Predictive Biomarkers in Non-Small-Cell Lung Cancer

Several predictive genetic biomarkers have been identified for NSCLC. Somatic genome alterations known as "driver mutations" are usually transformative variants arising in cancer cells in genes encoding for proteins important in cell growth and survival. Randomized controlled trials have demonstrated improved efficacy, often in conjunction with decreased toxicity, of matching targeted therapies to patients with specific driver mutations. Several such targeted therapies are approved by the Food and Drug Administration (FDA) for NSCLC. Guidelines generally suggest the analysis of either the primary NSCLC tumor or of metastasis for the presence of a set of driver mutations to select an appropriate treatment.

Genetic Biomarkers With FDA-Approved Targeted Therapies

The list of targeted therapies approved for NSCLC is evolving. Currently, there are FDA-approved targeted therapies for epidermal growth factor receptor (EGFR) variants, anaplastic lymphoma kinase (ALK) translocations, ROS1 translocations, and BRAF variants for NSCLC. Companion diagnostics using tissue samples have also been FDA-approved to identify the associated driver mutations for the targeted therapies. The evaluation of molecular analysis of tissue samples for targeted therapy of NSCLC is 'Molecular Analysis for Targeted Therapy for Non-Small-Cell Lung Cancer' (Policy #053 in the Pathology Section).

EGFR Variants

Specific EGFR variants confer sensitivity to treatment with tyrosine kinase inhibitors (TKIs), such as erlotinib, gefitinib, afatinib, and osimertinib; the most common variants are deletions in exons 19 and an exon 21 substitution variant (L858R). These variants are referred to as TKI-sensitizing variants and are found in approximately 10% of white patients and up to 50% of Asian patients. The prevalence of EGFR variants is not well characterized in other ethnic or racial groups but is estimated to be 10% to 15% in studies including general U.S. populations. TKIs are indicated as first-line treatment for patients with sensitizing variants; progression-free survival is improved with the use of TKIs. Patients receiving TKIs have fewer treatment-related adverse events than patients receiving cytotoxic chemotherapy.

ALK and ROS1 Translocations

ALK rearrangements confer resistance to TKIs. Approximately 4% of patients have ALK rearrangements. The TKI crizotinib, an inhibitor of ALK, ROS1, and mesenchymal-epithelial transition (MET) tyrosine kinases, is indicated in patients with ALK-positive tumors. In randomized trials comparing crizotinib with standard chemotherapy in ALK-positive patients, crizotinib has been associated with improved progression-free survival, response rates, lung cancer symptoms, and quality of life. ROS1 rearrangements develop in 1% to 2% of patients. For such patients, crizotinib has been shown to be effective, with response rates of about 70%.

BRAF Variants

RAF proteins are serine/threonine kinases that are downstream of RAS in the RAS-RAF-ERK-MAPK pathway. In this pathway, the BRAF gene is the most frequently mutated in NSCLC, in 1% to 3% of adenocarcinomas. Unlike melanoma, about 50% of the variants in NSCLC are non-V600E variants. BRAF or MEK inhibition with TKIs (eg, vemurafenib/dabrafenib or trametinib) was originally approved by the FDA for treatment of unresectable or metastatic melanoma with BRAF V600 variants but the combination of dabrafenib and trametinib was expanded to include treatment of metastatic NSCLC in 2017.

Genetic Biomarkers With Off-Label Targeted Therapies

Proposed targeted therapies may be used off-label for genetic alterations in human epidermal growth factor receptor 2 (trastuzumab, afatinib), MET (crizotinib), and RET (cabozantinib). Human epidermal growth factor receptor 2 is a member of the HER (EGFR) family of TK receptors and has no specific ligand. When activated, it forms dimers with other EGFR family members. Human epidermal growth factor receptor 2 is expressed in approximately 25% of NSCLC. RET (rearranged during transfection) is a proto-oncogene that encodes a receptor tyrosine kinase growth factor. RET fusions occur in 0.6% to 2% of NSCLCs and 1.2% to 2% of adenocarcinomas. MET amplification is one of the critical events for acquired resistance in EGFR-mutated adenocarcinomas refractory to EGFR TKIs. MET amplification occurs in 2% to 4% of treatment-naive NSCLC and MET and EGFR commutations occur in 5% to 20% of NSCLC tumors with acquired resistance to EGFR TKIs.

Genetic Biomarkers Without Targeted Therapies

The most common predictive variant in North American populations is KRAS, occurring in 20% to 25% of NSCLC. Patients with KRAS variants have shorter survival than those without KRAS variants, and thus KRAS is a prognostic marker. It also predicts a lack of TKI efficacy. Because KRAS variants are generally not found with other tumor biomarkers, KRAS testing might identify patients who would not benefit from further molecular testing. Targeted therapies are under investigation for KRAS-variant NSCLC.

Tyrosine Kinase Inhibitor-Resistance Variants

EGFR Variants

The EGFR variant T790M has been associated with acquired resistance to TKI therapy. When the T790M variant is detected in tissue biopsies from patients with suspected resistance to TKI therapy, osimertinib is recommended as second-line therapy. However, the use of osimertinib as first-line therapy for patients who have EGFR-sensitizing variants is emerging and may prevent the development of T790M resistance.

Treatment Selection

Tissue Biopsy as a Reference Standard

The standard for treatment selection in NSCLC is biomarker analysis of tissue samples obtained by biopsy or surgery. However, a lung biopsy is invasive with a slow turnaround time for obtaining results. Tissue biopsy may also be an imperfect reference standard due to inadequate sampling, tumor heterogeneity, or other factors.

Technologies for Detecting Circulating Tumor DNA

Cell-free DNA in blood is derived from nonmalignant and malignant cell DNA. The small DNA fragments released into the blood by tumor cells are referred to as ctDNA. Most ctDNA is derived from apoptotic and necrotic cells, either from the primary tumor, metastases or circulating tumor cells.^1, Unlike apoptosis, necrosis is considered a pathologic process, generating larger DNA fragments due to incomplete and random digestion of genomic DNA. The length or integrity of the circulating DNA can potentially distinguish between apoptotic and necrotic origins. The ctDNA can be used for genomic characterization of the tumor and identification of the biomarkers of interest.

Detection of ctDNA is challenging because cell-free DNA is diluted by nonmalignant circulating DNA and usually represents a small fraction (<1%) of total cell-free DNA. Therefore, methods up to 500 to 1000 times more sensitive than standard sequencing approaches (eg, Sanger) are needed.

Sensitive and specific methods are available to detect ctDNA and identify single nucleotide variants, duplications, insertions, deletions, and structural variants. Examples of methods are as follows:

• Denaturing high-performance liquid chromatography involves polymerase chain reaction (PCR) followed by denaturing plus hybridization and then separation.
• Peptide nucleic acid-locked nucleic acid PCR suppresses wild-type EGFR followed by enrichment for mutated EGFR.
• Amplification refractory mutation system PCR generates different-sized PCR products based on the allele followed by separation of PCR fragments to determine the presence of variants.
• BEAMing combines emulsion PCR with magnetic beads and flow cytometry.
• Digital genomic technologies, such as droplet digital PCR, allow for the enumeration of rare variants in complex mixtures of DNA.

Regulatory Status

In June 2016, cobasÒ EGFR Mutation Test v2 (Roche Molecular Systems), a real-time PCR test, was approved by the FDA through the premarket approval process (P150047).^2, This plasma test is a real-time PCR test approved as a companion diagnostic aid for selecting NSCLC patients who have EGFR exon 19 deletions, and L858R substitution variants, for treatment with erlotinib. A premarket approval supplement expanded the indication to include the test as a companion diagnostic for treatment with gefitinib in 2018 (P120019). Patients who test negative for the variants detected should be referred for (or "reflexed" to) routine biopsy with tissue testing for EGFR variants. A previously approved version 2 of this test, which used tissue biopsy specimens, was also approved for the detection of T790M variants in tissue, which are used to select patients to receive osimertinib. Approval of version 2 of the plasma test did not include detection of T790M variants.

No other ctDNA tests have FDA approval. Guardant Health (Guardant 360Ò) and Foundation Medicine (FoundationACT™) were granted Expedited Access for Premarket Approval and De Novo Medical Devices Intended for Unmet Medical Need for Life Threatening or Irreversibly Debilitating Diseases or Conditions in 2018. FoundationACT™ is currently marketed as FoundationOne Liquid.

Clinical laboratories may develop and validate tests in-house and market them as a laboratory service; laboratory-developed tests must meet the general regulatory standards of the Clinical Laboratory Improvement Amendments. Several companies market tests that detect tumor markers from peripheral blood, including TKI-sensitizing variants for NSCLC. Laboratories that offer laboratory-developed tests must be licensed by the Clinical Laboratory Improvement Amendments for high-complexity testing. To date, the FDA has chosen not to require any regulatory review of this test. Clinical laboratories accredited through the College of American Pathologists enroll in proficiency testing programs to measure the accuracy of the test results. There are currently no College of American Pathologists proficiency testing programs available for ctDNA testing to ensure the accuracy of ctDNA laboratory-developed tests.

Related Policies

• Miscellaneous Genetic and Molecular Diagnostic Tests (Policy #114 in the Pathology Section)
• Proteomic Testing for Targeted Therapy in Non-Small Cell Lung Cancer (Policy #115 in the Pathology Section)
• Circulating Tumor DNA and Circulating Tumor Cells for Cancer Management (Liquid Biopsy) (Policy #038 in the Pathology Section)
• Molecular Analysis for Targeted Therapy for Non-Small-Cell Lung Cancer (Policy #053 in the Pathology Section)

I. EGFR TESTING

A. At diagnosis, analysis of somatic variants in exons 18 through 21 (eg, G719X, L858R, T790M, S6781, L861Q) within the epidermal growth factor receptor (EGFR), using the cobas EGFR Mutation Test v2, Guardant360 test, OncoBEAM test or InVisionFirst-Lung test with plasma specimens to detect circulating tumor DNA (ctDNA), is considered medically necessary as an alternative to tissue biopsy (see Policy Guidelines) to predict treatment response to an EGFR tyrosine kinase inhibitor therapy (eg, erlotinib [Tarceva], gefitinib [Iressa], afatinib [Gilotrif], or osimertinib [Tagrisso]) in members with advanced lung adenocarcinoma, large cell carcinoma, advanced squamous cell non-small-cell lung cancer, and non-small-cell lung cancer not otherwise specified.

B. At progression, analysis of the EGFR T790M resistance variant for targeted therapy with osimertinib using ctDNA using the cobas EGFR Mutation Test v2, Guardant360 test, OncoBEAM test or InVisionFirst-Lung test with plasma specimens to detect circulating tumor DNA (ctDNA), is considered medically necessary in members with advanced lung adenocarcinoma, large cell carcinoma, advanced squamous cell non-small-cell lung cancer, and non-small-cell lung cancer not otherwise specified when tissue biopsy to obtain new tissue is not feasible, e.g., in those who do not have enough tissue for standard molecular testing using formalin-fixed paraffin-embedded tissue, do not have a biopsy-amenable lesion, or cannot undergo biopsy. (see Policy Guidelines).

C. Analysis of other EGFR variants within exons 22 to 24, or other applications related to NSCLC, is considered investigational.

Analysis of somatic rearrangement variants of the ALK gene using plasma specimens to detect ctDNA or RNA is considered investigational as an alternative to tissue biopsy to predict treatment response to ALK inhibitor therapy (eg, crizotinib [Xalkori], ceritinib [Zykadia], alectinib [Alecensa], or brigatinib [Alunbrig]) in members with NSCLC.

Analysis of the BRAF V600E variant using plasma specimens to detect ctDNA is considered investigational as an alternative to tissue biopsy to predict treatment response to BRAF or MEK inhibitor therapy (eg, dabrafenib [Tafinlar], trametinib [Mekinist]) in members with NSCLC.

Analysis of somatic rearrangement variants of the ROS1 gene using plasma specimens to detect ctDNA or RNA is considered investigational as an alternative to tissue biopsy to predict treatment response to ALK inhibitor therapy (crizotinib [Xalkori]) in members NSCLC.

Analysis of somatic variants of the KRAS gene using plasma specimens to detect ctDNA is considered investigational as a technique to predict treatment nonresponse to anti-EGFR therapy with tyrosine kinase inhibitors and for the use of the anti-EGFR monoclonal antibody cetuximab in NSCLC.

Analysis of alterations in the HER2, RET, and MET genes using plasma specimens to detect ctDNA for targeted therapy in patients with NSCLC is considered investigational.

· The Individual has either recurrent, relapsed, refractory, metastatic, or advanced stages III or IV cancer; and
· The Individual has either not been previously tested using the same NGS test for the same primary diagnosis of cancer or repeat testing using the same NGS test only when a new primary cancer diagnosis is made by the treating physician; and
· The Individual has decided to seek further cancer treatment (e.g., therapeutic chemotherapy).

The diagnostic laboratory test using NGS must have:
o Food & Drug Administration (FDA) approval or clearance as a companion in vitro diagnostic; and,
o an FDA-approved or cleared indication for use in that individual’s cancer; and,
o results provided to the treating physician for management of the individual using a report template to specify treatment options.

· At diagnosis when results for EGFR single nucleotide variants (SNVs) and insertions and deletions (indels); rearrangements in ALK and ROS1; and SNVs for BRAF are not available AND when tissue-based CGP is infeasible [i.e., quantity not sufficient (QNS) for tissue-based CGP or invasive biopsy is medically contraindicated],

· At progression:
o For patients progressing on or after chemotherapy or immunotherapy who have not been tested for EGFR SNVs and indels; rearrangements in ALK and ROS1; and SNVs for BRAFs, and for whom tissue-based CGP is infeasible; or
o For patients progressing on EGFR tyrosine kinase inhibitors (TKIs).

Genetics Nomenclature Update

The Human Genome Variation Society nomenclature is used to report information on variants found in DNA and serves as an international standard in DNA diagnostics. It is being implemented for genetic testing medical policy updates starting in 2017 (see Table PG1). The Society's nomenclature is recommended by the Human Variome Project, the Human Genome Organization, and by the Human Genome Variation Society itself.

The American College of Medical Genetics and Genomics and the Association for Molecular Pathology standards and guidelines for interpretation of sequence variants represent expert opinion from both organizations, in addition to the College of American Pathologists. These recommendations primarily apply to genetic tests used in clinical laboratories, including genotyping, single genes, panels, exomes, and genomes. Table PG2 shows the recommended standard terminology-"pathogenic," "likely pathogenic," "uncertain significance," "likely benign," and "benign"-to describe variants identified that cause Mendelian disorders.

Table PG1. Nomenclature to Report on Variants Found in DNA

Previous	Updated	Definition
Mutation	Disease-associated variant	Disease-associated change in the DNA sequence
	Variant	Change in the DNA sequence
	Familial variant	Disease-associated variant identified in a proband for use in subsequent targeted genetic testing in first-degree relatives

Table PG2. ACMG-AMP Standards and Guidelines for Variant Classification

Variant Classification	Definition
Pathogenic	Disease-causing change in the DNA sequence
Likely pathogenic	Likely disease-causing change in the DNA sequence
Variant of uncertain significance	Change in DNA sequence with uncertain effects on disease
Likely benign	Likely benign change in the DNA sequence
Benign	Benign change in the DNA sequence

ACMG: American College of Medical Genetics and Genomics; AMP: Association for Molecular Pathology.


[RATIONALE: This policy was created in 2018 and has been updated with searches of the MEDLINE database. The most recent literature update was performed through August 1, 2019.

Evidence reviews assess whether a medical test is clinically useful. A useful test provides information to make a clinical management decision that improves the net health outcome. That is, the balance of benefits and harms is better when the test is used to manage the condition than when another test or no test is used to manage the condition.

The first step in assessing a medical test is to formulate the clinical context and purpose of the test. The test must be technically reliable, clinically valid, and clinically useful for that purpose. Evidence reviews assess the evidence on whether a test is clinically valid and clinically useful. Technical reliability is outside the scope of these reviews, and credible information on technical reliability is available from other sources.

Selecting Targeted Therapy

Clinical Context and Test Purpose

The purpose of identifying targetable oncogenic "driver mutations" such as epidermal growth factor receptor (EGFR) variants in patients who have non-small-cell lung cancer (NSCLC) is to inform a decision whether patients should receive a targeted therapy vs another systemic therapy. Patients have traditionally been tested for driver mutations using samples from tissue biopsies.

Figures 1 and 2 show how liquid biopsy could be used to select first-line and second-line treatments in patients with advanced NSCLC with reflex to tissue biopsy and how it would potentially affect outcomes. The testing strategy in Figure 1 is based on the reflex testing strategy suggested in the U.S. Food and Drug Administration (FDA) approval for the cobas test. Some guidelines have suggested a different testing strategy wherein testing with a liquid biopsy is considered when testing with a tissue biopsy is not feasible.

The questions addressed in this evidence review are:

1. How accurately does liquid biopsy detect driver or resistance variants of interest in the relevant patient population (clinical validity)?
2. Does a strategy including liquid biopsy in patients with NSCLC improve the net health outcome compared with standard biopsy?

Patients

The target population consists of patients with NSCLC where tumor biomarker testing is indicated to select a treatment. Patients may be treatment-naive, or being considered for a treatment change due to progression, recurrence, or suspected treatment resistance.

Treatment recommendations for patients with advanced NSCLC are usually made in the tertiary care setting ideally in consultation with a multidisciplinary team of pathologists, thoracic surgeons, and oncologists.

Routine surveillance or periodic monitoring of treatment response as potential uses of the liquid biopsy were not evaluated in this policy.

Interventions

The technology considered is an analysis of tumor biomarkers in peripheral blood (liquid biopsy) to determine treatment selection. Several commercial tests are available and many more are in development. In contrast to tissue biopsy, guidelines do not exist establishing the recommended performance characteristics of liquid biopsy.^3,

The evidence is considered separately for the different biomarkers. Studies have evaluated liquid biopsy for biomarkers that detect EGFR tyrosine kinase inhibitor (TKI) sensitization, concentrating on the EGFR exon 19 deletion and EGFR L858R variants. Studies have also evaluated separately biomarkers associated with TKI resistance, concentrating on the EGFR T790M variant.

Studies have also assessed a liquid biopsy for detection of the EML4-ALK fusion oncogene and its variants, translocation between ROS1 and other genes (most commonly CD74), BRAF variants occurring at the V600 position of exon 15, and other variants.

Comparators

The relevant comparator of interest is testing for variants using tissue biopsy.

The testing strategy in Figure 1 is based on the reflex testing strategy suggested in the FDA approval for the cobas test. Some guidelines have suggested that testing with a liquid biopsy should be used when testing with tissue biopsy is not feasible.

Figure 1. Liquid and Tissue Biopsy in the Selection of First-Line Systemic Therapy for Advanced NSCLC

EGFR: epidermal growth factor receptor; NSCLC: non-small-cell lung cancer; PD-L1: programmed death-1 ligand; PFS: progression-free survival; ORR: objective response rate; OS: overall survival.

Figure 2. Liquid and Tissue Biopsy in the Selection of Second-Line Systemic Therapy for Advanced NSCLC

NSCLC: non-small-cell lung cancer; PFS: progression-free survival; ORR: objective response rate; OS: overall survival.

Outcomes

The outcomes of interest are OS and cancer-related survival. In the absence of direct evidence, the health outcomes of interest are observed indirectly as a consequence of the interventions taken based on the test results.

In patients who can undergo tissue biopsy, given that negative liquid biopsy results are reflexed to tissue biopsy, a negative liquid biopsy test (true or false) does not change outcomes compared with tissue biopsy.

Similarly, in patients who cannot undergo tissue biopsy, a negative liquid biopsy test (true or false) should result in the patient receiving the same treatment as he/she would have with no liquid biopsy test so a negative liquid biopsy test does not change outcomes.

The implications of positive liquid biopsy test results are described below.

Potential Beneficial Outcomes with Positive Result

For patients who can undergo tissue biopsy, the beneficial outcomes of a true-positive liquid biopsy result are the avoidance of tissue biopsy and its associated complications. In the National Lung Screening Trial, which enrolled 53454 persons at high- risk for lung cancer at 33 U.S. medical centers, the percentage of patients having at least 1 complication following a diagnostic needle biopsy was approximately 11%.^4,

For patients who cannot undergo tissue biopsy, the beneficial outcomes of a true-positive liquid biopsy result are receipt of a matched targeted therapy instead of chemotherapy and/or immunotherapy. The benefits of targeted therapy for patients with driver mutations in NSCLC are discussed in a separate policy.

Potential Harmful Outcomes with Positive Result

The harmful outcome of a false-positive liquid biopsy result is incorrect treatment with a targeted therapy instead of immunotherapy and/or chemotherapy. In a meta-analysis of randomized controlled trials (RCTs) of EGFR TKIs vs chemotherapy in patients without EGFR-sensitizing variants, the overall median progression-free survival (PFS) was 6.4 months in patients assigned to chemotherapy vs 1.9 months in patients assigned to EGFR TKIs (hazard ratio [HR], 1.41; 95% confidence interval [CI], 1.10 to 1.81). The advantage of chemotherapy over EGFR TKIs for patients without EGFR-sensitizing variants was true in both the first- and second-line settings.^5,

In the AZD9291 First Time In Patients Ascending Dose Study (AURA 1), single-arm, phase 1 trial of osimertinib, among 61 patients with EGFR-sensitizing variants who had progressed on an EGFR TKI but who did not have the EGFR T790M resistance variant, the response rate was 21% (95% CI, 12% to 34%) and median PFS was 2.8 months (95% CI, 2.1 to 4.3 months).^6, There was no concurrent control group in AURA 1 for comparison of osimertinib with other second-line treatments among T790M-negative patients. However, in the IMpower 150 trial, the addition of the immunotherapy atezolizumab to the combination chemotherapy of bevacizumab, carboplatin, and paclitaxel improved PFS in a subset of 111 patients with EGFR-sensitizing variants or ALK translocations who had progressed on a prior targeted agent (median PFS, 9.7 months vs 6.1 months; HR=0.59; 95% CI 0.37 to 0.94).^7,

Due to the poor prognosis of advanced NSCLC, the duration of follow-up for the outcomes of interest issix months and one year.

Study Selection Criteria

For the evaluation of the clinical validity of each test, studies that met the PICO criteria described above and the following eligibility criteria were considered:

• Reported on the performance characteristics (sensitivity and specificity) of the marketed version of the technology or included data sufficient to calculate sensitivity and specificity
• Included a suitable reference standard (tissue biopsy)
• Patient/sample clinical characteristics were described and patients were diagnosed with NSCLC
• Patient/sample selection criteria were described.
• At least 20 patients are included.

Assessment of technical reliability focuses on specific tests and operators and requires a review of unpublished and often proprietary information. Review of specific tests, operators, and unpublished data are outside the scope of this policy and alternative sources exist. This policy focuses on the clinical validity and clinical utility.

Clinically Valid

A test must detect the presence or absence of a condition, the risk of developing a condition in the future, or treatment response (beneficial or adverse).

BCBSA staff performed a systematic review, including 55 studies reporting clinically validity of liquid biopsy compared with tissue biopsy for detection of EGFR TKI-sensitivity variants or resistance variants through February 2017. Details of that systematic review are found in Appendix 1. In brief, most studies were conducted in Asia, using tests not currently being marketed in the U. S.. There was high variability in performance characteristics, with sensitivities ranging from close to 0% to 98% and specificities ranging from 71% to 100%. Therefore, evidence will not be pooled across tests going forward and instead reviewed separately for tests marketed in the U. S. A systematic review by Wu et al (2015) noted sensitivity might be lower in studies including non-Asian ethnicities (55%; 95% CI, 33% to 77%) compared with Asian ethnicities (68%; 95% CI, 57% to 79%), although the difference was not statistically significant.^8, Therefore, studies in the U. S. or similar populations will be most informative regarding the clinical validity of tests marketed in the U. S.

As previously described, there are multiple commercially available liquid biopsy tests that detect EGFR and other variants using a variety of detection methods. Given the breadth of molecular diagnostic methodologies available and the lack of guidelines regarding the recommended performance characteristics of liquid biopsy^3,, the clinical validity of each commercially available test must be established independently. The market is changing rapidly and all available tests may not be represented in the appraisal below.

Several clinical validity studies comparing liquid biopsy with tissue biopsy in patients who had advanced NSCLC for marketed tests have been published. Characteristics of the studies are shown in Table 1. Most have included testing for EGFR variants but a few included testing for less prevalent variants as well.

Evidence for the different variants is reviewed separately. Performance characteristics for detecting one type of variant (e.g., point mutations) may not represent performance to detect other types of variants (e.g., gene fusions).^9,

Table 1. Characteristics of Clinical Validity Studies of Liquid Biopsy With Tissue Biopsy as the Reference Standard

Study		Study Population	Design	Variants Includeda	Timing of Reference and Index Tests
Cobas EGFR test
Jenkins et al (2017)10,	Patients with advanced NSCLC who had progressed on EGFR TKI therapy enrolled in AURA extension or AURA2 studies in U.S., Europe, Asia, and Australia		Retrospective	EGFR resistance	Both tissue and blood samples collected at screening/baseline
FDA SSED (2016)11,	Patients with stage IIIb/IV NSCLC enrolled in a phase 3 RCT in Asia between 2011 and 2012		Retrospective	EGFR	Both tissue and blood samples collected at screening
Karlovich et al (2016)12,	Patients with newly diagnosed or relapsed patients with advanced (stage IIIB, IV) NSCLC in U.S., Europe, and Australia between 2011 and 2013		Prospective	EGFR, BRAF	Plasma was collected within 60 d of tumor biopsy
Thress et al (2015)13,	Patients with NSCLC enrolled in a multinational (including U.S.) phase 1 study who had progressed on an EGFR TKI therapy		Prospective	EGFR	Blood and tissue collected after progression and before next-line treatment; time between not specified
Mok et al (2015)14,	Patients enrolled in a phase 3 RCT in Asian with stage IIIB/IV NSCLC		Prospective	EGFR	Tissue samples from diagnosis or resection or biopsy 14 d before first study dose. Blood collected within 7 d prior to first study dose
Weber et al (2014)15,	Patients in Denmark with NSCLC (84% stage IV) from 2008 to 2011		Retrospective	EGFR	Blood samples collected a median of 10.5 mo after diagnostic biopsy
Guardant360
Leighl et al (2019)16,	Patients with biopsy-proven, previously untreated, nonsquamous NSCLC (stage IIIB/IV) enrolled in the NILE study (Non-invasive versus Invasive Lung Evaluation at one of 28 North American centers between 2016 and 2018		Prospective	EGFR, ALK, ROS1, BRAF	Unclear
Schwaederle et al (2017)17,	Patients with lung adenocarcinoma (86% with metastatic disease) from academic medical center in California between 2014 and 2015		Retrospective, consecutive	EGFR, ALK, ROS1, BRAF	Median time was 0.8 mo, range not given
Thompson et al (2016)18,	Patients with NSCLC or suspected NSCLC (96% stage IV) from Pennsylvania between 2015 and 2016		Prospective, consecutive	EGFR, ALK, ROS1, BRAF	Time between tissue and blood collection ranged from 0 d to >2 y
Villaflor et al (2016)19,	Patients in Chicago with NSCLC (68% stage IV) who had undergone at least 1 ctDNA test at a single commercial ctDNA laboratory in 2014 and 2015		Retrospective, selection unclear	EGFR, ROS1, BRAF	Time between biopsy and blood draw ranged from 0 d to 7 y (median, 1.4 y)
OncoBEAM
Ramalingam et al (2018)20,	Patients with locally advanced or metastatic NSCLC from the AURA study conducted in U.S., Europe, and Asia		Prospective	EGFR	Plasma was collected at baseline, time of tissue sample not specified
Karlovich et al (2016)12,	Patients with newly diagnosed or relapsed patients with advanced (stage IIIB, IV) NSCLC in U.S., Europe, and Australia between 2011 and 2013		Prospective	EGFR, BRAF	Plasma was collected within 60 d of tumor biopsy
Thress et al (2015)13,	Patients with NSCLC enrolled in a multinational (including U.S.) phase 1 study who had progressed on an EGFR TKI therapy		Prospective	EGFR	Blood and tissue collected after progression and before next-line treatment; time between not specified
GeneStrat
Mellert et al (2017)21,	Patients in the test utilization data had lung cancer; unclear whether the samples in the clinical validity data were from patients with advanced NSCLC, patient characteristics are not described		Retrospective, selection unclear	EGFR, ALK	Timing not described
ctDx-Lung
Paweletz et al (2016)22,	Patients in Boston with advanced NSCLC with a known tumor genotype, either untreated or progressive on therapy		Prospective	EGFR, ALK, ROS1, BRAF	Timing not described
InVision
Pritchet et al (2019)23,	Patients with untreated, advanced NSCLC; primarily from cohorts enrolled in 2 prospective US studies with 41 centers		Prospective	EGFR, ALK, ROS1, BRAF	Blood collected within 12 weeks of tissue biopsy and no therapy between tissue and blood samples
Remon et al (2019)24,		Patients with advanced NSCLC enrolled in single-center, prospective observational study in France. Patients were either treatment naıve for advanced disease or who had tissue-based molecular profile that failed or was not performed on the primary tissue sample (treated rescue cohort)	Prospective	EGFR, BRAF	Time between tissue biopsy and blood collection less than 100 days; median time between tissue biopsy and liquid biopsy collection was 34 days.

ctDNA: circulating tumor DNA; EGFR: epidermal growth factor receptor; FDA: Food and Drug Administration; NSCLC: non-small-cell lung cancer; RCT: randomized controlled trial;SSED: Summary of Safety and Effectiveness Data; TKI: tyrosine kinase inhibitor.
^a Noting EGFR, ALK, ROS1, and BRAF variants only.

Table 2 summarizes the results of clinical validation studies of liquid biopsy compared with tissue biopsy as a reference standard. Although tissue biopsy is not a perfect reference standard, the terms sensitivity and specificity will be used to describe the positive percent agreement and negative percent agreement, respectively. For detection of EGFR-sensitizing variants, the cobas several tests have multiple clinical validation studies of sufficient quality and the performance characteristics are well characterized with generally high specificity (>96%). For the detection of EGFR-resistance variants, fewer studies are available and estimates of specificity are more variable. For the detection of less prevalent driver mutations, such as ALK and ROS1 translocations and BRAF variants, few publications are available and, in these publications, only a very few variants have been identified.

Table 2. Results of Clinical Validity Studies of Liquid Biopsy With Tissue Biopsy as the Reference Standard

Study	Initial N	Final N	Excluded Samples	Sensitivity (95% CI)	Specificity (95% CI)
Cobas EGFR test
Jenkins et al (2017)10,
EGFR exon 19 deletion (sensitizing)	710	551	No plasma sample	85 (81 to 89)	98 (95 to 100)
EGFR exon 21 substitution (L858R, sensitizing)				76 (69 to 82)	98 (96 to 99)
EGFR exon 20 (T790M, resistance)	710	551		61 (57 to 66)	79 (70 to 85)
FDA SSED (2016)11,
EGFR-sensitizing variants	601	431	Insufficient plasma; invalid test result	77 (71 to 82)	98 (95 to 99)
Karlovich et al (2016)12,
EGFR-sensitizing variants	174	110	No matching tumor and plasma or inadequate tissue	73 (62 to 83)	100 (86 to 100)
EGFR exon 20 (T790M, resistance)	174	110		64 (45 to 80)	98 (91 to 100)
Thress et al (2015)13,
EGFR exon 19 deletion (sensitizing)	NR	72	Inadequate tumor tissue	82 (63 to 94)	97 (83 to 100)
EGFR exon 21 substitution (L858R, sensitizing)				87 (66 to 97)	97 (85 to 100)
EGFR exon 20 (T790M, resistance)	NR	72		73 (57 to 86)	67 (45 to 84)
Mok et al (2015)14,
EGFR-sensitizing variants	397	238	Insufficient plasma or tissue; invalid test result	75 (65 to 83)	96 (92 to 99)
Weber et al (2014)15,
EGFR-sensitizing and -resistance variants	199a	196	Inadequate tumor tissue	61 (41 to 78)	96 (92 to 99)
Guardant360
Leighl et al (2019)16,	307		No pretreatment ctDNA (4), no tissue genotyping (4), received prohibited treatment (8), metastatic disease not confirmed (4); squamous cell (5)
EGFR exon 19 deletion (sensitizing)		223		81 (60 to 95)c	100 (98 to 100)c
EGFR exon 21 substitution (L858R, sensitizing)		223		90 (56 to 100)c	100 (98 to 100)c
ALK fusion		215		63 (24 to 91)c	100 (98 to 100)c
ROS1 fusion		153		0 (0 to 84)c	100 (98 to 100)c
BRAF v600E		92		100 (16 to 100)c	100 (96 to 100)c
Schwaederle et al (2017)17,
EGFR variants	88	34	No tissue	54 (25 to 81)	90 (70 to 99)
Thompson et al (2016)18,	102	50	Insufficient tissue
EGFR-sensitizing				79 (58 to 93)c	100 (87 to 100)c
EGFR-resistance				50 (7 to 93)c	87 (74 to 95)c
ALK fusion				None identified	None identified
ROS1 fusion				None identified	None identified
BRAF v600E				100 (2.5 to 100)c	100 (93 to 100)c
Villaflor et al (2016)19,	68	31	No tissue
EGFR-sensitizing				63 (24 to 91)c	96 (78 to 100)c
ROS1				None identified	None identified
BRAF V600E				None identified	None identified
OncoBEAM
Ramalingam et al (2018)20,	60	51	Tissue or plasma not available
EGFR exon 19 deletion (sensitizing)				82 (60 to 95)	100 (88 to 100)
EGFR exon 21 substitution (L858R, sensitizing)				63 (41 to 81)	96 (81 to 100)
EGFR exon 20 (T790M, resistance)				100 (40 to 100)	98 (89 to 100)
Karlovich et al (2016)12,
EGFR-sensitizing variants	174	77	No matching tumor and plasma or inadequate tissue	82 (70 to 90)	67 (9 to 99)
EGFR Exon 20 (T790M, resistance)	174	77		73 (58 to 85)	50 (26 to 74)
Thress et al (2015)13,
EGFR exon 19 deletion (sensitizing)	NR	72	Inadequate tumor tissue	82 (63 to 94)	97 (83 to 100)
EGFR exon 21 substitution (L858R, sensitizing)				87 (66 to 97)	97 (85 to 100)
EGFR exon 20 (T790M, resistance)	NR	72		80 (65 to 91)	58 (36 to 78)
GeneStrat
Mellert et al (2017)21,
EGFR exon 19 deletion (sensitizing)				95.9 (NR)	100 (NR)
EGFR exon 21 substitution (L858R, sensitizing)				100 (NR)	100 (NR)
EGFR exon 20 (T790M, resistance)				86.7 (NR)	100 (NR)
ALK fusion				~85 (NR)	100 (NR)
ctDx-Lung
Paweletz et al (2016)22,	NR	48	NR
EGFR exon 19 deletion (sensitizing)				89 (65 to 99)c	100 (88 to 100)c
EGFR exon 21 substitution (L858R, sensitizing)				67 (9 to 99)c	100 (92 to 100)c
ALK fusion				67 (9 to 99)c	100 (92 to 100)c
ROS1 fusion				100 (16 to 100)c	100 (92 to 100)c
BRAF V600E				0 (0 to 98)c	100 (92 to 100)c
InVision
Pritchet et al (2019)23,	264		Missing tissue or ctDNA testing
EGFR exons 18-21)		114		100 (75 to 100)b,c	100 (96 to 100)b,c
ALK/ROS1 fusions		234		40 (5 to 85)b,c	100 (98 to 100)b,c
BRAF V600E		109		100 (48 to 100)b,c	100 (97 to 100)b,c
Remon et al (2019)24,	156		Missing tissue or ctDNA testing
EGFR exons 18-21)		78		88 (47 to 100)	98 (91 to 100)
BRAF V600E		75		50 (1 to 100)	100 (95 to 100)

CI: confidence interval; ctDNA: circulating tumor DNA; EGFR: epidermal growth factor receptor; FDA: Food and Drug Administration; NR: not reported; SSED: Summary of Safety and Effectiveness Data.

^a Unclear how many samples were eligible but not included

^b Only included the subset of patients with at least one mutation detected by liquid biopsy

^c Not reported; calculated based on data provided

The purpose of the limitations tables (see Tables 3 and 4) is to display notable limitations identified in each study. This information is synthesized as a summary of the body of evidence and provides the conclusions on the sufficiency of the evidence supporting the position statement.

Table 3. Relevance Limitations of Clinical Validity Studies of Liquid Biopsy With Tissue Biopsy as the Reference Standard

Study	Populationa	Interventionb	Comparatorc	Outcomesd	Duration of Follow-Upe
Cobas EGFR test
Jenkins et al (2017)10,
FDA SSED (2016)11,	4. Performed in Asia
Karlovich et al (2016)12,
Thress et al (2015)13,
Mok et al (2015)14,	4. Performed in Asia
Weber et al (2014)15,
Guardant360
Leighl et al (2019)16,
Schwaederle et al (2017)17,
Thompson et al (2016)18,
Villaflor et al (2016)19,
OncoBEAM
Ramalingam et al (2018)20,	4. Performed in Asia
Karlovich et al (2016)12,
Thress et al (2015)13,
GeneStrat
Mellert et al (2017)21,	3. Patient characteristics unclear
ctDx-Lung
Paweletz et al (2016)22,	2. Unclear if same as current marketed version
Invision
Pritchet et al (2019)23,
	4: Calculation of performance characteristics only included subset of patients with at least one mutation detected by liquid biopsy
Remon et al (2019)24,

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment.
FDA: Food and Drug Administration; SSED: Summary of Safety and Effectiveness Data.

^a Population key: 1. Intended use population unclear; 2. Clinical context is unclear; 3. Study population is unclear; 4. Study population not representative of intended use.

^bIntervention key: 1. Classification thresholds not defined; 2. Version used unclear; 3. Not intervention of interest.

^c Comparator key: 1. Classification thresholds not defined; 2. Not compared to credible reference standard; 3. Not compared to other tests in use for same purpose.

^d Outcomes key: 1. Study does not directly assess a key health outcome; 2. Evidence chain or decision model not explicated; 3. Key clinical validity outcomes not reported (sensitivity, specificity, and predictive values); 4. Reclassification of diagnostic or risk categories not reported; 5. Adverse events of the test not described (excluding minor discomforts and inconvenience of venipuncture or noninvasive tests).

^e Follow-Up key: 1. Follow-up duration not sufficient with respect to natural history of disease (true-positives, true-negatives, false-positives, false-negatives cannot be determined).

Table 4. Study Design and Conduct Limitations of Clinical Validity Studies of Liquid Biopsy With Tissue Biopsy as the Reference Standard

Study	Selectiona	Blindingb	Delivery of Testc	Selective Reportingd	Data Completenesse	Statisticalf
Cobas EGFR test
Jenkins et al (2017)10,
FDA SSED (2016)11,
Karlovich et al (2016)12,
Thress et al (2015)13,			1. Both samples collected after progression and before next treatment but time between blood and tissue sample collection not described			1. Precision estimates not reported but calculated based on data provided
Mok et al (2015)14,			1. Time between blood and tissue sample collection not described			1. Precision estimates not reported but calculated based on data provided
Weber et al(2014)15,	1,2. Unclear how patients were selected		2. Plasma not collected at time of tissue biopsy			1. Precision estimates not reported but calculated based on data provided
Guardant360
Leighl et al (2019)16,			2.Time between tissue and plasma sample unclear			1. Precision estimates not reported but calculated based on data provided
Schwaederle et al (2017)17,						1. Precision estimates not reported but calculated based on data provided
Thompson et al (2016)18,			1.Time between tissue and blood collection was up to >2 y, median not given			1. Precision estimates not reported but calculated based on data provided
Villaflor et al (2016)19,	1,2. Unclear how patients were selected		1.Time between tissue and blood collection was up 7 y, median 1.4 y			1. Precision estimates not reported but calculated based on data provided
OncoBEAM
Ramalingam et al (2018)20,			1. Time between blood and tissue sample collection not described
Karlovich et al (2016)12,
Thress et al (2015)13,			1. Both samples collected after progression and before next treatment but time between blood and tissue sample collection not described			1. Precision estimates not reported but calculated based on data provided
GeneStrat
Mellert et al (2017)21,	1,2. Unclear how patients were selected		1. Time between blood and tissue sample collection not described			1. Precision estimates not reported cannot be calculated based on data provided
ctDx-Lung
Paweletz et al (2016)22,	1,2. Unclear how patients were selected		1. Time between blood and tissue sample collection not described			1. Precision estimates not reported but calculated based on data provided
InVision
Pritchet et al (2019)23,						1. Precision estimates not reported but calculated based on data provided
Remon et al (2019)24,

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment.
FDA: Food and Drug Administration; SSED: Summary of Safety and Effectiveness Data.

^a Selection key: 1. Selection not described; 2. Selection not random or consecutive (ie, convenience).

^bBlinding key: 1. Not blinded to results of reference or other comparator tests.

^cTest Delivery key: 1. Timing of delivery of index or reference test not described; 2. Timing of index and comparator tests not same; 3. Procedure for interpreting tests not described; 4. Expertise of evaluators not described.

^d Selective Reporting key: 1. Not registered; 2. Evidence of selective reporting; 3. Evidence of selective publication.

^e Data Completeness key: 1. Inadequate description of indeterminate and missing samples; 2. High number of samples excluded; 3. High loss to follow-up or missing data.

^f Statistical key: 1. Confidence intervals and/or p values not reported; 2. Comparison with other tests not reported.

A summary of the previously described published evidence assessing the clinical validity of the specific commercial tests is shown in Table 5. The cobas test has at least 6 studies (n>1500), Guardant360 has at least 4 studies (n>450), OncoBEAM has at least3 studies (n>200), and InVision has at least2 studies (n>400), with the majority being of adequate quality to demonstrate the performance characteristics relative to a tissue test with tight precision estimates for specificity for EGFR TKI-sensitizing variants. Other tests have promising preliminary results but none of the remaining available tests other than the cobas, Guardant360, OncoBEAM and InVision tests have multiple studies of adequate quality to estimate the performance characteristics with sufficient precision for EGFR TKI-sensitizing variants.

Table 5. Summary of Published Evidence^a Assessing the Clinical Validity of Commercial Liquid Biopsy Tests for EGFR TKI-Sensitizing Variants

Test (Method)		Comparison With Tissue Test			Study Quality
	Studies Using Specific Commercial Test (95% CI) and/or Range, %			Available Studies
	Sens		Spec
Roche cobas EGFR Mutation Test v2	75 (69 to 80)b 61-87		97 (95 to 98)b 96-100	6	Very few limitations identified (Jenkins10,; FDA SSED11,; Karlovich12,; Thress13,; Mok14,; Weber15,)
Guardant360 (NGS)	79 (58 to 93) 54-79		100 (87 to 100) 90-100	4	Long time between tissue and ctDNA tests (Thompson18,; Villaflor19,); unclear patient selection (Villaflor19,); Very few limitations with Schwaederle17,
OncoBEAM	63-82		67-100	3	Few limitations identified (Karlovich12,; Thress13,; Rmalingam20,) Only a few negatives in Karlovich for estimating specificity.
Biodesix GeneStrat (ddPCR)	95.9 (NR)21,		100 (NR)21,	1	Patient characteristics and selection unclear; timing of blood and tissue samples unclear; precision estimates not provided (Mellert21,)
Resolution Bio ctDx-Lung	89 (65 to 99)b		100 (88 to 100)c	1	Several limitations identified (Paweletz22,)
Biocept (real-time PCR)	NA		NA	0	NA
Circulogene (Theranostics) liquid biopsy test (NGS)	NA		NA	0	NA
InVIsion (Inivata) (NGS)	88 - 100		98 - 100	2	Few limitations identified (Pritchett23,, Remon24,)

CI: confidence interval; ddPCR: digital droplet polymerase chain reaction; EGFR: epidermal growth factor receptor; FDA: Food and Drug Administration; NA: not applicable; NGS: next-generation sequencing; NR: not reported; PCR: polymerase chain reaction; Sens: sensitivity; Spec: specificity; SSED: Summary of Safety and Effectiveness Data; TKI: tyrosine kinase inhibitor.

^a Meeting selection criteria

^b For EGFR deletion 19.

Section Summary: Clinical Valid

The cobas test has very high accuracy (area under the receiver operating characteristic curve, 0.96), a sensitivity of about 75%, and a specificity above 95% for detection of EGFR TKI-sensitizing variants using tissue biopsy as the reference standard; these estimates are consistent across several studies performed using the test. The studies were performed in Asia, Europe, Australia, and the U.S., primarily in patients with advanced disease of adenocarcinoma histology. The Guardant360 test has 3 studies using tissue biopsy as the reference standard performed in the U. S. in the intended-use population. Estimates of specificity are consistently 90% or higher. Likewise, the OncoBEAM test has 3 studies using tissue biopsy in Asia, Europe, Australia, and the U. S. in the intended-use population, 2 of which provide precise estimates for specificity that are very high (>95%). The InVision test has 2 studies using tissue biopsy as the reference standard in the U. S. and France in the intended-use population, both provide precise estimates for specificity (>95%).

For tests other than the cobas test, Guardant360, OncoBEAM, and InVision for detecting EGFR TKI-sensitizing variants, few studies were identified that evaluated the clinical validity of these commercially available tests for EGFR variants in NSCLC.

For tests of other, less prevalent, variants, such as ALK and ROS1 translocations and BRAF V600E variants, few studies were identified that evaluated the clinical validity of any commercially available tests, and in these studies, very few variants were detected; therefore, performance characteristics are not well-characterized.

Fewer studies have examined the performance of liquid biopsy for the detection of T790M variants associated with EGFR TKI resistance and several different tests were used in the studies. Detection of these variants is potentially important for liquid biopsy because this variant is of interest after the initiation of treatment, when biopsies may be more difficult to obtain.Unlike the high specificities compared with tissue biopsy demonstrated for EGFR variants associated with TKI sensitivity, the moderate specificity means that liquid biopsy often detects T790M variants when they are not detected in tissue biopsy. Sacher et al (2016) suggested that these false-positives might represent tumor heterogeneity in the setting of treatment resistance, such that the T790M status of the biopsied site might not represent all tumors in the patient.^25,

Clinically Useful

A test is clinically useful if the use of the results informs management decisions that improve the net health outcome of care. The net health outcome can be improved if patients receive correct therapy, or more effective therapy, or avoid unnecessary therapy, or avoid unnecessary testing.

Direct Evidence

Direct evidence of clinical utility is provided by studies that have compared health outcomes for patients managed with and without the test. Because these are intervention studies, the preferred evidence would be from RCTs.

No RCTs comparing management with and without liquid biopsy were identified.

Evidence on the ability of liquid biopsy to predict treatment response similar to, or better than, a tissue biopsy is also of interest. If the two tests are highly correlated, they are likely to stratify treatment response similarly overall. To understand the implications of "false-positive" and "false-negative" liquid biopsies for outcomes, patients who have discordant results on liquid biopsy and standard biopsy are of particular interest. If patients who are negative for EGFR-sensitizing or -resistance variants on liquid biopsies but positive for those variants on standard biopsies respond to EGFR TKIs (ie, erlotinib, gefitinib, afatinib, osimertinib), it would suggest that the standard biopsy was correct and the liquid biopsy results were truly false-negatives. If patients with positive liquid biopsies and negative tissue biopsies for EGFR variants respond to EGFR TKIs, it would suggest that the positive liquid biopsies were correct rather than false-positives.

Chain of Evidence

Indirect evidence on clinical utility rests on clinical validity. If the evidence is insufficient to demonstrate test performance, no inferences can be made about clinical utility.

The clinical utility might alternatively be established based on a chain of evidence. Assuming that tissue biomarkers are the standard by which treatment decisions are made, an agreement between liquid and tissue biopsies would infer that treatment selection based on liquid or tissue biopsies is likely to yield similar outcomes. Also, a liquid biopsy would reduce the number of patients undergoing tissue sampling and any accompanying morbidity.

Depending on the analytic method, compared with a tissue biopsy, liquid biopsy appears somewhat less sensitive with generally high specificity in detecting an EGFR TKI-sensitizing variant that can predict outcomes. This finding suggests that an EGFR TKI-sensitizing variant identified by liquid biopsy could be used to select a treatment with reflex to tissue biopsy. However, evidence directly demonstrating the predictive ability of liquid biopsy would be most convincing. Also, outcomes in patients who have discordant results on liquid and tissue biopsy are of particular interest.

Therefore, BCBSA also considered evidence on the ability of liquid biopsy to predict treatment response. Liquid biopsy could improve patient outcomes if it predicts treatment response similar to, or better than, tissue biopsy. Treatment response as measured by OS outcomes would be most informative. PFS can be difficult to interpret because of confounding influences in retrospective observational subgroup analyses. Response rate may be more informative than PFS.

Some studies were nested in nonrandomized designs or RCTs. This structure potentially permits comparing associations between liquid biopsy and tissue biopsy results with outcomes. Because it has already been demonstrated by the prior studies that liquid biopsy and tissue biopsy are moderately correlated, they should both be associated with either prognosis of disease or prediction of treatment response as has been demonstrated for tissue biopsy. However, if liquid biopsy results are more strongly associated with outcomes, it might be considered better than tissue biopsy (considered the reference standard). Although liquid biopsy had a high specificity for EGFR-sensitizing variants (>90%) in almost all studies, false-positives could be a concern in patient populations with a low prevalence of treatable variants. Known variability of tumor tissue sampling raises concern whether false-positive liquid biopsies represent cases in which the tissue biopsy is falsely negative.

Sufficient numbers of patients have not been studied in which all possible combinations of liquid biopsy and tissue biopsy results have been analyzed for associations with patient outcomes. Available patient outcomes data for studies evaluating EGFR TKI-sensitizing and EGFR TKI-resistance variants are shown in Tables 6 and 7, respectively.

Table 6. EGFR TKI-Sensitizing Variants: Treatment Response Stratified by Liquid and Tissue Biopsy

Study/Patient Group	Country	Disease Stage	Technology Used to Detect ctDNA	Sample Sizes	Treatment Response
				n	Outcomes				p
Zhang et al (2017)26,; EGFR-positive and -negative patients treated with EGFR TKIs	China	IIIB, IV	ddPCR	PFS (95% CI), d
				n	Erlotinib		Chemotherapy		p
				Tissue positive vs tissue negative
				215	342 (291 to 393)		60 (0 to 124)		215
				Tissue positive and liquid positive vs liquid negative
				80	334 (298 to 371)		420 (100 to 740)		80
				Tissue negative and liquid positive
				3	133, 410, and 1153
FDA SSED (2016)11,; phase 3 ENSURE RCT in tissue EGFR-positivea	China, Malaysia, Philippines	IIIB, IV	cobas	PFS HR (95% CI) for Chemotherapy vs Erlotinib
				Overall (ie, tissue positive)
				179	0.33 (0.23 to 0.47)
				Patients with positive tissue and liquid
				137	0.29 (0.19 to 0.45)
				Patients with positive tissue and negative liquid
				42	0.37 (0.15 to 0.90)
Karachaliou et al (2015)27,; EURTAC trial in tissue EGFR-positivea	France, Italy, Spain	IIIB, IV	Multiplex 5´ nuclease rt-PCR (TaqMan)	OS (95% CI) for Erlotinib vs Chemotherapy, mo
				n	Erlotinib	Chemotherapy			p
				Overall (ie, tissue positive)
				97	25.8 (17.7 to 31.9)	18.1 (15.0 to 23.5)			0.14
				All patients with exon 19 deletion in tissue
				56	30.4 (19.8 to 55.7)	18.9 (10.4 to 36.2)			0.22
				Patients with exon 19 deletion in both tissue and ctDNA
				47	34.4 (22.9 to NR)	19.9 (9.8 to 36.2)			0.23
				Patients with exon 19 deletion in tissue but not ctDNA
				9	13.0 (8.9 to 19.8)	15.5 (0.3 to NR)			0.87
				All patients with L858R variant in tissue
				41	17.7 (6.3 to 26.8)	17.5 (8.2 to 23.5)			0.67
				Patients with L858R variant in both tissue and inctDNA
				29	13.7 (2.6 to 21.9)	12.6 (7.1 to 23.5)			0.67
				Patients with L858R variant in tissue but not in ctDNA
				12	29.4 (8.6 to 63.0)	25.6 (16.1 to NR)			0.64

CI: confidence interval; ctDNA: circulating tumor DNA; ddPCR: droplet digital polymerase chain reaction; EGFR: epidermal growth factor receptor; FDA: Food and Drug Administration; HR: hazard ratio; NR: not reported; OS: overall survival; PFS, progression-free survival; RCT: randomized controlled trial; rt-PCR: real-time polymerase chain reaction; SSED: Summary of Safety and Effectiveness; TKI: tyrosine kinase inhibitor.

^a Exon 19 deletion or L858R variant.

In Table 6 (sensitizing variants), the SSED document supporting the approval of the cobas EGFR Mutation Test v2 reported clinical outcome data derived from a randomized phase 3 trial of erlotinib vs gemcitabine plus cisplatin as first-line treatment of NSCLC.^11, However, only patients with EGFR variants detected from tissue biopsies were enrolled. In the overall study, erlotinib showed substantial improvement in PFS over chemotherapy (HR=0.33; 95% CI, 0.23 to 0.47), consistent with the known efficacy of erlotinib in patients with a sensitizing EGFR variant. Among the subset of patients with positive liquid biopsy results (77% [137/179]), erlotinib showed a similar improvement in PFS (HR=0.29; 95% CI, 0.19 to 0.45). However, the finding has limited meaning because all patients had positive tissue biopsies, thus showing a similar result. Those with negative liquid biopsies (n=42) also showed a similar magnitude of benefit of erlotinib (HR=0.37; 95% CI, 0.15 to 0.90), which would be consistent with liquid biopsies being false-negatives.

In Zhang et al (2017), PFS in the subset of patients treated with EGFR TKIs (114/215) was compared for groups of patients with biomarker status determined by tissue biopsy and by liquid biopsy.^26, The patients were primarily treated with gefitinib (n=94); 18 patients received erlotinib, 1 received icotinib, and 1 received afatinib. When patients were stratified by tissue biopsy EGFR status, PFS for EGFR-positive subjects was 342 days vs 60 days for EGFR-negative subjects (p<0.001). Among the tissue biopsy-positive patients, there was no difference in PFS between those with positive (334 days) and negative liquid biopsies (420 days), consistent with the liquid biopsies being false-negatives. Three patients were tissue biopsy-negative, but liquid biopsy-positive; they had PFS with TKI treatment of 133, 410, and 1153 days, respectively. Although the numbers are small, the PFS values are consistent with a response to TKIs and might represent tissue biopsies that did not reflect the correct EGFR status.

Table 7. EGFR TKI-Resistance Variants: Treatment Response Stratified by Liquid and Tissue Biopsy

Study/Patient Group	Country	Disease Stage	Technology Used to Detect ctDNA	Treatment Response
				n	Outcomes
Oxnard et al (2016)28,; AURA phase 1 trial of patients who progressed on EGFR TKI	Multinationalb	Advanced	BEAMing	ORR (95% CI) (Osimertinib)
				Liquid positive, tissue positive
				108	64% (54% to 73%)
				Liquid positive, tissue negative
				18	28% (10% to 53%)
				Liquid negative, tissue positive
				45	69% (53% to 82%)
				Liquid negative, tissue negative
				40	25% (13% to 41%)
				PFS (95% CI), mo
				Liquid positive, tissue positive
				111	9.3 (8.3 to 10.9)
				Liquid positive, tissue negative
				18	4.2 (1.3 to 5.6)
				Liquid negative, tissue positive
				47	16.5 (10.9 to NC)
				Liquid negative, tissue negative
				40	2.8 (1.4 to 4.2)
Thress et al (2015)13,; phase 1 AURA RCT in tissue EGFR-positivea with progression on EGFR TKI	Multinationalb	Advanced	cobas; BEAMing ddPCR	ORR (Osimertinib)
				Tissue positive vs tissue negative
				65	61% vs 29%
				Liquid positive vs liquid negative
				72	59% vs 35%
				Liquid positive, tissue biopsy negative
				8	38%
Karlovich et al (2016)12,; patients from observational study and a phase 1 dose-escalation part and a phase 2 study of roceiletinib	U.S., Australia, France, Poland	Advanced	BEAMing	ORR (95% CI) (Rociletinib)
				Liquid positive, tissue positive
				15	73 (51 to 96)
				Liquid positive, tissue negative
				4	25 (0 to 67)
				Liquid negative, tissue positive
				6	50 (10 to 90)
				Liquid negative, tissue negative
				3	33 (0 to 87)
Helman et al (2018)29,; patients who were tissue EGFR T790M-positive from the TIGER-X and TIGER-2 studies of roceiletinib	U.S.	Advanced or metastatic	Guardant	ORR (95% CI) (Rociletinib)
				Tissue positive
				77	29.9% (20.0 to 41.4)
				Liquid positive
				63	28.6% (17.9 to 41.3)
				PFS (95% CI), mo
				Tissue positive
				77	4.2 (3.9 to 5.7)
				Liquid positive
				63	4.1 (3.9 to5.6)

BEAM: beads, emulsions, amplification, and magnetics; CI: confidence interval; ctDNA: circulating tumor DNA; ddPCR:droplet digital polymerase chain reaction; EGFR: epidermal growth factor receptor; NC: not calculable; ORR: objective response rate; PFS: progression-free survival; RCT: randomized controlled trial; TKI: tyrosine kinase inhibitor.

^a Exon 19 deletion or L858R variant.
^b U.S, Australia, France, Germany, Italy, Japan, Korea, Spain, Taiwan, U.K.

For EGFR-resistance variants, Thress et al (2015) examined the response to the experimental therapeutic AZD9291 (osimertinib) by T790M status, determined using a tissue or liquid biopsy (see Table 7).^13, Patients were not selected for treatment based on T790M status, and there was only moderate concordance between tissue and liquid biopsies. Response rates by tissue biopsy variant identification (61% for positive variants vs 29% for negative variants) were qualitatively similar to the response rates by liquid biopsy variant identification (59% for positive variants vs 35% for negative variants). Formal statistical testing was not presented. However, the authors did report response rates for patients who had positive liquid biopsies but negative tissue biopsies. In these 8 patients, the pooled response rate was 38%. The number of patients is too small to make definitive conclusions but the response rate in these patients is closer to those for patients with negative variants than with positive variants. A source of additional uncertainty in these data is that the therapeutic responses to this experimental agent have not yet been well characterized.

Oxnard et al (2016) compared outcomes by T790M status for liquid biopsy and tissue biopsy in patients enrolled in the escalation and expansion cohorts of the phase 1 AURA study of osimertinib for advanced EGFR-variant NSCLC.^28, Some patients may have overlapped with the Thress et al (2015) study.^13, Among patients with T790M-negative ctDNA, objective response rate (ORR) was higher in 45 patients with T790M-positive tissue (69%; 95% CI, 53% to 82%) than in 40 patients with T790M-negative tissue (25%; 95% CI, 13% to 41%; p=0.001), as was median PFS (16.5 months vs 2.8 months; p=0.001), which is consistent with false-negative ctDNA results. Among patients with T790M-positive ctDNA, ORR and median PFS were higher in 108 patients with T790M-positive tissue (ORR=64%; 95% CI, 54% to 73%; PFS=9.3 months) than in 18 patients with T790M-negative tissue (ORR=28%; 95% CI, 10% to 53%; p=0.004; PFS=4.2 months; p=0.0002) which is consistent with false-positive ctDNA results. The authors concluded that a T790M-variant ctDNA assay could be used for osimertinib treatment decisions in patients with acquired EGFR TKI resistance and would permit avoiding tissue biopsy for patients with T790M-positive ctDNA results.

Karlovich et al (2016) compared outcomes by T790M status for liquid biopsy and tissue biopsy in patients enrolled in the TIGER-X phase 1/2 clinical trial of rociletinib and an observational study in patients with advanced NSCLC.^12, Rociletinib was an EGFR inhibitor in development for the treatment of patients with EGFR T790M-mutated NSCLC but the application for regulatory approval was withdrawn in 2016. The ORR was provided by cross-categories of results of tissue and ctDNA testing (see Table 8). Although CIs overlapped substantially and sample sizes in the cross-categories were small, the ORR was quantitatively largest in patients positive for T790M in both tissue and ctDNA and smaller in patients who were T790M negative in tissue regardless of ctDNA positivity.

Helman et al (2018) compared outcomes in patients with positive T790M status for liquid biopsy and tissue biopsy in patients enrolled in the TIGER-X and TIGER-2 trials of rociletinib.^29, The ORR and PFS were provided for patients who were tissue positive and for patients who were liquid positive (see Table 9). Both ORR and PFS were similar for the 77 patients who were identified as positive for T790M by tissue biopsy and the 63 patients identified as positive by ctDNA. Thus, 63 of 77 patients (81.8%) who had been identified as positive by tissue biopsy were also identified as positive by liquid biopsy, and this did not affect outcomes for treatment with rociletinib. As noted above, the application for regulatory approval of rociletinib was withdrawn, limiting interpretation of the effect of rociletinib.

Merker et al (2018) reported a joint review on circulating tumor DNA for the American Society of Clinical Oncology and College of American Pathologists.^30, The review was not specific to lung cancer but did make the following statements regarding the clinical utility of ctDNA testing for lung cancer:

• "At present, one PCR-based ctDNA assay for the detection of EGFR variants in patients with NSCLC has received regulatory approval in the United States and Europe, and PCR-based ctDNA assays for EGFR in NSCLC and KRAS in colorectal cancer are available for commercial use in Europe. These assays have demonstrated clinical validity, but the clinical utility in this setting is based on retrospective analyses."
• "Evidence demonstrated that, although positive EGFR testing results may effectively be used to guide therapy, undetected results should be confirmed with analysis of a tissue sample, if possible. Cases in which the variant is not detected in the ctDNA but is detected in the tissue sample are relatively common, so undetected ctDNA assay results should be confirmed in tumor tissue testing."
• "The challenges of demonstrating clinical utility are illustrated in NSCLC. A major potential issue is that the patient population selected for study inclusion may not be representative of those targeted for the intended clinical use of the ctDNA assay.."

The diagnostic characteristics of liquid biopsy for the detection of T790M variants associated with EGFR TKI-inhibitor resistance, an indication for treatment with osimertinib, has shown that liquid biopsy is moderately sensitive and moderately specific and thus overall concordance is moderate. Using tissue testing of negative liquid biopsies would increase sensitivity, but because liquid biopsy is not highly specific, it would result in false-positives. Because not enough data are available to determine whether these false-positives represent a faulty tissue reference standard or are correctly labeled as false-positives, outcomes for these patients are uncertain. In 1 study, 8 patients with negative tissue biopsies but positive liquid biopsies had low response rates consistent with those with negative tissue biopsies; and in the AURA study, 18 patients with liquid-positive, tissue-negative results had a low response rate, also consistent with negative tissue biopsy. In the TIGER-X study, three patients who were liquid-positive, tissue-negative had low response rates to rociletinib, similar to the other tissue-negative patients. However, although there is higher discordance in the liquid vs tissue results for the resistance variant, retrospective analyses have suggested that patients positive for T790M in liquid biopsy have outcomes with osimertinib that appear to be similar overall to patients positive by a tissue-based assay.

Section Summary: Clinically Useful

There is little evidence on the comparative validity of tissue and liquid biopsies in discordant cases for EGFR TKI-sensitizing variants. Based on the apparent response to EGFR TKIs in patients with negative liquid biopsies and positive tissue biopsies in the FDA approval study, these results are consistent with false-negative liquid biopsies. It is unclear whether false-positive liquid biopsies represent errors in the liquid biopsy or inadequacies of a tissue biopsy reference standard. In one study, three patients with negative tissue biopsies but positive liquid biopsies for biomarkers indicating EGFR TKI sensitivity had apparent responses to EGFR TKIs, consistent with the tissue biopsies being incorrectly negative.

A chain of evidence based on the sensitivity and specificity of liquid biopsy for the detection of EGFR TKI-sensitizing variants for tests with established clinical validity such as the cobas EGFR Mutation Test v2, Guardant360, or OncoBEAM, can support its utility. The body of evidence has demonstrated sensitivity generally between 60% and 80%, with high specificities (>95%). If a liquid biopsy is used to detect EGFR TKI-sensitizing variants with reflex testing of tissue samples in those with negative liquid biopsies, then the sensitivity of the testing strategy will be equivalent to tissue biopsy, and the specificity will be high. Therefore, outcomes should be similar, but tissue testing of biomarkers would be avoided in approximately two-thirds to three-quarters of patients with EGFR TKI-sensitizing variants.

For the other marketed tests that include detection of EGFR TKI-sensitizing variants and for liquid biopsy testing of other driver mutations, sufficient evidence of clinical validity is lacking, and thus a chain of evidence cannot be linked to support a conclusion that results for other ctDNA test methods will be similar to those for tissue biopsy.

For EGFR TKI-resistance variants, there is little evidence on the comparative validity of tissue and liquid biopsies in discordant cases. Based on the apparent response to osimertinib from the AURA study with liquid-negative, tissue-positive results, these results are more consistent with false-negative liquid biopsies. It is unclear whether false-positive liquid biopsies represent errors in the liquid biopsy or inadequacies of a tissue biopsy reference standard. In 3 studies, patients with negative tissue biopsies and positive liquid biopsies appeared not to have a high response to osimertinib or rociletinib..Sample sizes are very small for both scenarios of discordance. Although the evidence is limited, the College of American Pathologists , the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology published joint guidelines endorsed by American Society of Clinical Oncology with an expert consensus opinion that "Physicians may use plasma cfDNA methods to identify EGFR T790M mutations in lung adenocarcinoma patients with progression or secondary clinical resistance to EGFR targeted TKIs; testing of the tumor sample is recommended if the plasma result is negative." The National Comprehensive Cancer Network guidelines also state that at progression on erlotinib, afatinib, gefitinib or dacomitinib when testing for the T790M resistance variant, plasma-based testing should be considered and when plasma-based testing is negative tissue-based testing is strongly recommended.

For tests of other, less prevalent, variants, such as ALK and ROS1 translocations and BRAF V600E variants, few studies were identified that evaluated the clinical validity of any commercially available tests and in these studies, very few variants were detected; therefore, performance characteristics are not well characterized. Because sufficient evidence of clinical validity is lacking, a chain of evidence cannot be linked to support the conclusion that results for other variants using ctDNA test methods will be similar to those for tissue biopsy.

Summary of Evidence

For individuals with advanced NSCLC who receive testing for biomarkers of EGFR TKIs sensitivity using ctDNA with the cobas EGFR Mutation Test v2 (liquid biopsy), the evidence includes numerous studies assessing the diagnostic characteristics of liquid biopsy compared with tissue. The relevant outcomes are OS, disease-specific survival (DSS), and test validity. Current evidence does not permit determining whether cobas or tissue biopsy is more strongly associated with patient outcomes or treatment response. BCBSA identified no RCTs providing evidence of the clinical utility of cobas. The cobas EGFR Mutation Test has adequate evidence of clinical validity for the EGFR TKI-sensitizing variants. The Food and Drug Administration has suggested that a strategy of liquid biopsy followed by referral (reflex) tissue biopsy of negative liquid biopsies for the cobas test would result in an overall diagnostic performance equivalent to tissue biopsy. Several additional studies of the clinical validity of cobas have shown it to be moderately sensitive and highly specific compared with a reference standard of tissue biopsy. A chain of evidence demonstrates that the reflex testing strategy with the cobas test should produce outcomes similar to tissue testing while avoiding tissue testing in approximately two-thirds of patients with EGFR TKI-sensitizing variants. Patients who cannot undergo tissue biopsy would likely otherwise receive chemotherapy. The cobas test can identify patients for whom there is a net benefit of targeted therapy vs chemotherapy with high specificity. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

For individuals with advanced NSCLC who receive testing for biomarkers of EGFR TKI sensitivity using ctDNA (liquid biopsy) with the Guardant360, OncoBEAM or InVision tests, the evidence includes several studies assessing the diagnostic characteristics of liquid biopsy compared with tissue. The relevant outcomes are OS, DSS, and test validity. Current evidence does not permit determining whether liquid or tissue biopsy is more strongly associated with patient outcomes or treatment response. BCBSA identified no RCTs providing evidence of the clinical utility of these tests. The Guardant360, OncoBEAM, and InVision tests have adequate evidence of clinical validity for the EGFR TKI-sensitizing variants. A strategy of liquid biopsy followed by referral (reflex) tissue biopsy of negative liquid biopsies for the tests would result in an overall diagnostic performance similar to tissue biopsy. A chain of evidence demonstrates that the reflex testing strategy with the Guardant360, OncoBEAM or InVision tests should produce outcomes similar to tissue testing while avoiding tissue testing in approximately two-thirds of patients with EGFR TKI-sensitizing variants. Patients who cannot undergo tissue biopsy would likely otherwise receive chemotherapy. These tests can identify patients for whom there is a net benefit of targeted therapy vs chemotherapy with high specificity. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

For individuals with advanced NSCLC who receive testing for biomarkers of EGFR TKI sensitivity using ctDNA with tests other than the cobas EGFR Mutation Test v2, Guardant360, OncoBEAM or InVision tests, the evidence includes studies assessing the diagnostic characteristics of liquid biopsy compared with tissue reference standard. The relevant outcomes are OS, DSS, and test validity. Given the breadth of molecular diagnostic methodologies available to assess ctDNA, the clinical validity of each commercially available test must be established independently. None of the commercially available tests other than the cobas, Guardant360, OncoBEAM and InVision tests have multiple studies of adequate quality to estimate the performance characteristics with sufficient precision. Current evidence does not permit determining whether aliquid biopsy or tissue biopsy is more strongly associated with patient outcomes or treatment response. BCBSA found no RCTs providing evidence of the clinical utility of those methods of liquid biopsy. The evidence is insufficient to determine the effects of the technology on health outcomes.

For individuals with advanced NSCLC who receive testing for biomarkers other than EGFR using a liquid biopsy to select a targeted therapy, the evidence includes studies assessing the diagnostic characteristics of liquid biopsy compared with the tissue biopsy reference standard. The relevant outcomes are OSDSS, and test validity. Given the breadth of molecular diagnostic methodologies available to assess ctDNA, the clinical validity of each commercially available test must be established independently. None of the commercially available tests have multiple studies of adequate quality to estimate the performance characteristics with sufficient precision for variants other than EGFR. We found no RCTs providing evidence of the clinical utility of those of methods of liquid biopsy. The evidence is insufficient to determine the effects of the technology on health outcomes.

For individuals with advanced NSCLC who progressed on EGFR TKIs who receive testing for biomarkers of EGFR TKI resistance using liquid biopsy, the evidence includes studies assessing the diagnostic characteristics of liquid biopsy. The relevant outcomes are OS, DSS, and test validity. For variants that indicate EGFR TKI resistance and suitability for alternative treatments with osimertinib, liquid biopsy is moderately sensitive and moderately specific compared with a reference standard of tissue biopsy. Given the moderate clinical sensitivity and specificity of liquid biopsy, using liquid biopsy alone or in combination with tissue biopsy might result in the selection of different patients testing positive for EGFR TKI resistance. It cannot be determined whether patient outcomes are improved. However, although there is higher discordance in the liquid vs tissue results for the resistance variant, retrospective analyses have suggested that patients positive for T790M in liquid biopsy have outcomes with osimertinib that appear to be similar overall to patients positive by a tissue-based assay. The evidence is insufficient to determine the effects of the technology on health outcomes. Although the evidence is limited, the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology published joint guidelines endorsed by American Society of Clinical Oncologywith an expert consensus opinion that physicians may use liquid biopsy (cell-free DNA) to identify EGFR T790M variants in patients with progression or resistance to EGFR targeted TKIs and that testing of the tumor sample is recommended if the liquid biopsy result is negative. Similarly, the National Comprehensive Cancer Network guidelines also state that at progression on erlotinib, afatinib, gefitinib or dacomitinib when testing for the T790M resistance variant, liquid biopsy should be considered and when a liquid biopsy is negative tissue-based testing is strongly recommended.

SUPPLEMENTAL INFORMATION

Practice Guidelines and Position Statements

National Comprehensive Cancer Network

National Comprehensive Cancer Network guidelines (v.3.2019) discuss the role of liquid biopsy in the management of non-small-cell lung cancer.^3, The guidelines state that cell-free/circulating tumor DNA testing should not be used in lieu of tissue diagnosis. They also state that cfDNA testing can be used if the patient is not medically fit for tissue sample or there is insufficient tissue for molecular analysis and follow-up with tissue-based analysis will be done if plasma-based analysis is negative. The guidelines also state that at progression on erlotinib, afatinib, gefitinib or dacomitinib when testing for T790M, plasma-based testing should be considered and when plasma-based testing is negative tissue-based testing is strongly recommended.

International Association for the Study of Lung Cancer

The International Association for the Study of Lung Cancer (2018) published a statement paper on liquid biopsy for advanced non-small-cell lung cancer.^31, The work preparing the statement was supported by unrestricted grants from Guardant Health, Astra Zeneca, Biocept, and Roche. The statement made the following recommendations:

• "The criteria used to select treatment-naive patients for molecular testing of ctDNA [circulating tumor DNA] is the same used for molecular testing using DNA isolated from tissue."
• "Liquid biopsy can be considered at the time of initial diagnosis in all patients who need tumor molecular profiling, but it is particularly recommended when tumor tissue is scarce, unavailable, or a significant delay potentially greater than 2 weeks is expected in obtaining tumor tissue."

• Cobas EGFR MutationTest v2.
• droplet digital polymerase chain reaction next-generation sequencing panels
• Multiplex panels using next-generation sequencing platforms could be considered to detect EGFR, ALK, ROS1, or BRAF variants and a positive result would be adequate to initiate first-line therapy.

College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology

The College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology (2018) published a guideline on molecular testing for the selection of lung cancer patients for treatment with targeted tyrosine kinase inhibitors.^30,The American Society of Clinical Oncology also endorsed the joint College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology guidelines with minor modifications.^32,

The guidelines noted the following recommendation regarding liquid biopsy for activating EGFR mutations and a consensus opinion regarding liquid biopsy for the T790M resistance mutation.

• Recommendation: "In some clinical settings in which tissue is limited and/or insufficient for molecular testing, physicians may use a cfDNA assay to identify [activating] EGFR mutations."
• Expert Consensus Opinion: "Physicians may use plasma cfDNA methods to identify EGFR T790M mutations in lung adenocarcinoma patients with progression or secondary clinical resistance to EGFR targeted TKIs; testing of the tumor sample is recommended if the plasma result is negative."
• No recommendation: "There is currently insufficient evidence to support the use of circulating tumor cell molecular analysis for the diagnosis of primary lung adenocarcinoma, the identification of EGFR or other mutations, or the identification of EGFR T790M mutations at the time of EGFR TKI resistance."

The National Insititute for Health and Care Excellence (2018) issued an innovation briefing on plasma EGFR mutation tests for adults with locally advanced or metastatic NSCLC.^33, The briefing reviewed seven ctDNA tests available in Europe and concluded:

• "The intended place in therapy would be as an alternative to tissue EGFR testing or before tumour testing to inform decisions about prescribing EGFR-TKIs. Plasma testing may be particularly useful for people whose disease has developed resistance to an EGFR-TKI and who could be offered second-line EGFR-TKIs, if appropriate, without having further tissue testing."
• "The main points from the evidence summarised in this briefing are from 7 non-UK-based prospective studies with 2,106 adults. They show that the diagnostic accuracy of plasma EGFR mutation testing has a similar specificity, but lower sensitivity, compared with tissue EGFR mutation testing in adults with locally advanced or metastatic NSCLC."
• "Key uncertainties around the evidence or technology are that tests for identifying EGFR-TK mutations are rapidly evolving and there is no established gold-standard test against which to evaluate them."

Not applicable.

Ongoing and Unpublished Clinical Trials

Some currently ongoing and unpublished trials that might influence this review are listed in Table 8.

Table 8. Summary of Key Trials

NCT No.	Trial Name	Planned Enrollment	Completion Date
Ongoing
NCT03116633a	An Observational Multicenter Study to Evaluate the Performance and Utility of Inivata Liquid Biopsy Analysis Compared With Tissue Biopsy Analysis for Detection of Genomic Alterations in Patients With Lung Cancer	260	Oct 2018 (last updated Apr 2018)
NCT02906852a	Prospective Observational Study to Evaluate the Performance of Inivata Liquid Biopsy Analysis Compared With Standard Tissue Biopsy Analysis for Detection of Genomic Alterations in Patients With Advanced Non-small Cell Lung Cancer	530	Nov 2018 (last updated Apr 2018)
NCT01930474	Analysis of plasma tumor DNA in lung cancer patients	200	Dec 2018 (unknown)
NCT02140463	Next generation personalized therapy with plasma DNA Trial 2 in refractory solid tumors (The NEXT-2 Trial)	260	Dec 2018 (last updated Jan 2018)
NCT02894853a	Lung Cancer Early Molecular Assessment Trial	1297	Dec 2019
NCT02160366	Profile Related Evidence to Determine Individualized Cancer Therapy (PREDICT) Program in Advanced Cancer Patients	2000	Sep 2019
Unpublished
NCT02620527a	Study of Concordance Between Circulating Tumor DNA Assay and Foundation One Tissue Analysis For Genomic Alterations	1400	Dec 2017 (completed)
NCT02418234	T790M Mutation on ctDNA in patients with NSCLC after EGFR-TKI failure	314	Nov 2017 (completed)
NCT02284633a	Blood sample monitoring of patients with EGFR mutated lung cancer	250	Dec 2018

NCT: national clinical trial.
^a Denotes industry-sponsored or cosponsored trial.]
________________________________________________________________________________________

Horizon BCBSNJ Medical Policy Development Process:

This Horizon BCBSNJ Medical Policy (the “Medical Policy”) has been developed by Horizon BCBSNJ’s Medical Policy Committee (the “Committee”) consistent with generally accepted standards of medical practice, and reflects Horizon BCBSNJ’s view of the subject health care services, supplies or procedures, and in what circumstances they are deemed to be medically necessary or experimental/ investigational in nature. This Medical Policy also considers whether and to what degree the subject health care services, supplies or procedures are clinically appropriate, in terms of type, frequency, extent, site and duration and if they are considered effective for the illnesses, injuries or diseases discussed. Where relevant, this Medical Policy considers whether the subject health care services, supplies or procedures are being requested primarily for the convenience of the covered person or the health care provider. It may also consider whether the services, supplies or procedures are more costly than an alternative service or sequence of services, supplies or procedures that are at least as likely to produce equivalent therapeutic or diagnostic results as to the diagnosis or treatment of the relevant illness, injury or disease. In reaching its conclusion regarding what it considers to be the generally accepted standards of medical practice, the Committee reviews and considers the following: all credible scientific evidence published in peer-reviewed medical literature generally recognized by the relevant medical community, physician and health care provider specialty society recommendations, the views of physicians and health care providers practicing in relevant clinical areas (including, but not limited to, the prevailing opinion within the appropriate specialty) and any other relevant factor as determined by applicable State and Federal laws and regulations.

___________________________________________________________________________________________________________________________

Index:
Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
Circulating Tumor Cell Testing for Non-Small-Cell-Lung Cancer
Oncotype SEQ
Guardant360 for Non-Small-Cell-Lung Cancer
Circulogene Theranostics' Liquid Biopsy
Liquid Biopsy for Non-Small-Cell-Lung Cancer
CancerIntercept
FoundationACT
GeneStrat
cobas® EGFR Mutation Test v2
OncoBEAM
ALK Testing, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
BRAF V600E Testing, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
ROS1 Testing, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
KRAS Testing, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
HER2, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
RET, Circulating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
METCirculating Tumor DNA for Management of Non-Small-Cell-Lung Cancer (Liquid Biopsy)
InVisionFirst-Lung Test

References:
1. Alix-Panabieres C, Pantel K. Clinical applications of circulating tumor cells and circulating tumor DNA as liquid biopsy. Cancer Discov. May 2016;6(5):479-491. PMID 26969689

2. Food and Drug Administration (FDA). cobas EGFR Mutation Test v2 (P150047). 2016; https://www.accessdata.fda.gov/cdrh_docs/pdf15/P150047a.pdf. Accessed April 11, 2019.

3. National Comprehensive Cancer Network (NCCN). NCCN Clinical Practice Guidelines in Oncology: Non-small Cell Lung Cancer. Version 6.2019. https://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf. Accessed Aug 29, 2019.

4. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. Aug 4 2011;365(5):395-409. PMID 21714641

5. Lee JK, Hahn S, Kim DW, et al. Epidermal growth factor receptor tyrosine kinase inhibitors vs conventional chemotherapy in non-small cell lung cancer harboring wild-type epidermal growth factor receptor: a meta- analysis. Jama. Apr 9 2014;311(14):1430-1437. PMID 24715074

6. Janne PA, Yang JC, Kim DW, et al. AZD9291 in EGFR inhibitor-resistant non-small-cell lung cancer. N Engl J Med. Apr 30 2015;372(18):1689-1699. PMID 25923549

7. Socinski MA, Jotte RM, Cappuzzo F, et al. Atezolizumab for first-line treatment of metastatic nonsquamous NSCLC. N Engl J Med. Jun 14 2018;378(24):2288-2301. PMID 29863955

8. Wu Y, Liu H, Shi X, et al. Can EGFR mutations in plasma or serum be predictive markers of non-small-cell lung cancer? A meta-analysis. Lung Cancer. Jun 2015;88(3):246-253. PMID 25837799

9. Supplee JG, Milan MSD, Lim LP et al. Sensitivity of next-generation sequencing assays detecting oncogenic fusions in plasma cell-free DNA. Lung Cancer, 2019 Jul 20;134:96-99. PMID 31320002

10. Jenkins S, Yang JC, Ramalingam SS, et al. Plasma ctDNA analysis for detection of the EGFR T790M mutation in patients with advanced non-small cell lung cancer. J Thorac Oncol. Jul 2017;12(7):1061-1070. PMID 28428148

11. Food and Drug Administration. Summary of Safety and Effectiveness Data (SSED) cobas EGFR Mutation Test v2. 2016; https://www.accessdata.fda.gov/cdrh_docs/pdf15/P150047b.pdf. Accessed April 11, 2019.

12. Karlovich C, Goldman JW, Sun JM, et al. Assessment of EGFR mutation status in matched plasma and tumor tissue of NSCLC patients from a phase I study of rociletinib (CO-1686). Clin Cancer Res. May 15 2016;22(10):2386-2395. PMID 26747242

13. Thress KS, Brant R, Carr TH, et al. EGFR mutation detection in ctDNA from NSCLC patient plasma: A cross- platform comparison of leading technologies to support the clinical development of AZD9291. Lung Cancer. Dec 2015;90(3):509-515. PMID 26494259

14. Mok T, Wu YL, Lee JS, et al. Detection and dynamic changes of EGFR mutations from circulating tumor dna as a predictor of survival outcomes in NSCLC patients treated with first-line intercalated erlotinib and chemotherapy. Clin Cancer Res. Jul 15 2015;21(14):3196-3203. PMID 25829397

15. Weber B, Meldgaard P, Hager H, et al. Detection of EGFR mutations in plasma and biopsies from non-small cell lung cancer patients by allele-specific PCR assays. BMC Cancer. Apr 28 2014;14:294. PMID 24773774

16. Leighl NB, Page RD, Raymond VM et al. Clinical Utility of Comprehensive Cell-free DNA Analysis to Identify Genomic Biomarkers in Patients with Newly Diagnosed Metastatic Non-small Cell Lung Cancer. Clin. Cancer Res., 2019 Apr 17;25(15). PMID 30988079

17. Schwaederle MC, Patel SP, Husain H, et al. Utility of Genomic Assessment of Blood-Derived Circulating Tumor DNA (ctDNA) in Patients with Advanced Lung Adenocarcinoma. Clin Cancer Res. Sep 1 2017;23(17):5101-5111. PMID 28539465

18. Thompson JC, Yee SS, Troxel AB, et al. Detection of therapeutically targetable driver and resistance mutations in lung cancer patients by next-generation sequencing of cell-free circulating tumor DNA. Clin Cancer Res. Dec 01 2016;22(23):5772-5782. PMID 27601595

19. Villaflor V, Won B, Nagy R, et al. Biopsy-free circulating tumor DNA assay identifies actionable mutations in lung cancer. Oncotarget. Oct 11 2016;7(41):66880-66891. PMID 27602770

20. Ramalingam SS, Yang JC, Lee CK, et al. Osimertinib as first-line treatment of EGFR mutation-positive advanced non-small-cell lung cancer. J Clin Oncol. Mar 20 2018;36(9):841-849. PMID 28841389

21. Mellert H, Foreman T, Jackson L, et al. Development and clinical utility of a blood-based test service for the rapid identification of actionable mutations in non-small cell lung carcinoma. J Mol Diagn. May 2017;19(3):404-416. PMID 28433077

22. Paweletz CP, Sacher AG, Raymond CK, et al. Bias-corrected targeted next-generation sequencing for rapid, multiplexed detection of actionable alterations in cell-free DNA from advanced lung cancer patients. Clin Cancer Res. Feb 15 2016;22(4):915-922. PMID 26459174

23. Pritchett MA, Camidge DR, Patel M, et al. Prospective Clinical Validation of the InVisionFirst-Lung Circulating Tumor DNA Assay for Molecular Profiling of Patients With Advanced Nonsquamous Non–Small-Cell Lung Cancer. JCO Precision Oncology 2019 :3, 1-15.

24. Remon J, Lacroix L, Jovelet C, et al. Real-World Utility of an Amplicon-Based Next-Generation Sequencing Liquid Biopsy for Broad Molecular Profiling in Patients With Advanced Non–Small-Cell Lung Cancer. JCO Precision Oncology 2019 :3, 1-14.

25. Sacher AG, Paweletz C, Dahlberg SE, et al. Prospective validation of rapid plasma genotyping for the detection of EGFR and KRAS mutations in advanced lung cancer. JAMA Oncol. Aug 01 2016;2(8):1014-1022. PMID 27055085

26. Zhang Y, Xu Y, Zhong W, et al. Total DNA input is a crucial determinant of the sensitivity of plasma cell-free DNA EGFR mutation detection using droplet digital PCR. Oncotarget. Jan 24 2017;8(4):5861-5873. PMID 28052016

27. Karachaliou N, Mayo-de las Casas C, Queralt C, et al. Association of EGFR L858R mutation in circulating free DNA with survival in the EURTAC Trial. JAMA Oncol. May 2015;1(2):149-157. PMID 26181014

28. Oxnard GR, Thress KS, Alden RS, et al. Association between plasma genotyping and outcomes of treatment with osimertinib (AZD9291) in advanced non-small-cell lung cancer. J Clin Oncol. Oct 01 2016;34(28):3375- 3382. PMID 27354477

29. Lindeman, NN, Cagle, PP, Aisner, DD, Arcila, MM, Beasley, MM, Bernicker, EE, Colasacco, CC, Dacic, SS, Hirsch, FF, Kerr, KK, Kwiatkowski, DD, Ladanyi, MM, Nowak, JJ, Sholl, LL, Temple-Smolkin, RR, Solomon, BB, Souter, LL, Thunnissen, EE, Tsao, MM, Ventura, CC, Wynes, MM, Yatabe, YY. Updated Molecular Testing Guideline for the Selection of Lung Cancer Patients for Treatment With Targeted Tyrosine Kinase Inhibitors: Guideline From the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology. J Mol Diagn, 2018 Feb 6;20(2). PMID 29398453

30. Merker JD, Oxnard GR, Compton C et al. Circulating Tumor DNA Analysis in Patients With Cancer: American Society of Clinical Oncology and College of American Pathologists Joint Review. J. Clin. Oncol., 2018 Mar 6;36(16). PMID 29504847

31. Rolfo C, Mack PC, Scagliotti GV, et al. Liquid biopsy for advanced non-small cell lung cancer (NSCLC): a statement paper from the IASLC. J Thorac Oncol. Sep 2018;13(9):1248-1268. PMID 29885479

32. Kalemkerian GP, Narula N, Kennedy EB et al. Molecular Testing Guideline for the Selection of Patients With Lung Cancer for Treatment With Targeted Tyrosine Kinase Inhibitors: American Society of Clinical Oncology Endorsement of the College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology Clinical Practice Guideline Update. J. Clin. Oncol., 2018 Feb 6;36(9). PMID 29401004

33. National Institute for Health and Care Excellence. Plasma EGFR mutation tests for adults with locally advanced or metastatic non-small-cell lung cancer. Medtech innovation briefing. 2018. www.nice.org.uk/guidance/mib137. Accessed August 23, 2019.

Codes:
(The list of codes is not intended to be all-inclusive and is included below for informational purposes only. Inclusion or exclusion of a procedure, diagnosis, drug or device code(s) does not constitute or imply authorization, certification, approval, offer of coverage or guarantee of payment.)

CPT*

81235
81479
86152
86153