Plasma-based genetic testing can effectively be used to determine whether a tissue biopsy is necessary for EGFR mutation analysis in patients with non-small cell lung cancer.
Geoffrey Oxnard, MD
Plasma-based genetic testing can effectively be used to determine whether a tissue biopsy is necessary for EGFR mutation analysis in patients with nonsmall cell lung cancer (NSCLC), according findings from 3 studies presented during the 2016 European Lung Cancer Conference.
Each study looked at the concordance between mutation findings with tissue- and plasma-based testing for sensitizing EGFR alterations and the T790M acquired resistance mutation. Under the new proposed testing paradigm, patients with tumors that tested positive by liquid biopsy could forego further testing while negative findings would call for a confirmatory tissue biopsy.
“The data demonstrates that the responses are equivalent, which hopefully will ultimately lead us to a point where we no longer have to do a biopsy in every patient,” said study investigator Pasi A. Jänne, MD, PhD, professor of medicine at the Dana-Farber Cancer Institute. “I think we will see more and more plasma testing for genetic alterations in lung cancer, where we are trying to treat a genetically defined patient population.”
The largest of the studies,1which was known as ASSESS, examined 1162 patients from Europe and Japan with locally advanced or metastatic NSCLC, regardless of histology. The primary endpoint of the study was concordance between EGFR mutation statuses by tissue- and plasma-based testing. Additionally, the study sought to uncover clinical characteristics or patient demographics that may influence detection rates from plasma samples.
For plasma testing, a trend toward increased sensitivity was seen in those with a greater number of metastatic sites and higher metastatic grade. Additionally, age correlated with plasma-based testing sensitivity. Gender, ethnicity, smoking status, and PS did not influence the sensitivity of plasma-based mutation detection. Across all groups, the specificity of testing was >95%.
“If plasma testing is more reliable for some patients with certain characteristics, this may have implications in the way that we conduct mutation testing for patients with NSCLC, and ultimately impact upon treatment decisions,” said lead investigator Nicola Normanno, MD, chief of the Cell Biology and Biotherapy Unit, INT-Fondazione Pascale, Naples, Italy.
In those aged <65 years, plasma testing showed a sensitivity of 63.5% compared with 37.3% in those ≥65 years (P = .0002). For distant metastases, plasma testing had a sensitivity of 22.8% in the M1a group versus 63.4% in the M1b arm (P = .0915).
“The link with age is more difficult to understand,” said Normanno. “Evidence suggests that the biological features of certain tumors change with age. However, the specific biological mechanisms underlying the correlation between the success of plasma analysis and age will need to be investigated further.”
The sensitivity of plasma-based testing increased with the number of organs with metastasis (P = .2114). In those with no organ involvement, the sensitivity was 11.1% versus 69.4% in those with ≥3 organs involved. The sensitivity was 35.9% and 60.5% in those with 1 and 2 organs with metastasis, respectively.
“Our data suggests that for the majority of patients with metastatic disease a plasma test could be sufficient to determine EGFR mutation status, particularly when a robust and reliable methodology is used,” said Normanno. “Due to the low sensitivity of plasma genotyping, a biopsy will still be recommended in plasma negative cases.”
Liquid-based biopsy appears to have the most utility for detecting acquired mutations, such as T790M. In many cases, the ability to perform a rebiopsy is limited and the quantity of tissue obtained makes traditional testing challenging.
To uncover the agreement between liquid biopsy and traditional tissue-based tests at detecting T790M alterations, findings were assessed from a pooled analysis of two phase II studies.2Findings from the studies, AURA and AURA2, were instrumental in the FDA approval for osimertinib (Tagrisso) as a treatment for patients with T790M-mutant NSCLC following an EGFR-targeted TKI.
The testing analysis looked at the concordance between tissue testing using the approved companion diagnostic for osimertinib, the cobas EGFR Mutation Test v2, and a cobas plasma-based assay. The specificity of the test was confirmed using next-generation sequencing (NGS) on data from the AURA2 study alone.
In the pooled analysis, the positive agreement between the tissue and plasma test was 61.4% for the detection of EGFR T790M alterations. The negative agreement between the two tests was 78.6%. The positive agreement between the plasma test and NGS was 91.5%. The response rates to osimertinib were similar in those identified by tumor tissue and plasma.
“A non-invasive blood test appears to have the ability to find T790M-positive patients very effectively,” said Geoffrey Oxnard, MD, from the Dana-Farber Cancer Institute, who led a similar separate study. “The blood test only has a sensitivity of 70% or 80% so there are false negatives. In other words, if you have a negative result in the blood test it may be that the mutation was present but not detected.”
Oxnard led an analysis that looked exclusively at data from the phase I AURA trial of osimertinib in advanced NSCLC. This study retrospectively provided further information on the efficacy of the targeted agent in patients identified by tissue testing or plasma analysis using the BEAMing technique.
In 216 patients from the AURA trial,3there was a 70% concordance rate between tissue and plasma testing. In 179 patients detected by tissue samples, the objective response rate (ORR) with osimertinib was 62% and the median progression-free survival (PFS) was 9.7 months. In the plasma arm, the ORR was 63% and the median PFS was 9.7 months.
A subset of patients showed an interesting trend toward better outcomes when tested by plasma testing. For the plasma group (n = 104), the ORR was 46% and the median PFS was 8.2 months; however, in the tissue group (n = 58), the ORR was 26% and the median PFS was 3.4 months. An analysis of this group was able to identify subgroups of patients based on molecular characteristics, using the codetection of common sensitizing EGFR mutations. When including testing for common EGFR mutations, the concordance jumped to 80%.
In a group of patients with EGFR sensitizing mutations but no T790M alteration detected by plasma testing, the ORR was 38% and the median PFS was 4.4 months with osimertinib. Interestingly, when both T790M and a sensitizing EGFR mutation were not detected by plasma, better outcomes were seen. In this group, which did test positive for T790M by tissue, the ORR was 64% and the median PFS was 15.2 months.
“When we studied the tumor results on patients who were T790M-negative in the blood we could differentiate those who do better or worse on osimertinib, meaning that a biopsy is an effective fall back to clarify who should and who shouldn’t get the drug,” said Oxnard. “We conclude that a two stage approach is needed, starting with the blood test. Patients who test positive for T790M on the blood test can receive osimertinib. Those who test negative should have a biopsy test to clarify their T790M status.”
The high level of specificity seen across the studies, which ranged from 90% to 100%, is promising; however, the low sensitivity rate (70% to 80%) suggests that tissue-based testing will remain the standard of care. However, there are scenarios when tissue testing is not plausible, in which plasma testing is the only choice.
“These studies confirm the potential clinical utility of using circulating tumor DNA EGFR genotyping in routine practice and give information on the magnitude of false negatives,” said Jänne. “Plasma testing will not routinely replace tissue biopsy for mutation testing, which should still be regarded as the gold standard. It would be a complementary test, and may be a replacement in some patients, for example those in whom a tissue biopsy is not possible.”
Jänne went on to emphasize that blood could easily be drawn from all patients, whereas a biopsy can be difficult. With this in mind, plasma testing could expand genetic testing to a broader population of patients. However, one limitation of the results was the retrospective nature of the investigations, according to the researchers. Findings from the studies still need to be validated in real-world trials.