Circulating lung cancer biomarkers: From translational research to clinical practice

Abstract

Fundamental studies on biomarkers as well as developed assays for their detection can provide valuable information facilitating clinical decisions. For patients with lung cancer, there are established circulating biomarkers such as serum progastrin-releasing peptide (ProGRP), neuron-specific enolase (NSE), squamous cell carcinoma antigen (SCC-Ag), carcinoembryonic antigen (CEA), and cytokeratin-19 fragment (CYFRA21-1). There are also molecular biomarkers for targeted therapy such as epidermal growth factor receptor (EGFR) gene, anaplastic lymphoma kinase (ALK) gene, KRAS gene, and BRAF gene. However, there is still an unmet need for biomarkers that can be used for early detection and predict treatment response and survival. In this review, we describe the lung cancer biomarkers that are currently being used in clinical practice. We also discuss emerging preclinical and clinical studies on new biomarkers such as omics-based biomarkers for their potential clinical use to detect, predict, or monitor subtypes of lung cancer. Additionally, between-method differences in tumor markers warrant further development and improvement of the standardization and harmonization for each assay.

1 Introduction

Lung cancer is the leading cause of cancer-related deaths worldwide, with an estimated 238,340 new cases and 127,070 deaths in 2023 in the United States. [1] Non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) are the most common subtypes of lung cancer. [1] Although the morbidity and mortality rates of lung cancer vary by region and ethnicity, the overall prognosis of lung cancer is dismal. [2] For patients with lung cancer diagnosed at stage I-II, the overall survival rate is approximately 50%, while for those diagnosed at regional stage, it is approximately 20%, and for distant stage patients, it decreases to 4%. [3, 4] Clinically, the application of low-dose computed tomography (LDCT) has improved the detection rate of early-stage lung cancer. [5] However, the limitations of LDCT are its high false-positive rate (FPR), radiation exposure, and high cost. [6, 7] Therefore, establishing an efficient and cost-effective way to detect lung cancer remains a clinical challenge.

While traditional tumor biomarkers continue to play a role in patient management, the role of biomarkers for lung cancer has shown its advantage in molecularly targeted therapy and immune-checkpoint inhibitors for patients with advanced-stage lung cancer but is not satisfactory for early lung cancer screening. Emerging liquid biopsies, including protein or molecular biomarkers, provide an option to improve lung cancer diagnostics. In this paper, we discuss the clinical utility of established tumor markers for lung cancer and new candidates for further discovery. Importantly, standardization of these tumor markers in the clinical laboratory would be able to facilitate their value in clinical practice.

2 Established circulating protein biomarkers and their clinical use

2.1 Circulating protein biomarkers and their correlations with subtypes of lung cancer

While LDCT remains a sensitive screening tool for lung cancer diagnosis, circulating tumor markers specific to lung cancer and its histological type have been evaluated in clinical studies for their sensitivity and specificity (reviewed by Huang et al.). [8] Serum carcinoembryonic antigen (CEA), carbohydrate antigen 15.3 (CA15.3), squamous cell carcinoma– associated antigen (SCC-Ag), cytokeratin-19 fragment (CYFRA 21-1), neuron-specific enolase (NSE), and pro– gastrin-releasing peptide (ProGRP) have shown their value in facilitating patient monitoring. CEA, NSE and ProGRP are characterized as biomarkers for SCLC. CYFRA 21-1 and SCC-Ag are characterized as biomarkers for NSCLC. Furthermore, CEA is sensitive for lung adenocarcinoma (LUAD), while CYFRA 21-1 and SCC-Ag are sensitive for lung squamous cell carcinoma (LUSC). [8] An increase in serum cancer antigen 125 (CA125) is also seen in lung cancer, and its value in the diagnosis and prognosis of lung cancer warrants further study. [9] Interestingly, in a recent report based on a large-scale proteome analysis to identify circulating proteins biomarkers for risk of imminent lung cancer diagnosis demonstrates that 36 proteins including CA125/MUC-16 and CEACAM5/CEA were associated with imminent- but yet-to-be diagnosed lung cancer. [10]

2.2 Protein panels

To date, a panel of protein tumor markers demonstrates a better diagnostic or prognostic value than a single tumor marker. For example, a diagnostic prediction model based on CEA, CYFRA21-1 and NSE can discriminate patients with differentiated lung cancer from healthy controls and benign lung diseases, which are better than that of a single tumor marker. [11] Additionally, the prognostic prediction model also had good performance in predicting overall survival in lung cancer patients. [11] In another study, a combination of serum CEA, CA15.3, SCC-Ag, CYFRA 21-1, NSE, and ProGRP was assessed in a large cohort with the clinical suspicion of lung cancer, which had a sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of 88.5, 82, 83.7, and 87.3%, respectively. [12] However, the NPV of this panel in patients with small nodules less than 1 cm was 91.4%, while in those with intermediate-sized nodules (1-3 cm), it was 60.8%. Thus, this panel may add value for lung cancer risk stratification but not for diagnostics. These studies demonstrate the combined value of serum tumor markers. In addition, the role of tumor markers in the differentiation of lung cancer subtypes was also explored. NSE and proGRP had higher expression levels in SCLC than in other histological types. [12, 13] For lung squamous cell carcinoma (LUSC), CYFRA21-1 shows a high sensitivity, while SCC-Ag shows a high specificity but less sensitivity. [14, 15] In a recent study, a combination of CYFRA21-1, CEA, and human epididymis protein 4 (HE4) had a better performance on NSCLC, while a combination of ProGRP, NSE and HE4 was the best pattern for SCLC. [16] However, these serum biomarkers show less sensitivity for lung cancer diagnostics. On the one hand, the elevated expression level of tumor markers is seen not only in lung cancer but also in other malignancies. However, elevated levels of tumor markers are also seen in healthy people or noncancerous diseases, which is critical and limits their application in early lung cancer screening. [8]

Although the sensitivity of serum tumor markers used in lung cancer screening is generally low, the application of these tumor markers for the prediction of treatment response and prognosis may add value in disease monitoring. [8, 17] Notably, the reduction in serum CYFRA21-1 and NSE levels was associated with the disease control rate in patients with NSCLC who were treated with nivolumab. [18] This finding indicates that more clinical studies are necessary to further evaluate serum tumor markers for the prediction of the treatment response to immunotherapy.

3 Emerging biomarkers for liquid biopsy

Tissue tumor markers can be used for targeted therapy and immunotherapy when tumor tissues are available from biopsy at the time of diagnosis or during the whole treatment period. [19, 20] Examples include molecular tests of EGFR mutations, ALK gene rearrangements, ROS1, BRAF, NTRK, MET and RET for targeted therapies, tumor PD-L1 expression and tumor mutational burden for immune checkpoint inhibitors. For more detailed information, please refer to a recent review by Thai et al. [19] When the tumor tissues are difficult to obtain, circulating tumor DNA (ctDNA) extracted from the blood would be the compliment to tissue biopsy to identify patients who may benefit from targeted therapy or immunotherapy (reviewed by Duffy et al.). [21] The shorter half-life of ctDNA in the circulation made them more sensitive in reflecting tumor recurrence or response to treatments compared to the above established protein tumor markers. [21] Detecting ctDNA minimal residual disease (MRD) has shown value in improving survival outcomes (reviewed by Pellini et al.). [22] An ideal biomarker developed for clinical practice is simple, fast, cost effective, high sensitivity and specificity, and has reproducible detection. There are also limitations of current ctDNA assays due to their relatively long turn-around time and poor standardization. [21] Future large, randomized studies are still necessary to validate the clinical utility of ctDNA in different clinical settings.

Extracellular vesicles (EVs), including exosomes, microvesicles and apoptotic bodies, are membrane-delimited particles that are released from various types of cells. EVs derived from cancer cells carry information from their parental cells to receiving cells and play important roles in modulating the tumor microenvironment and promoting tumor progression; thus, EVs are attractive candidates for liquid biopsy. The cargo contents and corresponding functions of EVs were also studied in lung cancer (reviewed by Li et al.). [23] Accumulating evidence from mechanism-based studies of lung cancer cell-derived EVs has shown that cargo contents, including protein, DNA and various types of RNA, such as mRNA, lncRNA and microRNA, have the potential to serve as biomarkers for tumor diagnosis and prognosis. For example, the combination of serum exosomal PLA2G10 mRNA and protein showed better diagnostic power than CEA, CA125 and NSE for NSCLC in a small cohort study. [24] Three circular RNAs (circRNAs) in serum exosomes (circ_0047921, circ_0056285, and circ_0007761) have good discrimination ability for early-stage NSCLC from healthy controls, chronic obstructive pulmonary disease, or tuberculosis. [25] Plasma exosomal miR-181-5p, miR-30a-3p, miR-30e-3p and miR-361-5p were found to be adenocarcinoma specific, while miR-10b-5p, miR-15b-5p, and miR-320b were SCC-Ag specific. [26] These findings warrant further investigations in a large cohort validation study. To establish clinically useful biomarkers based on EVs, more reproducible and less labor- or equipment-intensive methods for EV isolation and characterization need to be developed. To that end, nanodevices and microfluidics-based isolation or detection methods for EVs are currently an active field of research, and promising results have been reported (reviewed by Li et al.). [27]

4 Implementation for early-stage lung cancer diagnostics

Early-stage lung cancer diagnosis remains a major challenge. In a recent meta-analysis, the diagnostic performances of ctDNA (Sensitivity 0.50, specificity 0.98, AUC = 0.64), DNA methylation (Sensitivity 0.72, specificity 0.82, AUC = 0.84) and CTC (Sensitivity 0.68, specificity 0.85, AUC = 0.82) in the stage I disease were calculated. [28] Additionally, the concentrations of circulating cell-free DNA (cfDNA) were lower in stage I NSCLC compared to stages II and III (10.28 ng/ml vs. 12.72 ng/mal, 12.34 ng/ml, respectively). [28] As the authors concluded, the diagnostic accuracy would be limited by the lower concentration of cfDNA in plasma and influenced by clonal hematopoiesis of undetermined potential during aging leading to a fasle-postive ctDNA detection rates.

Recent studies highlight that metabolomics in cooperation with artificial intelligence detectors has the potential for the early detection of lung cancer (reviewed by Mariën et al.). [29] With the advent of single-cell RNA sequencing (scRNA-seq), early lung cancer was featured with dysregulation of lipid metabolism in a study, and further detection on plasma untargeted lipidomics identified nine lipid-based models that reached more than 90.00% sensitivity and 92.00% specificity in two validation cohorts. [30] In addition, a study on salivary metabolomics identified differential metabolites that can distinguish early lung cancer patients from healthy controls with a sensitivity of 97.2% and a specificity of 92%. [31] Importantly, the power of protein- or metabolite-based disease classifiers has the potential for rapid translation into clinical practice.

5 Achieving consensus for result comparison by standardization of tumor markers

Achieving a qualified analytical result with comparable analyte measurements is the main goal of clinical laboratories. [32] However, most tumor markers being used are non-FDA or other regulatory agency approved (are indeed physician-prescribed laboratory-developed tests (LDTs)). Non-FDA approved LDT is referred to as an in vitro diagnostic device (IVD) that has risen quickly in recent years. Moreover, the results are varied from lab to lab and even FDA-approved LDT conducted in a Clinical Laboratory Improvement Amendment (CLIA)-certified laboratory are not consistent and comparable. The use of different standards, calibrators, reagents, antibodies (even targeting different epitopes), etc., is likely attributed to the variation in the results. In this context, standardization of tumor markers by improving internal quality control (IQC) processes and participating in external quality assessment (EQA) programs or proficiency testing (PT) are important variations to improve the inconsistency of results across different instruments or lab procedures. [33, 34]

For serum tumor markers, the lack of international standards remains an obstacle to achieving harmonization when the method changes or different laboratory measurement procedures are used. [35, 36] Among those established serum tumor markers for lung cancer, there is an established WHO International Standard (IS) and International Reference Reagents (IRR) for CEA but not the others. [36] In a study of the EQA assessment of eight tumor markers (AFP, CEA, t-PSA, CA125, CA153, CA199) in China, there was an increasing trend of improved testing performance with an overall decrease in the robust coefficient of variability (CV) from 2006 to 2013. [33] However, the EQA of CEA did not show further improvement from 2010 to 2013. [33] Thus, to avoid bias, continuous test performance in the same laboratory would be helpful to monitor patients. Making efforts to standardize or harmonize tumor markers requires multiple stakeholders, such as governmental agencies, regulatory agencies, the IVD industry, labs, and professional organizations, to be involved and engaged. [37]