Liquid Biopsy Offers Alternative to Predicting Osimertinib Response in NSCLC

Geoffrey R. Oxnard, MD

Plasma genotyping can, in most cases, identify T790M-positivity in non-small cell lung cancer (NSCLC), which gives patients the option of receiving the targeted therapy osimertinib (Tagrisso) without the need for a tumor biopsy, according Geoffrey R. Oxnard, MD, at the 2016 European Lung Cancer Conference (ELCC).

The analysis investigated the effectiveness of osimertinib in 216 pretreated patients with NSCLC with acquired resistance to first-line EGFR inhibitors who had both both plasma and tumor genotyping results from the phase I AURA trial.¹ Of the 179 patients who tested positive for T790M in the tumor, the overall response rate (ORR) was 62% and the median progression-free survival (PFS) rate of 9.7 months. In the 167 patients who tested positive for T790M in plasma, results were similarly positive, with a 63% ORR and an 8.2-month median PFS.

However, outcomes were unexpectedly favorable in 104 patients who tested negative by plasma for T790M, with a 46% ORR and an 8.2-month PFS. This was compared with 58 patients who tested negative for T790M in the tumor who exhibited a 26% ORR and 3.4-month PFS.

OncLive: What were the goals of the study?

What were the most significant findings?

To better understand how plasma genotyping may change the use of osimertinib and the significance of some tumors testing positive for T790M in the plasma, but not in the tumor, OncLive interviewed Oxnard, assistant professor of medicine, Harvard Medical School, Dana-Farber Cancer Institute, who presented the analysis at ELCC. Oxnard: Patients develop resistance to an EGFR inhibitor, and at that moment we have a decision to make. Do we start chemotherapy, or do we put them through a biopsy to try to characterize their resistance and use that information to guide their treatment? If the biopsy shows T790M, we can give them osimertinib, and they could have a dramatic response and a durable PFS with little toxicity. However, to do this we are currently required to do a biopsy. So the question we asked in this study was, can we use a blood test, and in which situations can it be used for T790M resistance genotyping?We performed an analysis of hundreds of patients treated on the phase I AURA trial of osimertinib. We looked back at their tumor results and their plasma results. We found out that the plasma genotyping, which used a technology called beaming, a digital pCR technology, could detect about 60% to 80% of the EGFR-mutations in the plasma. It could not detect all cases; some cases just don’t have any DNA. When it is positive for a sensitizing mutation, it can be believed. The specificity is very high. Interestingly, when it is positive for a T790M mutation the specificity is more modest. About 30% of the plasma-positives did not have a T790M-positive tumor; there was a disconnect between the plasma result and the tumor result.

How did that play out in the outcome? The response rate of patients receiving osimertinib with a positive T790M in their tumor was high—over 60% and they had a long PFS. The response was similarly high in the plasma-positive patients. Those who had a T790M mutation in their plasma had a great response rate and long PFS. It looks like plasma can be used to identity patients who can benefit from this drug, and a simple blood test can be used to to decide who should get the drug instead of needing to go through the biopsy.

What are the biggest advantages of avoiding a biopsy?

The catch is the negatives. With T790M-negative in a tumor it means a low response rate in the range of 20% with a short median PFS. Those patients don’t benefit from osimertinib and they should get chemotherapy. That is part of our standard approach now. However, if the plasma is negative for T790M, that includes some true negatives and some false negatives. There were some cases were the blood test missed T790M. Those patients who were plasma-negative but tumor-positive still had a reasonable response rate in the range of over 40%, and a median PFS of 8-plus months. Those patients need a biopsy. A biopsy can tell us which of those plasma-negatives are in fact tumor-positive and should get osimertinib, and which are in fact tumor-negative and should get chemotherapy.With the approval of osimertinib in Europe, the United States, and Japan, it is now standard to perform a biopsy, when feasible, for T790M, and doing that makes patients candidates for osimertinib. But it’s not feasible for all patients. A biopsy involves risk and some tumors are not obtainable by a biopsy. Bone lesions and brain lesions, for example are difficult to biopsy, and therefore not all patients can get a biopsy. That means that some patients cannot capitalize on this new drug. A liquid biopsy gives patients who can’t get a biopsy the opportunity to receive osimertinib, and it’s much quicker and easier than a tumor biopsy.

What does it mean that some patients tested positive to T790M in the plasma, but negative in the tumor? Should they still receive osimertinib?

Getting a biopsy for tumor genotyping is complicated. It can take weeks to schedule it and to wait for the results to come back—patients can get sicker during this time. A blood test that can provide the oncologist some direction is attractive and will help get some patients onto the right drug quickly. As long as the blood test is negative, we fall back and proceed with the biopsy as planned. That way we don’t miss cases that were falsely negative in plasma.One of the trickiest challenges we found in this analysis is that there was a subset of patients that were T790M-positive in plasma, but were negative in the tumor. The question is, does this represent a false positive, an error of the assay, or is this in fact a true result? Is it possible that there is T790M in the blood, but not in the one site of the tumor where the biopsy was taken? Resistance can be heterogeneous. Maybe, if we did multiple biopsies we would find it in some sites but not others.

What should oncologists take away from this analysis?

When we looked at patients’ relative amount of T790M, the patients who were positive-only in plasma had a relatively low amount of T790M, suggesting they may have had a heterogeneous positively for T790M. The idea is that maybe T790M can be present in greater or lesser degrees, and if you have more of it, you have a better chance of benefiting from osimertinib, and if you have less of it, you have a less chance of benefiting. This phenomenon of plasma-positive, tumor-negative may be a representation of the heterogeneity. When we looked at the magnitude of response based on their relative magnitude of T790M, we did find that for those that have less than 10% T790M out of all their tumor DNA in the plasma, the amount of response to osimertinib seemed to be less. I think this is a question that needs to be investigated further. Moving this test from just a positive- negative test into a gradation of positive test, may give us a better understanding of who gains more benefit and who might benefit more from chemotherapy or combination therapy.There are plasma genotyping tests for T790M that are available today. There are commercial laboratories all over that are offering this, and they can detect T790M and with that result you can potentially start a drug for your patients. It may be off-label or on-label depending on what country you live in. In Europe, it is now part of the approved label that the drug is available based on a plasma or tumor finding of T790M. One caveat to keep in mind is that not all of these tests are well validated. So we need to ask the companies to go through a validation process, so that we can trust their results, and use those results for our patients. Clearly, the need is real, the assays are available, and we are just waiting for the final pieces of scientific data to come together so that we know which assays are best for our patients to help them get the treatments they need.

Oxnard G, Thress K, Alden R, et al. Plasma genotyping for predicting benefit from osimertinib in patients with advanced NSCLC. Presented at the European Lung Cancer Conference, April 13-16; Geneva, Switzerland. Abstract 135O

<<<

View more from the 2016 European Lung Cancer Conference