Matched tissue and liquid biopsies for advanced non-small cell lung cancer patients A potentially indispensable complementary approach

Highlights

• •Next generation sequencing from matched tissue and liquid biopsies can have an add value at diagnosis and at progression in advanced non-small cell lung patients in order.
• •To allow the detection of additional genomic alteration.
• •To identified clonal hematopoiesis of indeterminate potential.
• •To reduce the turnaround time to obtain the results.
• •To have an alternative in case of insufficient tissue material.
• •To develop translational and clinical research programs in prospective clinical trials.

Abstract

The introduction of liquid biopsies (LB) has brought forth a number of therapeutic opportunities into the domain of thoracic oncology. Many of which have been adopted for care of patients presenting with advanced non-squamous non-small cell lung cancer (aNS-NSCLC). For example, one of the most frequent indications to perform a LB in these patients, at least in Europe, is for patients treated with tyrosine kinase inhibitors (TKIs) targeting EGFR and ALK genomic alterations when the tumor progresses. A tissue biopsy (TB) must then be taken, ideally from a site of a tumor that progresses, in particular if the LB does not permit detection of a mechanism of resistance to TKI. A LB from a patient with aNS-NSCLC is recommended before first-line therapy if no tissue and/or cytological material is accessible or if the extracted nucleic acid is insufficient in amount and/or of poor quality. At present a LB and a TB are rarely performed simultaneously before treatment and/or on tumor progression. This complementary/matched testing approach is still controversial but needs to be better evaluated to determine the true benefit to care of patients. This review provides an update on the complementarity of the LB and TB method for care of patients presenting with aNS-NSCLC.

Graphical abstract

Introduction

The management of tissue biopsies (TB) in thoracic oncology is a critical step in the histological diagnosis, prognostic evaluation, and the study by immunohistochemical analysis and/or in situ hybridization and molecular biology of biomarkers predictive of the response to immunotherapy or targeted therapies. TB is nowadays the prevailing gold standard for the diagnosis of non-small cell lung cancer (NSCLC) among patients suspected on a CT scan to have a lung tumor. Therefore, TB is commonly the specimen used for tumor testing to determine the current guideline-recommended biomarkers [1,2]. In the last few years the number of these biomarkers has rapidly evolved in thoracic oncology and it is now necessary to analyze in advanced non-squamous NSCLC (aNS-NSCLC) patients the status of PD-L1 and EGFR, ALK, ROS1, BRAF, RET, MET, NTRK, KRAS and HER2 [1,2]. The panel of genes to be tested can be enlarged to other genes if the patient can be included into a clinical trial targeting a rare genomic alteration such as rearrangements in NRG1 [1,3]. Due to the long list of genes to be evaluated, molecular testing using next-generation sequencing (NGS) is certainly the optimal approach [1,4]. Additionally, some molecules under development or used in early stage clinical trials target proteins such as LAG3, VISTA or TIGIT, which can be expressed and evaluated in the tumor microenvironment or on tumor cells [5,6].

More and more often the diagnosis of lung cancer is made using flexible endoscopes of small caliber that give access to peripheral lesions or with cytological samples obtained by transbronchial puncture. However, as a consequence only small-sized samples can be obtained in this setting. So, the new challenges of the pathologists are to perform all the mandatory analyzes mentioned above with this limited amount of biological material [7]. To provide optimal care to patients, according to the international guidelines, the delay to providing a result must not exceed 10 days [8]. However, to perform the morphological, immunohistochemical and molecular biological analyzes, in particular NGS from fixed tissue, within the given delay, can be problematic. Finally, when the tumor progresses the location of the site for biopsy or the performance status of the patient can make it impossible to obtain a TB. For all these reasons, a LB can represent an alternative or a complementary approach to a TB [1,9,10].

This review deals with the current issues concerning the role of LB at diagnosis or on progression of the lung cancer. We discuss the potential interest of systematically and simultaneously performing both a TB and LB from patient with an aNS-NSCLC.

International recommendations and good practices for the use of liquid and tissue biopsies in thoracic oncology

Different recommendations have recently been established in Europe [from the European Society of Medical Oncology (ESMO)] and the USA [from the National Comprehensive Cancer Network (NCCN) and from the International Association for the Study of Lung Cancer (IASLC)] concerning sampling by LB in thoracic oncology [11], [12], [13], [14]. Thus, the current algorithms recognize that systematic use of a LB before treatment is not recommended at present, except if it is impossible to obtain access to a TB [8]. In this context, it is important to highlight that in around 15 to 40% of NSCLC cases, NGS TB is not feasible due to an insufficient tissue sample and/or extracted nucleic acid [15], [16], [17]. So, the NCCN guidelines stipulate that a LB must be done at baseline if the amount of nucleic acid from TB is insufficient to do molecular testing or if the patient is unfit for invasive tissue sampling or when the tumor tissue specimen is inadequate or unobtainable [14]. When a tumor progresses under treatment, notably in tyrosine kinase inhibitors (TKIs) treated patients, a LB can first be performed, which may detect a targetable mutation of resistance [18]. In fact, a plasma-first approach can be performed to test for EGFR T790M in patients who have developed resistance to first- or second-generation TKIs. Subsequently, in the absence of an identified molecular target in plasma, a TB is highly recommended and certainly mandatory [12]. This is aligned with the recent recommendations from the IASLC for LB for NSCLC, where LB is recommended for cases for whom the tissue sample is unavailable (plasma first approach), or in cases where a TB is inadequate to conduct comprehensive tissue genotyping [14]. Finally according to these IASLC recommendations, for cases with oncogene-addicted NSCLC progressing after initial targeted therapy, a plasma first approach should also be considered as standard of care [14].

One gene sequencing test looking for activating mutations in EGFR or resistance mutations to TKI have been, for a long time in clinical practice, the only reason for doing a LB for patients with aNS-NSCLC, at least in Europe and in a lot of community hospitals worldwide [1,10,14,19]. Other one-gene sequencing tests targeting different genes have also been proposed using a LB but are rarely used in thoracic oncology [20,21]. The tests looking specifically for EGFR mutations in plasma have the advantage of obtaining easily CE-IVD labeling, to be easy to perform and interpret in most of the laboratories, and some are recognized in the USA as companion diagnostics. However, these blood targeted sequencing tests have limitations. First, RT-PCR tests do not detect all mutations on a gene, in particular on EGFR, for which insertions in exon 20 are not detected in around 50% of cases, while they are detected with NGS. Secondly, sequential testing with several targets can also result in increasing costs associated with the progressive increase in the number of genes to be analyzed. Thus, NGS LB is more and more recommended at diagnosis and at tumor progression. Panels of genes of various size (from 11 to 500 genes) can be proposed with the possibility of performing the analyzes in the laboratory (“in house” testing) or by an external laboratory developed commercially (“outsource” testing). The delay in obtaining the results from a LB must be mastered and not exceed 10 days, as for molecular testing performed from a TB. Thus, NGS plasma testing is more often recommended for best clinical practices using a LB from aNS-NSCLC patients, depending on the different international recommendations described above.

Toward new algorithms and good practices for the use of liquid biopsies for advanced non-squamous non-small cell lung cancer

Different new algorithms for the use of LB in thoracic oncology might be proposed or discussed, but are not yet part of the international recommendations. A TB performed before treatment is the gold standard for NS-NSCLC patients. However, an increasing number of challenges are facing the use of TB in thoracic oncology (Table 1) [22]. Systematic NGS analyzes from these biopsies have been proposed due to the increased number of genes to be evaluated [2]. One of the challenges consists in performing these analyzes as reflex testing without waiting for the prescription from the clinician, so as to better manage the time required to provide the results [[19], [23], [24]] (Fig. 1). In fact, the majority of laboratories wait for the prescription to perform molecular tests (“be spoke testing pathway”), in particular NGS analyzes, which increases the delay in transmitting results (Fig. 1). Moreover, the length of time between the scheduling of the TB and the procedure itself (both by bronchial endoscopy or by transthoracic puncture) can vary widely according to the organization and the availability of the physicians and can add many weeks to an already long wait for a laboratory diagnosis. Additionally, the risk of complications after a TB is another concern, up to 60% according to studies using transthoracic needle biopsies, notably inducing the risk of pneumothorax in older patients with chronic obstructive pulmonary disease and emphysema [25,26]. This is the reason why implementation of a “plasma-first” approach in healthcare settings is recommended and discussed by some investigators and might significantly reduce the reporting time, enabling patients to begin targeted therapy earlier [27], [28], [29].

Table 1.

The technical, logistical and biological limitations of using alone a tissue biopsy for patients with advanced non-small cell lung cancer.

Technical issues	Logistical issues	Tumor biology issues
- Novel biopsy procedures in thoracic oncology lead to smaller samples with diminished tumor cellularity (eg. endobronchial ultrasound, flexible endoscope instrument with small caliber) - Different molecular tests with the potential for discordant results (IHC vs FISH vs RT-PCR vs NGS) - Technology specific failure due to differences in sensitivity/specificity according to amplicon-based versus hybrid capture sequencing approaches and different gene panel sizes - Quality assurance in genomic medicine and certification (inter laboratory validation of results)	- Getting molecular testing results in a clinically relevant turnaround time - Desirability of centralized vs distributed or « in house » testing approaches - Regular training of laboratory and clinical staff into new technologies developed from tissue biopsies - Compile a report from the voluminous data into a standardized report easily usable by physicians - Data storage capacity for NGS within a health-care system	- Intra-tumor and inter-tumor site heterogeneity Sampling bias Differential responses to treatment - Tumor evolution and resistance in response to treatment - Need for longitudinal sampling using invasive methods - Evolving treatment paradigms in immuno-oncology and new biomarkers to assess with smaller tissue samples - Increasing complexity of detectable genomic variations in tumor (eg. epigenetic changes and non-coding variants)

Fig. 1.

The two different pathways requesting genomic testing in routine clinical practice. In the bespoke testing pathway, only the oncologist takes the decision to do a molecular test (both on tissue and/or on liquid biopsies) according to the clinical presentation and to the patient's performance status as well as to availability of the results of the PD-L1 expression. In the reflex testing pathway, the pathologist does not wait for the physician to request for molecular testing. In the latter algorithm the pathological and molecular reports are systematically available for treatment decision making.

The “plasma-first” approach without a simultaneously TB (i.e. at the same time) was proposed initially mainly in USA [28,29]. The aim of this proposal is to rapidly detect actionable genomic alterations without waiting for the results and/or performing a TB [29], [30], [31]. This concept is associated with the fact that in certain situations, such as for the organization of some community hospitals, for example in the USA, the delay to obtaining a TB and the results of molecular-histology are sometimes lengthy before proposing an urgent therapeutic indication [32]. Therefore, the LB can be done several days and even several weeks before the TB and save precious time to obtaining the results of molecular testing [27,31]. Another option can be to first perform a LB first and only then to take a TB if no suspect driver or resistance genomic alterations are detected in ctDNA or if the ctDNA quantity is too low (which leads to a non-informative LB result) [27]. However, there are certainly limitations in setting up these algorithms, notably according to the specificity and sensitivity of NGS from cfDNA, which can lead to some false positive and negative results, in comparison to the results obtained from a TB. False positive findings may result in incorrect treatment, which can be deleterious to patients, and thus increase the financial burden of healthcare. In fact, false positive results obtained from the molecular analysis of cfDNA is rare, occurring in cases of germline mutations or in cases of somatic mutations associated with a phenomenon of clonal hematopoiesis (CHM) [33], [34], [35], [36]. The assessment of the variant allelic frequency, as well as the identification of the type of mutation, are pivotal to eliminate a germline or a CHM associated mutation (even if the genomic alterations associated with CHM on a gene of interest to the clinician are exceptional). Ideally, NGS of matched PBMCs could be of interest to check the presence of a germline and/or an CHM-associated mutation. However, in daily practice this latter approach is not possible for a routine set up, due to an associated increase in the financial burden and the laboratory workload. More importantly, it is now admitted that LB NGS has a sensitivity of around 50–75% when compared to TB NGS, notably depending on the genomic alteration [37,38]. This lower sensitivity can be due to a technical and/or a biological issue. In this regard, it is essential to take into consideration the tumor fraction in circulating DNA, which correlates in the large majority of cases with the possibility of detecting actionable genomic alterations [39]. Therefore, the levels of ctDNA vary according to the tumor stage (IIIB versus IV), the tumor size, and the tumor burden. In addition, the amount of ctDNA also correlates with the different sites of metastases. So, liver metastases have a higher plasma-tissue concordance for actionable mutations compared to those with M1a disease or brain metastases [40,41]. The NILE study, a multicenter North American prospective observational study, demonstrated non-inferiority of cfDNA based tumor geneotyping compared to tissue-based genotyping [42]. Notably gene fusion detection was quite similar with LB and TB [42]. However, some studies showed that detection in plasma of actionable fusions in ALK, ROS1, RET and NTRK, and of actionable amplifications, notably in MET, has a lower sensitivity compared to detection with a TB of these fusions and amplifications [27,37]. In the recent study reported by Sugimoto and colleagues a paired analysis of more than one thousand NSCLC plasma-tissue samples was performed and some discrepancies were noted between LB and TB for oncogenic driver detection with a sensitivity of 72% for LB [37]. Other studies showed a variable sensitivity (from 66% to 80%) of LB according to the NGS assays [27,31,[43], [44], [45]]. For all these reasons, negative results following “plasma-first” NGS require further investigation from TB NGS to rule out the possibility of false negatives. One consequence of the discrepancies in sensitivity between NGS LB and NGS TB, could be that the therapeutic strategies based only on the NGS LB results lead to incorrect treatment decision making, such as giving first-line chemotherapy and/or immunotherapy while an actionable genomic alteration present in the tumor is undetectable in a LB. Moreover, it is obvious that immunotherapy or immuno-chemotherapy treatment decision making is only based on the percentage of tumor cells evaluated in TB expressing PD-L1 in patients without associated actionable mutations [46,47]

To summarize, the delays in obtaining an appointment with a specialist as well as the time to sending a TB to a pathology laboratory and then to the molecular pathology laboratory, are sometimes not in the best interests of the patient, depending on the local care organization. Moreover, most of the clinical trials, notable phase 1 clinical trials, require a tumor tissue sample to be obtained to assess some associated predictive biomarkers. However, getting access to tissue samples at tumor progression is often difficult, which could lead to limitations to patient inclusion, notably of elderly patients [48]. In this context, a “plasma-first” strategy can accelerate access to these molecular results and to a higher number of patients to be tested [48]. Caution should be expressed when adopting the practice of a “plasma-first” approach, notably with regard to the general public or even some uninformed clinicians, to not give the impression that a LB can definitively replace a TB, in particular at tumor diagnosis. It is necessary to continue to say in thoracic oncology that « the tissue is still the issue » and that the diagnosis of a lung cancer is only based on a TB. Finally, in the domain of immunotherapy, it is also essential to assess only on tissue samples the status of PD-L1 on a percentage of positive tumor cells. Therefore, despite some clinical research data, the evaluation of the PD-L1 status with LB is not done in daily practice [49,50].

Another strategy to set up could include routine implementation at diagnosis of complementary NGS with plasma to be performed simultaneously with NGS TB [39,43,44,[51], [52], [53], [54], [55], [56], [57]]. Sampling for a single tumor lesion (primary tumor or metastatic site) may not capture the complete genomic landscape due to the molecular heterogeneity of lung cancers [58]. Therefore, a complementary LB aims to detect actionable genomic alterations that are not present in a TB. So, rather than substituting TB with LB, adding a concurrent plasma NGS test to tissue NGS testing should improve the chances of detecting actionable mutations in aNS-NSCLC, of improving prognosis in addition to choosing and timing initiation of treatment [43,51]. However, combining at the same time NGS with TB and LB can have limitations in routine clinical practice. First, it is necessary to obtain all the results without too much difference between the timeline of doing NGS with TB and LB, to assess the presence of actionable genomic alterations at around the same time. It is for this reason that new NGS technologies can provide results from TB in a reasonable period of time [59]. Moreover, for this, the different TB and LB samples should be taken at the same health care institution and should probably be mastered in the same clinical and molecular laboratory. Secondly, doing NGS reflex testing with TB and LB results in a dramatic increase in the cost as well as a substantial impact on the workload of both the technical staff and molecular pathologists. So, the new algorithm using matched TB and LB needs to be discussed with the different administrative and laboratory staff members before starting, so as to be sure of the sustainability over time. Moreover, the possibility of reimbursement by the public health care system of simultaneous NGS with TB and LB for the same patient needs to be taking into consideration. In this regards, doing systematically comprehensive genomic profiling (CGP) using large panels both on matched TB and LB may not be economical sustainable and even more may not be useful for the therapeutic strategy [44]. In this regards, recent studies have shown that tissue CGP may only be considered as a follow-up when an oncogenic driver is not identified by LB CGP [44,60]. In addition, due to the increase in the number of samples, the availability of the NGS sequencer (s) depending on the different workflows has to be managed. As already mentioned above, the detection of actionable genomic alterations with a LB alone and not with a TB must be interpreted with precaution. In fact, the decision to administer a targeted therapy based on the results of a blood analysis alone must be taken after discussion with the clinician and the molecular pathologist, while integrating the allelic frequency percentage associated with the genomic alteration. Moreover, the ctDNA tumor fraction (TF) should be evaluated since there exists a correlation between the TF and detection of actionable genomic alterations [39]. One issue certainly concerns the possibility of using the same NGS panels and the same sequencing technology for LB and TB analyzes, since discrepant results may result from technological bias.

When the tumor progresses, both TB and matched LB obtained from an individual patient for a NGS analysis can certainly provide a map of the ideal molecular portrait of the tumor, in particular in the event of tumor heterogeneity, and improve the chance of detection of a mechanism of resistance that can be targeted with a new drug. Nonetheless, a TB provides biological information that cannot be obtained from a LB, in particular the identification of histological transformation into a small cell or a squamous cell lung carcinoma [61]. Moreover, some mechanisms of resistance associated with amplification in MET that emerge with different TKI treatments are more easily identified with a TB than with a LB [1,10]. Finally, at tumor progression the best therapeutic strategy is to propose to include patients into early phase clinical trials, in this case a LB is a short-term challenge. However, inclusion of a patient into a clinical trial based on a LB can only be envisaged if the evaluation of potential therapeutic targets is performed with large panels of genes. This can be associated with technological and biological limitations of cfDNA or with the costs of care depending on the case.

What are the perspectives?

Advanced squamous cell non-small cell lung cancer

In the past patients with advanced squamous NSCLC did not benefit from the same therapeutic progress as those with NS-NSCLC, in particular in the domain of targeted therapies. However, recently molecules associated with the expression of a number of biomarkers are being evaluated in clinical trials and may soon be used on these patients in daily practice [62]. Thus, inhibitors targeting FGFR are being tested on patients presenting with either over-expression or fusion of FGFR [63]. Detection of genomic alterations in EGFR or ALK, even rarely identified in squamous NSCLC, can allow a few patients to benefit from targeted therapies [64,65]. It is legitimate to consider that patients with advanced squamous NSCLC will rapidly benefit from molecular analyzes using TB and LB with NGS panels. Therefore, algorithms similar to those discussed above may be developed for advanced squamous NSCLC.

Small cell lung cancer

LB may soon become complementary and sometimes even an alternative to TB for small cell lung cancer (SCLC), notably in cases of small-sized TB that are difficult to interpret, with poor sample quality due to necrotic areas, or are not feasible [66]. LB may play a diagnostic and prognostic role and may also be useful to predict the therapeutic response to novel molecules that are being used in clinical trials [66], [67], [68]. So, different DNA methylation profiles have been recently evaluated and may be associated in the near future with new therapeutic strategies [60]. In addition, since many circulating tumor cells (CTCs) are usually present in a LB from SCLC patients, it may be possible to assess the expression of certain proteins in CTCs, such as DLL3, in a perspective manner [69]. So, knowing that some of these proteins could be evaluated as a companion diagnostic, this could open up new opportunities to use LB in SCLCs as a complement to a TB.

Early stage non-small cell lung cancer

Early screening

The screening for lung cancer with low dose CT (LDCT) scans can detect small-sized tumors that are operable [70]. At the moment, the usefulness of LB has been evaluated together with LDCT in research projects screening populations at risk of developing a lung cancer [71]. The results of LB and then of TB in the case of a significant lung image could then be compared. So, LB can become part of an early lung cancer diagnosis when the nodule detected on LDCT is not sufficiently suspicious of a cancer. At present this approach is not part of the national or international recommendations and thus is not deployed in routine clinical practice. One major challenge is certainly to be able to detect ctDNA in early-stage lung cancer, notably in stage I lung cancer, due to the low level of shedding of tumor cells into the blood stream. So, very sensitive methods of ctDNA detection need to be developed in the near future [72]. Some studies using CTCs in early-stage NSCLC gave controversial and discrepant results, emphazing the variability of the specificity and sensitivity of the methods used [73]. However, recent studies based on new technologies, such as the targeted methylation method or the fragmentomic assay, reported promising results for early detection of malignant tumors, showing a good sensitivity and specificity [74], [75], [76]

Detection of minimal (or molecular) residual disease

One of the major challenges in thoracic oncology is to reduce the number of relapses post-surgery. Relapse is mainly due to shedding of CTCs into the blood that gives rise to tumor progression and/or emergence of metastases. Early detection of minimal residual disease (so called molecular residual disease) (MRD) with blood should allow treatment to be modified thus leading to an increase in the different adjuvant treatments in thoracic oncology [77], [78], [79], [80], [81], [82]. The detection of MRD could lead to a body scan looking for a tumor site for a possible TB. However, at present, only a few tests are available for the detection of MRD and are used in some prospective analyzes [82,83]. Multi-centric studies are necessary before setting up MRD detection according to some new international recommendations [82,83]. Recently, the results of the TRACERx study made from ctDNA on more than one thousand samples from 197 NSCLC patients having a complete surgically resection demonstrated very promising results [84]. However, some questions are currently being discussed: i) when should blood samples be taken for MRD assessment, ii) how many time points for MRD assessment, iii) what should be the cutoff used for the amount of ctDNA and, iv) how should the specificity and sensitivity of the different technologies and NGS panels be harmonized and compared [83]. Finally, in this setting LB will certainly be complementary to TB, but also for radiological imaging.

Contribution of the liquid biopsy to the era of immunotherapy

Several studies have shown the interest in performing a LB for patients treated with immunotherapy, but a LB is not performed at present in routine clinical practice in this context [49]. Comparison of the status of PD-L1 or c-MET with CTCs and tumor cells from a TB have shown discordant results [85,86]. These studies are in fact difficult to reproduce in daily practice. Other circulating biomarkers could be integrated into thoracic oncology as predictive factors of response to immunotherapy [49,[87], [88], [89]]. Thus, the analysis of the tumor mutational burden (TMB), of the repertoire of T or B lymphocyte receptors, of the level of expression of some cytokines, or of the extracellular vesicle PD-L1 dynamics are currently being studied [49,88,90]. Additionally, LB can be used to assess ctDNA clearance in advanced NSCLC patients treated with immune-chemotherapy. Therefore, recent studies demonstrated that ctDNA metrics collected across longitudinal time points can be used to risk stratify patients and predict survival or a durable treatment response in patients with metastatic nonsquamous NSCLC treated with immunotherapy or chemo-immunotherapy combinations [87,89]. Finally, comparative TMB assessment from matched TB and LB taken at the same time could be of interest in aNSCLC, since tumors show a TMB heterogeneity depending on the TB site [91].

Conclusion

The set-up of matched TB and LB both at diagnosis and at tumor progression for a NS-NSCLC patient currently faces many challenges (Table 1). Nowadays, international recommendations advise, under certain precise indications, that a LB can be performed for patients with aNS-NSCLC. Whatever the situation, a TB should be given priority and a LB should only be performed if the TB cannot be done or is not contributive due to a low number of tumor cells and/or degraded nucleic acid [92,93]. The respective contribution of TB and LB have been established but the interest in performing systematically and simultaneously the two biopsies before treatment and/or on tumor progression at an advanced tumor stage is presently controversial. Different algorithms must integrate the costs (including the workload in the laboratories) and potential reimbursement of tests, as well as future in vitro diagnostic regulations (IVDR) that will soon be associated with the different NGS tests [94], [95], [96]. Several existing and future challenges must be mastered for a broader use of LB in daily routine and in particular if a LB is systematically performed in association with a TB at diagnosis and/or at tumor progression. The recent emergence of novel antibodies and their use in immunohistochemistry on sequential tissue sections from a TB limit the results of molecular tests, in particular NGS, the use of which will become out of control given the increase in the number of genes for analysis [5]. However, future development of multiplex immunohistochemistry could enable biological material to be saved [97]. However, the combination of NGS TB with NGS LB may make up for the lack of information obtained from some tissue samples (Table 2).

Table 2.

Opportunities and issues of performing matched tissue and liquid biopsies at diagnosis and at tumor progression for patients with advanced non-small cell lung cancer.

At tumor diagnosis	Opportunities	Increase the detection of the number of actionable genomic alterations Decrease the TAT to detect actionable genomic alterations Allow a global view of the tumor molecular portrait to be obtained Higher possibility of detecting primary resistance mechanisms Allow simultaneous tissue and blood TMB to be evaluated Allow new clinical research project to be initiated
	Issues	Not recommended nowadays by international guidelines Costly Need to be validated for procedures of reimbursment Need to have definitive proof of the added value Challenge to use the same panels/same sequencing technologies Need to master well the TAT to obtain both results Difficulties in controlling the tissue and blood sample workflow at the same time Interpretation of germline and somatic mutations linked to the CHM phenomenon Increased workload for the technician, molecular pathologist, geneticist and bioinformatician
At tumor progression	Opportunities	Increase the possibilities of detecting secondary mechanisms of resistance Decrease the TAT to obtain results of mechanisms of secondary resistance Increase the possibility of including patients into clinical trials
	Issues	Matched TB is not always possible depending of the patient’ status and the site of tumor progression Same issues as those cited for at tumor diagnosis

To conclude, a LB in thoracic oncology must be considered while taking into account the latest international recommendations. However, LB and TB can be systematically and simultaneously performed in thoracic oncology before treatment, within the framework of evaluation, but are not part of the current recommendations. The molecular tests that use LB and TB should ideally be based on similarly NGS methods with the same panel of genes. Globally, the identification of gene fusions, but even more of gene amplification is less sensitive with a LB than a TB when taken at the same time. Finally, on tumor progression of a NS-NSCLC a LB and a TB together could identify, in an ideal setting, the global molecular portrait of the tumor, which would allow rapid detection of a mechanism of treatment resistance and thus rapid adaptation of the second-line treatment.

CRediT authorship contribution statement

Paul Hofman: Conceptualization, Writing – original draft, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.