Precision Medicine in Oncology: From Testing to Therapy

In this Breaking Down Health Care conversation, John Hennessy, MBA, and Michael Kolodziej, MD, explore the distinction between precision and personalized medicine, the challenges and nuances of next-generation sequencing (NGS) testing in oncology, and the clinical implications of targeted therapies.

Read the transcript: 

John Hennessy, MBA: Welcome to Breaking Down Health Care, where we'll discuss evolving topics in health care in the United States. I'm John Hennessey, a principal with Valuate Health Consultancy. I'm here again with my colleague, Mike Kolodziej, who has a Substack called Decoding Health Care and is always involved in things going on in oncology.

Mike, I reread your substack before we talked today, and I noticed that you and I both have grad school FOMO. I was certainly convinced that Duke University would never have me again if the opportunity came for me to apply for a degree at that institution.

Michael Kolodziej, MD: Every time I head back to St. Louis for a reunion, I experience imposter syndrome.

Hennessy: Exactly. I try not to go back for reunions and it doesn't bother me too much. We're talking about precision medicine, and a challenge for everyone in oncology—whether you're a patient, an institution, or a physician—is that we talk a lot about next-generation sequencing (NGS) testing. So, when we think about NGS testing being covered, what does that mean? What does it not mean?

Dr Kolodziej: Yes. Let's start at the beginning. I think payers, like physicians, accept the premise that cancers can have a mutation that drives their growth and if you block the mutation, you can either hinder the growth or eliminate the cancer. I think they like that.

There is, in fact, a long-standing FDA paradigm for that called companion diagnostics. In simple terms, you do a test for an EGFR mutation, and if the EGFR mutation is present, there's a companion drug that is indicated. That paradigm has been embraced by payers forever and ever.

NGS got complicated because NGS involved looking for a ton of mutations. This became possible because of technological advances in DNA sequencing. The original leader in the commercial space was Foundation Medicine. There are now many other companies that do these panels of mutations.

I think we can say, without being critical, that the panel generally consists of pretty much every gene ever implicated in cancer. That's how the panels were created. Some of them are important now, some of them are likely to be important at some point, and some of them likely will never be important. The payers have always had a problem doing tests where the result doesn't improve the health outcomes of patients.

We've just conceded that the NGS panels have lots of "tests" that don't impact how that patient is going to be treated. Not only was this an issue for commercial health plans, but it was also an issue for Medicare. Ultimately, Medicare, in collaboration with the FDA, took the bull by the horns, issued a national coverage decision in parallel with FDA approval of the Foundation Medicine Panel, and basically said that if you're a Medicare beneficiary with advanced cancer, you can get the test done.

Now there were stipulations—you can get it done once in your lifetime. But they paved a way for Medicare beneficiaries to get the test done, to get a panel done and paid for. So in America, if you have Medicare or Medicare Advantage (since they're held to the same coverage rules), you can order the test and get the result.

That does not mean you can necessarily act on that result because the drug indicated by the NGS panel results may or may not be covered by Medicare. Medicare coverage of therapies is dictated by the FDA approval of the therapy or by compendia, especially National Comprehensive Cancer Network (NCCN) compendial lists.

It has been several years since that national coverage decision was released, and an increasing number of Medicare beneficiaries have been tested. Medicare will pay for the test as long as the testing entity has an FDA-approved test. The reason they do that is quality control; the FDA looks very carefully at how well the test is being done. They do it very well. Many other commercial entities, such as Guardant, have FDA-approved tests.

So Medicare is easy. Commercial health plans, however, are not easy because they are not bound by CMS coverage decisions, drugs aside. We've talked about drugs and FDA approval—commercial payers can't say no to therapeutics, but they can say no to diagnostics. There has continued to be a big debate about whether NGS panels should be covered for all patients with cancer, basically a parallel to the Medicare decision. As far as I know, not a single commercial health plan in America that does this. There may be 1 or 2 small Blue Cross Blue Shield plans that do, but none of the major national plans.

They do it for certain cancers. For example, a lot of them will pay for an NGS panel in lung cancer. Why do they pay for it in lung cancer? Because there are enough mutations that matter, where there's a companion diagnostic paradigm and a therapy available if you identify the mutation. But if you look, for example, at colon cancer, there is a very limited number of mutations that you can do anything about. Therefore, health plans use the language “investigational” and “experimental” when they talk about the coverage decisions for NGS panels in advanced cancers outside of a select few.

You may say, "Oh, they're not really investigational and experimental," but the health plan will pay for the single mutation tests that matter. Not RAS, for example, in colon cancer. The problem is there may be limited tissue, and it's a real pain in the neck to get it done. So the coverage decision remains unsettled. If you are a physician in America with a Medicare beneficiary, you have a green light to order the test. If it's a commercial plan, you have to see what the commercial plan will let you do.

Let me just add that this has nothing to do with the cost of the test. These tests cost $3000 to $5000, somewhere in that ballpark. Part of their hesitation to open the floodgates, if you will, is because they're afraid doctors are going to get those results, they're going to see a mutation and think, “Oh that means I can use this drug.” Now, with Medicare, I've already told you, sometimes you can, sometimes you can't. But the commercial health plans are afraid that they're going to get a lot of off-label prescribing, which would be expensive. Although the test may only cost $3000 to $5000, the targeted therapy is $15 000 a month for however long, and they don't like that. There's a little bit of data that that actually doesn't really occur, but they're hesitant to move forward.

Hennessy: You bring up the Pandora's box issue which is: Is it good news when I find out with one of these tests, if I'm a patient, that I have one of these mutations? Maybe we can talk about where the mutations make a difference and where they don't. Maybe talk a little bit, too, about how just because these are targeted medications doesn't mean that they are free from adverse events and side effects. I’ve got a target—are they actionable or not? And what about the challenge of when they aren't or when the medicines are targeted but still pretty challenging to administer?

Dr Kolodziej: The easiest case, obviously, is the context-specific mutations, such as EGFR mutations in non-small cell lung cancer, for example. Getting that news is actually good news for a patient, because, usually, you can get a targeted therapy that's going to prolong your survival—although it likely won't cure you—and the toxicities will be manageable.

Then we've got a group of mutations—we’ll call them polytumor mutations (that's a word I made up, polytumor). People often talk about pantumor mutations. Those are mutations where it doesn't matter what the context is. Neurotrophic tyrosine receptor kinase (NTRK) fusions, for example, are pantumor, but there are not that many pantumor ones. Most of them are polytumor. HER2, for example. HER2 mutations appear to be important in lung cancer and colon cancer, and they are obviously important in breast cancer. Interestingly, however, in breast cancer, it's amplification, whereas in colon and lung, it's usually not—it’s usually point mutation. But that doesn't matter; they're activating mutations.

So we're stuck in this quandary where the science is still developing, and it's almost hard to believe we don't know more than we already know. Real-world evidence should give us a ton of information, but we haven't really been able to collect that in a fashion that helps us move things forward.

There have been a couple of attempts, the NCI-MATCH trial and ASCO's TAPUR trial, to determine whether mutations are exclusively context-dependent, polytumor, or pantumor, where a group of patients, irrespective of where the cancer originated, are put into a cohort if they have the same mutation, and a specific targeted therapy is tested.

I will say that my reading of the results of all those trials—and there's been like a zillion abstracts and publications—is that, for the most part, they're context-specific. Sometimes they're polytumor and pantumor, but many of them are not. Everybody always goes back to the example of the BRAF mutation in colorectal cancer. If you have a colon cancer that has BRAF mutated, that is a bad colon cancer. It doesn't mean you use the melanoma drugs for BRAF in colon cancer because they don't actually work very well, they have plenty of side effects, and they're expensive as hell.

If you have a context-specific mutation for which there's a good therapy, such as ALK or EGFR—and the list is growing and growing—then yes, you have another therapeutic armamentarium that is likely to benefit you, but not necessarily likely to cure you.

Hennessy: As we bring this episode to a close, you titled this Substack episode The Precision Medicine. Sometimes that is thrown around in the same context as personalized medicine, but are they different? Is it the same thing, or is one a little bit unique compared with the other?

Dr Kolodziej: That's a wonderful question, John. I heard a wise person once say, "Your prognosis is more dependent on your zip code than your genetic code."

I think precision medicine just means this is what your mutational analysis shows—this is what's driving your cancer or making your cancer escape the therapeutic options that are available. But personalized medicine takes all kinds of other things into account, like you 4000 r social determinants of health, your personal wishes after informed consent and shared decision-making, and a million other factors that determine what the right therapy is for you.

Personalized medicine is very hard to codify in the sense that these are all the things that you need to consider as you're writing a personalized prescription for a patient. However, experienced clinicians do this every day. If it were as simple as looking at the mutation, looking for the drug that matches, and just prescribing that, artificial intelligence could take oncologists' place tomorrow. But it is far more complicated than that. For that, we should all be thankful—since we're in the season of Thanksgiving—and feel good about our job security, and continue to hone our craft in doing the best things for our patients.

Hennessy: Thanks for joining us this afternoon on Breaking Down Health Care. We have more episodes to come, so please join us as we bring new episodes to you. Feel free to drop a line to Mike or me if you've got some ideas of things you'd like us to talk about.