Editor, Rebekah Jordan, sat down with Dr. Gen Li, president and co-founder of Phesi, to learn more about how the company’s AI-driven clinical database is allowing insights and observations to be drawn from big clinical trial data.
Key insights:
- AI can ensure that all aspects of a clinical trial are in sync, giving a higher level of assurance for success.
- Breast cancer was the most studied clinical area in 2022, despite the fact that the overall clinical trial development activity was declining.
- The number of clinical trial sites only able to recruit a single patient for a particular trial is increasing, with an increasing number of attrition rates.
- Dencentralised trials could minimise some of the issues associated with clinical trials, and may encourage more participants for trials.
Q. Could you tell us about the kind of work Phesi does?
DEVLIN
A. Let me start by saying there is nothing more objective than data. The industry has many mechanisms to collect large amounts of data, but they have shown little value to us.
The way in which Phesi works is that we have huge mechanisms for collecting, structuring and interpreting data. Using different types of big data technology and artificial intelligence (AI), we have been able to assemble a huge database, from patients to clinical trials.
The history of Phesi over the past 15 years has always been about innovation. We are innovative and intuitive in the ways we collect, process, and interpret data.
We can identify the most important investigative sites around the world in a way that creates a much more detailed, quantifiable profile, through an evidence-supported process.
Q. How does the AI work to gather and analyse the data?
A. Our method of interpreting the data is patented and protected, but we can share that we gather the data with our AI mechanism.
The AI gathers data by using a formula that is much faster than human processing. It also has a kind of self-improving process - just like a human being - minimising the need for manual input and handling.
By interpreting the data and looking at the patterns, we started building algorithms. These algorithms are then fed into AI to interpret the data. The results feed back to the mechanism, encouraging a self-enhancing, self-improving process – achieving better and more accurate results each time.
Q. Breast cancer was highlighted as the most studied area of 2022, does Phesi research the reasons behind such data?
A. We found that breast cancer was the most studied area in 2022, even though in the same year, the entire clinical trial development activity was declining.
One of the factors at play here, not surprisingly, is that breast cancer is the most profound cancer type impacting women around the world, at almost all ages. It affects a large amount of the global population, which could explain why it came out on top.
The second is the complexity of the disease. Over time, our understanding of breast cancer has been improving and getting more sophisticated. Previously, we were treating breast cancer as though it was one disease. Now, we know there are many different subtypes and each of them becomes one of its own kind of disease. I think we have identified around 20 different diseases all under the umbrella of breast cancer.
Although we can treat some of these subtypes, other aggressive subtypes prove more difficult to treat, which is why we are intensely working on them to understand how they work and act, to develop better treatments. I have no doubt we'll achieve those things in the coming years.
Q. What reasons do you think are behind the increasing clinical trial attrition rate?
A. Oddly, it's been anticipated. With physical restrictions preventing national and international travel, the interaction between physicians and patients became very limited. Even so, it became much more difficult to recruit patients for clinical trials.
A recently published article in Applied Clinical Trials showed that the number of sites only able to recruit a single patient for a particular trial is increasing.
Even in a normal situation, recruiting cancer patients is a difficult process – and 40% of the trials being conducted now are cancer trials. With the factors derived from Covid-19, along with the situation of the Ukraine invasion, there were physical restrictions on certain populations that we could reach. As a result, the recruitment figures certainly declined – which affects the reliability of the data that is being gathered.
In terms of the causal analysis behind the high attrition rate, we know Covid-19 had an impact – we can see that from the timing of the increase. But what is more interesting is whether this impact is going to be temporary or have a lasting effect. At this point, we don’t have a clear answer yet.
Q. What effect could decentralised trials have on the current attrition rate?
A. The decentralised trials concept has been here before the pandemic, but what the pandemic forced us to do was to find solutions in a situation that was not optional. I believe they are here to stay. Even with Covid-19 behind us, the effects of applying these trials will make the resources for pharma companies more productive, and it may also encourage more patients to participate in clinical trials.
However, it would be naive to think decentralised trials are going to solve all our problems. There are certain situations, and certainly many diseases, where you simply cannot apply decentralised trials. For example, the patient-doctor relationship could be negatively impacted which affects the success of the trial. It’s a similar situation to working from home. Although it opened new opportunities, it didn’t come without difficulties.
Additionally, the pharma industry is a conservative one. It is hesitant to adapt to anything new, no matter how much advantage is offered. If we can get more people to understand the advantage and efficiency of decentralised trials and AI, and how they can be brought in to help handle big data, we will be able to see the application of these technologies grow within clinical development. AI can ensure that all aspects of a clinical trial are in sync, giving a higher level of assurance for success.
Q. What trends in clinical development do you expect to see this year?
A. I'm hoping we will see the level of clinical activity coming back to the way it was in previous years. As we find better ways to manage and handle certain diseases, the level of impact will continue to be minimised.
What we are going to see, and what we are already seeing since working with our clients, is that more people are understanding the value behind big data, and what data science can bring into this space. In turn, the industry can create smarter clinical trials, bringing faster cures to patients – I hope to see this trend continue.
Generally speaking, chronic challenges within the whole clinical development industry are going to continue. There's no such thing as a perfectly designed clinical trial, we can categorically say. The difficulty of recruiting patients will take some time to resolve.
Finally, we are swimming in a mass amount of data and trying to understand what data is important to us. We aim to build a good structure which will make it easier to spot certain patterns and trends. With a better understanding, a more data-driven approach, and a collaborative effort, we can address challenges across the industry.