bs_bs_banner
Outcomes of an Australian testing programme for epidermal growth factor receptor mutations in non-small cell lung cancer M. J. Peters,1 J. J. Bowden,2 P. Carpenter,3 J. Lewis3 and B. Solomon4 1
Concord Repatriation General Hospital and 3AstraZeneca Pty Ltd, Sydney, New South Wales and 2Flinders Medical Centre, Adelaide, South Australia
and 4Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia
Key words carcinoma, non-small-cell lung, receptor, epidermal growth factor, mutation rate, receptor, epidermal growth factor/antagonist and inhibitor. Correspondence Matthew Peters, Department of Respiratory Medicine, Concord Repatriation General Hospital, Hospital Road, Concord, NSW 2139, Australia. Email: [email protected] Received 23 February 2014; accepted 1 April 2014. doi:10.1111/imj.12449
Abstract Background: Molecular characterisation of non-squamous non-small-cell lung cancer (NSCLC) is required to direct optimal treatment. Treatment of NSCLC with inhibitors of epidermal growth factor receptor (EGFR) tyrosine kinase (EGFR-TKI) should be guided by the presence of activating mutations of the EGFR gene. Aim: To gain insight into the rate of testing, the range of tissues samples, test utility and outcome when cost of testing as a barrier to access is removed in the Australian setting. Methods: In October 2010, a sponsored programme was commenced to gather data on EGFR gene mutation testing in Australia. Partnering laboratories were funded for provision of de-identified results. For participating patients, the programme supported the test charge. Mutation testing was performed using Sanger sequencing of exons 18–21 of the EGFR. Results: Samples 2013 were submitted from 2012 patients. Full sequencing was achieved in 1717 (85%). Failure of full sequencing was more likely in samples derived from fine needle aspiration(FNA) biopsy than tissue biopsy or pleural/pericardial fluid cell blocks OR 3.1 (95% CI 1.9–5.2). There were 359 mutations seen in 337 patients. 14.5% of cases had a classical mutation conferring sensitivity to EGFR-TKI. In addition there was a range of less common mutations – some predicting responses and others of uncertain significance. 1.4% of cases had mutations associated with non-responsiveness to EGFR-TKI. Conclusions: EGFR gene mutation testing is feasible on local and interstate lung cancer samples. The rate of valid test outcomes is high, but FNA samples are associated with more frequent test failure.
Introduction Lung cancer is most commonly diagnosed at an advanced stage. Compared with some other solid tumours, progress on improved treatments for lung cancer has been slow,1 and lung cancer remains the most frequent cause of cancer death in both men and women in Australia.2 In the past two decades, there have been incremental improvements in cytotoxic treatments for the patients
Funding/Conflict of interest: This programme was funded by AstraZeneca. J. Lewis and P. Carpenter are employees of AstraZeneca. None of the other authors has been recompensed for time involved in evaluating and reporting these data. However, B. Solomon, M. J. Peters and J. J. Bowden have received honoraria for Continuing Medical Education presentations and B. Solomon and M. J. Peters have previously served on Advisory Boards in relation to gefitinib. © 2014 The Authors Internal Medicine Journal © 2014 Royal Australasian College of Physicians
with better performance status. However, an increase in mean age at diagnosis3 requires that patients be matched to the maximum extent possible to treatment with the highest ratio of clinical benefit to likely toxicity. Specific genomic alterations, identified in a subset of patients with advanced, non-small-cell lung cancer (NSCLC), confer a high likelihood of a treatment response with targeted therapies. The rarer of these is the EML4-ALK translocation4 that is associated with responses to crizotinib.5 The more common group of mutations is detected in exons 18–21 of the epidermal growth factor receptor (EGFR) gene that encodes its tyrosine-kinase domain. While these mutations generally result in activation of the EGFR independent of ligand, some, but not all, mutations in these exons predict a response to EGFR-TKI, afatinib,6 erlotinib7 and gefitinib.8 Deletions in exon 19 and the L858R point mutations are 575
Peters et al.
the common predictors of a good response to EGFR-TKI. Point mutations in exon 18, at G719, and in exon 21, at L861, may be associated with less treatment benefit,9 whereas exon 20 insertions or complex insertion/ deletions are associated with lack of response. EGFR mutations and EML4-ALK translocations are generally not found in the same tumour.10 Well-defined clinical characteristics, including Asian ethnicity, female gender and never smoker or light smoker status,11 are associated with the presence of EGFR mutations in NSCLC. However, even a full set of these characteristics does not ensure that a mutation will be present, and mutations can be detected in patients with none of these clinical features. Activating EGFR gene mutations is rare in squamous cell carcinoma, and molecular testing is funded only for non-squamous NSCLC.12 Even though EGFR mutation testing is now funded, it remains important for clinicians to know whether a requested test will produce a valid result and how likely it is that a treatment-changing mutation will be detected. Representative Australian data are lacking for both. The aims of this programme, conceived and executed before testing was funded, was first to determine the rate of a failed or incomplete test from different tissues and sample types. Second, when cost of the pathology test is removed from clinical decision taking, what is the frequency of testing? Last, what is the spectrum and frequency of EGFR gene mutations seen that could influence decisions about optimal treatment?
Methods This programme began in October 2010. At that time, three National Association of Testing Authorities, Australia accredited laboratories provided EGFR gene mutation testing services to most of Australia except for Western Australia. These provided de-identified information on mutation testing outcomes performed using Sanger sequencing of exons 18–21 of the EGFR13 from formalin fixed paraffin embedded tissue, where required after tissue micro-dissection. As compensation for the provision of mutation information, laboratories received sufficient funding so that patients were not charged for the mutation test. The programme was widely publicised to Medical Oncologists, Respiratory Physicians individually and through multi-disciplinary teams. It was available in all states except for Western Australia where alternative mechanisms for testing were in place. Participation in the programme did not require that a patient be eligible for publicly funded or compassionate access to gefitinib. 576
The extent of information provided about the sample supplied for the programme varied between laboratories. Information that was provided did include tissue of origin, sample type (tissue biopsy, fine needle aspiration (FNA) cell block or fluid cell block) and location of the referring. For purpose of analysis, tissue biopsy was all tissue consisting of a bronchial or pleural biopsy, tissue core biopsy or larger. Fluid samples were cell blocks obtained from pleural or pericardial effusions. FNA samples were cell blocks obtained from a simple needle aspiration. Where it was not clear, the tissue or sample type was called ‘unspecified’. There was no information on subsequent treatment or clinical outcomes on any patient. First, primary outcomes were reported as full sequencing completed or incomplete. Second, test outcome is noted as mutation characterised or not detected. One laboratory provided no information on incomplete tests; only the number of such tests is known. It is thus possible only to estimate the failure rate by tissue or biopsy type for the whole sample. We compared first the rates of biopsy, cell block and FNA in the sets of completely and incompletely sequenced samples. Differences were compared with chi-squared tests. To estimate the overall failure rate by biopsy type, we assumed that the distribution of biopsy types was similar in the unreported results as for those where biopsy type was provided. The rates of testing, in the past 12 months of the programme period from ACT/New South Wales, South Australia/Northern Territory, Tasmania and Victoria, were compared to the number of new cases of NSCLC expected based on the most recent AIHW data.14 We used the number of valid tests as the numerator as we could not determine the number of incomplete tests by State.
Results Samples of 2013 were submitted from 2012 patients for analysis. One patient had both brain and liver biopsies sent for analysis. In calculating the yield of a valid test result, these samples are considered separately, but are taken as one for reporting of the spectrum of mutations. The rate of testing per new case of lung cancer was highest in Victoria (Table 1). Sequencing was complete for exons of interest in 1717 (85%). In one additional case, an activating mutation was detected, but other exons were not fully sequenced. Reasons for an incomplete test were provided in 98 patients. In 17, no tissue was found in the blocks submitted, and in 25 no tumour tissue was detected. In a further 25, there was a failure of polymerase chain reaction (PCR) amplification related to poor quality DNA, and in 31 only partial sequencing was completed. In these cases, © 2014 The Authors Internal Medicine Journal © 2014 Royal Australasian College of Physicians
EGFR mutations in NSCLC in Australia
Table 1 Distribution of studied cases by State or Territory together with the rate of testing based on samples tested as a proportion of all new lung cancer cases diagnosed in the study period based on 2011 data from the Australian Institute of Health and Welfare14 State
ACT/New South Wales Queensland South Australia/NT Tasmania Victoria Victoria/Tasmania combined
Number of valid test results
Valid tests per annual lung cancer case
Rate of activating mutation detection (%)
501 156 79 25 586 611
0.157 0.037 0.200 0.089 0.232 0.217
17.4 20.3 14.7 12.0 13.5 13.4
The rate of activating mutations is the proportion of valid samples showing L858R point mutations or exon 19 deletions.
exon 19 was sequenced in 12 cases and exon 21 in 17 cases. The relationship between the biopsy procedure and test outcome is shown in Table 2. Fine needle biopsies were more likely to result in incomplete sequencing than either tissue biopsy (P
Outcomes of an Australian testing programme for epidermal growth factor receptor mutations in non-small cell lung cancer M. J. Peters,1 J. J. Bowden,2 P. Carpenter,3 J. Lewis3 and B. Solomon4 1
Concord Repatriation General Hospital and 3AstraZeneca Pty Ltd, Sydney, New South Wales and 2Flinders Medical Centre, Adelaide, South Australia
and 4Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia
Key words carcinoma, non-small-cell lung, receptor, epidermal growth factor, mutation rate, receptor, epidermal growth factor/antagonist and inhibitor. Correspondence Matthew Peters, Department of Respiratory Medicine, Concord Repatriation General Hospital, Hospital Road, Concord, NSW 2139, Australia. Email: [email protected] Received 23 February 2014; accepted 1 April 2014. doi:10.1111/imj.12449
Abstract Background: Molecular characterisation of non-squamous non-small-cell lung cancer (NSCLC) is required to direct optimal treatment. Treatment of NSCLC with inhibitors of epidermal growth factor receptor (EGFR) tyrosine kinase (EGFR-TKI) should be guided by the presence of activating mutations of the EGFR gene. Aim: To gain insight into the rate of testing, the range of tissues samples, test utility and outcome when cost of testing as a barrier to access is removed in the Australian setting. Methods: In October 2010, a sponsored programme was commenced to gather data on EGFR gene mutation testing in Australia. Partnering laboratories were funded for provision of de-identified results. For participating patients, the programme supported the test charge. Mutation testing was performed using Sanger sequencing of exons 18–21 of the EGFR. Results: Samples 2013 were submitted from 2012 patients. Full sequencing was achieved in 1717 (85%). Failure of full sequencing was more likely in samples derived from fine needle aspiration(FNA) biopsy than tissue biopsy or pleural/pericardial fluid cell blocks OR 3.1 (95% CI 1.9–5.2). There were 359 mutations seen in 337 patients. 14.5% of cases had a classical mutation conferring sensitivity to EGFR-TKI. In addition there was a range of less common mutations – some predicting responses and others of uncertain significance. 1.4% of cases had mutations associated with non-responsiveness to EGFR-TKI. Conclusions: EGFR gene mutation testing is feasible on local and interstate lung cancer samples. The rate of valid test outcomes is high, but FNA samples are associated with more frequent test failure.
Introduction Lung cancer is most commonly diagnosed at an advanced stage. Compared with some other solid tumours, progress on improved treatments for lung cancer has been slow,1 and lung cancer remains the most frequent cause of cancer death in both men and women in Australia.2 In the past two decades, there have been incremental improvements in cytotoxic treatments for the patients
Funding/Conflict of interest: This programme was funded by AstraZeneca. J. Lewis and P. Carpenter are employees of AstraZeneca. None of the other authors has been recompensed for time involved in evaluating and reporting these data. However, B. Solomon, M. J. Peters and J. J. Bowden have received honoraria for Continuing Medical Education presentations and B. Solomon and M. J. Peters have previously served on Advisory Boards in relation to gefitinib. © 2014 The Authors Internal Medicine Journal © 2014 Royal Australasian College of Physicians
with better performance status. However, an increase in mean age at diagnosis3 requires that patients be matched to the maximum extent possible to treatment with the highest ratio of clinical benefit to likely toxicity. Specific genomic alterations, identified in a subset of patients with advanced, non-small-cell lung cancer (NSCLC), confer a high likelihood of a treatment response with targeted therapies. The rarer of these is the EML4-ALK translocation4 that is associated with responses to crizotinib.5 The more common group of mutations is detected in exons 18–21 of the epidermal growth factor receptor (EGFR) gene that encodes its tyrosine-kinase domain. While these mutations generally result in activation of the EGFR independent of ligand, some, but not all, mutations in these exons predict a response to EGFR-TKI, afatinib,6 erlotinib7 and gefitinib.8 Deletions in exon 19 and the L858R point mutations are 575
Peters et al.
the common predictors of a good response to EGFR-TKI. Point mutations in exon 18, at G719, and in exon 21, at L861, may be associated with less treatment benefit,9 whereas exon 20 insertions or complex insertion/ deletions are associated with lack of response. EGFR mutations and EML4-ALK translocations are generally not found in the same tumour.10 Well-defined clinical characteristics, including Asian ethnicity, female gender and never smoker or light smoker status,11 are associated with the presence of EGFR mutations in NSCLC. However, even a full set of these characteristics does not ensure that a mutation will be present, and mutations can be detected in patients with none of these clinical features. Activating EGFR gene mutations is rare in squamous cell carcinoma, and molecular testing is funded only for non-squamous NSCLC.12 Even though EGFR mutation testing is now funded, it remains important for clinicians to know whether a requested test will produce a valid result and how likely it is that a treatment-changing mutation will be detected. Representative Australian data are lacking for both. The aims of this programme, conceived and executed before testing was funded, was first to determine the rate of a failed or incomplete test from different tissues and sample types. Second, when cost of the pathology test is removed from clinical decision taking, what is the frequency of testing? Last, what is the spectrum and frequency of EGFR gene mutations seen that could influence decisions about optimal treatment?
Methods This programme began in October 2010. At that time, three National Association of Testing Authorities, Australia accredited laboratories provided EGFR gene mutation testing services to most of Australia except for Western Australia. These provided de-identified information on mutation testing outcomes performed using Sanger sequencing of exons 18–21 of the EGFR13 from formalin fixed paraffin embedded tissue, where required after tissue micro-dissection. As compensation for the provision of mutation information, laboratories received sufficient funding so that patients were not charged for the mutation test. The programme was widely publicised to Medical Oncologists, Respiratory Physicians individually and through multi-disciplinary teams. It was available in all states except for Western Australia where alternative mechanisms for testing were in place. Participation in the programme did not require that a patient be eligible for publicly funded or compassionate access to gefitinib. 576
The extent of information provided about the sample supplied for the programme varied between laboratories. Information that was provided did include tissue of origin, sample type (tissue biopsy, fine needle aspiration (FNA) cell block or fluid cell block) and location of the referring. For purpose of analysis, tissue biopsy was all tissue consisting of a bronchial or pleural biopsy, tissue core biopsy or larger. Fluid samples were cell blocks obtained from pleural or pericardial effusions. FNA samples were cell blocks obtained from a simple needle aspiration. Where it was not clear, the tissue or sample type was called ‘unspecified’. There was no information on subsequent treatment or clinical outcomes on any patient. First, primary outcomes were reported as full sequencing completed or incomplete. Second, test outcome is noted as mutation characterised or not detected. One laboratory provided no information on incomplete tests; only the number of such tests is known. It is thus possible only to estimate the failure rate by tissue or biopsy type for the whole sample. We compared first the rates of biopsy, cell block and FNA in the sets of completely and incompletely sequenced samples. Differences were compared with chi-squared tests. To estimate the overall failure rate by biopsy type, we assumed that the distribution of biopsy types was similar in the unreported results as for those where biopsy type was provided. The rates of testing, in the past 12 months of the programme period from ACT/New South Wales, South Australia/Northern Territory, Tasmania and Victoria, were compared to the number of new cases of NSCLC expected based on the most recent AIHW data.14 We used the number of valid tests as the numerator as we could not determine the number of incomplete tests by State.
Results Samples of 2013 were submitted from 2012 patients for analysis. One patient had both brain and liver biopsies sent for analysis. In calculating the yield of a valid test result, these samples are considered separately, but are taken as one for reporting of the spectrum of mutations. The rate of testing per new case of lung cancer was highest in Victoria (Table 1). Sequencing was complete for exons of interest in 1717 (85%). In one additional case, an activating mutation was detected, but other exons were not fully sequenced. Reasons for an incomplete test were provided in 98 patients. In 17, no tissue was found in the blocks submitted, and in 25 no tumour tissue was detected. In a further 25, there was a failure of polymerase chain reaction (PCR) amplification related to poor quality DNA, and in 31 only partial sequencing was completed. In these cases, © 2014 The Authors Internal Medicine Journal © 2014 Royal Australasian College of Physicians
EGFR mutations in NSCLC in Australia
Table 1 Distribution of studied cases by State or Territory together with the rate of testing based on samples tested as a proportion of all new lung cancer cases diagnosed in the study period based on 2011 data from the Australian Institute of Health and Welfare14 State
ACT/New South Wales Queensland South Australia/NT Tasmania Victoria Victoria/Tasmania combined
Number of valid test results
Valid tests per annual lung cancer case
Rate of activating mutation detection (%)
501 156 79 25 586 611
0.157 0.037 0.200 0.089 0.232 0.217
17.4 20.3 14.7 12.0 13.5 13.4
The rate of activating mutations is the proportion of valid samples showing L858R point mutations or exon 19 deletions.
exon 19 was sequenced in 12 cases and exon 21 in 17 cases. The relationship between the biopsy procedure and test outcome is shown in Table 2. Fine needle biopsies were more likely to result in incomplete sequencing than either tissue biopsy (P
