Liquid Biopsies Using Circulating Tumor DNA in Non-Small Cell Lung Cancer

Liquid biopsies for the diagnosis and treatment of lung cancer have developed rapidly, driven primarily by technical advances in sensitivity to detect circulating tumor DNA (ctDNA). Still, technical limitations such as the challenge of detecting low-level ctDNA variants and distinguishing tumor-related variants from clonal hematopoiesis remain. With further technical advancements, new applications for ctDNA analysis are emerging including detection of post-treatment molecular residual disease (MRD), clinical trial selection, and early cancer detection. This chapter reviews the current state of ctDNA testing in NSCLC, the underlying technological advances enabling ctDNA detection, and the potential to expand ctDNA analysis to new applications.

Key points

  • A liquid biopsy is a test done on a body fluid sample (usually blood) to assess cancer.

  • Circulating tumor DNA (ctDNA) can be measured by polymerase chain reaction (PCR) and next-generation sequencing (NGS)-based technologies.

  • ctDNA testing is approved in clinical practice for recurrent or metastatic NSCLC when tissue testing is inadequate or not feasible.

  • Major challenges of ctDNA technology include early-stage cancer detection and distinguishing tumor mutations from clonal hematopoiesis.


The development of targeted therapies has substantially improved outcomes for a subset of patients with non-small cell lung cancer (NSCLC). As treatments become more effective, they are also becoming more personalized, requiring novel testing strategies to identify driver mutations, track tumor evolution, and monitor disease recurrence. Liquid biopsies offer a noninvasive way to monitor and personalize treatment for NSCLC.

A liquid biopsy refers to a clinical assay performed on a sample of bodily fluid for cancer assessment. For solid tumor malignancies, these assays typically assess either circulating tumor cells or cell-free circulating tumor DNA (ctDNA) in the blood. In this article we will focus on ctDNA detection from blood plasma and clinical applications of this technology. ,

Cell-free DNA

In both healthy and diseased individuals, nonencapsulated extracellular DNA fragments called cell-free DNA (cfDNA) are found in different bodily fluids. Although the process is not well-studied, cfDNA is thought to be generated primarily through passive release from apoptotic and necrotic cells ( Fig. 1 ). As such, cfDNA fragments are usually 150 to 200 bp in length, roughly corresponding to the length of DNA wrapped around a single nucleosome. cfDNA has a short half-life ranging between 16 minutes and 2.5 hours, making it an ideal substrate for studying tumor mutational dynamics as discussed later in this article. In blood plasma, most cfDNA originates from nucleated peripheral blood cells and endothelial cells, and is present between ∼1 and ∼100 ng/mL.

Fig. 1

Cell-free DNA generated by healthy endothelial cells and leukocytes, and ctDNA released from solid tumor malignant cells. Circulating tumor cells are also depicted.

The amount of plasma cfDNA is dramatically increased in many disease states, often as a result of increased cell turnover. In patients diagnosed with solid malignancies, DNA from cancer cells also contributes to plasma cfDNA and is referred to as ctDNA. In lung cancer, ctDNA tends to be higher in patients with metastatic or locally advanced disease. Overall, ctDNA levels tend to increase with disease progression.

Circulating tumor DNA detection methods

Methods used to detect DNA mutations in plasma are similar to those used on biopsied tissue, but tailored to meet some of the unique challenges of cfDNA as a substrate. Detection techniques can be broadly categorized by their scope of genomic coverage, from targeted allele-specific polymerase chain reaction (PCR) to next-generation sequencing (NGS) techniques such as hybrid capture NGS, whole exome sequencing and whole genome sequencing ( Table 1 ). Common challenges for all detection methods include the low amount of ctDNA present in plasma, the abundant background of cfDNA arising from normal (nontumor) cells, and the uncertain origin of variants discovered.

Table 1

Comparison of Techniques Used for ctDNA Detection

Adapted from Chin RI, Chen K, Usmani A, et al. Detection of solid tumor molecular residual disease (MRD) using circulating tumor DNA (ctDNA). Mol Diagnosis Ther. 2019;23(3):311–31; with permission.

Technique Example Technologies Scale of Analysis Method Detection Limit (as % of cfDNA) Cost Assay Personalization Advantages Limitations
AS-PCR Cobas EGFR mutation test v2 (Roche) Single-locus or multiplexed assays Preferentially amplifies rare mutant DNA molecules ∼0.1%–1% $ Some required Ease of use; ideal for detecting recurrent “hotspot” point mutations Limited multiplexing ability; analytical sensitivity not as low as digital PCR
Digital PCR Digital PCR
Single-locus or multiplexed assays Partitions target DNA into different reactions for massively parallel quantitative PCR ∼0.04%–0.1% $$ Some required High sensitivity Limited multiplexing ability
Genome wide NGS of whole genome ∼10% $$–$$$ Not required Entire genome is interrogated Low sensitivity; mostly limited to SCNA detection; requires in-depth genomics and bioinformatics analysis
Retrotransposon-based Amplicon NGS FAST-SeqS
Genome-wide retrotransposon sites PCR amplification of retrotransposon insertion sites before NGS analysis ∼5% $$ Not required Rapid aneuploidy assessment with lower cost than WGS Limited to aneuploidy detection
WES WES Exome wide NGS of whole exome ∼5% $$$ Not required Entire exome is interrogated Typically low sensitivity; requires in-depth genomics and bioinformatics analysis
Multiplex PCR-based NGS TAm-Seq
Enhanced TAm-Seq
Targeted sequencing PCR amplification enriches targets before NGS analysis ∼0.01%–2.0% $$ Some required High sensitivity (modern methods) Less comprehensive than WGS or WES; requires bioinformatics analysis
Hybrid capture-based NGS CAPP-Seq
FoundationOne Liquid
Targeted sequencing Subset of exome is hybridized to biotinylated probes and captured for NGS analysis ∼0.001%–0.5% $$ Typically not required High sensitivity; detects multiple mutation types; broadly applicable without personalization Less comprehensive than WGS or WES; requires bioinformatics analysis
Combination approaches CAPP-Seq + GRP
Single locus to genome wide Combines different ctDNA detection methods sometimes including protein biomarkers Variable $$–$$$ Variable Improved detection compared with standard ctDNA analysis alone in certain settings More time and resource intensive than individual approaches

Abbreviations: ARMS, amplification-refractory mutation system; BEAMing, beads, emulsion, amplification and magnetics; CAPP-Seq, cancer personalized profiling by deep sequencing; ctDNA, circulating tumor deoxyribonucleic acid; FAST-SeqS, fast aneuploidy screening test-sequencing system; GRP, genome representation profiling; mFAST-SeqS, modified FAST-SeqS; PARE, personalized analysis of rearranged ends; PASA, amplification of specific alleles; Plasma-Seq, plasma sequencing; Safe-SeqS, safe sequencing system; SCNA, somatic copy-number alternation; TAm-Seq, tagged amplicon deep sequencing; TEC-Seq, targeted error correction sequencing; WALDO, within sample aneuploidy detection; WES, whole exome sequencing; WGS, whole genome sequencing.

An additional challenge shared by all ctDNA assays is the potential for leukocyte genomic DNA contamination of samples. Although some early ctDNA testing was performed on serum, where the total nucleic acid yield is higher, modern assays exclusively use plasma to avoid potential contamination from lysed blood cells. To further prevent leukocyte genomic DNA in plasma samples, the College of American Pathologists and the American Society of Clinical Oncology now recommend separating plasma from blood cells within 6 hours of collection in EDTA tubes, and the International Association for the Study of Lung Cancer recommends processing blood within 2 hours. Alternatively, leukocyte stabilization tubes allow specimens to sit for up to 48 hours before processing. No standard collection volume has been established for ctDNA assays. However, to increase analytical sensitivity, plasma volumes of 10-mL or higher are common, requiring that multiple EDTA vacutainers be drawn per assay.

Allele-Specific Polymerase Chain Reaction

PCR-based assays are typically used when detecting single mutations or small panels of well-characterized variants, to query for targeted therapeutic eligibility or evidence of acquired resistance. Allele-specific PCR (AS-PCR, also known as amplification refractory mutation system) can report allele frequencies for single nucleotide variants in plasma cfDNA. In conventional PCR, primers are designed to anneal outside of the sequence region of interest, while in AS-PCR, 3′ primers overlap the region of interest. AS-PCR primers uniquely complementary to either variant-containing or wild-type sequences are used in separate reactions, and amplification occurs only when the template DNA perfectly matches the 3′ primer. AS-PCR assays are an attractive option for rapid clinical testing, with a quantitative readout using standard real-time PCR equipment. The cobas epidermal growth factor receptor (EGFR) mutation test v2 (Roche, Basel, Switzerland) is an example of an AS-PCR assay that queries 42 different EGFR mutations. Most clinical laboratories already perform real-time PCR assays, reagents are relatively inexpensive, and interpretation of the spectrophotometric signal is straightforward. Important drawbacks of AS-PCR include limited multiplexing and sensitivities in plasma of only about 70% to 80%. ,

Droplet Digital Polymerase Chain Reaction

Digital droplet PCR (ddPCR) measures tens of thousands of PCR reactions in nanoliter-scale droplets, evaluating each reaction as a discrete measurement. ddPCR has superior analytical sensitivity to AS-PCR, with a detection limit in the range of 0.04% to 0.10%. However, currently the clinical benefits of increased sensitivity are unclear as ddPCR and AS-PCR performed similarly as confirmatory tests for patients positive for the EGFR T790M mutation in plasma. Both PCR methods have a substantially faster turnaround time than NGS-based methods, with most results returned within 72 hours compared to 1 to 2 weeks for massively parallel sequencing. Like AS-PCR, a drawback of ddPCR is the limited ability to multiplex the assay to query mutations in multiple genes simultaneously.

Targeted Next-Generation Sequencing

As clinically useful molecular targets continue to accumulate, NGS becomes an increasingly important testing method in NSCLC. Although whole exome and even whole genome sequencing can potentially offer more comprehensive genomic information, most ctDNA NGS assays in clinical use are more targeted, using either hybrid capture panels or amplicon-based NGS. Both of these methods, when applied to ctDNA detection, involve high-depth NGS to detect low allele frequency mutations in the plasma.

Mutant allele frequency of ctDNA in plasma varies with cancer stage and treatment effect, typically being significantly lower in early-stage disease and after cytotoxic treatment. Reliably detecting ctDNA mutations in lung cancer patients with NGS, especially for post-treatment molecular residual disease (MRD) detection, requires extraordinary sensitivity. , , This sensitivity can be achieved through personalized high depth sequencing, with raw nondeduplicated depths of 50,000× commonly achieved, along with digital error suppression techniques, enabling detection of variants with allele frequencies as low as approximately 10 −4 . , However, bridge amplification sequencing has an error rate of approximately 10 −3 , meaning that any rare variant discovered at this high depth is likely to originate from ex vivo sequencing errors than from a real, in vivo ctDNA mutation. To account for this, modern NGS-based techniques use barcoded sequencing adapters and bioinformatic error suppression to reduce this technical error rate significantly and enable ultra-low level ctDNA detection. , ,

Clinical use of liquid biopsies

The College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology recommend molecular testing from tumor tissue in patients with NSCLC. ctDNA testing should only be used when there is insufficient tissue for molecular testing or when tissue cannot be obtained. Of note, liquid biopsies may spare patients from major complications associated with computed tomography-guided lung biopsy that occur in approximately 5% of patients.

In clinical practice, ctDNA assessment from blood plasma is the mostly widely used liquid biopsy approach, with Zill and colleagues recently reporting results from 21,807 patients with advanced cancer assessed using the Guardant360 assay (Guardant Health, Redwood City, CA). ctDNA analysis can identify actionable tumor mutations at the time of diagnosis and/or recurrence of patients with advanced/metastatic adenocarcinoma of the lung. Additionally, ctDNA testing is recommended if a patient is diagnosed with squamous cell carcinoma and is a never smoker or young patient, or if a patient has a nonsquamous component on tumor histology. In clinical practice, we can usually forego the need to perform tissue NGS if the ctDNA molecular test is positive for an actionable mutation. This recommendation applies to patients being treated with targeted therapy and chemotherapy and/or immunotherapy. , In addition, different proof-of-concept studies have demonstrated ctDNA’s ability to detect post-treatment MRD, which was shown to be prognostic. , , , Still, ctDNA analysis for MRD detection is not yet approved for clinical practice.

As an example of ctDNA’s impact in clinical practice, PCR-based techniques are used to detect EGFR exon 19 deletions and L858R single nucleotide variants in plasma, reaching sensitivities as high as 100% (range, 78%–100%) while maintaining high specificity (93%–100%). Detection of EGFR mutations are of extreme clinical importance because their detection changes the management of NSCLC patients. If an EGFR exon 19 deletion or L858R mutation is detected, patients should be treated with an EGFR tyrosine kinase inhibitor in the first-line setting rather than chemotherapy or immunotherapy. Although PCR allows rapid detection of genetic alterations without extensive genomic analysis, thus translating to cost-effective techniques, it only enables monitoring of a limited number of known mutations. Conversely, NGS-based techniques can interrogate larger regions of multiple genes in a single run. Limitations of NGS-based technologies include longer processing and analysis time than PCR, and higher costs (see Table 1 ).

As of 2019, the only test approved by the US Food and Drug Administration for the molecular analysis of liquid biopsy specimens in NSCLC is the cobas test (Roche), an AS-PCR assay that queries EGFR mutations. Nevertheless, assay development and clinical implementation are outpacing regulatory approval. At leading cancer centers, ctDNA testing is being performed routinely in patients with solid malignancies using either in-house laboratory developed tests (ie, MSK-ACCESS ) or commercially available assays (ie, Inivata [Cambridge, UK] and Guardant Health , ) that query panels of potentially actionable mutations, which may inform clinical decision-making.

Tumor Profiling at Diagnosis

Even though ctDNA testing is being used in clinical practice, it is not the gold standard for actionable mutation assessment in newly diagnosed metastatic NSCLC owing to concerns regarding sensitivity. Sabari and colleagues illustrated ctDNA sensitivity limitations in a prospective study with 210 patients with advanced NSCLC. The detection rate of known tumor-derived somatic mutations in patients receiving systemic therapy was only 42.9% when using a 21-gene hybrid capture NGS assay. Even though ctDNA analysis detected a variety of mutations with a shorter turnaround time, their findings support that negative plasma genotyping requires confirmatory tumor testing to rule out a false negative result.

In an attempt to prove ctDNA noninferiority to physician discretion standard-of-care (SOC) tumor genotyping, Leighl and colleagues conducted a prospective study (the NILE study), which enrolled 307 patients with newly diagnosed stage IIIB and stage IV NSCLC, and compared physician’s choice tissue genotyping with plasma genotyping using the Guardant360 platform to detect somatic ctDNA mutations. The authors compared these tests’ performances in identifying 8 guideline-recommended biomarkers (EGFR mutations, ALK fusions, ROS1 fusions, BRAF V600E mutation, RET fusions, MET amplifications, MET exon 14 skipping variants and ERBB2 [HER2] mutation), in addition to KRAS mutations. ctDNA testing identified significantly more patients (77.0% vs 27.3%) with a guideline-recommended biomarker than physician-discretion SOC genotyping (60.0% vs 21.3%). Overall clinical sensitivity of ctDNA relative to tissue was 80% for the detection of any guideline-recommended biomarker. When restricting analysis to the targets approved by the US Food and Drug Administration, namely actionable genomic alterations in EGFR, ALK, ROS1, and BRAF, ctDNA concordance with tissue testing was greater than 98.2% with 100% positive predictive value. Also, although EGFR mutations and ALK fusions were tested in the majority of patients (83% and 80% of patients, respectively) by physician SOC tissue genotyping, the other guideline-recommended mutations were tested in only one-quarter to one-third of cases. Overall, ctDNA testing increased the detection of somatic mutations by 48%, including those with tissue that was not assessed or yielded insufficient or negative results. Another advantage of ctDNA testing was a significantly shorter turnaround time of a median of 9 days compared to 15 days for tissue genotyping.

This study’s results suggest that panel-based ctDNA testing is noninferior to physician-discretion SOC genotyping and has advantages including a shorter turnaround time, mutational assessment not reliant on a physician ordering each molecular test separately, and the potential to identify mutations missed by tissue genotyping. Thus, in the future, clinicians may implement parallel plasma genotyping for newly diagnosed metastatic NSCLC to avoid incomplete genotyping and diagnostic delays.

Another study of tissue genotyping in patients diagnosed with advanced NSCLC (stages IIIB and IV) in the community setting noted that only 59% of the 814 patients met the guideline recommendations for EGFR and ALK testing and only 8% underwent comprehensive tissue genomic profiling. Barriers to complete genomic profiling included lack of tissue for molecular analysis, and reliance on an oncologist’s orders to have each genotyping test performed. ctDNA testing may thus improve rates of molecular testing in the community setting, enabling the identification of more patients who are candidates for targeted therapies.

Circulating Tumor DNA Testing for Targeted Therapy

All patients diagnosed with advanced/metastatic nonsquamous NSCLC and younger patients, nonsmokers or light smokers diagnosed with squamous NSCLC should be tested for EGFR mutations, ALK fusions, ROS1 fusions, BRAF V600E mutation, RET fusions, MET amplifications, MET exon 14 skipping variants, and ERBB2 (HER2) mutations according to different society guidelines (American Society of Clinical Oncology, National Comprehensive Cancer Network, International Association for the Study of Lung Cancer). , , Molecular testing should be performed using tissue samples, however ctDNA testing can be performed if tumor tissue is inadequate or not obtainable in order to identify actionable mutations.

Epidermal growth factor receptor mutations

In 2015, Mok and colleagues published the results of a prospective study assessing the concordance rates between plasma and tissue genotyping for EGFR mutation in 238 patients diagnosed with NSCLC. The authors described a concordance rate of 88% between plasma and tissue, with a sensitivity of 75% and specificity of 96% for blood testing of EGFR mutations. Additionally, dynamic changes in ctDNA EGFR mutational status correlated strongly with clinical outcomes. Patients who were ctDNA-positive for EGFR mutation at baseline that later became undetectable by cycle 3 had significantly longer progression-free and overall survival than patients with persistently detectable EGFR mutations. These results suggest that EGFR ctDNA testing is both predictive and prognostic, and that dynamic changes in blood-based EGFR testing may be useful as a biomarker for treatment response.

Resistance to EGFR tyrosine kinase inhibitors is a well-described phenomenon. Approximately 60% of patients treated with a first-line EGFR tyrosine kinase inhibitor (erlotinib, afatinib, gefitinib) who develop treatment resistance, acquire an EGFR T790M mutation. In 2017, Jenkins and colleagues assessed the cobas plasma test to detect T790M mutations from plasma cfDNA in patients enrolled onto the AURA extension and AURA2 studies. Their results revealed that T790M mutations were detected in plasma of only 61% of the patients whose tumor harbored such a mutation. Although other techniques such as ddPCR may have a higher sensitivity (71%) to detect the T790M mutation, the discrepancy between plasma and tissue genotyping for this mutation may be more common than for exon 19 deletions or L858R variants. As such, tissue-based testing is preferable whenever feasible to query EGFR T790M mutational status. , , ,

Anaplastic lymphoma kinase (ALK) rearrangements

Although there are substantial data on EGFR plasma genotyping, data on ALK rearrangement assessment using ctDNA in treatment-naïve patients is more limited. In the NILE study, ctDNA sensitivity to detect ALK fusions in plasma was 62.5%. This prospective cohort of 282 patients with ctDNA testing that included assessment of ALK rearrangements is the largest to date. Retrospective data suggest that quantitative PCR is a marginal method for the detection of ALK rearrangements in plasma. Even though ddPCR is generally more sensitive than quantitative PCR at detecting mutations in cell-free DNA, a large prospective study using this technology has not been conducted for ALK fusions.

Circulating Tumor DNA Testing for Immunotherapy

Immune checkpoint inhibitors (ICIs) have changed the treatment landscape for patients with NSCLC. Recent studies suggest that high tumor mutational burden (TMB) is associated with clinical benefit from ICIs. Nonetheless, 30% of patients with NSCLC do not have adequate tissue available for standard biomarker testing. In this setting, Chaudhuri and colleagues and Gandara and colleagues reported on the development of novel assays to measure TMB from blood plasma ctDNA results. Gandara and colleagues used samples from patients enrolled on the POPLAR and OAK studies to evaluate and validate their test, respectively, which was based on the FoundationOne hybrid capture NGS assay (Foundation Medicine Inc, Cambridge, MA). The authors found that TMB extrapolated from blood plasma correlated with tissue TMB. Also, when variant allele fractions were more than 1%, the concordance rate between tissue and blood TMB was very high (r 2 = 0.998). Importantly, samples from patients enrolled on the POPLAR study demonstrated that a blood-derived TMB (bTMB) of 16 or greater was associated with improved progression-free survival (PFS) and overall survival (OS) with atezolizumab versus docetaxel, compared with the full biomarker-evaluable population. When tested in the OAK study, there remained a significant PFS benefit for patients with a bTMB of 16 or greater treated with atezolizumab; however, the OS benefit compared to the general biomarker-evaluable population was not seen until a higher cutpoint of bTMB≥24. Interestingly, a high bTMB did not significantly correlate with high tumor programmed death ligand 1 (PD-L1) expression; however, the authors showed that these factors independently contributed to the benefit of immunotherapy, with PFS outcomes being best in those with both high tumor PD-L1 and high bTMB. These findings suggest that blood plasma-derived TMB is a potential noninvasive biomarker to predict benefit from ICIs.

Finally, Goldberg and colleagues demonstrated a strong correlation between longitudinal ctDNA changes, radiographic response, OS, and PFS in patients with metastatic NSCLC treated with immune checkpoint inhibitors. Among 49 patients enrolled onto their study, 28 had somatic mutations that were identified in baseline plasma. The study focused on these 28 patients, who had serial ctDNA analysis during ICI therapy which was compared with tumor size measured by computed tomography analyzed by RECIST 1.1. ctDNA response was defined as more than a 50% decrease in mutant allele fraction from baseline followed by a second confirmatory measurement. Their results revealed that ctDNA response was seen more rapidly than radiographic response, occurring a median of 42.5 days earlier. Additionally, ctDNA response was associated with improved PFS and OS, with a hazard ratio of 0.29 and 0.17, respectively. Although compelling, the small size of this study and the lack of independent validation of its results make it challenging to translate into clinical practice yet.

Liquid biopsy challenges

ctDNA analysis is complicated by several factors including intertumoral and intratumoral heterogeneity, clonal hematopoiesis (CH) and the technical challenge of detecting variants at low allele fractions. , , With the constant clonal selection that tumors undergo during different lines of treatment, single-site biopsies are unable to reflect the complexity of the entire tumor genomic landscape. Patients may have intratumoral heterogeneity (variable genomic landscape within one site of disease), and intertumoral heterogeneity (molecular differences between the primary tumor and metastatic sites), which may reflect discordant genotyping results between plasma and tissue. , Furthermore, the sensitivity of ctDNA detection depends on ctDNA levels. , Patients diagnosed with advanced/metastatic NSCLC often possess 10% or higher levels of ctDNA, whereas individuals with localized disease typically harbor 1% or less ctDNA in their plasma. As such, highly sensitive assays are necessary to detect ctDNA in patients, especially in those with earlier stages of disease. , , , , ,

CH is another challenge facing ctDNA analysis. , CH arises when age-dependent acquired mutations accumulate in hematopoietic progenitor cells, leading to genetically distinct subpopulations that contribute disproportionately to the population of mature blood cells. , In the measurement of ctDNA, CH can result in false-positive results owing to detection of nonreference variants in the plasma, some of which are in genes like TP53 and KRAS that are frequently mutated in solid tumors. , Although the prevalence of CH has previously been reported to be approximately 30% in adults over 60 years of age, Swanton and associates presented findings from the ongoing Circulating Cell-free Genome Atlas Study at the 2018 American Society of Clinical Oncology meeting showing that CH was detected in 92% of patients with a variant allele fraction of more than 0.1%. Because most of plasma cfDNA is composed of DNA arising from nucleated hematopoietic cells, CH-related mutations can thus lead to false-positive results. This issue was reported by Hu and colleagues in a cohort of patients with advanced NSCLC. The authors performed paired genotyping of peripheral blood cells and plasma, and detected false-positive ctDNA genotyping for KRAS and TP53 mutations owing to CH. With modern NGS-based assays that possess high sensitivity, there is thus a high likelihood of identifying CH, which can be misinterpreted as tumoral mutations. Therefore, caution is advised when interpreting plasma genotyping results from commercial vendors, especially given that these assays do not filter variants present in matched peripheral blood leukocytes. , ,

Analytical aspects also pose a challenge to ctDNA analysis. Given the short half-life of ctDNA and the possibility of contamination of plasma cfDNA with DNA from healthy leukocytes, it is crucial that blood be collected in specialized tubes or be processed and stored quickly after collection. The isolation of cfDNA from plasma has yet to be completely standardized and multiple protocols and kits have been developed. , ctDNA fragments are often shorter than healthy cfDNA, and both are substantially smaller than genomic cellular DNA derived from blood leukocytes. , , , Thus, enriching for shorter cfDNA fragments may boost the sensitivity of ctDNA detection. Given that ctDNA typically is a small fraction of total cfDNA, very sensitive methods to reliably detect mutant DNA molecules are required. , Finally, discordances have been reported between ctDNA assay platforms, which could be related to different sets of mutations queried among other factors including the analytical techniques used. , Further cross-validation and standardization between assays and tissue testing need to be rigorously performed if we are to expand the clinical utility of ctDNA testing in standard clinical practice.

Future directions

ctDNA assays currently in clinical use are largely restricted to the diagnostic testing of recurrent and/or metastatic disease, to aid in the selection of targeted therapies when tissue is inadequate or unavailable. However, as ctDNA assay sensitivity increases, new applications will emerge, including testing in lieu of tissue even when it is available, and potentially screening and MRD monitoring. Better screening for lung cancers is especially important because the 5-year OS is approximately 70% at stage 1, whereas the OS is only 16% at stage IV. In addition, only 5% of patients receive annual low-dose computed tomography screening. So far, hybrid capture panels have shown insufficient sensitivity in small screening studies. Nevertheless, higher sensitivity assays have shown promise for MRD detection in small clinical studies. , , Optimizations to further lower the detection limit while maintaining specificity could potentially enable early cancer screening as well. We also envision that, with the expansion of mutational targets queried by ctDNA assays coupled with the proliferation of precision therapeutics, we can offer more patients molecularly targeted agents through umbrella clinical trials.


  • A liquid biopsy is a clinical assay performed on a sample of body fluid to assess cancer.

  • Liquid biopsy ctDNA assays commonly refer to the detection of tumor-specific mutations in the blood plasma of patients with cancer.

  • ctDNA detection can be performed using ultra-sensitive PCR- and NGS-based techniques.

  • ctDNA testing may not be sensitive enough to detect certain actionable mutations reliably like EGFR T790M.

  • ctDNA testing should be used in clinical practice for recurrent or metastatic NSCLC only when tissue cannot be obtained or is insufficient for molecular analysis.

  • ctDNA results may be confounded by CH, especially an issue for commercial assays where matched leukocytes are not sequenced.

  • Further cross-validation between ctDNA assays and with tissue testing needs to be rigorously performed.

  • Future directions include using ctDNA to molecularly guide umbrella trial enrollment, detect MRD, and potentially enable early cancer screening.


A.A. Chaudhuri is a scientific advisor/consultant for Geneoscopy, Roche Sequencing Solutions and Tempus Labs; has received speaker honoraria and travel support from Varian Medical Systems, Roche Sequencing Solutions, and Foundation Medicine; receives research support from Roche Sequencing Solutions ; and is an inventor of intellectual property licensed to Biocognitive Labs. The other authors have nothing to disclose.


Only gold members can continue reading. Log In or Register to continue

Aug 16, 2020 | Posted by in CARDIAC SURGERY | Comments Off on Liquid Biopsies Using Circulating Tumor DNA in Non-Small Cell Lung Cancer
Premium Wordpress Themes by UFO Themes