could be good candidates to move forward with. How do you do that? Take it from the millions to the... You know, hundreds or tens.
Luca Parisi:So there are a few phases within this target identification and validation, as well as molecular optimization phases. And really, initially, it starts from identifying the specific target through some, pretty much some desktop research, I may say, if I may say. So he based on in vitro. studies, we would find that the, you know, the multitude of millions molecules that could be suitable for a particular disease. And then after that, we would need to understand how well they modify. Uh, the bodily functions of particular organs, for example, that are involved in that disease and looking at effectively how well they absorb within the body or how well they act within the body to be able to provide the particular response that is of interest. And that effectively, what that means is that we are going through a funnel at every step in the way we are reducing the number of candidate molecules. And then after that, we really need to think about whether. the molecule that is potentially suitable because it is modifying the body's, the body's response to effectively a biochemical stimuli, uh, in an, in a way that they could treat a disease. But then we need to understand is the dose. Uh, really inducing the desired response. And what that means is on a disease specific basis. So for example, you could have a drug that would target diabetes and obviously the dose, the expected response would be a decrease in the glucose in the blood. And then obviously you would have other types of conditions, for example, kidney, kidney diseases, whereby Uh, you would expect that a particular drug would induce an improvement in the EGFR, which is the estimated glomerular filtrate, effectively how well our kidneys work in being able to filter out the waste through the body and, and really being able to understand, uh, not just a drug. With respect to the expected response, but also the dosage of that drug to induce the best response in the human body. And that is again, through research, pretty much and modeling computational modeling to understand how particular ensembles of molecules can act in the human body. And that, that is where we go into the next phase of the identification of that molecule, which is the molecule optimization stage. And that means really, we now narrow it down to those millions of initial molecules to a few thousands of molecules that have got that expected, um, you know, response with respect to the drug and dosage. But then we need to be able to understand, are they actually absorbing well within the body? Do they actually have the properties that we require them to be so that they can not just be efficacious, but they can be safe. Can they be tolerated? for a particular term, you know, as opposed to, for example, long term versus short term, what are the potential side effects? And again, looking at computational modeling simulations to be able to understand alongside in vitro studies, whether we can narrow that down, narrow that huge set of thousands of candidate molecules to probably a few hundreds. And that's where we have really, uh, to look at. Uh, properties around how the body absorbs that particular molecule, whether it's soluble, whether it's developable within the body and, and really ultimately inducing the best response wise being metabolized well by the human body. Uh, and that is still within the realm of optimizing for the molecules. And then a few of those candidate molecules, um, for example, it could be tens of those molecules would be tested in animal studies thereafter.
Chris Brandt:Right. Okay. So, so now we've identified the molecule we've used. And, and, and I, and I understand that like one of the big things here is that you guys just, it's the amount, you just have such an amount of data that you can apply to this, that, you know, you really have a good ability to, to find, find the right thing. But so then we, then we kind of take that into, to you. Yeah. Phase one trials. And so like these trials are really complex, putting them together and making them work and making sure they're done well and, and, and that they deliver the kind of results, right?
Luca Parisi:In that respect, we have a global landscape of clinical trials that, as I said, are being taken from public registries as well as curated by our expert analysts. So we do have teams of 400 plus of about 500 analysts worldwide. whose job is really to lend their expertise in their particular therapeutic areas and diseases, as well as patient segments, which are really niche sub cohorts of people with a particular condition to be able to help us understand how to advise clients as to how likely is that their drug is going to make it through the various phases of the clinical trial. And in phase one, really, what we're looking at is to consider Let's say six to 10 subjects and really understand the effect of that particular treatment in humans for the first time. So really after preclinical or animal studies, we are looking into understanding the effects in humans. But as I said, the, the number of subjects is relatively modest. So we're talking about 10 subjects there. And, and then therefore we leverage our history of clinical trials and we look to identify similar. Trials that have been run in the past to be able to infer the likelihood of success of that particular trial that is being designed in within phase one. And that's a crucial phase because even though it involves a few subjects, it's crucial because at every step in the way. the clinical trials can fail. And we know how much that costs. I've mentioned it at the beginning of the podcast. So, and one thing that is worth noticing that, and you might have heard this reference many times, but I think it's important to say that 80 percent of the trials fail. And going back to the cost, we're talking about 1. billion us dollars per drug of Obviously costs that are, uh, unfortunately not being invested in that way for that particular drug. And that means that delays also the treatment for the patients that need them. So it's very important to be able for us to advise clients to increase the likelihood of success of the trials from a design standpoint. You might have heard of it, you know, set up your, for example, in this case trial by by design to success and really setting it from the study characteristics. For example, should we choose a particular subset of patients as opposed to another? Should we choose specific countries to run the trial on as opposed to others? How about the sites, the hospital? Do they have the equipment, the investigators and the healthcare professionals that can actually help the patients? Take them through the journey of visits that the trial is entailing and really equipping them to make sure that ultimately they are educated, enrolled, obviously, after the screening and engage the most importantly throughout the trial and all of these aspects. are important. And that's why we have actually solutions that target each and every step that I've mentioned now.
Chris Brandt:Historically, we've had a lot of challenges in this area, particularly because, you know, a lot of drugs did not get tested against a broad Subset of the population, right? I mean, you have a lot of things that were just, you know, never tested on women, or they weren't tested on people of color, or they weren't tested on children, you know, so there's, there's all these, um, you know, drugs that have come to market that we use on a regular basis that, you know, for some of the population never really kind of went through a clinical study, you know, that targets their, uh, specific situation, right? So, like, what you're doing here is you're really, um, I mean, it's, it's two things. One, you, you guys have the, the ability to design these studies effectively. You have the ability to go out and find these populations effectively, and in addition, you have this massive cache of data about some of these molecules that have already been through the process and what happened in that process. So, You're shortcut, you're, you're, you're improving the quality of the study and you're shortcutting the, the process a bit, right? Is, is that a fair assessment of how you're
Luca Parisi:doing it? Exactly, Chris. That was very well put. And, uh, uh, as I said, we do have that global. database of clinical trials, but we also have to your point, complementary data around real world information about patients. And that is extremely important and not supplementary, it's complementary to what we know from the history. Because as you said correctly, what if a drug hasn't been tested yet on a particular patient population? Well, we can advise our clients as to where that particular patient population is and also Model of that likelihood of success based on similar trials that have occurred in the past and really tying up the present, the, uh, the most up-to-date patient availability with the past history about running clinical trials that are similar to that of interest and really understanding the likelihood of success of a trial in other cohorts that have never been tested before. For example, as you correctly said, diversity across all of the various facets that it means, right? Starting from gender. Really? And then obviously you leveraging age groups and then, and then also looking at race, ethnicity, and really understanding what if now as a client, I'd like to run a trial in a particular sub cohort that is diverse, specific for my study, because I know the incidence, the incidence has increased in a particular disease area. Will I be able to be successful? Well, The way whereby we are helping our clients there is not just based on the past history of similar trials, but also tying that story with the most up to date patient availability to date. And we can pinpoint as to where they are, particularly for the U S uh, where obviously we have a huge amount of data, uh, that really can supplement those, those insights.
Chris Brandt:So, like the process goes, you know, like we've tested it on animals, we've got it in a phase one trials, which is about 10 ish people kind of range, it sounds like. And then you go to phase two and what does, what does that look like?
Luca Parisi:So in phase two, um, the number of subjects is increasing, obviously at every step in the way, the number of subjects increases, but also the focus becomes more holistic. Okay. In a way that in phase two, we are no longer concerned about the effect of a treatment, but we're also ensuring that it's safe for humans that received it. So really looking at safety profiles of the drug at various dosages, for example, and looking at the same time, balancing it out with the efficacy of that treatment. So achieving a fine balance between safety and efficacy for those 20 to 50 subjects that typically are of interest in phase two studies.
Chris Brandt:Okay. So now, now you you're kind of. Like honing in on, on dosage here and you're, you're trying to, you know, like you're, you're, you're getting a bigger cohort. So now you go into phase three and what's, what happens in phase three?
Luca Parisi:After phase two, um, the, the optimal dosage has been, uh, effectively identified. And in phase two, it's the, the, the thing that is important to note is that in some cases there will be analysis against the competitor drugs as well. Okay. And that means that we're in phase three, we are starting with a candidate drug. That has been proven safe and efficacious on over, let's say, 30 to 50, 20 to 50 subjects. And it has also been proven potentially better than a competitor's drug. It could be because maybe it has the same safety profile, but, but it's more efficacious. So in that case, it is better than the competitive drug. So in phase three, we're looking at hundreds of patients really from 100 to 200 or even more in some cases, it could be a thousand depending on the indication. And we're looking to confirm that the same or similar safety profile and efficacy will hold on a larger population. So really to confirm that what we've seen in phase two is going to effectively hold with a larger statistically larger population.
Chris Brandt:Okay. And then so now we, now we're moving on to the next phase. And so what does that look like? Like we're on phase four now, right?
Luca Parisi:That's right, Chris. Yes. So we are going outside of the, um, more controlled setting, what we call randomized controlled setting, which depending on the study designs, obviously some, some of them are randomized control trials. Some others are, um, in a structure a bit differently, but really we are looking at taking the study out of the clinical trial, hospital based trial and looking at how it performs in the real world in the wild, literally where the patients are taking the medicine to be able to improve the quality of life or in some cases to save their lives. And really, uh, conducting observational studies to understand whether the medicine is actually safe and effective in the long term. So there are different follow up times that are being considered. And we are talking about hundreds, in some cases, even thousands of people being involved in observational studies that, as I said, they are not under a controlled environment like a clinical trial, but really. Right. Prospectively in observation so that we can understand the behavior of a drug and the safety and efficacy in the wild and really proving that the drug still holds and, and it's still safe and effective for that particular patient population. Note that in some cases when that is not the case, a drug can be taken off the market. So it's crucial to make sure that the drug's intended effects, including both safety and efficacy will hold. for a long time.
Chris Brandt:Yeah. And you don't want to, you don't want to be shutting down a trial in phase four because you've spent a lot of money by that point, I imagine.
Luca Parisi:Exactly. And that's why we are helping to predict the likelihood of success along the phases of a clinical trial so that they also understand projecting forward, what is phase four going to look like?
Chris Brandt:So then after phase four, where, where does it go from
Luca Parisi:there? In that case, it's confirmed that even in the wild, after observational studies, the drug is safe and effective. So it continues to be on the market and until obviously, uh, you know, it's, it's, it's possible. So it all depends again on whether the various regulatory bodies receive, uh, Uh, warnings from the public or that obviously have to go through, uh, reviews of pharmaceutical companies that have me have manufactured those drugs. So everything really can happen after that as well. But the likelihood of success is pretty much guaranteed after that.
Chris Brandt:You help these companies also like figure out how to manufacture this stuff as well. Is that, is that correct?
Luca Parisi:I would say partly in a sense that our. Uh, and I would say indirectly rather than directly, because we do have a database of drugs that as I said, it's called pharma projects, which has been used by, uh, sponsors. So pharmaceutical companies that are manufacturing the drugs, however, they are consuming it. And as a, let's say, source of R and D intelligence to be able to then. make up those relationships between molecules and understand how to best effectively design the drug. But we are mainly involved instead within the traditional clinical trial development within human studies. So really from phase one to two to phase four studies, looking at again, optimizing the likelihood of success given certain study characteristics, for example, particular protocol design criteria. or eligibility criteria of, of, uh, of the specific patient population to be enrolled. How to best enroll the subjects across various countries and sites and which investigators would be best equipped to actually drive the conduct the study as a principal investigator given not just their past history, but also the presence of, uh, patients that match those particular inclusion and exclusion criteria. So really we are mainly involved from phase one to phase four studies, I would say as a company.
Chris Brandt:Now, another thing that I think that kind of grows out of, like, all this data that you've collected and you have, um, is, is that sometimes drugs come to market and they're effective in the market and they're widely used, but, you know, we've, we discover that there's off label uses, right? They're, they're like, You know, diseases and things that these drugs can target and be beneficial for that they weren't initially created and the studies were designed to treat because you have all that clinical data. You can identify off label uses that may be effective, right?
Luca Parisi:So one way to do so is to actually understand the relationships between certain molecules and certain diseases and looking at really understanding the, and then tying that back to the historical success of certain drugs in those particular diseases. So it's the combination of this wealth of data that really ensures that. Clients can understand how to best repurpose the drugs to be able to treat additional diseases that they weren't even thought of before. And we've seen that with benevolent AI, for example, repurposing Illy Lilly's drug that was meant to treat rheumatoid arthritis, like baritisinib. To repurpose it to treat severe cases of hospitalized patients with COVID 19. And that was done in record time. And it would have never been possible unless AI would have been used to mine those patterns and those associations, you know, between, between the data points and really pinpoint was to which of those, which of those millions, potentially millions of candidates could have been viable for a trial in humans at the time where humanity was really hanging. So it was really important to use AI to be able to do that in record time to be able to face the pandemic.
Chris Brandt:Now you're talking about something that doesn't even have to go through that whole study process necessarily that you, you know, you just outlined because it's already gone through it. And, you know, the efficacy and the safety at least has been, um, identified. And now you're just saying, like, this will also, you know, it's kind of like a freebie. You know, we made this drug to solve this problem, but it also solves this problem. And, you know, that, that, that's, you know, the time to market for that and the, the ability to, um, You know, help people quicker is, is really compelling there. And using AI to get there is, is a really interesting piece of that.
Luca Parisi:Again, going back to the example I provided before, it took about two years and roughly three months to get to the stage of identifying that particular molecule and, and repurpose it for COVID 19 really. And that would have never been possible without AI. As I said, it typically takes 10 to 12 years. And as you said, correctly, the reason is that. The safety and efficacy profile were already proven in particular indications, like rheumatoid arthritis. So it took actually much less to be able to figure out and obviously facing the pandemic with an emergency use authorization from the FDA to be able to market the drug fast and Following successful outcomes in phase three studies. So they still go through obviously, uh, the clinical trial development, but with much more accelerated timelines. And obviously in terms of when we are looking at, uh, the need from a population health standpoint, that also helped to speed up the, the, you know, really the life cycle of a trial and ultimately achieve a 38 percent reduction in mortality considering that drug repurposed by AI.
Chris Brandt:That's amazing. But, you know, you can take that even a step further, right? Because I know that one of the things that you guys are looking at too is, is, um, the idea of, uh, you know, personalized medicine, right? You've got, you've got all this, this data and, you know, individuals have sort of unique kind of, you know, biology and, and situations and comorbidities and all that. And, and you can apply your data to, to help with, with that as well, correct?
Luca Parisi:So it ultimately goes back to. Tying up this multitude of data sets and for example, when you consider biomarkers lab from lab tests or medical imaging data, all of these tell a picture about the patient and really help us to pinpoint as to which other historical patients they may be more similar to and really to achieve that. Uh, both personalized medic medicine approach as well as understand whether certain treatments in the past for a similar patient have worked for that particular individual. So really not just modeling that subject in isolation, but in tandem with historical patterns that are best suited and more closely related to their individual needs and profile with respect to obviously their. Um, biochemical, um, biochemical, let's say, uh, nature that could be due to, uh, biomarkers as I said, that come in from various places. Uh, it could be from medical imaging, as I said, or lab tests.
Chris Brandt:It brings to mind another aspect of this, right? I mean, cause like one thing you're doing here is you're helping to move. You know, the development of drugs through the system faster, um, with greater success rates, um, and I got to imagine that also helps to reduce the cost of developing these things, which then makes me think, you know, with with all this that's going on, you know, one of the one of the areas that. of medicine that is often overlooked is sort of the rare disease front, right? And, and, uh, I can imagine that, you know, like by reducing the costs of developing these drugs, manufacturers can target maybe more rare diseases because, you know, they don't have a huge. Base of people who are going to be buying these drugs, but now now because they can reduce the cost, it becomes more economical to makes more economic sense to develop for these drugs. But then on the other side, too, because you have this sort of off label use kind of situation, you may be able to identify You know, treatments for rare diseases that, you know, nobody was really even investing in.
Luca Parisi:I would say that's an opportunity coming from this acceleration of clinical trials driven by AI. And really, it's so beautiful to think about that we are able to then, yes, Let them focus on those populations, niche populations that really need a treatment desperately because of the rarity of their condition and therefore really having the attention they deserve to be able to then ship a drug that really helps them. So in the end, AI is not just an enabler for the traditional clinical trial workflow, but also opens up new opportunities and avenues. to be able to be more equitable, I would say, in the, in the production and the privatization of certain treatments for even those, uh, niche, niche populations that need them.
Chris Brandt:It's good to see, you know, people who didn't get access to some of this getting access to, to some of this, could you, you know, you must have a lot of really, um, Interesting stories. I mean, I, I'm sure that, you know, so much of what you do is under deep NDAs, I'm sure, but I mean, can you kind of generally speak to some of like the really interesting success stories you've seen
Luca Parisi:in this incredible success stories that you have seen personally is around really, um. Site recommendation, investigative recommendation, as well as informing the country strategy as to which countries should be prioritized. And that is crucial because many trials just fail because sites don't recruit. So a lot of trials fail just simply because these hospitals, they were maybe historically, uh, potentially, uh, useful to recruit specific patient populations. They are not performing as they, they were meant to. And the way again, whereby we have helped clients to recruit in this cases is to tie that historical landscape of trials, trials data with the up to date patient availability and really understanding. You know, whether the site performed well in the past, but also whether the site in that specific principal investigator have access to the patients to be able to recruit in the study. And that was a massive success story really to paint the, the entire holistic picture going from study design, protocol optimization through to choosing the right countries for your study, including the sites, the hospitals that run them and the principal investigators that are conducting it.
Chris Brandt:That's a real shame that like, you know, potentially promising treatment is just never makes it to market because you just can't find enough people. You know, that's just, that's a seems. like a terrible outcome. Um, I, I got to imagine they were very happy to, you know, have you find, find those people for them, right?
Luca Parisi:That's the case. That, that is really the case when it comes to, uh, clinical trial success. It's, it mostly comes down to recruitment. If you can recruit enough patients in the right time, then you've got that statistical power in your study. And, and that is super important. So what I mentioned earlier, when, you know, it wasn't just a textbook definition of saying we need 10 patients in phase one. You know, 20 to 50 patients in phase two hundreds in phase three and even thousands. But that is because. We ultimately need to ensure that statistically the drug will be safe and effective. So recruiting the right adequate number of patients is crucial. And this is the real reason why many trials fail.
Chris Brandt:That just seems like a real, uh, real bad way to, to lose some, some good, good possibilities. Um, so, so tell me like what's next for sightline. I mean like, where do you guys go from here? You, you, you know, you're, you're, you're helping bring drugs to market faster, um, and better. Um, so what's, what's next for you guys?
Luca Parisi:To be honest, baking some solutions that are quite cutting edge that leverage, for example, generative AI, as well as the more traditional machine learning algorithms and statistical tools to be able to take a drug to the market faster at all of the phases of a clinical trial, but also. All of the phases of the, it's life cycle. So coming from, let's say protocol design through to optimization. So re optimizing those inclusion and exclusion criteria so that a study by design doesn't end up being a rare disease study because the specific inclusion and exclusion criteria are too strict. So, and that's also a pain point in the industry. We are helping our clients and we will help our clients with some very cutting edge solutions, uh, around specifically. Optimizing those inclusion, exclusion criteria and really helping them even before understanding which countries and which sites and which investigators to choose to set the trial up for success. So we really want to make sure that the study is designed with a way whereby it can recruit the patients that it needs to be able to ultimately bring that drug to the patients faster. So really that that's, that is a key component of our focus at the moment.
Chris Brandt:When you start introducing AI into all of this, just the possibilities get endless. And I think that, you know, at some point too, you know, there's going to be some really interesting, you know, modeling and things like that, that's going to come out of this, that's going to help maybe even shortcut some of these, you know, trial phases, right? You know, because we could do, you know, some of that analytical work in models, you know, Uh, ahead of actually trying it in humans to get a better, better understanding of how it's going to react, right?
Luca Parisi:Exactly, Chris. It's all about using technology to assist the clinical trial designers, planning and execution teams on the ground to ultimately foresee potential risks and meeting. And understanding which mitigation actions they need to put in place before they actually happen, before those risks actually become manifest. And that's how we are enabling our clients with the, our data, their data, as well as our AI capabilities that sit on top of them to be able to ultimately yield those insights before the risks become manifest.
Chris Brandt:I gotta say thanks for doing, doing the work as I get older and I, and I'm now on more things because, you know, getting old sucks. Um, it, it's just like, I, I'm, I'm excited to see, you know, organizations like yours, companies like yours really sort of speeding up the development of these things and, and, and just, and, and acting is kind of like a force multiplier on, on, on that discovery front, right? And I think that's just. That's just awesome. And, uh, you know, keep, keep doing it. I, I, I, I want to, I want to see where you guys go next. I'm excited to see what, what comes out of, you know, your efforts. And, uh, I just want to say thanks so much for being on and telling us all about this, what is sometimes a very confusing. you know, uh, landscape for, for the lay folks here. So, um, thanks so much for coming on and explaining it and, and thanks for driving us into the future. Thank you much,
Luca Parisi:Chris, for this great opportunity and great as well, uh, possibility to be able to ponder upon the journey that we've, we've done so far and where we're headed next, ultimately with the same mission to ultimately impact the millions of patients lives out there positively. Thank you so much, Chris. Yeah. Thanks
Chris Brandt:for watching. I'd love to hear from you in the comments. And if you could give us a like, think about subscribing and I'll see you in the next one.