Lack of Concordance Regarding Plasma Testing in Lung Cancer Must Be Overcome

Although tumor-based and plasma-based next-generation sequencing (NGS) tests have come a long way in detecting genetic abnormalities in patients with lung cancer, various factors can contribute to a lack of concordance, according to Andrew Mckenzie, PhD.

Tumor shedding in particular can lead to a lack of concordance, Mckenzie explained. When tumors shed DNA into the bloodstream, it is possible to detect mutations using a plasma-based assay. However, should a blood test yield a negative result for a certain mutation, Mckenzie cautioned that this could be because the tumor has not shed DNA. Better education is needed on this aspect of disease biology so that test results can be better understood.

“This is a really interesting area of biology, and we’re still learning a lot about it: how and when certain tumors shed DNA in the bloodstream,” said Mckenzie. “What are those mechanisms? Is it strictly necrosis or apoptosis that leads to that? What about these tumors that are in the frontline setting and haven’t been exposed to therapy? They’re still shedding DNA, so there’s something biologically going on. All of that plays a role in how you interpret the results that you get back.”

In an interview with OncLive® during a 2020 Institutional Perspectives in Cancer webinar on lung cancer, Mckenzie, the director of personalized medicine at Sarah Cannon Research Institute, discussed the differences between tissue- and plasma-based NGS tests, factors that can lead to a lack of concordance, and important areas of ongoing research.

OncLive®: Could you start off by providing insight into the use of tumor tissue testing versus plasma testing in lung cancer?

Mckenzie: Many people are under the assumption that one test can replace the other. What I want to get across is that they’re 2 very different assays that serve 2 different purposes and [can] even yield different results. At Sarah Cannon Research Institute, we see the use for plasma- based testing in patients who have really difficult-to-biopsy disease, bone-only disease, or when monitoring disease over time; [this helps us] detect the emergence or disappearance of certain mutations over time.

We still believe that the gold standard for NGS is tissue-based sequencing—particularly, tissue-based tests that can detect RNA fusion events that are a bit more difficult to pick up on the DNA side or the plasma side. We put them in 2 different buckets even though some [institutions] use them interchangeably. We’re trying to [teach them] that this isn’t really the way that we should go about it.

What are some differences in results that can be seen among the tests?

The FDA has approved 3 or 4 different tissue-based tests that are tumor agnostic, broad sequencing. This year, the FDA also approved 2 plasma-based tests. The differences among those tests are the genes included on those panels and whether those tests will detect RNA versus DNA. We believe that all those tests are great and do a wonderful job. [We just] really need to understand which genes are on that panel. There’s a difference between someone being negative for a mutation and just having never been tested for it. That difference between negative and null is something that gets lost on a lot of folks when they order these [tests]. They think, “I ordered NGS; therefore, I’ve looked for everything that [I feel is necessary].” The [tests all use] excellent technology and serve the purpose really well, but we just need to know what we’re looking for.

What factors should be taken into consideration to match patients with the right test?

We recommend that physicians take a look at a few factors when deciding which test to order. The first, as I mentioned earlier, is whether the patient has disease [that can be biopsied]. If so, the recommendation is to try to get a fresh biopsy and do the NGS on that tissue; that’s our first-line gold standard. If we don’t have that and the patient hasn’t been exposed to several lines of systemic therapy, we’re OK with going back to archival tissue and testing it; that’s our second option. If none of that works or you don’t have any archival tissue, or they have been through several lines of therapy, or they don’t want to rely on the archival tissue, then the third option is plasma-based testing.

The other option that is emerging in the field now is [focused on determining] whether the tumors are shedding DNA into the bloodstream for a plasma-based test. Some really elegant work has shown that the more a tumor sheds DNA into the bloodstream, the worse the prognosis is. [This approach represents a way] to gauge how well, or not, a patient might do just based on the presence or absence of that circulating DNA. As such, that’s another factor we take into consideration. If you do a blood draw and it comes back with no mutations, we don’t [necessarily believe] that those tumors don’t have any mutations; we just think that they’re not shedding into the bloodstream. That’s where we might want to go back and try to find tissue if we haven’t done so already.

Certainly, cost is another factor that we talk about when [determining] which test should be done for different patients; that’s always a consideration. Our clinicians have great relationships with our vendors. Usually, cost is not a barrier to getting this done. Our vendors are very dedicated to supplying those tests when they’re needed. That has become less of an issue over the years, but I know that early on, cost was a big factor when deciding which test to order.

Another factor that gets overlooked a bit is the support that vendors give for interpreting those results. Different vendors offer different levels of support and, depending on the sophistication of the clinician, they might require a bit more support in knowing what to do with those results when they come back. That’s something else that plays into [all of this].

How does cell-free DNA shedding contribute to a lack of concordance?

Just because a plasma-based test did not detect a mutation doesn’t mean that a mutation isn’t present; it might just mean that the DNA wasn’t shed into the bloodstream. That’s not a limitation of the test or [an indication that it has] low sensitivity—it’s just the result of the disease biology. That’s something that we need to educate more on. What does it mean to have a negative result on a plasma-based test?

How can heterogeneity create a lack of concordance?

A really interesting part of tumor biology is that not all the cells that make up a single tumor are genetically identical. There is a lot of variability in the genetics of these tumor cells, maybe even in the same tumor. Research from Charles Swanton, FRCP, BSc, PhD, of the Francis Crick Institute has shown that if you take a tumor and do several slices through it and sequence them independently, you get different mutations that appear on the different samples. This was a bit disheartening, to know that tumors are a bit more complex than we gave them credit for. Now we’re realizing that that is true not only for [the primary] tumor but also when a tumor metastasizes. Metastatic lesions can carry unique genetic abnormalities from the primary tumor.

It means that you need to know what you’re biopsying; if it’s a primary tumor versus metastases, you might expect there to be differences, and that’s OK. Early on, people saw discordant results but didn’t take into account that they were analyzing different lesions, even at different times in treatment. That’s another thing that we learned: These tumors change over time and through treatment. That all plays into heterogeneity and increases the complexity of interpreting these results. Our guiding principle is that if a tumor is driven by a particular mutation, that drive mutation usually persists. That’s something that we look for across different metastatic lesions and in the plasma-based versus the tissue-based testing; we want to identify those driver mutations that persist, even as the tumor landscape changes.

What other factors can lead to a lack of concordance?

Heterogeneity, sample type, and the test itself; in addition, there are different tests for different analytes. The fact that an RNA-based test picked up a fusion but your DNA-based test did not just means that you tested 2 different things. Maybe we should focus on changing that vocabulary; I don’t know that it’s fair [to call them] discordant; rather, they’re just testing different things with different tests that are yielding different results.

Where do you feel future efforts should be focused?

I am particularly interested in 2 big areas. The first is the early-detection space in cancer, [which is focused on] using the presence or absence of some genomic signature to indicate whether an otherwise healthy patient may develop cancer. There’s really excellent ongoing work in the plasma-based testing space. On the other side of that is how we monitor disease in a patient who actively has cancer. There’s some really excellent work in tracking how many patients change over time and through treatments with blood draws. Understanding how tumors evolve over time will really shed some light on why these different tests might be yielding different results. If we can [learn] a bit more about the underlying biology, it will help us understand the results that we’re getting now.

We’re getting these really great snapshots when we perform these NGS tests, but that snapshot is from whenever that tumor was excised. As we evolve in our understanding, we’ll be able to put that snapshot into better context.

What are the major takeaways regarding the current state of NGS?

NGS has opened up an incredible toolbox for the treatment of patients with cancer. However, just like with anything, you have to know what the tests are saying in order to make the most of the results that come back. We’re advocating not only for [further] adoption of NGS but also for seeking out the resources that are needed to fully understand results. [We need to] dig deep into these tests and really find out what they’re saying. Don’t suppose anything, because there’s a lot of underlying complexity that can lead to some of the unsuspected results.