Integrating ctDNA testing for EGFR analysis in advanced non-small cell lung cancer: strategies for clinical laboratories

Introduction

Precision oncology has profoundly influenced the management of lung cancer (LC), which was the most frequently diagnosed cancer in 2022, responsible for almost 2.5 million new cases worldwide and the leading cause of cancer-related mortality [1]. Identification of molecular mechanisms involved in LC has led to the development and clinical implementation of targeted therapies, especially for non-small cell lung cancer (NSCLC).

The epidermal growth factor receptor (EGFR) gene represents the most prevalent and first targetable oncogenic gene in NSCLC. EGFR mutations occur in about 15–40 % of cases. EGFR tyrosine kinase inhibitors (EGFR-TKIs) have become the first-line systemic therapy for patients with EGFR-activating mutations as they lead to improvements in survival and quality of life [2]. Therefore, molecular testing is mandatory for advanced NSCLC to identify patients eligible for EGFR-TKI therapy [3].

However, the use of this testing in the clinical routine is often limited by the difficulty in obtaining adequate tumor tissue, which is considered the gold standard for molecular studies. In up to 30 % of LC patients, sufficient material for EGFR molecular analysis is unavailable [4]. Additionally, the invasive nature of tissue biopsies makes continuous monitoring impractical. In this context, “liquid biopsy” (LB) strategies have emerged as non-invasive or minimally invasive methods for analyzing tumor characteristics from biological fluids like blood [5]. These methods detect tumor-derived analytes, including circulating tumor DNA (ctDNA), circulating tumor cells, circulating free RNA, and extracellular vesicles.

This review will primarily focus on ctDNA analysis due to its extensive validation and broad acceptance in clinical settings. Analyzing EGFR mutations in ctDNA is an established method for identifying patients who would benefit from EGFR-TKIs. ctDNA testing is particularly valuable when tissue samples are unavailable or when fast turnaround time (TAT) is needed for urgent treatment decisions. However, the routine implementation of ctDNA analysis in clinical laboratories still faces significant challenges. Laboratory professionals play a crucial role in ensuring the reliability and optimal interpretation of these results, making them essential for successful integration into clinical practice.

This review aims to provide a comprehensive synthesis of current knowledge regarding the clinical utility of EGFR molecular analysis in ctDNA. From a clinical laboratory perspective, the review will explore technical aspects, clinical applications, benefits, and ongoing challenges associated with ctDNA analysis, while offering practical guidelines to effectively integrate ctDNA testing into routine practice.

EGFR: structure, signaling and clinical implications

The EGFR is a 170-kDa transmembrane glycoprotein belonging to a family of four cell-surface receptors: (EGFR or HER1), HER2, HER3 and HER4. Early studies on the EGFR pathway were initiated with the discovery of epidermal growth factor (EGF) by Cohen and Levi-Montalcini in 1962 [6], [7]. Since then, biochemical and genetic studies have elucidated its role in regulating cellular growth, detailing the intricate cascade of intracellular signals following ligand-receptor interaction.

The oncoprotein EGFR is composed of three domains: an extracellular domain (rich in carbohydrates that facilitate ligand binding), a transmembrane domain (hydrophobic), and a cytoplasmic domain (with tyrosine kinase activity and phosphorylation sites). Ligands such as EGF and transforming growth factor α (TGFα) can bind to EGFR. Upon ligand binding, EGFR dimerizes, forming homodimers, and heterodimers with HER2 or other family members. This dimerization facilitates autophosphorylation, activating the downstream signaling [8]. The activation of this cascade induces cell proliferation and plays a significant role in cell survival and differentiation.

Therefore, EGFR overexpression or inappropriate activation can contribute to cancer development and progression. Increased EGFR signaling can result from EGFR gene amplification, protein overexpression, or specific activating mutations. EGFR is frequently overexpressed in various solid tumors, including lung, head and neck, breast, kidney, colon, ovarian, prostate, brain, and bladder cancers [9]. This overexpression has been associated with increased tumor aggressiveness and a reduced response to conventional therapies.

In NSCLC, EGFR deregulation is prevalent, characterized by protein overexpression in 85 % of cases, gene mutations in 10–40 %, and gene amplification in approximately 10 % Protein overexpression (assessed by immunohistochemistry) or gene copy number (assessed by fluorescence in situ hybridization) lack consistent prognostic or predictive value. In contrast, the detection of EGFR mutations has proven clinical utility as biomarkers for predicting response to EGFR-TKIs [10]. These somatic mutations, which result in constitutive activation of EGFR kinase activity, are more frequent in women, patients with adenocarcinoma, Asian ethnicity, and those who have never smoked. The most common activating EGFR mutations in NSCLC are exon 19 deletions (>50 %) followed by the p.L858R point mutation in exon 21 (∼40 %), both of which are associated with a positive response to EGFR-TKI therapy. Mutations in exons 18 and 20 account for the remaining 10 % of EGFR mutations in NSCLC [11] (Figure 1).

Figure 1:

Clinical relevance of key EGFR mutations in non-small cell lung cancer. The Figure illustrates the clinical relevance of key EGFR mutations in non-small cell lung cancer (NSCLC). Mutations are grouped by their location within specific exons (18–21) and their association with drug sensitivity or resistance. Mutations in blue boxes are linked to increased drug sensitivity to EGFR tyrosine kinase inhibitors (TKIs), with their corresponding prevalence percentages noted. Mutations in red boxes are associated with drug resistance. Exon 19 deletions and the L858R mutation in exon 21 are the most prevalent drug-sensitive mutations, while T790M in exon 20 is the most common resistance mutation. The domains of the EGFR protein (extracellular, transmembrane, intracellular, tyrosine kinase, and regulatory) are also shown for structural context. Adapted from: Sharma SV, et al.: Epidermal growth factor receptor mutations in lung cancer, Nat Rev Cancer 7: 169–181, 2007 and Bai Y, et al.: Molecular genetics of solid tumors, in Henry’s Clinical Diagnosis and Management by Laboratory Methods, 24th ed, pp. 1579–1587, Elsevier, 2022. Figure created with Biorender.

The p.T790M point mutation represents the most common mechanism of acquired resistance, occurring in 50–60 % of patients treated with first- and second-generation EGFR-TKIs. In cases of atypical EGFR mutations, treatment response may vary compared to classical mutations. For instance, exon 20 insertions are typically associated with resistance to standard EGFR-TKIs, although the response can be heterogeneous [12], [13].

Methodologies for molecular characterization of EGFR gene in circulating tumor DNA

Methodologies for detecting EGFR mutations in ctDNA can be categorized into two groups: targeted and non-targeted approaches (as detailed in Table 1). Targeted methods include real-time polymerase chain reaction-based techniques (RT-PCR), the amplification refractory mutation system (ARMS), and digital PCR (dPCR) methods such as droplet digital PCR (ddPCR) and BEAMing (beads, emulsions, amplification, and magnetics). These methods provide excellent specificity but are limited to detecting deletions and point mutations pre-defined in their design.

Regarding analytical sensitivity, ARMS and RT-PCR are known for their fast and precise detection capabilities, with a limit of detection (LOD) ranging from 0.1 to 5 % in most commercial assays. These methods generate qualitative or semi-quantitative results. In contrast, dPCR offers enhanced sensitivity, achieving a LOD of 0.1–1 % by partitioning the sample into numerous discrete reactions. This approach minimizes background interference and improves the detection of low-abundance mutations. Unlike the previous methods, dPCR provides a direct quantitative result, enabling precise determination of variant allele frequencies (VAF).

Next-generation sequencing (NGS) offers a non-targeted approach that enables comprehensive genomic profiling, allowing for the detection of a wide range of mutations in EGFR and other relevant genes while providing quantitative results. NGS also allows the identification of novel mutations not detected by targeted assays. Despite historically facing challenges with lower sensitivity and specificity due to the inherent error rates in DNA polymerase and sequencing processes, advancements like deep sequencing, molecular barcoding, and error-correction algorithms have significantly improved their performance. These developments have led to unprecedented levels of sensitivity (LOD from 0.1 to 0.5 %).

In terms of turnaround time (TAT), the broader scope of NGS results in a longer TAT and requires complex computational analysis compared to targeted approaches such as RT-PCR or dPCR. While NGS methods have a TAT of 1–2 weeks, Idylla and other PCR methods have a TAT of approximately seven days [31], [32], [33]. In summary, while targeted approaches are highly effective for detecting specific mutations with low detection limits and fast TAT, NGS provides a more comprehensive overview of the mutational landscape.

According to the 2022 report from the European Molecular Quality Network (EMQN), targeted approaches are used by the majority of laboratories, with RT-PCR being the predominant method (51 %). Among these, the cobas^® EGFR Mutation Test v2 (Roche) was the most commonly employed assay, used by 27 % of all laboratories. This test was the first LB assay approved by the U.S. Food and Drug Administration (FDA) in 2016. Meanwhile, NGS strategies are also commonly adopted, with the Oncomine™ Lung cfDNA Assay (Thermo Fisher Scientific) being the most frequently used panel [34].

The rapid advancement of NGS technologies in recent years has significantly reduced sequencing costs while enhancing accuracy. Moreover, as the list of actionable gene alterations with FDA-approved therapies continues to grow (EGFR, ALK, ROS1, BRAF, RET, MET, NTRK, KRAS, HER2), there is a corresponding increase in the recommendation for broader NGS panels. These panels are not only more cost-effective than sequential single-gene testing in tissue samples but are also gaining prominence in plasma-based testing. The International Association for the Study of Lung Cancer (IASLC) and the European Society for Medical Oncology (ESMO) recommend adopting broad-based plasma ctDNA analysis by NGS, in advanced NSCLC [34], [35], [36]. In regions with a high prevalence of EGFR mutations, an initial targeted analysis remains pertinent as the first step in molecular testing. However, if a negative result is obtained by single-gene testing, it is advisable to proceed with serial testing for additional actionable biomarkers.

Comparison of plasma ctDNA and tissue samples for EGFR molecular analysis

Tumor tissue has historically been considered the gold standard for EGFR molecular analysis. However, it has inherent limitations, including its invasive nature and challenges associated with obtaining biopsies from hard-to-reach locations or patients with complex clinical conditions. Additionally, tumor tissue samples may yield insufficient or degraded DNA, complicating the analysis. EGFR analysis in ctDNA obtained from plasma has shown high concordance with tissue samples, addressing some of these limitations. Plasma-based ctDNA analysis has several advantages: it is minimally invasive, demonstrates high concordance with tissue results, offers a rapid TAT and allows for repeated testing over time, providing a more comprehensive representation of tumor heterogeneity and clonal evolution. However, potential drawbacks include the risk of false negatives and higher costs when used alongside tissue testing [35].

When comparing plasma-based ctDNA testing with tumor tissue analysis, studies have shown a high level of concordance for EGFR mutations, as well as other guideline-recommended biomarkers [37], [38]. González de Aledo-Castillo et al. reported a concordance of 95.2 % for plasma cfDNA testing using the Cobas^® EGFR test in NSCLC patients at stages III and IV, compared with tissue analysis. In their prospective study, 39.9 % of patients lacked accessible tumor tissue for molecular characterization. In this scenario, plasma-based EGFR testing facilitated therapy initiation in 34.8 % of these patients, emphasizing the utility of ctDNA when tissue biopsy is not feasible [14].

Regarding the analytical performance, plasma-based ctDNA analysis generally demonstrates high specificity across platforms (>96 %). However, sensitivity varies considerably (58–100 %), depending on the platform and the mutations studied [35]. For instance, T790M mutations are typically detected at lower frequencies in plasma ctDNA compared to the primary sensitizing mutations (L858R or exon 19 deletions), leading to reduced sensitivity (41–90.5 %) [35]. A recent meta-analysis reported a pooled sensitivity of 68 % (95 % CI=60–75 %) and specificity of 98 % (95 % CI=95–99 %) [39].

Another significant advantage of plasma-based ctDNA analysis is the shorter TAT compared with tissue genotyping, which has a median TAT of 12 days (range 1–54 days) [40], [41]. Overall, employing a plasma-based strategy, significantly improves the availability of test results before the initial patient visit (85 vs. 9 %, p<0.0001), thus reducing time-to-treatment for patients with advanced NSCLC [42].

Plasma-based ctDNA analysis also offers the possibility of monitoring treatment response and disease progression longitudinally in patients receiving EGFR-TKIs [43]. Tissue re-biopsies can be risky due to the patient’s health status or tumor location; plasma analysis is a safer alternative to assess EGFR mutation status during treatment [44]. Additionally, ctDNA analysis can provide valuable insights into the mechanisms of resistance, such as MET amplification, revealing tumor heterogeneity and clonal evolution in osimertinib-resistant cases with EGFR mutations (AURA3 trial, NCT02151981) [45], [46]. Nevertheless, it is important to note that ctDNA analysis may not identify all resistance mechanisms, such as histologic transitions, which require morphological examination of tissue.

False negatives remain the weakness of ctDNA analysis; tissue biopsy remains necessary when LB tests yield negative results. Sensitivity may fluctuate based on both analytical and biological factors. Like other circulating markers, such as serum tumor markers (STMs), the biological characteristics of the tumor – such as stage, tumor burden, and vascularization – can influence the amount of ctDNA released into circulation, which impacts sensitivity. Tumors with low ctDNA shedding are observed in 15–32 % of NSCLC patients [47], and sensitivity is further reduced in early-stage NSCLC, dropping below 30 % [48]. False negative results can also arise from technique’s limitations, underscoring the increasing focus on high-sensitive methods like NGS. Some studies have explored alternative matrices, such as bronchial washing fluid, which has shown significantly higher sensitivity for detecting EGFR mutations than plasma [49], [50]. False positives are uncommon but can occur, particularly when using broad-spectrum methods like NGS, which analyze multiple genes. These false positives can arise from germline variants or non-tumor sources such as clonal hematopoiesis of indeterminate potential (CHIP). While CHIP can influence mutations in genes like TP53 and KRAS, it does not notably impact the analysis of EGFR mutations [51].

In conclusion, EGFR analysis in plasmatic ctDNA offers significant advantages in terms of non-invasiveness, accessibility, and monitoring capability, but it also has limitations. Plasma ctDNA analysis does not replace tissue genotyping and should be considered a complementary tool in the therapeutic decision-making for advanced NSCLC.

Preanalytical and postanalytical factors

The pre-analytical phase plays a critical role in influencing cfDNA results. Adhering to established guidelines for sample collection, handling, transport, and cfDNA extraction is essential to achieve optimal standardization and reliability [52]. The main challenges affecting the quality of cfDNA analysis include: (1) the heterogeneous content of blood, which complicates cfDNA isolation, (2) enzymatic degradation and clotting, (3) the instability of naked DNA in biological environments, (4) contamination of cfDNA by genomic DNA (gDNA), and (5) the difficulty in detecting the small, specific fraction of ctDNA within the cfDNA pool.

Regarding cfDNA isolation, plasma is the preferred specimen over serum to avoid contamination with gDNA from disrupted leukocytes during clotting [53]. Blood should be collected in tubes with cell stabilizer (e.g. Streck^®) or EDTA, which is the recommended anticoagulant. When EDTA is used, plasma isolation should be performed within 4 h [53]. When samples need to be transported to different centers, it is preferable to use commercial cell stabilizer tubes, as they prevent blood cell lysis and allow processing times to be extended for several days at room temperature. A two-step plasma centrifugation process is recommended: first, low-speed centrifugation to concentrate blood cells, followed by high-speed centrifugation to remove cell debris. For the extraction method, ready-to-use automated kits are preferred for routine applications [52]. It is also advised to perform quality control on the cfDNA extract to assess both cfDNA quantity and fragmentation levels [54].

In the post-analytical phase, the interpretation and reporting of EGFR variants in ctDNA should be consistent with standard criteria for somatic variant interpretation [55], [56]. It is also crucial to consider the unique characteristics of ctDNA and adhere to specific guidelines for ctDNA analysis [37], [57], ensuring accurate detection and reporting of EGFR mutations relevant to NSCLC.

Molecular reports for cfDNA should be clear and concise, emphasizing clinically significant information. Genetic alterations should be described using standardized nomenclature as per Human Genome Variation Society (HGVS) guidelines (http://varnomen.hgvs.org). Additionally, reports should include simplified, colloquial terms to enhance clarity and understanding. For instance, an EGFR gene alteration should be listed as “c.2369C>T (p.Thr790Met)” and can optionally be described as the “T790M mutation”. Reports should provide essential details to guide clinical decision-making, emphasizing the clinical significance of identified variants concerning FDA/EMA-approved therapies. Key methodological information should be included at the end of the report: the alterations tested, limitations, allelic frequencies, coverage, and other pertinent technical data. The language used in reporting should highlight the possibility of discrepancies with tumor testing, particularly in situations where a variant is not identified in plasma ctDNA [37]. Considering the potential risk of false negatives, EGFR results in ctDNA should not be reported merely as negative. Instead, they should be described as “non-informative” or “not detected”. This terminology reflects the possibility that an EGFR mutation may be present in the tissue but not detectable in the plasma due to biological factors or technical limitations.

Finally, Molecular Tumor Boards are crucial for interpreting complex cases, requiring the collaboration of multidisciplinary experts. These boards ensure a thorough evaluation of molecular findings, with laboratory professionals playing a crucial role in ensuring the correct interpretation of molecular alterations, especially when addressing variants of uncertain significance or conflicting results.

Clinical utility and guideline recommendations across diverse clinical scenarios

National and international guidelines recommend ctDNA genotyping as a surrogate for tissue genotyping in the molecular analysis of EGFR and other relevant biomarkers in advanced NSCLC [58], [59]. This approach is advised when tissue samples are unavailable, both for initial diagnosis and for monitoring resistance. Decision algorithms, provided by organizations such as IASLC, help guide the selection of plasma- or tissue-based testing strategies. IASLC recommends a “plasma-first” strategy when a tissue sample is unavailable, a “complementary approach” when tissue is available but is scant or of uncertain adequacy for genotyping, and a “sequential approach” when tissue genotyping is incomplete [35].

By contrast, tissue-based testing remains the preferred method for early stages NSCLC (stages I–III), which accounts for 25–30 % of patients. Osimertinib, a third-generation EGFR-TKI, was approved by the FDA in 2020 to reduce the recurrence rate post-surgery in patients with stage IB to IIIA NSCLC with EGFR activating mutations (EGFR+), based on the ADAURA trial, which demonstrated significantly improved disease-free survival (DFS) [60]. In the neoadjuvant setting, anti-EGFR therapies may reduce tumor size and improve resection rates, though, ongoing trials like NeoADAURA are critical to further refining treatment strategies in this context.

In early-stage NSCLC, ctDNA is not routinely recommended, and its clinical utility remains under investigation. Its potential lies in its value as a prognostic biomarker for assessing the risk of recurrence. Many patients experience recurrence following surgery, and post-surgical cfDNA monitoring, offers an opportunity for earlier detection, enabling prompt intervention or closer follow-up before radiological signs of recurrence appear. In NSCLC-EGFR+ patients, the presence or absence of ctDNA before and after surgery is a strong prognostic indicator, with ctDNA negativity or clearance being associated with better DFS. Notably, most patients with undetectable minimal residual disease through longitudinal ctDNA monitoring did not experience recurrence [61].

In the screening context, there is a critical need for non-invasive screening methods to improve early detection of LC in high-risk and general populations. Research into circulating biomarkers, including proteins, autoantibodies, gene expression profiles, and microRNAs, shows promise for early detection and distinguishing cancer from non-cancerous nodules [62]. Currently, none of these approaches are approved for clinical use. Blood tests utilizing LB techniques can identify various surgically resectable cancers by detecting cfDNA mutations in key oncogenes, such as EGFR, and analyzing circulating proteins. However, their sensitivity for detecting these alterations in early-stage lung cancer (LC) remains relatively low [63].

Combining different biomarkers and techniques with clinical features offers the most promising strategy to increase detection sensitivity and specificity. A recent study demonstrated that integrating cfDNA fragmentation analysis with clinical risk factors and STM concentration (CEA) significantly enhanced LC detection across various stages and subtypes, including 91 % of stage I/II and 96 % of stage III/IV, at 80 % specificity [64].

Correlation between EGFR status and serum tumor markers

Tumoral circulome represents the analysis of genetic biomarkers (cfDNA) in addition to other tumor-derived biomarkers, such as proteins and other molecules released into circulation. This novel multi-omic approach is gaining interest in the precision oncology field [65]. STMs are glycoproteins secreted into the bloodstream from the tumor cells or tumor microenvironment. STMs are routinely measured in clinical laboratories for the management of epithelial neoplasms; however, their clinical utility in LC remains controversial. Prospective studies have shown that combining six STMs (CEA, CYFRA 21.1, CA 15.3, SCC, NSE, and ProGRP) with clinical parameters represents the most accurate approach for detecting LC in symptomatic patients [66]. Additionally, recent findings indicate that integrating STM with ctDNA analysis is a promising strategy for early diagnosis of LC [63], [67]. The combination of STMs and ctDNA has predictive value, with elevated levels being linked to poorer prognosis and shorter progression-free survival following EGFR-TKI treatment [68]. However, there is currently insufficient evidence to include none of these blood biomarkers in clinical guidelines for LC diagnosis.

Regarding molecular features, several STMs have been linked to EGFR mutation status or other molecular alterations, as summarized in Table 2. Most studies have found an association between elevated levels of CEA and CA 15-3, and/or decreased levels of CYFRA 21-1 and SCC, with EGFR mutations [69], [70], [71], [72], [73], [74], [75]. This aligns with evidence suggesting that specific markers predict certain histology: CEA and CA 15-3 for adenocarcinoma, where EGFR mutations are more prevalent, and CYFRA 21-1 and SCC for squamous cell carcinoma [66]. By contrast, other studies found no significant differences between EGFR status and STM [76], [77], [78]; particularly in early-stage patients where expression in circulation and, therefore, the sensitivity is low [70]. During follow-up of EGFR-TKI therapy, dynamic STMs may predict the effectiveness of treatment [79], including the presence of secondary T790M mutation.

All this evidence suggests that STMs could be integrated with other clinical factors to predict EGFR status and combined ctDNA + STMs monitoring can offer insights into therapy response, enhancing personalized treatment decisions. However, there is considerable variability among studies highlighting the need for ongoing research into their clinical utility.

A promising strategy can be driven by recent advances in artificial intelligence and machine learning, enabling the efficient combination of markers into multi-marker models, integrating clinical and imaging data. This strategy facilitates the analysis of large volumes of clinical and imaging data to identify patterns that can predict the presence and progression of cancer [81], [82]. These advancements pave the way for the development of clinical decision support tools and open up new opportunities for more effective, personalized cancer treatments.

Conclusions

This review provides practical guidelines for integrating ctDNA analysis into real-world clinical settings. EGFR mutations are critical therapeutic targets in advanced NSCLC, and the limited availability of tissue samples highlights the need for non-invasive approaches, such as ctDNA analysis in plasma. Current guidelines recommend ctDNA genotyping as a surrogate for tissue genotyping when tissue samples are unavailable, with decision algorithms guiding the choice between plasma- or tissue-based testing strategies.

Clinical laboratories play a crucial role in ensuring reliable molecular analysis of EGFR in ctDNA, addressing both pre-analytical and post-analytical phases. While targeted methods remain the most common approach in clinical practice due to their robustness and rapid TAT, NGS based methods are becoming increasingly important for more comprehensive mutational profiling. This shift is driven by the evolving landscape of targeted therapies and the growing need for comprehensive testing to identify resistance mechanisms.

Beyond its established role in advanced NSCLC, ctDNA analysis is being explored for earlier stages of the disease. The potential use of EGFR-TKI therapies in both neoadjuvant and adjuvant settings highlights the need for non-invasive EGFR testing in early-stage NSCLC to guide targeted treatment. While ctDNA analysis has shown prognostic utility post-surgery, its application in early stages and LC screening is currently limited by sensitivity issues. Tumoral circulome analysis, which integrates molecular alterations, cfDNA fragmentation, proteins and other biomolecules released by tumor cells into the circulation, is emerging as a promising non-invasive method. This approach could significantly enhance early detection sensitivity and precision, as well as improve personalized treatment strategies.