CHICAGO (GenomeWeb) – When Biden Cancer Initiative President Greg Simon speaks in public, his attention-grabbing line is also an eye-opener.
"The cancer is the only part of the system that's completely focused on the patient," Simon said during the opening keynote of the annual American Medical Informatics Association (AMIA) informatics summit in late March.
That quip was meant to illustrate the myriad factors working against each cancer patient receiving the most appropriate care possible.
Healthcare systems and even physicians worry about how they will be reimbursed. Payors fret the cost of providing oncology services. Pharmaceutical companies often are looking to push the most expensive treatments, regardless of whether they are most appropriate. Even patients themselves struggle to prioritize work, life, and treatment, and Simon himself did when fighting chronic lymphocytic leukemia.
Befitting his appearance at AMIA, Simon emphasized the need for better data interoperability in building a more precise and effective cancer treatment apparatus to harness the power of genomics and informatics. That strategy also fits with recommendations he helped draft in 2016 while heading the White House Cancer Moonshot Task Force on behalf of then-Vice President Joe Biden.
That panel, charged with identifying priority research targets for precision oncology, in 2016 issued a list of recommendations including some focused on building and using molecular and genomic data to guide cancer care and develop early detection and intervention tools. Among the key data sharing tools the panel identified was the Genomic Data Commons, a cloud-based platform that the National Cancer Institute funded to make genomic and clinical data from large-scale projects available to researchers.
After the 2017 change in administration, Biden and his wife, Jill, created the private-sector Biden Cancer Initiative to further the work of the Cancer Moonshot. Simon, who has headed that organization since its founding, spoke to GenomeWeb this month about the challenges the oncology world faces and how informatics can address some of those issues.
Below is an edited transcript of the interview.
Your line about the cancer being the only part of the system that's completely focused on the patient stood out. I know it was tongue in cheek, but it's really true.
It is true. It's very sad and very true. That's why cancer is so hard. It's a disease that starts as "you" in most cases, though not every case. It starts as you, so it knows the neighborhood. That's why cancer's escape pathways are so difficult to block because they know all the trap doors and they live there. We don't.
Our minds really don't comprehend a cellular environment. We don't have the ability to have that point of view, so we're dealing with an enemy that knows the landscape better than we do, knows the escape routes better than we do, knows the biology better than we do.
I used to give a speech from the standpoint of a cancer cell. The worst thing that can happen to cancer cell is to find out that it's in a mouse because we kill cancer in mice all the time because it's a lot simpler.
Cancer as a biological phenomenon is a pure-form event. Cancer science is a very complex, complicated, distracted series of events based on the fact that it's part of society, and society cares about status and money and competition and collaboration. Even people who are highly motivated to fight cancer still have a hard time being collaborative, being divorced from the rewards system. It's a difficult thing.
Because everyone has their own agendas?
Yes, and they're trained to have their own agendas. Once I say that the only thing that's focused on the patient is the cancer cell, people get it. I can talk all day about [research] in publication and nobody gets it. It's a sad fact.
How do you overcome that?
We're trying to create new incentives and new patterns, so getting people to share data should be the default, not the exception, and developing standards for diagnosis and pathology. There's not a command-and-control button where you can realign all the magnetic fields to do it differently. It's often hard to change as compared to start something from scratch.
What was the true purpose of the Cancer Moonshot and how has that carried over to the Biden Cancer Initiative?
[The Cancer Moonshot] was an attempt to motivate and mobilize people and organizations of all kinds to take on the challenge in their own way. It wasn't so much to come up with grand ideas from the White House. In fact, we actively avoided trying to do white papers and recommendations and sort of thing. Instead, we focused on what can a [cancer] foundation started by children do? What can Deloitte do? What can a pharma company do? What can a university do? What can Airbnb do? What can Uber do?
We were putting out a call to action for people to figure out where they touch patients in the cancer journey, prevention through survivorship. Where do you touch the patient and how do you improve the impact of that touch with more money, more productivity, more people, more activity, or more collaborations? If you're doing X now how, do you do 2X? That was our challenge to people, and that's where we got 80 commitments in the Moonshot and we got another 57 at our [Biden Cancer Initiative] summit in September last year.
Is that going to be an annual summit?
No, we are avoiding annual events. We are doing a different kind of meeting this year, which will be focused more on our actual working groups. This will be much more of a working summit meeting and not the bells and whistles that we had last time.
On the informatics side, have you seen much progress since the Moonshot, since Vice President Biden left office?
Yeah. The good news is that [the Centers for Medicare and Medicaid Services and the Office of the National Coordinator for Health Information Technology] have been proposing the right things from our perspective. It's just taking a long time.
We have supported their proposal. We hope that it gets implemented relatively quickly. This is about interoperability and about getting your records. But more importantly, we don't think that the medical community needs the government to tell them what the right thing to do is when it comes to distributing medical records.
From our perspective, providers ought to be providing people their medical records as the default, the way your bank does. When you take out money or you write a check, you get a statement that says you took out money or you wrote a check. We need that in healthcare, and obviously all the providers know where we are because they send us a bill every month.
Do you think providers have gotten the message? Do you think vendors have gotten the message?
I certainly don't know if vendors have gotten the message. I think providers are getting the message. I think we're definitely getting more attention and we are actively working with cancer centers and others to help us figure out the best way to get [organizations] to sign up for providing people their records as a matter of default. There's actually been a lot of support from some key cancer centers, so I think that's going to happen. It will take a while.
As the issue becomes more visible and the patients become more vocal and groups like us help magnify that, I think that we can get to the point before the [federal] rules come out where there's a core of hospitals and cancer centers that have committed to providing people their records without being asked. Being asked is complicated. Where do I send a letter? I wouldn't even know. Where do I send a letter to Sloan Kettering to get my medical records? I have no idea.
Would data sharing include molecular tests?
It should include anything that you have done at a doctor. It's up to you. If there are certain things you don't want to know that they tested you for. Let's say that somebody sequences your genome as part of dealing with your cancer and there are some things you don't want to know, like [whether you have the] Huntington's gene. You're perfectly entitled to say, "I never want that." Some people don't want to know the sex of their baby, but they do want to know all the other statistics about the visit. The people making the decision should be the patients, not the hospital administrators.
What is your organization doing in the area of data integration? There's a reason why AMIA invited you to speak, and it wasn't because people were wondering if Joe Biden was going to run for president.
I am not a data scientist, and I don't pretend to be a data scientist. We have a whole working group of data scientists on data sharing that also involves how to use data repositories, how to create functional databases. We're not interested in just throwing data on the internet and saying that that's data sharing. It needs to be organized in a way that's useful to data scientists.
That's where interoperability standards come in?
Yeah. Also, no institution has "big data." They're not big enough. In order to have truly big data, you have to have systemwide-compatible, interoperable databases that can be call-based. We don't have to be moving things around, but you need to be able to do your studies and your algorithms on the data in the cloud aggregated from enough places that you truly have big data.
One of the places that does that is the Genomic Data Commons at the University of Chicago. They've aggregated the raw genomic data of about 35,000 people. They started with 14,000 people from the Cancer Genome Atlas. In that scenario, you can go to that site and manipulate the data and play with the data where it is, using tools that you bring or tools that they provide you. That's the kind of thing we're talking about when we say we want to let people do in medical data what they'd been able to do with financial data for decades.
Again, that's probably a cultural thing, and culture takes a long time to change. How do you change it?
It's not easy. You have to show, not tell. You're not going to change it by haranguing people doing it wrong. You change it by showing people how to do it right, and that's what we do with our work. We're trying to find models that are already working and scale them so that people know what they're shooting for, whether it's sharing their work or [following] standards for assays for immunotherapy drugs. You have to show people, "This is what I'm talking about." Then they do what they have to do to make it look like that. I wouldn't know enough to tell people how to change their internal systems because I'm not an expert in it. They are.
But people on your staff and some of your partners are experts.
Yes. Those are the people in our working groups that help us figure out what is the best option to get this at a scalable level. What do we have to do? It takes months, but it shouldn't take years. I'll put it that way.
You'd think it shouldn't take years but here we are 28 years after the first Institute of Medicine report calling for what were then known as "computer-based patient records" and we still don't have the system that people envisioned.
No, that's really frustrating, and this is another example of not concentrating on the patient. There's nothing patient-friendly about the development of electronic medical records [although] the concept was supposed to be patient-friendly.
Moving over to the idea of precision oncology, do you believe that all oncology should be precision oncology?
From a semantic point of view, yes. All oncology does tend to be precision because it gets very personal very quickly. The question is, what are we learning through precision medicine that is of general applicability? Every cancer is different. Every patient is a little different. But there are trends and there are ways to measure and predict who's going to respond to what.
Where precision oncology can really make a difference is if we can learn from everything going on in cancer clinics with every cancer patient every day, then we will be able to predict not to do some things for some people because it won't work and it will just cost a lot of money. That's why we're developing one of these assays to measure the effectiveness of checkpoint inhibitors because there is not a standard to measure whether people are responding or not.
Which study is this?
We have a working group focused on data standards that's working on an assay for PD-1 that would measure the responsiveness to the checkpoint inhibitor both for regulatory purposes but also for clinical purposes.
Beyond that, we also are creating a system for how to develop standard assays in the future for future checkpoint inhibitors that we don't know about now. That's what I get excited about.
When can we expect to see some results on that?
In the fall, we will have a pretty good idea where all of that is.
Data seems to be underpinning all of this, both on the research and the clinical sides.
That's right. Obviously, we're missing a lot of data on every patient because few patients have their family history. Few patients know their complete medical history. You're often shooting in the dark. The promise of precision medicine, electronic medical records, and data repositories is that we can start putting the information in context so that we can be predictive.
I'll give you an example. Randy Scott and Genomic Health came up with a test that was controversial at first — now it's not controversial — that says that women who have already been treated for breast cancer will not benefit from more chemotherapy. That's huge because chemo is damaging to your DNA, to your telomeres, to your nervous system. You can lose hearing. You can get neuropathy. The fact that we were able to now separate women who need additional chemo from women who don't is not just an economic savings, but it's an emotional and physical savings. That is the point of some of what we're doing in precision oncology, which is to know whom to treat and to leave alone and improve their quality of life.