Overview
We hosted a live discussion and Q&A with Mr. Mark Capone, former CEO and President of Myriad Genetics, on Tuesday, November 3, 2020. We covered the end markets (the TAMs - total addressable markets), the acquisitions of Grail (ILMN) and Thrive Early Detection (EXAS), and the overall momentum for liquid biopsy in the quest to improve outcomes for cancer patients. Additionally, we touched on key Position Monitor stocks, including Guardant (GH), Natera (NTRA), and Invitae (NVTA), among others, and provided a sample of our latest claims data analysis in the accompanying slides.
|
Health Care Call Replay & Notes - Health Care Subscribers: CLICK HERE for event details (includes video and materials link). |
CALL NOTES
Brief background/overview of the “major buckets” around liquid biopsy.
The topic of liquid biopsy is a large area. People often view these as monolithic, but each application is different and has unique challenges.
- Asymptomatic screening - multi-cancer test, whereby looking at a large number of potential cancers (some up to 50 different cancers).
- Asymptomatic screening focused on a specific cancer - a narrow panel for colon cancer, for example, which has a lot of interest. Smaller market, but easier process than #1.
- Asymptomatic high-risk screening - e.g., BRCA+, looking for specific types of cancer in patients that have been preselected or identified as at higher risk than the average population.
- Therapeutic selection - genetic mutations in circulating DNA that would guide therapeutic selection.
- Recurrence monitoring and MRD, which are kind of the same.
- Companion diagnostics for pharma/biotech that have various products under development.
From the opportunity set, do you have a favored player? Exact/Thrive, Illumina/Grail, and Guardant seem to be out front attacking the biggest TAM, but we see at least 40 private companies in the market now pursuing various end markets…
- As we look at the biggest TAM, the asymptomatic screening area, the spending is huge and competitive set small. Grail and Thrive have had the resources to look at multi-cancer screening - they are the two with [the most] data.
- Exact had its own data (w/ Mayo, that we just recently heard about), and with Thrive they are one of the two "best-positioned."
- When I look at specific screening for individual cancers, Freenome and Guardant have made nice progress on the colon cancer side - it’s an excellent market for blood-based testing if they can make it work. The progress has been impressive.
- Therapeutic selection and recurrence is a wider field. Guardant was one of the first movers, and Inivata/NeoGenomics has an opportunity to play in that space... there's a broader set of candidates in the therapeutic selection/recurrence segment because it's relatively easier to do. Guardant is a leader here, and Foundation is making a play - tumor and now liquid biopsy. Those are the companies to watch.
When we talk to colorectal surgeons or other specialists in the GI space that screen patients, they don’t like the blood-based tests. They don’t even like the fecal tests because colonoscopy is curative (clip polyps). Generally, it's harder to detect cancer early...
- The early work at Myriad shifted due precisely to those concerns - i.e., the ability to get sufficient sensitivity and PPV - positive predictive value.
- It's important to look downstream in the continuum of care for recurrence monitoring or therapeutic selection to demonstrate clinical utility. All of these areas have significant hurdles. With a multi-cancer screen vs. screening for specific cancer, the likelihood of being successful is higher if focused on one (can optimize the assay). There’s no question in the earlier stages that the amount of material circulating in the blood is lower, therefore, the only way to get an acceptable sensitivity = multi-omics. We see that with testing combining epigenetics, protein, and DNA (that combination gives a fighting chance to reach an acceptable sensitivity).
- What is that level? 74% for stage 1 or 2 colon cancer. Demonstrating anything in stages 3 and 4 is pointless because the patient has symptoms.
- Data generated by Freenome and Guardant suggest that the tests are now capable of delivering on that - meeting FDA guidance on #s - but those are highly select sample sets w/ a lot of "mathematical tuning." I think the jury is still out. And, colonoscopy remains the gold standard for a GI - anything else is a step back. But, from a pragmatic standpoint, maybe we can do good for patients with screening if they are reluctant to do a colonoscopy or try stool-based testing. Having an effective screening tech that leads to colonoscopy could be very helpful to have.
- We’ve seen some very nice progress with technology (e.g., the use of AI/machine learning and epigenetics, which is very important w/ early cancer detection). Combining all the technologies puts the industry in a different - perhaps better - position to succeed compared to earlier tries. That said, caution is warranted due to the early data being derived from studies on patients that were known to have cancer (even early stage). This is quite different than a truly asymptomatic screening setting, and we could see significant decreases in sensitivity – Thrive’s data is a good example of this.
- With Freenome's, Guardant's, and Exact's data, the algorithms are often highly tuned to the sample set collected - Guardant's initial data showed a low sensitivity, but a second "bite of the apple" included a "thermodynamic model" (it’s not clear what that means). So, there’s a lot of math behind arriving at a higher sensitivity.
- Nothing inappropriate has been done to date, but we should be careful about extrapolating results to a broader population because there's a big gap.
- When faced w/ a true asymptomatic screening study/population, stage 1 and 2 sensitivities after - numbers will be lower.
Is the regulator prepared to deal with this? Specific IP, know-how, trade secrets, methodology… will we need to see a bigger outcomes study?
- The regulators will be prepared to handle this. You just lock the signature prior to doing a large validation study - and it does need to be a large study. Everything must be laid out in advance (protocols). If done appropriately, the regulators can deal with it (they have in the past w/ BRCA testing - e.g., a 4k page submission to the agency was fine).
On Illumina's earnings call, management seemed pretty confident that they aren't expecting to see a deterioration in sensitivity (for Grail/Galleri) over time. Could they be immune?
- Maybe they are confident because they see data that we don’t? They are running some big studies and maybe they have it. As we look at the published data, collected on patients known to have cancer - stage 1 only 39% sensitivity in the 12 cancers prespecified. 18% for all cancers. Those aren’t big numbers. There was a significant increase going to stage 2, but it's hard to expect that in a true asymptomatic population.
- Thrive, 2 sets of data, first in that same case-control method, people known to have cancer - looked like 47% stage 1, 73% in stage 2 (sensitivity). But, with true asymptomatic -> 27% sensitivity (caveat, those numbers may not all be perfectly accurate, but the takeaway is the same). It's a good example of what happens in the "wild." All that said, they could have data we haven't seen too!
There’s still utility, right? With specificity 99.9%, if you find "it" - it's real, so just test frequently enough (even with the lower bound of sensitivity - similar to a FIT test) to overcome the weakness of the signal by doing it more frequently?
- That’s a good question - the FDA opined for colon on FIT - the bar isn’t insurmountable, but we aren’t anywhere near 74% with the data seen to date.
- Moving away from colon cancer, the FDA picked that 74% level because other technology can deliver on that, take a look at pancreatic. In cases where there are not good screens at all, the threshold could be lower and the test would have value. That level is TBD, but we can also look at prostate cancer and sensitivity w/ PSA testing, which data suggest = ~30%. That's not a high number, but it's a $50 test. The U.S. Preventive Services Task Force (USPSTF) removed the recommendation for routine PSA screening because of the lower performance.
- Part of the reason USPSTF came to that decision is the downstream action is pretty invasive (a biopsy). The analysis based on fasl positives and false negatives yielded a pretty negative result.
- Let’s think about where the test is ordered - primary care practices. A test with 40% sensitivity could miss ovarian cancer easily, and that leads to “extraordinarily difficult” conversations. So, there’s a regulatory threshold AND a practical adoption threshold which could be much different (a lack of doctor confidence in the test).
- While specificity may be high, the positive predictive values (PPV) could be low. In the Grail paper, the PPV was calculated based on 40% sensitivities & 99% specificities = 10-30% PPV. So, even when you do deliver a positive result, downstream testing finds nothing 70-90% of the time.
How would you position a highly specific pancreatic cancer test to a payer (with the downstream test being innocuous)?
- Pancreatic cancer is a good example... say you get 50% sensitivity, 99% specificity. Because pancreatic cancer is so deadly and not found until stage 3 or 4, you’d have an appropriate argument for the test. It’s a cancer-specific test, so the first thing to do is seek FDA approval. In the case of pancreatic, Medicare does not have the authorization to screen for that. So, you’d have to get that. Go with USPSTF, which gets you two things: the impetus for Medicare to cover it and, if the law is upheld, commercial payers are required to cover screening with no out-of-pocket to patients. The key to success, getting it through USPSTF.
- That’s why Exact was successful. IT was fortunate to get FDA approval two weeks prior to the ACA and was included, as a consequence, and rapidly able to get reimbursement across the board. Had that not happened, if approval was delayed, it could have taken years.
- An issue w/ USPSTF is they meet every five years - for a new innovation in pancreatic cancer, there may be enough reason to speed up the process.
- If that doesn’t work, must appeal to payers and make the case that there’s economic utility. Get it included in medical guidelines such that the test should be routine. The success rate in getting payers to cover is high b/c it's hard not to cover the standard of care.
- The third path, if 1 and 2 don't work, is hand-to-hand combat with each payer making an economic utility case. With an MRI or PET scan of the pancreas as the downstream cost, there's a question as to whether enough money can be saved via avoidance of pancreatic cancer - it costs $200k+ to treat it, so there may be an economic utility argument to make.
Doesn't FDA approval get you CMS coverage/payment?
- That’s true in the category of therapeutic selection or recurrence monitoring. They’ll give deference to the FDA. In this screening area, it’s problematic. By law, Medicare cannot pay for screening. There are five cancers that congress passed legislation to allow Medicare to cover - colon cancer is one of them. That’s why the USPSTF route can be successful. Pancreatic is not on that list - it would take an act of congress.
As an aside the risk factors for pancreatic = a large population you would want to screen. Could there be a hook - if you’re good on one, get a “ride-along”?
- No, I don’t think regulators would view it that way. Pancreatic is ripe for innovation and there are clear options/pathways to get a test approved.
- With a pan-cancer test - if the pancreatic cancer portion met the criteria for sensitivity and specificity and you're making a claim about 11 other cancers - each one individually would have to meet whatever sensitivity and specificity claims are required by regulators.
Galleri is expected to launch next year. What could the launch trajectory look like?
- First, who is the test being marketed to? It’s got to be primary care, right? Those PCPs are very busy and it’ll take time for the trajectory to develop. As of now, the tests aren’t endorsed by any professional societies, the studies haven’t been run with true asymptomatic populations, and broad payer coverage isn't likely.
- Some tests might go the cash-pay route, and that’ll play into the adoption/uptake.
- The whole process will only improve with better data. Liquid biopsy – especially for screening – is in its infancy and data are not consistent with the claims you’d want to make. That said, [Illumina] may have the sensitivity and specificity data and we don’t know it.
Ok, so how long could it take to reach a “good” level of adoption for a pan-cancer test? Is it 3-5 years or 5-10 years?
- It’s going to take multiple years. Exact Sciences and Cologuard in the early days are a good proxy. It was a very shallow trajectory. Using the most recent experience doesn’t make sense. It’s going to be a long runway. The real question is: “When is the next round of data coming for true asymptomatic screening?” Everything hinges on data.
- The real near-term opportunity is for Natera’s Signatera, Guardant, and others focused where Medicare has a good track record of reimbursement.
- One thing watch - how quickly commercial payers come on w/ coverage decisions - how many commercial payers paying for a full panel test for lung cancer - recurrence or treatment selection - small % of commercial payers.
- Paying for lung cancer, but EGFR - or a glorified EGFR test with other stuff in it - clinical utility data generally is NOT available. A commercial payer will look at that skeptically, and if we are able to detect residual disease or recurrence, does that patient live longer? Showing higher progression-free survival by using test relative to the current standard of care will be needed. That's a significant hurdle - drugs have a hard time showing that.
- Medicare has relaxed the need to see clinical utility - covering them w/ out. Significant change vs. 5 years ago… Medicare seems interested in getting the tests out there, really being pushed hard, trying to professionalize, fraud misuse, LDTs vs. approval process, etc.
ClinVar
- The FDA has made it clear that they’d love to have a central repository for variant classification that facilitates review. ClinVar is not new, and the issue has always been the quality of clinical value and differences. This is why Myriad owns a database and set of classifications. There are material differences between ClinVar and Myriad’s databases.
- The concept is good, but what’s required to ensure variant classification is underappreciated – a ton of technical work is needed, and there’s a large number of variants w/ no accompanying evidence in ClinVar. To classify one variant, you might need to find 30 others that carry the variant, map it over w/ cancer in those families to see if that variant tracks, etc. It can take decades to classify variants – none of that is in ClinVar. Who is going to do the work to ensure variants in ClinVar are accurate?
What’s the near-term impact of COVID-19 on the space?
- It’s just getting patients getting back into the system, for Guardant or Foundation – or blood vs. tissue. That’s the biggest issue. Patients going in for initial appointments or follow-up… the front-end of the funnel for oncology care. The less invasive interaction, the better. If patients are concerned, they stay out of hospitals, so that could benefit blood-based tests.
Other Topics Discussed
- With billions-of-dollars spent developing liquid biopsies, we want companies to have IP that’s protected at the end of the day. If Grail comes up with an algo and anyone can copy it, what’s the incentive for anyone to invest in the kind of discovery work that’s needed. Big ideas will not be funded unless there’s IP. Also, Mr. Capone is bullish on the oncology space and the [large] bets that have been/are being placed there.
- There’s a lot of waste in the treatment of cancer(s) globally. Preventing cancer, to begin with, or focusing treatment based on higher probability outcomes could prevent a lot of waste.
Links to relevant articles/data referenced herein
1. Original article: Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA
a. Follow-ups (in the last one, it is acknowledged that the claims in this data set cannot be extended to asymptomatic):
i. https://www.annalsofoncology.org/article/S0(20)39312-1/fulltext
ii. https://www.annalsofoncology.org/article/S0(20)39870-7/fulltext
iii. https://www.annalsofoncology.org/article/S0(20)39895-1/fulltext
2. Payer-related JNCCN article: https://jnccn.org/view/journals/jnccn/18/7/article-p866.xml
Speaker Bio
Mark Capone is presently the CEO of Precision Medicine Advisors, a consulting company focused on molecular diagnostics, pharmaceuticals, and biotechnology, which he founded in March 2020 following an 18-year career with Myriad Genetics. Mr. Capone joined Myriad in 2002 and oversaw its growth from an early-stage pioneer in molecular diagnostics with $14MM revenue, 300 employees, and 1 product to the largest and only profitable global precision medicine company with $790MM in revenue, 2,800 employees, and 11 products. He served as Myriad’s CEO and President from 2015-2020, and prior to joining the company, Mark spent 17 years with Eli Lilly and Company in positions across the entire value chain.
Please send us your questions or feedback at .
Thomas Tobin
Managing Director
Twitter
LinkedIn
