Saving Lives and Costs with Next-Generation Sequencing in Non-Small Cell Lung Cancer

Podcast Transcript

Dale Shepard, MD, PhD: Cancer Advances, a Cleveland Clinic podcast for medical professionals exploring the latest innovative research and clinical advances in the field of oncology. Thank you for joining us for another episode of Cancer Advances. I'm your host, Dr. Dale Shepherd, a medical oncologist here at Cleveland Clinic directing the Taussig Early Cancer Therapeutics Program and co-directing the Cleveland Clinic Sarcoma Program. Today I'm very happy to be joined by Dr. Nathan Pennell, director of the lung Cancer medical oncology program here at Cleveland Clinic. He's been a guest on this podcast in the past to discuss the Crystal One Lung Cancer trial, and that episode is still available for you to listen to. He's here today to talk about modeling costs and life years gained by population-wide, next generation sequencing or single gene testing in non-squamous, non-small cell lung cancer. So welcome back, Nate.

Nathan Pennell, MD, PhD: Thanks, Dale. Always a pleasure to come and chat with you. And I am always excited to talk about increasing testing for lung cancer patients.

Dale Shepard, MD, PhD: Excellent. So, remind us of a little bit about what you do here at the Cleveland Clinic.

Nathan Pennell, MD, PhD: Well, I am a medical oncologist and I see patients with lung cancer and other thoracic malignancies. I run the lung cancer program for the Cancer Institute and I'm also the vice chair of clinical research, so overseeing kind of our clinical trials program here.

Dale Shepard, MD, PhD: Excellent. So, we're going to talk about gene testing and how that's kind of changed really how we think about lung cancer. Lung cancer has been divided now into a whole bunch of little individualized diseases, right? So, to start off, tell us a little bit about just the big picture, how non-small cell lung cancer is treated and how you approach therapy from a medical oncology side.

Nathan Pennell, MD, PhD: Yeah, my career over the last 15 years has really kind of paralleled a vast change in how we both approach and think of lung cancer. And so non-small cell lung cancer 15 years ago was treated like a single disease. So young, old squamous cell carcinoma, adenocarcinoma, large cell carcinoma, smokers, non-smokers, everybody got the same thing, which was chemotherapy. And the median survival was 10 months, two-year survival was 10 percent, so more than half of people died in the first year and the other half died in the second year, and it was pretty dismal.

What changed probably 15 to 20 years ago was the recognition that there are actually important different subtypes of non-small cell lung cancer, and those are defined by different genetic alterations in the DNA of the tumors. The first one was epidermal growth factor receptor mutations identified in 2004. There were oral therapies that worked quite well for those patients that have gotten better and better and help people live longer.

10 years ago, we had two, we had EGFR and then something called ALK gene fusions, and there were a couple oral therapies that were quite effective for that, which have gotten even better. And so that kind of brought us into the realm of needing to do biomarker testing for every patient with non-small cell lung cancer, especially the non-squamous adenocarcinoma subtypes. And we actually did a pretty good job. So, by the middle of the 2010s, we were actually testing probably 80 percent of newly diagnosed people in the US for EGFR and ALK.

The problem with that is in 2023, it's not just two biomarkers that we're looking for, we're now at about 10 and we're getting at least one new one every year. And that's really made it challenge because the same tests that people invested so much time and effort into educating people on and the infrastructure to do these tests for EGFR and ALK is not adequate to test for 10 or more different alterations now. You actually have to do a completely different kind of test if you're going to do that. And not everyone has yet kind of bought ... It's not so much that they haven't bought into the importance of it, it just hasn't happened. Ultimately, we're missing a lot of people who could be benefiting from testing and treatment.

Dale Shepard, MD, PhD: Just to clarify, when we talk about the testing part, traditionally a lot of these tests you had morphology and then you had immunohistochemistry, but these are really more on the genomic level.

Nathan Pennell, MD, PhD: Right. So, we still need the pathologist to do the standard histology to tell us what subtype of lung cancer it is. We still use immunohistochemical testing for things like PD-L1 to help guide immune therapy. We can even use it for some genes like ALK. You can actually do immunohistochemical staining, which is cheap and available everywhere.

The earliest single gene tests for EGFR and ALK and Ross and some of these we're using things called PCR, which can detect amplification of single mutations as long as you know what to look for. Fluorescence in situ hybridization or FISH, which can look for chromosomal translocations for single gene fusions like ALK and Rett and Ross.

And now of course, multiplex tests use something called next-gen sequencing or NGS, which is kind of a catchall term and there's a lot of different technologies that fall under that. But essentially what it does is it allows us to test for potentially hundreds and thousands of different genes at the same time and lots of different types of alterations including mutations, gene amplifications and gene fusions, and translocations as well all at the same time.

Dale Shepard, MD, PhD: And so, you did some modeling work and you looked at comparisons of next generation sequencing, which you just described versus single gene testing. So seems a little bit intuitive what single gene testing would be but give us a quick comparison of those two.

Nathan Pennell, MD, PhD: So, 10 years ago, single gene testing for EGFR and ALK was totally appropriate. It was fast, relatively inexpensive and could be done as I said, and about 80% of people were getting that done. The problem is as we started doing more and more markers, what we found is that once you get beyond about three of these biomarkers with a single gene test, you start running out of tissue because you use up material from the biopsy when you do this. And because you have to pay for each individual test, once you get to more than three, it starts becoming prohibitively expensive to actually do all these tests. And now again, we're testing 10 or more of these. And so, you can use a single multiplex test that can cover hundreds of different genes and certainly all 10 of the ones that we need for FDA approved treatments. It does cost more than a single gene test, but it costs less than say four single gene tests. And so ultimately, we've shown, my group has looked at this and shown that once you get to that crucial four or higher, it's actually more cost-effective to use a single NGS test than it is to use multiple gene tests.

Unfortunately, if you look nationwide, what you find is lots of places are still using panels of single gene tests because that's just what they've set up over the last decade. That's what their infrastructure is designed to do. The other issue that comes into it is the payers. The payers are used to paying for approved single gene tests. And when you start getting to these larger multiplex tests like why am I paying for hundreds of genes when you only need eight or 10 of them? And they don't really understand that it's the same test with the same cost to cover both of those things. And so, it's taken quite a lot of time and work to get people to recognize the importance of NGS in cancer. And it's still debatable for patients with lots of different advanced cancers how important it's to do comprehensive testing outside of a clinical trial or research. But for lung cancer, it's absolutely the standard of care to do it. We're just not doing it nearly often enough.

Dale Shepard, MD, PhD: And I guess the question, change is hard and setting up new tests and things like that. We've seen this before years ago as new mutations became important, that lag for people. So, when people are doing single gene testing, on average how long does it take? I mean, we have drugs for KRAS G12C now. I mean have people appropriately added that to single gene tests or is that still sort of a pain point in terms of getting testing?

Nathan Pennell, MD, PhD: Well, the good news is because KRAS was the first oncogene even before EGFR, most of the multiplex tests already included that. And most of the emerging genomic targets that have drugs that are in development are included in all the kind of commercial panels that people would use. So, we're not really talking about experimental discoveries for different genes here. Most of the things that we need to detect are available on commercially available panels that already exist out there.

Dale Shepard, MD, PhD: But if people are using single gene testing, are they also adding things as needed for the most part?

Nathan Pennell, MD, PhD: They can. Sure. For something like KRAS, that's easy because it's a single point mutation test. And so, it's easy to add extra primers to your multiplex PCR test. You're checking for EGR and say BRAF, GAD, KRAS. That's not a big deal.

Dale Shepard, MD, PhD: Right.

Nathan Pennell, MD, PhD: Now when you start talking about something like MET exon 14 skipping mutations or you want to talk about gene fusions like Rett, Ross, NTRK, you can't do DNA PCR tests for those, you actually need RNA panels, fusion panels to detect that. Although you can do NGS panels and that can cover everything, every type of alteration. And so, you can actually theoretically do a series of single gene tests that covers everything, EGFR, ALK, Ross, MET, BRAF, NTRK, Rett, MET, KRAS, HER2. But as I've said, once you get to that fourth test, you've dropped down to probably 40 percent or fewer of patients have material left to do it. And I don't think anyone short of a massive surgical resection would have enough material to do all 10 that we currently need.

Dale Shepard, MD, PhD: And I guess that becomes a question particularly within lung cancer. How often is that tissue itself, the amount of tissue a problem?

Nathan Pennell, MD, PhD: This is a major barrier to getting testing done. The more complex the test is, the more material that's needed in order to run it. And that becomes even more complicated when you start talking about doing RNA type tests.

And what I can say is there isn't really a one size fit all answer for this. Most patients in the US who get biopsied for advanced lung cancer get relatively small biopsies, often cytology specimens, cell block from a malignant effusion or an FNA. And all of that can be used for testing. The key is you have to get enough material in order to do it, and in order to do that, you have to be communicating with the people getting the biopsy that they know this is probably lung cancer. You can't let the patient up until you're sure you've gotten enough material.

At the Cleveland Clinic, we'll use something called rapid onsite evaluation or ROSE where a cytopathologist is actually there in the room looking at it is saying, "Yes, we've got lots of cancer, just get me some more." And then even with something as simple as a bronchoscopy, we can get more than enough material. We can test more than 90 percent of our patients internally diagnosed here with all of the approved tests that we need. But that type of resource is simply not available everywhere.

Dale Shepard, MD, PhD: And then we're going to talk about, I swear we're going to talk about your modeling thing here in a second. One more thing about the sampling though within lung cancer and these tests within lung cancer. How many blood sample tests, cell-free DNA in order to get samples for analysis, how has that been sort of introduced at this point?

Nathan Pennell, MD, PhD: Yeah, that has actually worked its way into standard practice really fast because people can see quickly an impact on patient care. Basically, every patient has blood, and the test takes about a week-ish to get the results, and it's pretty good. It's very specific, which means if you get a positive result, you can act on that without needing any tissue. It hasn't quite caught up to the same sensitivity as tissue testing, but it's not bad. 70 to 80 percent sensitive to pick up things. And for people who have lots of tumor burden, bone metastases, liver metastases are pretty high.

And so, blood testing has really broadened the availability of genetic testing because of these issues we talked about with tissue availability. And of course, I see people commonly who get their biopsies in another state and another health system, and then they come to see me, and I don't quickly have access to that tissue. I don't know if it's adequate. I don't know if someone's ordered testing. Even if it is, I don't know how to get access to those results. And so, blood testing is invaluable for that. And when we start talking about applying this worldwide where it's almost unimaginable that expensive next-gen sequencing could be available to patients in countries of low and middle income countries where they're really still working on basic population health, blood testing I think is going to be a way to extend these life extending therapies to a much larger number of people.

Dale Shepard, MD, PhD: So, you did some research. You looked at some modeling, looking at costs, life years gained based on results from genomic testing. Tell us a little bit about what you did, what you found.

Nathan Pennell, MD, PhD: Yeah, you read the title of the paper and it sounds a little dry. Oh gosh, Modeling Life Years Gained with Next-Gen Sequencing Versus Single Gene Testing. But basically, what it boils down to is a thought experiment. We know approximately how many people there are with lung cancer every year. We know approximately what the distribution is of how many of them have these driver oncogenes. And we know what the percentages of single gene testing versus NGS. So, we can estimate every year how many people are identified as the people who should be getting identified and treated. And then you can estimate how long they might live if they're effectively treated, and how much that would cost because we know how much the drugs cost, how much the test cost.

And so ultimately, we were able to build a theoretical model where we could just alter different percentages of testing and we could estimate, well, how many people are being missed and how many years of their lives could we save and extend at what cost by increasing the percentage of people that are tested? And ultimately, if 100 percent of people were tested, how many lives would that save?

Dale Shepard, MD, PhD: And so, what were your findings?

Nathan Pennell, MD, PhD: Well, the baseline is, as I'd said, so essentially 80 percent of people getting EGFR and ALK testing. And now of course that has actually changed a bit since we first made the model, and more people are getting NGS. But if we use that as our baseline, what we showed was for every 10 percent shift from using single gene testing to broader NGS, where instead of detecting just the two alterations, we're now detecting all 10, and those patients are going on to treatment. We could actually increase the life years gained by about 2,600. If we replaced all of those people getting single gene testing with NGS so that 80 percent of people with lung cancer were being tested, we could save an additional 21,000 life years gained, and it would actually save money about 600 dollars per life year's gain less than using single gene testing partly because of avoiding use of expensive ineffective therapies for these patients.

And then of course, our best-case scenario, if 100 percent of people, which we'll never achieve, but if 100 percent of people who should be tested actually got tested with NGS, we could save an additional 15,000 life years gained. And ultimately it would only cost a little over 16,000 dollars per life years gained. So, from cost-effectiveness analysis, that's like 10 percent of the limit of what you would consider to be a cost-effective intervention. So, it's clearly doable. It's not out of reach. It's clearly affordable from a public health standpoint, at least in the US, maybe not everywhere. We just need to do it. And I just want to make sure that I credit Dr. Chris Lemmon, who is a thoracic oncologist now at the University of Cincinnati, who did a lot of this work and really was responsible for driving this project during his fellowship. I really, really appreciate the work that he's done on this.

Dale Shepard, MD, PhD: What's the biggest barrier? I mean, if we know that you can capture more things more easily by doing a whole panel, why would people still not be doing it? Why are people hesitant to do testing? Why are testing rates lower than they should be?

Nathan Pennell, MD, PhD: That is such an incredibly important question. If you ask medical oncologists and I give talks about this topic all the time, 100 percent of them say, "Of course I have to order NGS on my patients. I do it on every single patient." And you talk to pathologists and they're like, "Well, of course you have to do it. I mean, it's in the guidelines. You have to test every patient."

And then when you see publications where they look at actual rates of testing and Medicare billing for this and which patients actually end up getting treated, what you find is probably about a third of people who are eligible for testing and treatment actually get all the way to receiving the drugs that would benefit them. And it can go wrong in so many ways. I mean, some patients get sent home from the ED with a lung mass being told you've got lung cancer, put their affairs in order and never get a biopsy. Some people get a biopsy that just creates a smear that says non-small cell lung cancer, and that's it. And that's all they ever get. And you can't do testing on that. Sometimes the pathologists aren't stewarding the tissue in a way that will preserve enough for biomarkers. So, they do lots of different tests even though it's likely lung cancer. And so that's an incredibly important part of this, is making sure that you're aware of what needs to be done with what you've got so that you make your diagnosis, but you also preserve enough for testing.

But then someone has to actually order the test. Here at the Cleveland Clinic, we use what's called reflex testing where our pathologists make the diagnosis of adenocarcinoma and then they just order molecular testing directly from that. But most places it just goes in the computer with a diagnosis, and it waits for someone to come back and ask them to run additional tests, which wastes more tissue, takes extra time, and some people fall through the cracks. There's a high failure rate, as we talked about with small samples. Not every assay can be run on the material that's available for patients. And so, you may wait a few weeks and get a quantity not sufficient result from your vendor that you've sent it to, and now you're stuck with a patient who's sick and needs treatment. It's been weeks. You just move on to treatment rather than go for another biopsy.

And then finally, the reports have to get back to the doc. They have to be able to interpret it. The doc has to know how to prescribe the right medication, and then that has to be affordable and accessible to the patient, which is actually not trivial anymore because of how expensive they are and changes in insurance coverage. So ultimately, there are so many things that can go wrong that it ends up being two thirds of people are missed.

Dale Shepard, MD, PhD: One thing that seems really complex still is one of the later points where things could go wrong. And that's being able to find the data, use the data in a meaningful way, the whole bioinformatics piece. Any thoughts about that? A lot of these reports come in, they're not really searchable forms. They end up in all kinds of random places in a medical record. How do we make that better?

Nathan Pennell, MD, PhD: Oh, every time I have someone who recurs after a surgery years ago, and I'm scrambling through the EMR, hoping someone uploaded a PDF of a test that was done five years ago that I don't even know if it exists or not. It's really remarkable how much information is out there that is hard to access. And I think increasing interoperability of electronic medical records is going to make a big difference. Sharing data across different data sets so that it's searchable. I think the move into AI may make a big difference as we're able to access more data. And if we can get around concerns about data privacy so that you'll be able to access all of the stuff. So, it is getting better, and it is getting easier, but there is still a lot of room for improvement.

Dale Shepard, MD, PhD: What do you think is going to be the next big step? As we look at genomic incorporation of therapies, what's going to be that next plateau that really changes how we treat lung cancer? Different biomarkers? I mean or even just availability of them?

Nathan Pennell, MD, PhD: So, I'm going to be really boring here, and I'm going to say there is an increasing recognition as companies continue to develop these incredibly, incredibly effective treatments for tiny populations of people, that they're losing money because no one is being identified to get their drugs that they're developing. And they're starting to recognize that. And there are partnerships out there, public-private partnerships with the government, with private organizations like the International Association for the Study of Lung Cancer with multiple pharmaceutical companies. And they are going to really start investing a lot of time and money into solving many of these choke points and barriers that we have talked about. And if we can just apply the existing treatments that we already have to the patients that we already are treating, we could, as I've tried to illustrate in this paper, be saving tens of thousands of lives per year just with what we already have.

And if you think about most of the people with lung cancer in the world are not in the United States. I mean, there are almost as many smokers in China as there are people in the United States. That's not going to change for a long time. And most people with lung cancer in the world right now are never getting genomic testing to identify targeted alterations. They're lucky if they're getting chemotherapy. We need cheap, high reliability, widely accessible molecular testing. And so, blood testing I think is a great place for that. I mean, technology just keeps getting cheaper and cheaper and works better and better, and it should be within reach to apply this kind of testing worldwide.

Dale Shepard, MD, PhD: So, are there any other changes that you see as making a big impact?

Nathan Pennell, MD, PhD: Yeah, I think the other big impact, especially in the targeted treatment field, is in applying all of these really exciting, targeted treatments that we spend a little bit of time talking about into the earlier stages of patients. So now that we're getting better early detection, lung cancer screening is hopefully going to continue to take off. We're finding people with earlier stages, and we have started to see trials showing survival benefits by applying testing and targeted treatments to curable early-stage patients. And so, the big one was the recent ADAURA study using Osimertinib for EGFR+ mutant lung cancer, stage 1B through 3A after surgery had a 10 to 12 percent five-year improvement in survival with three years of adjuvant treatment. And that's only, I think, the very first step into a much larger world of personalized tailored treatment for people with lung cancer across all stages.

Dale Shepard, MD, PhD: Well, it looks like lung cancer is an enormous problem. It looks like this is a good way to make some progress toward helping out. Appreciate your insights today.

Nathan Pennell, MD, PhD: Thank you.

Dale Shepard, MD, PhD: To make a direct online referral to our Taussig Cancer Institute, complete our online cancer patient referral form by visiting clevelandclinic.org/cancerpatientreferrals. You'll receive confirmation once the appointment is scheduled.

This concludes this episode of Cancer Advances. For more podcast episodes, visit our website, clevelandclinic.org/canceradvancespodcast. Subscribe on Apple Podcasts, Spotify, or wherever you listen to podcasts. Don't forget, you can access real time updates from Cleveland Clinic's Cancer Center experts on our Consult QD website at consultqd.clevelandclinic.org/cancer.

Thank you for listening. Please join us again soon.