OR WAIT 15 SECS
Tzvia Bader, CEO of TrialJectory, discusses how the company’s AI-powered technology platform is democratizing cancer care and expanding access to advanced new treatments.
Pharmaceutical companies are investing billions of dollars in clinical development of new drugs, yet 90% of oncology trials are behind on their timelines because they have a difficult time finding eligible patients. Some companies are bridging the gap between actively enrolling oncology trials and patients by acting as a trusted advisor to patients, going through the entire journey with them every step of the way. In this interview, Tzvia Bader, CEO of TrialJectory, will discuss how the company’s AI-powered technology platform is democratizing cancer care and expanding access to advanced new treatments.
Moe Alsumidaie: Can you explain the current issues with clinical trial recruitment and the patient-matching process? What’s not working for patients?
Tzvia Bader: There are several problems with the current recruitment process. Many patients in the community are not offered trials by their oncologists due to the accessibility challenges oncologists face. Patients who are looking for clinical trials independently, either after their initial diagnosis or after failing other lines of treatment, have very few outlets to help them identify relevant options. Existing resources, such as clinicaltrials.gov, are difficult to navigate due to the volume and complexity of the trials. Eligibility criteria such as genetic mutations, biomarkers, treatment history, and overall health, as well as the medical jargon used in trial descriptions, all make it hard for patients to identify which trials are actually right for them. Therefore, the recruitment ends up being done only by the big research centers, and the majority of the patients in community oncology centers have limited access to those lifesaving trials.
MA: How can technology make this process more efficient and effective?
TB: TrialJectory’s technology mimics the mind of an oncologist. First off, it sifts carefully through the volume of trials. There are over 100,000 trials in oncology, with 17,000 that are recruiting-and this number is only increasing. Another issue is the complexity resulting from the number of ‘attributes’ or data points that come into play. We look at all the attributes in every clinical trial and uses those data points to match a patient to a trial.
The power of AI takes this incredible volume of data and reads through it in a way that a single oncologist does not have the time to do on his or her own. AI also translates the very complex descriptions of the eligibility criteria into patient-friendly language.
Best of all, AI finds clinical trial options for patients in a matter of minutes. As a result, patients can partner with their doctors to access all relevant information at the right time, with the process beginning with the patient, as opposed to the oncologist dictating the treatment plan.
MA: How is AI improving efficiency? How is machine learning being used?
TB: We don’t just need a strong machine that can process a lot of data points; we need a smart machine. If you think about the innovation coming into play in the area of oncology-such as genetic mutations, biomarker mechanisms, new drugs, new treatment protocols, and more-we need an AI solution that is able to identify all the available clinical trials that match the patient’s profile.
We want a machine that can read unstructured text, like a treatment protocol-something that only machine learning can do. The machine needs to understand the context of the wording and the different terms, whether it’s a biomarker, drug, or side effect, for example. We need a machine that is constantly learning via the text to identify, extract, and standardize relevant information.
We want the machine to have the superpower of an oncologist who can read through all those documents and understand them for what they are, while also making them applicable to each specific patient. With AI, the platform can discover all available information in a way that a single doctor does not have the time to do on his or her own.
Aside from the patient-matching side of the platform, we also provide decision support for the patient. We share patient-friendly descriptions and data regarding the trials that empower the patient. We can track multiple trials and see, over time, the results that are coming from a specific type of trial. We can take different data points and rate them to make relevant information accessible to the patient and oncologists to support them in the decision making.
AI platforms will not replace doctors. The technology is not making any actual decisions. They are simply filtering, analyzing, and processing the information so that it can be made available for patients to make the right decisions for themselves.
MA: Now that we’ve talked about machine learning and what goes on behind the mechanics, let’s talk about the interface. How can patients take their health into their own hands?
TB: Patients need to take an active role. Studies have shown that the more active the patient, the better the outcome is. To achieve this, we must first give them smart tools to make relevant information accessible. We cannot expect them to go to clinicaltrials.gov, review all the trials, automatically know what’s relevant for them.
We must use smart tools to filter through all the data by analyzing information and only giving patients trials that are relevant for them-and we must facilitate this process. We must mitigate the barriers, including accessibility and understanding. We can empower patients to have an intelligent discussion with their physicians and make an intelligent decision from a place of control.
As one oncologist once described to me, “I spend 10 minutes with the patient saying no about a lot of things that he dug out of Google, rather than talking about what is relevant.” We can overcome this by making the relevant information accessible and easy to understand so that patients can sit as an equal in front of their oncologist. This is key to having a productive and hopeful patient-doctor discussion that leads to making critical decisions.
MA: Lastly, you mentioned that this would work in oncology. What other disease indications do you think this would work for?
TB: You can teach a machine anything, but it takes time. You must train it with text, correct errors, and let it mature. This can be applied to many chronic diseases, especially where there is a challenge-for example if there are more variables in the eligibility criteria and, therefore, more attributes that come into play. When we’re looking for a more specific medical profile of a patient, it makes sense. There’s a lot of data saved.
When I look at this industry and why it needs to be changed, I see the role of the patient becoming significantly more important. The patient is not eliminating the role of an oncologist; instead, he or she is taking a lead role in terms of the search and understanding of his or her options. If we continue to give tools only to the research institutions, we will stay at a standstill right where we are. If we provide tools and access to patients and community oncologists in a smart way, we can overcome those barriers.
By doing that, we can push the industry forward. Patients can now ask and find out the answer to, “What’s available out there now, and what’s right for me?” and get access to the most advanced life-saving treatments no matter where they live or where they are being treated. That’s what excites me the most.
Moe Alsumidaie, MBA, MSF, is a thought leader and expert in the application of business analytics toward clinical trials, and Editorial Advisory Board member for and regular contributor to Applied Clinical Trials.