Exploring the Potential of Non-Coding RNAs as Liquid Biopsy Biomarkers for Lung Cancer Screening: A Literature Review

Simple Summary Low-dose CT scan screening will be widely implemented on a large-scale base, aiming to reduce lung cancer related mortality in the high risk smoking population as already reported in multiple trials, in several countries. Recent evidence has suggested that the identification of liquid biopsy biomarkers may improve its accuracy in lung cancer early detection, reducing the false positive rate as well as overdiagnosis issues and potentially addressing one of the major obstacles in the implementation of Low-dose CT scan alone in this context. RNAs, particularly non-coding RNAs, are for sure the most studied and promising circulating biomarkers in this setting. Abstract Lung cancer represent the leading cause of cancer mortality, so several efforts have been focused on the development of a screening program. To address the issue of high overdiagnosis and false positive rates associated to LDCT-based screening, there is a need for new diagnostic biomarkers, with liquid biopsy ncRNAs detection emerging as a promising approach. In this scenario, this work provides an updated summary of the literature evidence about the role of non-coding RNAs in lung cancer screening. A literature search on PubMed was performed including studies which investigated liquid biopsy non-coding RNAs biomarker lung cancer patients and a control cohort. Micro RNAs were the most widely studied biomarkers in this setting but some preliminary evidence was found also for other non-coding RNAs, suggesting that a multi-biomarker based liquid biopsy approach could enhance their efficacy in the screening context. However, further studies are needed in order to optimize detection techniques as well as diagnostic accuracy before introducing novel biomarkers in the early diagnosis setting.


Introduction
Lung cancer remains nowadays the leading cause of cancer mortality accounting for 12% of overall cancer deaths worldwide.This is certainly linked to the peculiar biological behavior of this disease as well as to a significant diagnostic delay leading to advancedstage diagnoses in about 50% of cases.For this reason, several efforts over the last years have been focused on the development of effective secondary prevention strategies, with different studies and metanalysis [1] showing that low-dose computed tomography (LDCT) is able to reduce lung cancer-related mortality in high-risk smoking subjects.
In detail, the National Lung Cancer Screening Trial (NLST) and The Dutch-Belgian Randomized Lung Cancer Screening Trial (NELSON) randomized studies demonstrated a significant reduction (around 20%) of lung cancer-related mortality in smoking populations undergoing LDCT as compared to either thorax RX or clinical observation [2,3], leading to the introduction of lung cancer screening in the United States since 2013.Among the different barriers limiting LDCT screening implementation in Europe, the high rate of overdiagnosis and false positive cases represent a relevant unmet need significantly impacting the subject management in real world scenarios.In addition to that, the potential exposure to the imaging radiation and the risk of overtreatment for indolent lung nodules further reduce subjects' compliance to the LDCT screening.In this context, the integration of tumor biomarkers through liquid biopsy could improve the diagnostic accuracy of LDCT screening in a non-invasive manner aiming to identify the high-risk population requiring further investigation, personalizing screening intervals and likely increasing subjects' compliance to the screening procedures.Furthermore, the possibility to perform a liquid biopsy in the peripheral hospitals near rural areas could allow to reach a larger smoking population who is usually recalcitrant to the LDCT, thus increasing the access rate to lung cancer screening in a different way and promoting personalized approaches.
A liquid biopsy is able to identify circulating tumor biomarkers that can be considered surrogates of the primary tumor as circulating tumor cells (CTCs), circulating tumor DNA (ctDNA), microRNA (miRNA), and exosomes.Liquid biopsies are already playing an important role in the clinical management of metastatic lung cancer patients through the evaluation of the tumor molecular profiling by ctDNA analysis, while also progressively extending to the early-stage disease in terms of minimal residual disease monitoring as well as cancer interception [4].
The role of CTCs in lung cancer screening has been investigated in several trials since cell dissemination is a relatively early event in tumor progression.These trials showed that CTCs detected in high-risk patients are able to anticipate the diagnosis of lung cancer, even years earlier than CT scans [5].One of the main problems with using CTCs as a biomarker is represented by their rarity in peripheral blood.
Several studies have investigated the role of CtDNA and cell-free DNA (cfDNA) in screening encountering a fundamental issue: the concentration of cfDNA correlates with the disease burden of the tumor that is very low in early-stage disease, making it difficult to isolate.Despite this limitation, different studies have investigated the role of cfDNA in early diagnosis.To distinguish tumor from non-tumor cfDNA, they have looked at its concentration, genetic changes, or methylation as possible biomarkers [6].In 2020 a study proposed the use of a cfDNA-based machine-learning method to improve the specificity of LDCT screening with interesting results [7].
Another interesting finding is the development of a blood-based multi-cancer early detection (MCED) test targeting a screening population.The test has been developed for the early detection of more than 50 types of cancer.The MCED test analyzed the methylation patterns of CtDNA and demonstrated high specificity (99.1%) and a positive predicted value of approximately 40% [8].
Only a small fraction (approximately 3%) of the genetic transcript is able to encode proteins, while the remaining part is defined as non-coding RNAs (ncRNAs).The definition of "coding" encompasses RNAs that encode proteins from DNA-derived information, such as mRNAs.Noncoding RNAs have a different role since they act as cellular regulators of gene expression at different transcriptional, post-transcriptional, and epigenetic levels.A few exceptions to this definition include some ncRNAs binding ribosomes encoding peptides exerting a modulator function on cellular activities [9].It has become clear that ncRNAs also play an important role in the communication processes between cancer cells and tumor microenvironment and are crucial for regulating tumor growth [9].In recent years the knowledge about ncRNAs roles in cancer process has exponentially grown [10], including diagnostic, prognostic, predictive, and therapeutical applications across different cancers and settings [11].
ncRNAs can be classified into two macro-categories: housekeeping ncRNAs and regulatory ncRNAs.Housekeeping ncRNAs regulate basic cellular functions and are ubiquitously expressed.Regulatory ncRNAs play a pivotal role in gene expression regulation and protein translation both at transcriptional and post-transcriptional levels.Increasing evidence points out their role in cancer development, regulation, and growth, making them a precious tool for cancer management at different stages [12,13].
The ncRNAs can be found in different biofluids either as freely or encapsulated in extracellular vesicles [12,14], making them potentially stable biomarkers for clinical use, which have been explored also within LC screening clinical trials [13,15,16].
To date, among the different liquid biopsy biomarkers under clinical investigation in the lung cancer early detection setting, ncRNAs are for sure one of the most promising biomarkers to be implemented in the context of lung cancer screening.For this reason, this review will specifically focus on ncRNA role, describing biological function, available evidence, and clinical trials ongoing in this emerging setting.

Methods of Literature Search
An extensive literature search was performed on PubMed, using as cut-off date of 16 March 2023.Keywords included: noncoding RNA, lung cancer, screening.A total of 2538 articles were found and screened for eligibility, taking into account three major criteria of inclusion: (1) Single ncRNAs or ncRNA-based genomic signature; (2) biomarker analysis performed on blood or other biofluids; and (3) biomarker involvement in non-small cell lung cancer (NSCLC) screening or early diagnosis.
Both prospective and retrospective studies were considered.Only studies performed on humans were included, but they could include an in vitro/in vivo validation part.Only studies that involved a control group were considered.The control group could also have other pulmonary conditions or pulmonary nodules (PNs) that were retrospectively prospectively identified as benign lesions.Published abstracts without associated full articles were excluded from the analysis.Three independent reviewers collected data from the included articles, and another one subsequently reviewed all of the information.
To draw a clear overview of all clinical trials concerning LC screening that included a liquid biopsy part, we also performed research on clinicaltrials.govusing the following keywords: lung cancer and screening.A total of 601 trials were found and, among them, only those involving LDCT screening and encompassing biofluids collection for biomarkers research were included.

Micro RNAs (miRNAs)
MiRNAs are fragments of single-stranded non-coding RNA with a length of approximately 20 ribonucleotides.Since they remain stable in biofluids, unlike other free RNA molecules, they can be detected in both serum and plasma.They regulate gene expression at the post-transcriptional level and are involved in the regulation of cell proliferation and apoptosis.In fact, their targets include oncogenes and tumor suppressor genes and their dysregulation can lead to malignant cell transformation across different tumor types [6,17].
miRNAs are one of the pivotal biomarkers explored in phase III trials of lung cancer screening, and our literature search identified studies that used both multi-miRNA signatures (>2 miRNA) and single miRNA approach for early NSCLC detection.In detail both the miR-test, a serum-based 13 miRNA signature, and the micro-RNA signature classifier (MSC), a plasma-based 24 miRNA risk score, showed very promising data for clinical use [18].
Overall, 32 studies evaluating the expression of multi-miRNA in early-stage NSCLC patients compared to healthy controls have been identified through our PubMed research.The identified signatures ranged from 2 to 24 miRNAs and were all validated on biological fluids that could be used for liquid biopsies purposes (Table 1).Almost all of the evaluated studies used plasma or serum for miRNA detection and the most used sequencing tech-nique was quantitative real-time PCR (qRT-PCR).Among all of these studies matching our inclusion criteria, only six included a CT scan integrated to a specific miRNA signature for LC diagnosis.
Sozzi G. et al. [19] conducted a large retrospective analysis in this setting analyzing plasma samples from 939 participants to the Italian randomized MILD LC screening study (69 patients diagnosed with LC and 870 healthy individuals) using a 24-miRNA classifier.They identified a higher sensitivity (87%) and a similar specificity (81%) for LC detection, compared to LDCT alone (79% and 81%, respectively) with a false-positive rate of 3.7% vs. 19.4% with and without MSC integration.
In another prospective analysis conducted in the BIOMILD study, patients were [16] stratified into four different subgroups based on a miRNA signature classifier (MSC): 2 MSC+ with or without a positive CT scan and 2 MSC− with or without a positive CT scan.Individuals with a positive CT scan and an MSC− had a lower incidence of LC and individuals with both CT and MSC negative had a lower overall LC incidence at four years, interval cancer, stage I, and advanced stages diagnosis, as well as the lowest LC mortality rate at five years as compared to all other subgroups.So, the authors found out that the combined use of LDCT and MSC at baseline was able to predict individual LC incidence and mortality, with a major effect of MSC for LDCT-positive individuals.
Shun J et al. [20] applied a 3 miRNAs (miRs-21, 210, and 486-5p) plasma signature on healthy subjects, patients with benign pulmonary nodules (PNs), and malignant PNs.This approach achieved an area under the curve (AUC) of 0.855 for lung cancer detection in the testing cohort.The panel of the three mi RNA was then validated in an independent cohort of 156 patients who had solitary PNs, and this miRNAs signature produced a 76.32% sensitivity and 85% specificity in differentiating malignant from benign solitary PNs.
The same group [21] screened 10 miRNA differently expressed by LC and healthy smokers sputum and built a logistic regression model on a 2 miRNA combination (miR-31 and miR-210).This model generated an AUC of 0.83 in distinguishing LC patients from healthy smokers, moreover, the combination of CT scans and the 2 miRNA combination achieved an AUC of 0.95.In the validation cohort, the AUC dropped to 0.79, but the combination of miRNA and CT scans improved the specificity and sensitivity compared to CT scan alone.
Another plasma-based approach was conducted by Zheng D. et al. [22], who evaluated circulating small extracellular vesicle (EV) microRNAs in 208 patients with CT-detected PNs.Five miRNAs (let-7b-3p, miR-125b-5p, miR-150-5p, miR-101-3p, and miR-3168), included within the CirsEV-miR model were firstly tested in a small training cohort of 47 patients and then validated in a testing-cohort of 62 patients achieving an AUC for lung cancer detection of 0.920 and 0.760, respectively.This model was then validated in an external cohort of 92 patients (20 patients with benign PNs and 79 with malignant PNs), reaching an AUC of 0.781.
EVs and miRNAs were also tested in this setting by using NGS analysis [23].They analyzed plasma from patients who had Lung-RADS4 PNs then confirmed as LC, versus over-diagnosed Lung-RADS4 PNs or high-risk Lung-RADS2 screening controls.They identified different expression levels of let-7b-5p, miR-184, and miR-22-3p as biomarkers for potentially discriminating cancer patients from high-risk controls.The multiple logistic regression analyses of the 3 EV miRNAs showed a combined ROC AUC value of 92.4%.
Other pulmonary pathological conditions, such as chronic obstructive pulmonary disease (COPD) and asthma, were included in some of the other studies and represent an interesting approach to eliminate some biases that could be created by smoking-related or pre-existing pulmonary conditions in the implementation of liquid biopsies within lung screening programs.Halvorsen A.R. et al. [24] used serum also from 16 COPD subjects to build their miRNA signature for their prediction model, showing a good performance in discriminating lung cancer from the control groups (AUC 0.89).Yang X et al. [25] used also serum of COPD, in their logistic regression model obtaining not only good performance in discriminating lung cancer patients from controls, but also a higher accuracy for ade-nocarcinoma (AC) patients rather than squamous cell carcinoma (SCC).Zaporozhchenko I.A. et al. [26] analyzed 179 miRNA in plasma samples obtained from patients with a non-cancerous lung disease (hyper-or metaplastic endobronchitis (EB)) and a cancer-free group of healthy volunteers.They found a 14 miRNA signature discriminating LC group and controls, but interestingly the performance of the model was largely unaffected by the presence of samples from patients with endobronchitis.A similar approach was led by Nadal E. et al. [27] analyzing also serum samples of patients with COPD and identifying a 4 miRNA signature for LC diagnosis, clustering also the discovery set into 2 different groups, characterized by different metastasis-free survival (MFS) and overall survival (OS).Fehlmann T. et al. [15], instead led a large multicenter retrospective cohort study, analyzing 3046 samples of LC patients (including NSCLC and small cell lung cancer, SCLC), and patients with other lung conditions (mostly COPD).A 14-miRNA signature derived from the training set was used to distinguish patients with lung cancer from patients with nontumor lung diseases both in the testing set (accuracy of 92.5%, sensitivity of 96.4%, and specificity of 88.6%) and in the validation set (accuracy of 95.9%, sensitivity of 76.3%, and specificity of 97.5%).
Some of the studies listed in Table 1, tested another interesting use of LB-based approach in the LC early-diagnosis setting which is LC histological subtype prediction.Lu S. et al. [28] conducted a miRNA analysis on plasma samples of a large cohort of patients (1132 samples, including healthy individuals and patients with NSCLC or SCLC) collected from five medical centers, developing a plasma miRNA panel capable to discriminate LC patients from healthy individuals, and SCLC from NSCLC (AUC 0.878 and 0.869 for training and validation cohort, respectively).Instead, a study by Powrózek T [29], et al., showed that miR-944 had a high diagnostic accuracy for operable squamous cell carcinoma detection (AUC 0.982), whereas miR-3662 for operable adenocarcinoma diagnosis (AUC 0.926).Jiang Y. et al.
[30] used an NGS-based approach in analyzing plasma-derived EVs from healthy individuals, patients with early-stage SCLC, and patients with early-stage NSCLC, finding out that miRNA-483-3p derived from plasma EVs could be a potential biomarker for early-stage SCLC diagnosis, while both miRNA-152-3p and miRNA-1277-5p could be used for early-stage NSCLC diagnosis.

Long Non Coding RNAs (lnc-RNAs)
LncRNAs look quite promising since they have been demonstrated to be stable in biofluids [88,89] and to be frequently dysregulated in NSCLC pathogenesis [90].According to our literature search, a multi-lncRNA approach was conducted across 4 studies.Gupta C et al. [91], analyzed lncRNAs in the sputum of LC patients and cancer-free individuals demonstrating a good ability in discriminating the two groups through a panel containing SNHG1, H19, and HOTAIR (AUC 0.90).The second multi-lncRNA approach was conducted by Yuan S. et al. [92], who collected 528 plasma samples of patients with either LC, other lung conditions, or healthy volunteers.They identified a 4-lncRNA panel (RMRP, NEAT1, TUG1, and MALAT1) with a high diagnostic value for NSCLC (AUC 0.85 for AC and 0.93 for SCC in the expansion cohort).An alternative approach conducted by Li X et al. [93], aimed to search for lncRNAs in tumor-educated platelet (TEP), where a combined use of linc-GTF2H2-1, RP3-466P17.2, and lnc-ST8SIA4-12 achieved an AUC of 0.895.Ultimately in the analysis by Kamel L.M. Et al [94], the combination of GAS5 and SOX2OT showed an AUC of 0.95 for distinguishing LC patients from healthy controls.

Circular-RNAs (Circ-RNAs)
Circ-RNAs can be freely detected in biofluids (plasma and saliva) as well as in exosomes [109], and are aberrantly expressed in early-stage lung adenocarcinoma, making them a good biomarker for LC early detection [110].Even though Yang X. et al. [111] metaanalysis, comparing circRNAs' expression in tissue and plasma/serum samples, showed that the diagnostic accuracy of tissue was higher (AUC 0.85 vs. 0.79), other evidence points out in the opposite direction.Falin C. et al. [112] validated a combination of circRNAs (hsa_circ_0001492, hsa_circ_0001346, hsa_circ_0000690, and hsa_circ_0001439) that were significantly upregulated in plasma exosomes of AC patients as compared to healthy controls.Hang D. et al. [113] adopted RNA sequencing (RNA-seq) and qRT-PCR approaches to explore cancer-related circRNAs expression, showing that circFARSA was increased in cancerous tissues, and was more abundant in the plasma of LC patients than controls.Other three circRNAs were tested as potential biomarkers for LC early detection with liquid biopsy showing a good diagnostic accuracy: hsa_circ_0023179 [114], hsa_circ_0006423 [115] and circFOXP1 [116].

Other Non-Coding RNAs and Combined Approaches
For what concerns small-nuclear RNAs we found three studies that tested the differences between LC patients and controls.Köhler J. et al. [117], determined RNU2-1f in the serum of patients with LC, chronic lung disease, and healthy controls, showing the ability to discriminate the LC group from others (AUC of 0.91).Moreover, the two isoforms of RNU2 (RNU2-1 and RNU2-2) were also tested in another study by Mazières J et al. [118], who demonstrated that miR-U2-1 was able to discriminate between patients with COPD and patients with COPD and lung cancer (AUC of 0.866).Dong et al. [119] used a tumor-platelet educated approach, finding out that TEP U1, U2, U5 were decreased in early-stage lung cancer patients compared with those in healthy subjects.
For what concerns piwiRNAs we found a study by Li J. et al. [120] demonstrating that piR-hsa-26925 and piR-hsa-5444 had a significantly higher level in serum exosome samples of AC patients than healthy controls.
No studies matching our inclusion criteria were found about ribosomal RNA (rRNA), transfer RNA (tRNA), and small nucleolar-RNAs (sno-RNAs) in the context of lung cancer screening.
Few studies were conducted using a combined ncRNAs approach, according to our inclusion criteria.In detail Peng H et al. [121] constructed a miRNA and MALAT1 noncoding RNA panel showing a good performance also in detecting stages I/II/III NSCLC.A panel of seven small ncRNA pair ratios was tested by Dou Y. et al. [122] and could differentiate AC patients from other lung diseases of high-risk controls.

Ongoing Clinical Trials on Liquid Biopsy in Lung Cancer Screening
We also performed a study of ongoing clinical trials on clinicaltrials.govusing the keywords "lung cancer" and "screening".The data collection was completed on 16 March 2023 and identified a total of 601 ongoing trials related to lung cancer screening.Out of the 601 clinical trials identified, we selected 55 trials incorporating liquid biopsy and the analysis of biological samples for the detection of predictive biomarkers in the setting of LC screening (Table 2).The selected trials did not exclusively include healthy individuals at high risk of developing lung cancer, but also those with lung nodules, CT suspicion or pathologically confirmed lung cancer, as well as other benign lung diseases.Furthermore, a particularly noteworthy study included only never-smokers (defined as individuals with a lifetime exposure of less than 100 cigarettes) and Asian women (NCT05164757).
Among the 55 clinical trials shortlisted based on our inclusion criteria, 25 of them involved the use of chest CT or LDCT scans as a diagnostic tool for lung cancer screening.The HANSE trial (NCT04913155) also investigated other indicators such as coronary calcium score and emphysema score.One of the selected studies involved the use of chest MRI to assess the concordance of imaging features of nodules between LDCT and MRI in the study population (NCT05699213).Concluding, a small portion of these studies incorporates pulmonary function testing within their research protocols.
These selected trials also involved the collection and analysis of various biological samples to identify possible biomarkers for the early detection of lung cancer.Specifically, they included blood samples, different airways samples (bronchoalveolar lavage, BAL, bronchial biopsy and brushing samples, nasal swab, and brush samples), sputum samples, buccal swab samples, urine samples, and feces samples.
Among the 55 clinical trials that met our inclusion criteria, blood samples were collected in 52 trials, but only 33 of these explicitly state the specific biomarkers that were intended to be analyzed, including miRNA, epigenetic biomarkers, circulating free DNA (cfDNA), circulating tumor DNA (ctDNA), circulating tumor cells (CTC), Associated Macrophage-Like cells (CAMLs), exosome antigens, methylation changes in peripheral blood mononuclear cells (PBMC) and circulating tumor DNA, RNA integrity number (RIN), protein signatures, DNA methylation, whole-genome methylation, tumor antibodies, circulating nucleic acids, proteins, and genetic variation single nucleotide polymorphisms (SNPs), as well as DNA and RNA for germline analysis and whole-exome sequencing (WES).Specifically, only eight trials have a clear focus on the identification and analysis of miRNA.Moreover, 3 out of these 52 clinical trials involve the storage of blood samples in biobanks for potential future studies.
A small part of the clinical trials that met our inclusion criteria have already published results.We have already discussed the results of The Multicentric Italian Lung Detection (MILD) study, a prospective randomized controlled screening trial that compared the diagnostic performance of two different LDCT screening intervals in high-risk smoking populations.After a median active screening period of 6.2 years, the MILD trial concluded that biennial LDCT screening for lung cancer in individuals with a negative baseline LDCT can achieve a comparable clinical outcome to annual LDCT screening.The study, as already said, highlights the potential of circulating miRNAs as biomarkers for cancer detection and prognosis [19].In this scenario, we previously illustrated also the results of the BioMILD trial [16] which is a large prospective study that aims to optimize the screening intensity for lung cancer through a combination of LDCT and a blood-based microRNA assay (MSC).The participants underwent baseline LDCT examination, spirometry, and miRNA profiling, and were followed for a median duration of 5.3 years.The study discovered that participants who were double-negative for LDCT and MSC had very low rates of lung cancer incidence and mortality.As a result, they were recommended to undergo LDCT screening once every three years.The results of the study confirmed that the combined use of LDCT and blood miRNAs at baseline can predict individual lung cancer incidence and mortality.
The New York University Lung Cancer Biomarker Center (NYULCBC) enrolled highrisk smokers and lung cancer patients into a screening cohort and a "rule-out lung cancer" cohort with the aim of identifying and validating biomarkers for the early detection of lung cancer.The participants completed a medical and respiratory symptom questionnaire, underwent pulmonary function testing, blood sampling, chest CT, and were followed up for nodule stability.Greenberg et al. [123] conducted a study to evaluate the levels of serum S-Adenosylmethionine (AdoMet) in participants enrolled in the NYULCBC trial from February to August 2004.The study found that patients with lung cancer had higher levels of serum AdoMet compared to healthy non-smokers and high-risk smokers with small noncalcified nodules.AdoMet level alone was able to differentiate patients with lung cancer from smokers with benign nodules with high sensitivity and specificity.When combined with nodule size, AdoMet level showed a sensitivity and specificity of 100% and 94%, respectively.The elevated AdoMet level in lung cancer patients may relate to the role of AdoMet in DNA methylation, as hypermethylation of the promoter regions of tumor suppressor genes in lung cancer and other malignancies has been reported.AdoMet could be a promising marker for early-stage lung cancer detection, but further studies are needed to confirm its efficacy in larger populations and its clinical utility for recurrence diagnosis.
Several clinical trials are investigating the effectiveness of incorporating new blood tests along with LDCT for lung cancer screening.The NCT01925625 trial tested whether using the EarlyCDT-Lung test and subsequent CT scanning to identify individuals at high risk of lung cancer could reduce the incidence of patients with advanced-stage lung cancer at diagnosis compared to standard clinical practice.The EarlyCDT-Lung test used an enzyme-linked immunosorbent assay (ELISA) to measure seven distinct autoantibodies, each having specificity for different tumor-associated antigens, including p53, NY-ESO-1, CAGE, GBU4-5, HuD, MAGE A4 and SOX2.At 2 years, the test showed high specificity (90.4%) and moderate sensitivity (32.1%) with a higher number of early-stage lung cancers detected in the intervention arm.However, no significant differences were observed in lung cancer and all-cause mortality between the intervention and control groups [124].The study suggested that blood-based biomarkers followed by LDCT can detect earlystage lung cancer, but more research is required to determine the long-term impact and increase engagement.
Another test is Lung EpiCheck (Nucleix, Modi'in, Israel), which has been designed to detect hypermethylation status across six markers that are associated with lung cancer, by using cfDNA analysis.Recently, this test has been validated in European and Chinese patients samples and has demonstrated high accuracy rates, as well as an independent predictive capability for lung cancer detection, suggesting potential utility for improving screening access and compliance among high-risk populations [125].In this scenario, the NCT04968548 trial is an observational study aimed at collecting blood samples and clinical data from individuals undergoing LDCT for lung cancer screening and those with confirmed lung cancer to determine and validate the Lung EpiCheck.
Furthermore, the NCT03452514 trial aims to validate the HMBDx microRNA Test by collecting blood samples from 400 individuals who are undergoing LDCT screening.The study plans to analyze microRNA signatures using a novel lung cancer test, compare the results with those obtained through CT scan findings and follow-up tests, and maintain a minimum follow-up period of 12 months post-enrollment.
Lastly, the primary objective of the NCT05306288 clinical trial is to validate the DELFIbased test for detecting lung cancer among individuals eligible for routine screening, using a genome-wide analysis technique called "DNA evaluation of fragments for early interception" (DELFI) to detect abnormalities in cfDNA [126].Participants have blood collected and undergo medical record review at baseline and two additional time points.Presently, no conclusions are available as these last two clinical trials are still ongoing.

Discussion
Given the elevated incidence of overdiagnosis and false positive cases associated with LDCT screening, the identification of reliable biomarkers capable of improving the diagnostic accuracy, represents an unmet need.In this scenario, ncRNAs might be a potential reliable tool to stratify populations into precise categories of lung cancer risk.To date, microRNAs are those most investigated in large prospective trials for lung cancer screening purposes.As reported in the bioMILD trial, the implementation of miRNAs in NSCLC screening can reduce false positive rates and improve diagnostic accuracy of LDCT, thus opening the way for personalized screening approaches.
Furthermore, what emerged from our literature research is an extreme heterogeneity of the conducted studies using different methodologies of analyses and selecting various risk populations.This inevitably can be seen as a positive aspect, as in most of the studies presented, the results were consistent with the ability of ncRNAs to distinguish populations with lung cancer from those that were negative or might face overdiagnosis if subjected to LDCT.However, from a methodological point of view, it clearly constitutes a major issue to be addressed with further research in order to standardize a potential application of ncRNAs liquid biopsy in a real-world setting and safely implement them into our clinical practice.
Methodological limitations of analysis also emerged from this wide literature search, including the heterogeneity of ncRNA detection methods used across the different studies, mostly based on q-RT-PCR, but also on NGS limited panels, digital-droplet PCR, and RNAseq, pointed out the issue of standardization methods to make ncRNAs part of clinical practice.In fact, from a practical point of view, detecting and sequencing this genetic material might be challenging for different reasons.Next-generation sequencing (NGS) is one of the high-throughput screening methods that can be implemented more efficiently into clinical research to validate panels of ncRNAs that can be used for LC screening research programs [127].
Some limits need also to be considered once we propose ncRNAs as a biofluid-based biomarker for LC screening but also in general for other purposes.First of all, the overall quantity of ncRNAs is generally lower in the intracellular, extracellular ambient, as well as in plasma or serum as compared to other genetic material, so it might be a challenge to detect them in patient-derived blood samples [128].Another potential issue is related to the post-transcriptional modifications of ncRNA sequence making them similar to other ncRNAs of the same family (such as for micro RNAs, miRNAs), as well as to mRNAs sequence, making it difficult to distinguish each other.Another issue related to the use of ncRNAs liquid biopsy in LC screening context is the cost-effectiveness benefit that the realworld application of these techniques could imply.For now, large studies demonstrated a clear clinical benefit of LDCT-based screening programs [1], but it is not clear if the implementation of LB in this setting will be feasible from this point of view.In addition to that, further clinical trials testing the role of LB ncRNAs detection in non and light-smokers should be conducted.
Moreover, lung cancer heterogeneity is well-known and established across the board [129][130][131], limiting the use of single-biomarker based approaches.Conversely, the use of multiple biomarkers of the same class or multiple ncRNA class panels could improve diagnostic accuracy within screening programs, since the genetic variability among differ-ent tumors and individuals could be covered by different biomarkers working together at the same time.

Conclusions
In conclusion, the implementation of ncRNAs for LC screening purposes is one of the most promising biomarkers to integrate LB in the prevention setting.For this reason, standardization of protocols for LB ncRNAs detection and further prospective clinical trials with larger cohorts are needed to validate and introduce these novel biomarkers in the clinical arena.In addition, we believe that the use of ncRNAs belonging to multiple subcategories can further improve the ability to discriminate between negative and positive subjects, and therefore using expanded ncRNA panels for LC early detection should be one of the next implementations for LB studies in this research context.To date, clinicians should carefully interpret LB results coming from the early diagnosis studies and policymakers should push research to focus also on the implementation of liquid biopsy in the real-world setting.

Table 2 .
Ongoing clinical trials investigating liquid biopsy non-coding RNA for lung cancer screening.