5 min read

Personalized data could bring physIQ to the forefront of clinical trial evolution

Originally published in PharmaVoice

As personalized medicine gets even more personal, patient-centric data is becoming more of an asset in the life science industry. And many experts are wondering, just how far can new technologies go to deliver these tailored therapies?

For physIQ, a digital medicine company developing remote monitoring analytics, that question is a driving force. The company uses wearable biosensors to collect patient physiologic data, like heart rate, respiratory rate and temperature, and feeds the information into its AI analytic platforms, accelerateIQ for clinical trials and pinpointIQ for healthcare settings, to create a “digital twin” of the patient. That synthetic twin acts as a personalized baseline that can be used to detect physiological deviations and predict adverse events.

“A lot of people claim that they have AI. It’s not really AI. It’s often just a decision support tree. And I think that’s really where I would suggest our key differentiation occurs,” physIQ Chief Strategy Officer John Varaklis says about the company’s proprietary technology.

The digital health field has become increasingly crowded and is expected that it will grow 17.9% between 2021 and 2030. But Varaklis suggests that physIQ’s ability to translate continuous biosensor data into analytics in near real time sets it apart and provides partners with greater control of decentralized clinical trials.

“We see the development of an ecosystem of a variety of different partners as critical to helping change the way that clinical trials and ultimately healthcare is also provided and managed,” he says.

physIQ announced in 2021 a multi-year licensing partnership with J&J’s Janssen Pharmaceuticals, allowing the pharma giant to expand its use of biosensors across clinical trials. The partnership built on a 2018 oncology study to identify biomarkers that could predict adverse reactions to cancer treatments. More recently, physIQ has forged partnerships with clinical research organizations including Syneos, Science 37 and CellCarta.

Here, Varaklis discusses the company’s latest partnership to study immune responses to vaccines, how it hopes to shift the paradigm of drug safety and why its therapy agnostic approach is so important. 

This interview has been edited for length and clarity. 

PharmaVoice: How is physIQ’s approach different from other platforms on the market?

John Varaklis: There are a couple of dimensions that are differentiators. The first is, typically what has been used in my experience in clinical development is a comparison to a general population. So you were linked to a dataset that likely looked like a 55-year-old white male who weighed a certain amount and had a certain health status. Well, not everybody is the 55-year-old white male with comorbidities. The approach that physIQ uses is to compare you to you. We create the digital twin that understands how you react, how you behave, how your activities are focused on daily activities, and then it compares that with those changes in near real time to you. 

And so detecting changes, detecting and understanding how therapies or how your activities of daily living may be improving or worsening, is much more relevant. It’s much more informative for clinicians, health care providers or even yourself to better understand what you would need to do to sustain or improve your health. I think that’s one of the key parameters that differentiates.

The second link to that is, I’m convinced we have probably the most unique data trove in the world. We have over 5 million hours of on-body multivariate bio sensor data. And I think that type of foundation clearly allows us to differentiate from others in the space. It gives us the opportunity to test the analytics on very relevant data and significant volumes of data so that the analytics can be as precise as possible. We can limit the false positives and false negatives that have plagued a lot of the solutions that are out there. 

The last piece is that we’ve actually delivered on the promise of digital biomarkers. We have six that have been cleared. We have several more in the pipeline now that are going to be going forward to the FDA for clearance. And we have successfully used these in clinical trials and in health care management. 

Your partnership with CellCarta is focused on monitoring differences in how people react to vaccines, even some for COVID-19. Could you tell me more about that trial?

Our hypothesis was that when you receive the COVID-19 vaccine, for example, it’s likely that your body produces and reacts with a release of inflammatory cells, cytokines and potentially other cells. We wanted to see whether we could demonstrate this with our physiology analytics that we’re developing to look for these very, very small changes in the human body linked to an inflammation response. We noted that CellCarta has developed some sophisticated ability to test for these types of cells in blood or other types of tissue. So we partnered together to create a program where we would monitor patients after a baseline. We would put a sensor on the patient and we would understand what their behavior and their physiology look like for a period of five days. We would then have that patient receive the vaccine. And then we would collect blood and monitor how the physiology changed and look for what the changes could be linked to in the patient’s blood. We’re hoping to understand some of the immune response whether that’s due to interleukins, interferons, T cells, etc.

Are there any areas of medicine where your technology works best, or doesn’t work? 

This may sound arrogant, but no. I think what we’re excited about is that the clever individuals that started to put a lot of this foundation in place early on at physIQ realized that being therapy area and sensor agnostic is the right way to go forward. It took a lot of effort. I always jokingly say that they did the hard part first in enabling this unique platform to be agnostic to most of the inputs and the outputs. Which means that when looking for partnerships, it becomes a question of identifying, for example, parts and pieces of ecosystems that work well with that model.

You collect tons of different data points — heart rates, temperature even voice recording data. What is the impact of collecting voice data?

For voice, I think, just listening to my voice now you can probably tell I’m tired. And so there are many, many signals that one can extract from the voice quality, the speed, the reaction time to answer questions, the monotony or the tone of voice, whether it’s dynamic or whether it’s very flat in its effect. All of those can be used to point to different types of conditions and change in a patient’s status. For example, if a patient is depressed, they may speak more slowly, they may speak more softly, they may not be speaking as much. We’ve seen some very interesting research looking suggesting that the way people speak and pronounce certain words may be a very early indicator of some neurological disorders, potentially Alzheimer’s or multiple sclerosis, Parkinson’s etc. So there is also that longitudinal opportunity to use voice as well.

What mountains are there left to climb in terms of collecting patient data?

One area that I see as an important consideration for the future is to be able to collect as much information as possible using passive approaches. So looking at different ways to use some of the sensors that are being developed to collect information without necessarily influencing the patient and potentially changing that patient during those measurements is an area that we’re considering as an important next step.

A lot of effort is spent in clinical development, studying whether a product is safe. One would argue that you can almost not have any efficacy, but as long as it’s safe, then it’s OK to put on the market and the FDA will be OK with that. What we’re trying to do with our analytics, is provide a different paradigm for pharmacovigilance and that is to be able to potentially predict when events may be occurring and provide opportunities for mitigation.

We’ve got some great early results linked to the vaccine program that I mentioned earlier, about detecting inflammation and being able to see, for example, in certain oncology therapies, the prodromal signature of inflammation occurring much sooner than what is typically used now which is basically detecting a fever. Regarding efficacy, again, I think with the way that we’re collecting and providing insights, I think we’re going to be able to help a number of the pharma companies that we’re working with now identify new ways to measure the impact or the effect of their therapy in real world settings.