Dr Parikh on the Future of Oncology: AI, Data, and Patients at the Center

Delivering better cancer care requires more than advances in treatment options. Improving outcomes also depends on how effectively clinicians identify gaps in care, incorporate patients’ experiences, and use evidence from routine clinical practice to inform treatment decisions. As the oncology landscape continues to evolve, strengthening collaboration among clinicians, researchers, health systems, and patients will be increasingly important.

These challenges and opportunities were explored at the 44th Annual Chemotherapy Foundation Symposium (CFS) in New York City, where a panel brought together perspectives from clinicians, researchers, and a patient representative to discuss the future of oncology practice.

In an interview with Targeted Onc^TM prior to the panel, Ravi B. Parikh, MD, MPP, panelist, medical oncologist, and associate professor at Winship Cancer Institute of Emory University, shared his vision for the future of oncology practice, including the evolving role of technology, opportunities to better leverage real-world data, and policy priorities that could shape the delivery of cancer care.

Targeted Onc: From your perspective as an oncologist and researcher, what does the future of oncology practice look like?

Ravi B. Parikh, MD, MPP: I think the future of oncology practice is going to be characterized by a lot of integration of oncology clinicians with emerging technology. I focus a lot in my academic work on uses of technology like artificial intelligence [AI] and how they can apply in clinical practice. I think the AI wave is sort of inevitable, and it’s a little [unclear] where it’s going to end up, but I think the answer is going to be somewhere in between what we have now and full autonomous practice of AI; I don’t think we’re going to get there.

What it means practically is that AI-based tools that are [being] fine-tuned for oncology practice are going to be serving as clinical decision support, not only for pulling up literature like I use it now, but increasingly for making individualized patient decisions and potentially even simulating different courses of therapy for a patient, and what might happen if we chose treatment A vs treatment B.

When it comes to research, I think that there are probably 2 [approaches to] using AI in the future. One is happening right now, which is using AI to help speed [up] the existing clinical trial infrastructure, whether that be enrolling patients in trials faster or getting through trials faster. But I think increasingly, we’re going to be using AI to not only design trials that are more targeted towards particular patients, but also to design therapeutics and drugs and radiation regimens. That’s where I see the future of oncology lying, at least from an AI lens.

What do you think needs to change in the way researchers, clinicians, health systems, and patients collaborate to improve cancer care?

I think that there is going to be an increasing need for us to collect patient voices, patient symptoms, and patient-reported data from where they’re sitting, as opposed to in the healthcare clinic. When you think about the current state of where we listen to patients or where we collect patient data, it’s largely in the 4 walls of the hospital; even for a patient who’s interacting with the health system pretty frequently, that’s only a small minority of where patients are spending their time.

There’s this huge black box in what happens when patients are at home, and most of the calls that I get have to do with patient uncertainty about their management plan when they’re at home. I think increasingly, we’re going to be relying on technology-driven methods to allow patients to ask questions whenever and to whatever extent they want to, and to get a pretty rapid response, whether that be from a human or from an automated tool that’s just reinforcing what their plan has to say. I think that’s going to require a lot of digital literacy among patients, physicians, and other oncology clinicians, and it’s also going to require enough trust in those automated systems such that we can be comfortable deferring that type of triage to AI or other smartphone-based technologies, as opposed to what happens now, which is patients having to call in the clinic and waiting for a [physician’s] response.

How can real-world data be better leveraged to identify gaps in care, improve outcomes, and inform treatment decisions?

When we talk about real-world data, we’re largely talking about data that are collected through routine practice, whether that be from an electronic health record [EHR] in terms of what’s coded, or as it pertains to emerging claims data that might be coming about from billing data. I’m always going to trust less [in] real-world data that comes from a relatively clean, well-done clinical trial to answer which treatment is better, treatment A or treatment B. But where I think real-world data can serve a big purpose is helping to uncover areas where there are certain gaps in guideline-based practice, whether that be [determining] what percentage of clinicians are ordering next-generation sequencing, or what percentage are using a biomarker-based therapy for a particular mutation. As we use more genetic and molecular data to guide our decisions, more of that is being collected in the real world and can be used to help influence those quality metrics. I think we’re going to be able to see a lot better refinement of those measures of quality for oncologists.

What I hope is that real-world data are going to include more of the patient voice as well. Right now, we don’t capture much of the patient voice because we don’t have robust ways of integrating patient-reported outcomes into our [EHR]. But that’s happening more [frequently]; more text-based messaging and wearable data can feed in directly to the EHR, and so perhaps in the future, some of [these] data will not only be used to help collect patient data, but also to inform how we ought to be practicing better and tailoring chemotherapy and systemic therapy regimens better.

What are some of the most consequential or current topics in oncology policy that you think oncologists should be aware of?

One is around removing insurance barriers to guideline-based practices in oncology, particularly things like genetic testing, next-generation sequencing, and comprehensive biomarker screening. There has been so much…where these types of sequencing technologies have been integrated into our routine practice that we need to be removing payer-based barriers. That fits into the…prior authorization reform movement as a whole, but I see that particular angle as an area where there’s a lot of consensus and where we need to essentially be accepting that comprehensive genomic sequencing is a part of good cancer care and can eventually lead to downstream cost-effectiveness.

I think the second policy area that I would be attuned to is legislation that [is] allowing patients and other providers of healthcare data to be in charge of what happens to their data. The whole AI revolution is predicated on having access to large amounts of data. Whether it be data that are obtained from voice recordings of their interactions with their physician during the clinic, their billing data, or EHR data, more and more you keep on hearing from patients, “I don’t know where my where my data [are] going.” As a clinician, I’m very concerned about that. I’m also concerned about where my own practice data [are] going, and I generally advocate for policies that ensure better safeguards for how companies are using our data and whether that’s disclosed to their eventual end users.

What perspectives are you most interested in hearing from your fellow panelists?

Not only is this a clinical oncology conference, but so much of the future of oncology practice is essentially how care is going to be shaped around the patient, whether we’re going to be empowering patients more, or whether a lot of these AI-based technologies are going to be removing the patient-doctor relationship. I personally am very interested to be hearing from both of my co-panelists, particularly our patient representative on the panel, about the mood in the air about this emerging technology.

How have patients seen technology positively influence vs negatively influence their life, and what do they feel like is going to be the role of a doctor and an oncologist in 5 to 10 years if more of our clinical intuition is being possessed by these AI tools? Who would patients trust more? And I think those are the types of meaty questions that are going to be interesting to hear more about during the panel.