Expert Q&A: Trends in Precision Medicine

While still in its early stages, precision medicine has great potential to improve the effectiveness of treatments for patients and reduce overall healthcare costs. However, many challenges remain to be addressed before this potential can be fully realized. Christian Olsen is the associate vice president, industry principal of Biologics at Dotmatics, and a leader in R&D scientific software connecting science, data, and decision-making. In this Q&A with G2 Intelligence managing editor Rachel Muenz, he discusses recent trends in precision medicine, the role of lab tests and diagnostics, key challenges in the field, and the role of artificial intelligence in solving these issues.

Q: What are some of the key trends you’ve seen in precision medicine that involve laboratory testing/diagnostics?

A: We see the increasing use of molecular diagnostics, which identify genetic changes in a patient’s DNA. These tests are becoming increasingly important in precision medicine, as they can be used to identify patients at risk for certain diseases or who could benefit from specific treatments. As molecular diagnostics become more widespread, doctors will be able to develop personalized treatment plans (e.g., treatment expansion) for patients based on their genetic makeup. This should lead to more effective and targeted treatments that improve outcomes for the patient. We are expecting AI to play a role in this space in a couple of ways: 1), where AI is used to develop new laboratory tests and to improve the accuracy and efficiency of existing tests, and 2), where AI is used to analyze large sets of patient data to find patterns and trends that can help doctors make better-informed clinical decisions.

Q: What are some of the main challenges related to these trends, and what are some of the solutions to these challenges?

A: So far, we see challenges in two distinct groups. The first group is cost and access. Precision medicine can be expensive since it requires new and specialized lab tests. Precision medicine is still in its early days, which means everyone won’t have access to it at the same time. The second group of challenges we see AI playing a big role in addressing. Precision medicine relies heavily on the accuracy of lab tests. If the tests aren’t accurate, it will lead to misdiagnosis and wrong treatment recommendations. Interpretation of the data is also in this group. And this is where we see AI being leveraged to sift through the large amounts of data, which would be difficult for doctors who aren’t trained in precision medicine.

Q: What developments in precision medicine do you find most interesting/exciting? Why?

A: There are a few things that excite us about precision medicine, starting with the patient. We’re excited to see personalized treatment plans that lead to improved patient outcomes. These treatments will consider the diversity and nuance that patients have instead of treating them in a “cookie cutter” fashion. We are also excited to see new therapies and drugs get discovered or repurposed into the market more quickly than the typical ~12–15 years it has taken in the past. Most of the larger pharma and biotech companies we work with are genetically validating their targets, which means R&D that used to take years is getting reduced to months. Not only is the return on investment (ROI) better for the company, but patients won’t have to wait as long to get the treatments they need. That is a total win-win.

Q: How do you expect these developments to progress as we move farther into the future?

A: We’re very optimistic here. We expect to see the ROI improve over the next five years. We also expect to see the more difficult chronic diseases (e.g., Alzheimer’s disease) receive better treatments if not outright cures. The key to these developments is data-centric—generating, managing, analyzing, and connecting massive data sets in ways that couldn’t be accomplished before.

Q: What are some ways that lab and other healthcare leaders can stay on top of these changes?

A: To stay on top of precision medicine developments, we recommend first and foremost to be patient. This is a new field that will take time getting worked into our healthcare system. There’s the discovery and proof-of-concept stage for precision medicine and then there’s the implementation stage. Look no further than the adoption of electronic health records as an example of some of the challenges for rolling out what seems like a “no-brainer” improvement for patient care into our healthcare system.

There are quite a few annual conferences and workshops that focus on personalized medicine and it’s really easy to meet up to learn from the presentations and to talk with other experts or researchers at the meetings. We also recommend getting involved in research on precision medicine, which is often shared via poster presentations. Talk with the poster presenters during the conference happy hour and there will be no shortage of ideas and opportunities.

Q: Was there anything you wanted to add that my above questions missed, or that you just wanted to mention?

A: Using AI in drug discovery is a potential game changer for precision medicine, however reaching a point where customers can do so is a multi-step process. First, scientists and researchers need clean data—and lots of it. This sounds simple but the reality for biotech and pharmaceutical companies is that they have mountains of data that are related and connected, but aren’t structured in a way that affords discovery easily—mind-blowing figures, like hundreds of terabytes produced every week. And it’s stored in all types of locations that can’t be shared effectively among each other, including various locally installed software tools, on instruments, in Excel spreadsheets, as images, etc.

Once the data are organized, you’re able to do useful things like run reports and ad hoc queries, which help scientists begin to analyze and connect the dots in ways they would have missed before. Using interactive visualization is the next level of that process because it enables scientists to actually see patterns within their own discovery data.

From there we start getting into the real possibilities of AI with things like predictive modeling and prescriptive analytics. Imagine if AI could discover that patients with a specific genetic mutation have a higher likelihood of responding positively to a certain class of drugs, much like how we know that mutations to BRCA1/BRCA2 cause increased risk of breast and ovarian cancer. That would lead to personalized treatment plans that could significantly improve patient outcomes and reduce adverse reactions. Predictive modeling and prescriptive analytics make this possible.

Ultimately, you’re maximizing the power of AI when you reach the point of automated decision-making, where AI is able to do all kinds of exciting things—to automatically identify patterns, trends, and correlations within the data, generate predictions, provide recommendations for potential drug candidates, optimize experimental designs, and even propose novel hypotheses for further investigation. That’s where we’re headed in precision medicine.