Molecular Testing of Metastatic Non-Small Cell Lung Cancer in the Asia-Pacific Region

The management of advanced non-small cell lung cancer (NSCLC) patients has become more complex with the identification of many oncogenic drivers starting with epidermal growth factor receptor (EGFR) mutation and the availability of therapies specifically targeting these molecular alterations resulting in improved treatment outcomes. Molecular biomarker testing of NSCLC is now considered standard of care and part of the diagnostic algorithm to identify subsets of patients for molecular-targeted treatment. With the increasing use of small biopsies and cytological specimens for diagnosis and the need to identify an increasing number of predictive biomarkers, proper management of the limited amount of sampling materials available is important. A limited diagnostic workup is recommended to preserve sufficient tissue for molecular testing. Molecular testing of lung adenocarcinoma for alterations in the EGFR, ALK and ROS1 genes is now considered standard of care while emerging predictive biomarkers such as mutations in BRAF, HER2, MET exon 14, RET, and NTRK may be included in an expanded panel testing. The extent of therapy-predictive biomarker testing is dependent on the approval and reimbursement status of the recommended therapeutic agents in each country or institution. When selecting a testing method, timeliness of results availability for treatment decision-making needs to be considered. Turnaround time should not be more than 10 working days from receipt of sample to reporting of all results. This position statement recommends molecular profiling of metastatic NSCLC at diagnosis of and at disease progression following first-line targeted therapy as this impacts on the treatment outcomes of patients.

Lung cancer is the leading cause of cancer death for both men and women worldwide.¹ Of an estimated 2.09 million new lung cancer cases worldwide annually, 58.5% occur in Asia and 60.7% of the 1.76 million annual lung cancer deaths worldwide occur in Asia.² Non-small-cell lung cancer (NSCLC) accounts for about 85% of lung cancers.³ The majority of patients present with locally advanced or metastatic disease.^1,4,5 In recent years, several somatic mutations that contribute to the carcinogenesis of NSCLC have been identified.^3,6 Lung cancer in never smokers, which accounts for about a quarter of lung cancer in Asia,⁷ is considered a different disease with specific molecular and genetic differences from that of smokers.⁸ In never smokers with lung cancer, there is a higher prevalence of adenocarcinomas and women, especially Asian women.⁹

EGFR activating mutations, one of the most commonly mutated genes in NSCLC, are harboured by up to 15% of adenocarcinoma and less than 5% of squamous cell carcinomas (SCC) in Caucasian patients^3,10 but are much more common in Asian populations, with prevalence rates reported in up to 59% of adenocarcinoma and 5.4% of SCC cases.^11,12,13 EGFR mutations are also more common in never-smokers and female lung cancer patients.^13,14 Improved progression-free survival, response rate, symptom control, and quality of life are observed in EGFR-mutant patients treated with an EGFR-tyrosine kinase inhibitor (TKI) compared with platinum doublet chemotherapy in many phase III clinical trials.^{11,15,16,17,18,19,20,21} EGFR mutations are now accepted as the predictive biomarker of treatment outcome to first-line EGFR-TKIs.^22,23 In recent years, many more gene alterations that predict responsiveness to targeted therapy have been discovered in NSCLC and molecular testing has now expanded to include targets such as anaplastic lymphoma kinase (ALK) and ROS proto-oncogene 1 (ROS1) translocations, B-Raf proto-oncogene (BRAF) mutation, mesenchymal-epithelial transition (MET) amplification and mutation, RET proto-oncogene (RET) fusion, Erb-B2 receptor tyrosine kinase-2 (ERBB2) [or human epidermal growth factor receptor-2 (HER2)] mutation, neurotrophic receptor tyrosine kinase (NTRK) and Kirsten rat sarcoma 2 viral oncogene homologue (KRAS) mutation.^24,25,26 In addition, immunotherapy has emerged as part of standard of care treatment for patients whose tumours do not harbour driver mutations. As such, the management of advanced NSCLC now requires testing for multiple biomarkers to guide the treatment strategy prompting the need to establish a consensus statement and recommendations for predictive biomarker testing and molecular profiling of metastatic NSCLC in the Asia-Pacific region. Although there are national practice guidelines for lung cancer treatment^27,28 and an Asian adaptation of the European Society of Medical oncology (ESMO) guidelines for the management of metastatic NSCLC,^25,29,30 there are no specific guidelines or recommendations on molecular testing of NSCLC for the region. Wide variations in testing are to be expected across the region because of variations in the availability of testing facility and reimbursement for both testing and lung cancer treatment in the healthcare systems of different countries.^4,31

This position statement paper by the Lung Cancer Assembly of the Asian Pacific Society of Respirology (APSR) highlights important aspects of molecular and biomarker testing for patients with metastatic NSCLC in the Asia-Pacific region and is not a treatment guideline. In addition, immunotherapy with checkpoint inhibitors has emerged as part of standard of care treatment for advanced stage NSCLC patients.

CKL, as the head of the APSR Lung Cancer Assembly, chaired a panel of respiratory physicians (except YSL) who were selected for their expertise in various aspects of lung cancer diagnosis and management. This position statement is the result of face-to-face discussions at APSR Congress meetings and exchange of views via email communications among these experts to develop the scope and draft the recommendations.

To understand the practice of EGFR mutation and other molecular testing in NSCLC across countries in the Asia-Pacific region the APSR membership was surveyed using a web-based online survey tool (SurveyMonkey) during the period 18 August to 3 October 2018.³² The survey questionnaire is as shown in Appendix A and the results are presented in Appendix B.

All authors reviewed and approved the final manuscript which is the outcome of the expert group discussion and consensus arrived on 5 December 2019. The panel did not receive any commercial sponsorship and all panel members worked on an honorary basis. This position paper is endorsed by the APSR and is posted at the APSR website as a resource document. The clinical relevance of the document will be reviewed after a maximum of 5 years.

The recommendations for molecular testing of metastatic NSCLC in this position statement are outlined in Table 4a and Table 4b which also show a comparison with the recommendations by other scientific societies.

Types of specimen suitable for molecular testing
Formalin-fixed paraffin-embedded (FFPE) core tissue biopsy is the preferred sample for molecular testing (Table 4a). However, cell blocks or any cytology sample including smear preparations with adequate cellularity and preservation may be tested for lung cancer molecular biomarkers.²⁴

Specimen sampling procedures
The testing for predictive biomarkers prior to treatment initiation in patients with newly diagnosed metastatic NSCLC requires close coordination between the respiratory physician, interventional radiologist, pathologist, and oncologist to ensure that the specimen is used judiciously to obtain the required information for treatment decision-making.³³ Sufficient tissue must be obtained to make a histological diagnosis of lung cancer and for molecular testing.³⁴

Factors to be considered when choosing the optimal diagnostic approach include the anticipated diagnostic yield, diagnostic accuracy, adequacy of the amount of tissue specimen obtainable, invasiveness and risk of procedure, accessibility and timeliness of procedure, and technologies and expertise available. Patients with central lesions and suspected endobronchial involvement should undergo bronchoscopy³⁰ (Table 4a) which allows endobronchial forceps biopsy, bronchial washing, bronchial brushing, as well as bronchial and transbronchial needle biopsy, with a diagnostic yield of 65%-88%.^35,36,37 Patients with suspected hilar and mediastinal nodal disease should be biopsied by endobronchial ultrasound (EBUS) and/or endoscopic ultrasound (EUS), or mediastinoscopy.³⁸ By combining direct bronchoscopic visualization of the airway and ultrasound-guided biopsy of the lesion, EBUS has a diagnostic yield of 75%-85% in large central lesions.^39,40 Transthoracic core needle biopsy (TTNB) or fine needle aspiration (FNA) under CT imaging guidance is indicated in mid to peripheral lesions.^25,30 Thoracoscopic evaluation and biopsy of the pleura should be considered if cytology results are negative on repeated thoracocenteses in patients with pleural effusions. Rapid on-site evaluation (ROSE) by the cytopathologists and/or cytotechnologists should be carried out when available to increase the yield at the time of specimen sampling which may reduce the need for additional diagnostic procedures.

When limited tissue is obtained, laboratories should use techniques to maximise the tissue, including "up-front" slide sectioning, for diagnostic and molecular biomarker testing.²⁶ The use of immunohistochemistry (IHC) should be limited and molecular testing should be prioritised. In small biopsy specimens, a limited number of immunostains with one lung adenocarcinoma marker [thyroid transcription factor 1 (TTF-1) or napsin A] and one squamous carcinoma marker (p40 or p63) should suffice for diagnosis.²⁶

A European expert group recommends⁴¹

• At least five endobronchial/transbronchial forceps biopsies with consideration for an additional five forceps biopsies or two cryobiopsies to maximise the amount of tissue obtained,
• At least four EBUS/EUS needle aspiration passes per target lesion, and
• At least two percutaneous core needle biopsies with an 18-20 gauge needle to ensure sufficient tissues while three to six core needle biopsies may be considered to maximise the amount of tissue obtained.

Molecular testing methods

Testing methods must be able to detect molecular alterations in specimens with as little as 20% cancer cells.²⁴ Polymerase chain reaction (PCR)-based methods are more sensitive than Sanger sequencing and can reliably detect low frequency mutations in small samples.

EGFR testing methods should cover EGFR mutations in exons 18–21 (Table 4a).^24,25,26 If resources or specimen material are limited, at least the most common activating mutations, i.e., exon 19 deletion and exon 21 L858R point mutation should be tested.^24,25

For ALK testing, IHC with high-performance ALK antibodies and validated assays [e.g., FDA-approved IHC (ALK [D5F3] CDx Assay] is accepted as equivalent to fluorescence in situ hybridisation (FISH)²⁴ with a shorter turnaround time.

Since ROS1 rearrangement is relatively rare and its occurrence is mutually exclusive with other oncogenic drivers including EGFR and ALK, sequential testing of EGFR and ALK followed by ROS1 is a pragmatic approach. If ROS1 IHC is used for screening, a positive result should be confirmed by FISH or real-time (RT)-PCR or next-generation sequencing (NGS). The high sensitivity of IHC means tumours that stain negative for ROS1 can be considered ROS1-negative.²⁴

RT-PCR, Sanger sequencing (ideally paired with tumour enrichment) and NGS are the commonly used methods for testing BRAF mutation.²⁶

When choosing a testing method, it is important to consider the timeliness of results availability. Turnaround time should not be more than 10 working days from sample receipt to reporting of all results.^24,26

With the growing list of targetable gene alterations, there is increasing demand for multiplexed and simultaneous testing for many genes at once rather than sequential single-gene testing at the time of diagnosis. Compared to sequential single gene testing, NGS targeted gene panels with its ability to test multiple genes of interest on limited material and lower cancer cell concentration from small biopsies and cytological samples, offer a good option for rapid, cost-efficient and sample preserving genomic profiling.⁴² However, in most clinical settings sequential testing may be more cost effective. If available, NGS panels are preferred over single-gene assays for expanded testing beyond EGFR, ALK and ROS1.^25,30

As shown in the results of our survey (Appendix B), the access to and the performance of predictive biomarker testing are generally driven by the availability, affordability and reimbursement status of tests and treatments.³² High volume practices are more likely to perform such tests.³² Given the diversity of healthcare infrastructure and economies, variations in practice, resource availability, healthcare financing and standards of care are to be expected across countries in the Asia-Pacific region. Oncology practice in lower income countries often lags behind the rest of the world. As much as practicing clinicians are enthusiastic about embracing evidence-based advances and recommendations by international guidelines, lower income countries face financial constraints and other challenges in adopting such recommendations.

Molecular testing at the initial diagnosis of metastatic NSCLC
Molecular testing are largely driven by the availability of targeted therapy (Table 4b). The recommended algorithm for molecular testing in metastatic NSCLC patients is shown in Figure 1. Since the vast majority of oncogene-addicted lung cancers are adenocarcinomas, all patients with metastatic, possible, probable or definite, adenocarcinoma should be tested for oncogenic drivers which include EGFR mutations, and ALK and ROS1 rearrangements.^24,25,26 Although oncogenic drivers are much more common in adenocarcinomas in never-smokers (i.e., those who never smoked or who smoked < 100 cigarettes in lifetime), light-smokers (< 15 pack-years) or long-time ex-smokers (quitted > 10 years) they can also be found in current smokers and patients who smoke more heavily.

For squamous cell carcinoma, testing for EGFR mutations, and ALK and ROS1 rearrangements are recommended in never- or light-smokers or long-time ex-smoker, those with small biopsy specimens in which an adenocarcinoma component cannot be excluded, and those with adenosquamous carcinoma.³⁴

Upfront testing should include EGFR mutations and rearrangements of ALK and ROS1. BRAF mutation testing maybe included as part of larger testing panels performed either initially or when routine EGFR, ALK, and ROS1 testing are negative and if BRAF/MEK inhibitors are available for use.^24,26 Additional genetic testing of emerging biomarkers such as MET, HER2, RET and NTRK may be carried out if broader molecular profiling by NGS is performed^24,25,26 to identify patients for treatment with specific inhibitors of these targets if such treatments are available or to direct patients to clinical trials. Individual testing for EGFR mutation, ALK and ROS1 can usually be done within 2 weeks while broader molecular profiling by NGS often takes 3 weeks or longer. In practice, the immediately druggable EGFR mutation, ALK and ROS1 are tested upfront for the timely commencement of appropriate first-line targeted therapy. If these three genetic alterations are not detected, first-line chemotherapy can be started while waiting for results of NGS testing for emerging biomarkers with potentially effective therapies.

Molecular testing in patients who develop disease progression on targeted therapy
Molecular testing for patients who develop disease progression on first-line and/or second-line targeted therapy should take into consideration the results of previous molecular testing and treatment received (Table 4b) (Figure 1)

It is strongly recommended that in patients with sensitizing EGFR-mutant lung adenocarcinoma who have progressed while on treatment with a first- or second-generation EGFR-TKI, acquired EGFR T790M exon 20 substitution mutation testing should be done to select them for therapy with osimertinib, a third-generation T790M-mutant specific EGFR-TKI.^24,25,26

Peripheral blood for plasma cell-free DNA (cfDNA) or circulating tumour DNA testing is an acceptable approach to detect T790M at relapse but the test lacks sensitivity and there is a 20%-30% chance of missing T790M mutation.⁴³ The presence of the original sensitising EGFR mutation in the liquid biopsy will serve as an indicator of sufficient cfDNA in the plasma sample. If plasma testing is negative for T790M, tumour tissue rebiopsy is strongly recommended to determine T790M status by tissue-based testing.^24,25,26,44 If T790M mutation is not detected by a sensitive assay, testing for alternative mechanisms of resistance (MET amplification, HER2 amplification) may be carried out to direct the patients for therapies targeting these other molecular mechanisms of acquired resistance.²⁶ A repeat tissue biopsy may also be more effective in identifying other resistance mechanisms which may require alternative treatment which include MET amplification, HER2 alterations, and small cell transformation.²⁵

For patients with ALK rearranged NSCLC who have failed first-line treatment with crizotinib, routine identification of specific tyrosine kinase domain mutations is not recommended to determine the appropriate next line treatment,²⁷ since most crizotinib-resistant tumours do not develop ALK resistance mutations and are sensitive to more potent second-generation ALK-TKIs. As newer-generation ALK-TKIs are increasingly being used in the first- and second-line settings, ALK mutation testing by plasma cfDNA or tumour rebiopsy may become a routine at relapse to select the next ALK-TKI as secondary ALK tyrosine kinase domain resistance mutations are more common after treatment with second-generation ALK-TKIs and because of the differential sensitivity to different second- or third-generation ALK-TKIs depending on the specific ALK tyrosine kinase resistance mutations.^25,26,46,47

Biomarker testing for therapy with immune checkpoint inhibitors in patient with metastatic NSCLC
Programmed death-ligand 1 (PD-L1) expression on tumours by IHC should be determined before first-line treatment in patients with metastatic NSCLC without common driver mutations (Table 4b) (Figure 1).³⁴ PD-L1 testing is required as a companion diagnostic for pembrolizumab monotherapy in all lines of treatment but not required for prescribing nivolumab or atezolizumab monotherapy in second-line although it may be informative as a complementary diagnostic.^25,30

In different clinical trials, tissue- and blood-based tumour mutational burden (TMB) has been evaluated as a predictor of clinical benefit from immunotherapy with nivolumab, ipilimumab, atezolizumab and durvalumab but definitive cut-offs are yet to be defined.^48,49,50,51

The recommendations in this position statement are summarised in Table 2. In the near future, more cost effective broad molecular testing with shorter turnaround time is anticipated to increase as more druggable biomarkers are identified and as technology advances.

Acknowledgements: The submission of this manuscript is made possible by funding from the APSR Assembly Fund.

Disclosure statement