Dr Miruna Sasu Discusses the Transformative Power of Real-world Data

,

Miruna Sasu, PhD, MBA, president and CEO of COTA Healthcare, discusses the medicines discovery process and how real-world data can be used to evaluate patient and tumor response to treatment regimens.

Miruna Sasu, PhD, MBA, president and CEO of COTA Healthcare, discusses the medicines discovery process and how real-world data can be used to evaluate patient and tumor response to treatment regimens.

Transcript

Can you discuss how real-world data can shake up the traditional approach to research and development?

I could talk about this for hours. In fact, I did spend a lot of time in life science. I started as the head of digital health and innovation at Bristol Myers Squibb, where we incubated companies like COTA. So they came to us with a glimmer of an idea and we essentially productized and helped them to develop and perform the strategy where those companies would be producing things for life science companies. The first thing we did was to attempt to bring in the data—both the real-world data and the data science—into every aspect of product development and product launch at a pharmaceutical company. That starts in discovery.

So, discovering medicines with data, what does that mean? First of all, it can't be just with data; it has to be done in a lab; it also can be augmented with datasets. What are they looking for in discovery? They're looking for signals to figure out if a particular person's tumor responds better or worse, given that they are positive for a particular biomarker or biomarker set. So, you can actually do that work in the lab, right, in a wet lab, on tumor tissue. You can actually also do that work as a data science project in a database. You can take data from different genetic screens on tumor types in real people that have had it done. So, someone who has a genetic panel that's been done on their particular tumor, you can take that data—take many, many of those—and run your analysis on what happened. How does it work? Do they actually respond to a PD-L1 [programmed death-ligand 1] inhibitor, for example?

So that's in discovery. Then you move on to the development organization. So, once a compound has been discovered, it gets pushed into a sort of machine that is clinical trial evidence generation. There are lots of teams that will take that compound, they will sort of productize it, send it out to sites, and then perform clinical trials [that] put it in humans. There are ways to do this in an actual database as well. So, before the trial even starts, you can look for patients that match particular criteria. You can imagine in a dataset, you can say, “I want patients with X, Y, and Z, this type of tumor, so show me where those patients might be.” So, you’re looking on a map where those patients might be, and you choose the sites that we're going to work with based on where those patients are. Instead of making those patients come to your sites, you can figure out which sites are getting the patients and go to them. So that's a really big deal, because before this was ever thought about, it wasn't done that way. It was always site selection first; you selected your sites and then you took in patients. Well, that's what makes the patient have to go 150 miles to the site. So doing it in the opposite way, sort of patient driven, I would say it's more like patient matching, not site matching, not site selection.

Doing it in a database this way really makes the development in a clinical trial much easier and less burdensome on the patient, and just better for the trial altogether. Because you can also look for whatever population of interest there might be, say, for example, in multiple myeloma. You want to enroll Black patients. Well, of course you do. So go to them; you can look in the database first and go to them.

So say the drug has launched, we’ve done all the clinical trials, it’s launched and out there, post marketing the FDA always asks for more trials to show how it's working in the real world data. The clinical trial is a very small population that was meant to get the drug to market. Well, how's it working in the real world? Before, we would have had to do an actual trial, we would have had to go out, open a trial and say, “OK, let's see how these patients are doing. Let's bring them into the clinic.” You don't have to do that anymore—because you can just take one of the databases from COTA and see how those patients are doing based on their medical records. You can literally see all the patients that are, for example, on [pembrolizumab]. How are they doing after the drug has launched? What's going on? What sort of events are they having? Are there any quality-of-life issues? These types of things can be garnered directly from the electronic medical record.