You are here

ctDNA Determination of EGFR Mutation Status in European and Japanese Patients with Advanced NSCLC: The ASSESS Study

Journal of Thoracic Oncology, Volume 11, Issue 10, October 2016, Pages 1682 - 1689

Commentary by Tom Stinchcombe

There are logistical, practical, and technical barriers to performing biopsies in patients with advanced NSCLC, and even when successfully performed some specimens do not contain enough tumor cells for molecular testing This problem has become more relevant as there is increased need for more testing, and the need to assess patients with an EGFR mutation who have progressed on an EGFR TKI for the acquired resistance T790M mutation. This non-interventional study assessed the performance of circulating free tumor-derived DNA (ctDNA) for EGFR mutation testing. Of the 1311 patients enrolled 1288 were eligible, and the prevalence of EGFR mutations was 16% for evaluable tissue/cytologic samples, and 9% for plasma samples.  The concordance of mutation status in the 1162 matched samples was 89%. The sensitivity was 46%, specificity was 97%, and the positive predictive value was 78%, and the negative predictive value was 90%. Some the false positive plasma results were related to the use of insensitive methods of testing the tumor sample. These data suggest ctDNA is a feasible method of testing for EGFR mutations. Previous studies had been done as part of a clinical trial with standardized testing methods and a central laboratory, and this study was performed outside the context of a clinical trial and may be more applicable to routine clinical practice. The ability of plasma tests to detect T790M mutations not detected on tumor samples due to tumor heterogeneity or problems with tumor testing methods has been observed in other studies as well. Patients with extra-thoracic metastases (M1b disease) are more likely to have a mutation detected. If an EGFR mutation is detected in ctDNA then the patient can initiate therapy with an EGFR TKI. However, if a mutation is not detected and another oncogenic driver (e.g. KRAS) is not detected then repeat biopsy should be considered.

Commented by Nir Peled

EGFR TKIs have been approved recently upon ctDNA for 1st line therapy. This regulatory breakthrough may lead to a shorter turnaround time and to a higher rate of detection of EGFR mutations. In 2nd line, ctDNA may even offer a better reflection of the total body still the main barrier for wide implementation is accuracy and expense. Reck et al have performed an international study (ASSESS study) reported recently in the recent TJO, showing the feasibility of ctDNA and its good concordance with tissue sampling in 1162 NSCLC patients. Different platforms were used showing a concerning range of sensitivity to detect EGFR mutation through the ctDNA (40%-75%). As expected though, with a relatively good specificity and concordance to tissue sampling. This study emphasizes the need for a careful platform selection and appropriate QA processes to ensure higher rate of detection for both naïve patients and on progression. Until a better sensitivity will be presented, it is our duty to make sure we do not leave any patient behind. Negative ctDNA EGFR testing should be repeated or validated by a tissue biopsy until proven otherwise. 



To offer patients with EGFR mutation–positive advanced NSCLC appropriate EGFR tyrosine kinase inhibitor treatment, mutation testing of tumor samples is required. However, tissue/cytologic samples are not always available or evaluable. The large, noninterventional diagnostic ASSESS study (NCT01785888) evaluated the utility of circulating free tumor-derived DNA (ctDNA) from plasma for EGFR mutation testing.


ASSESS was conducted in 56 centers (in Europe and Japan). Eligible patients (with newly diagnosed locally advanced/metastatic treatment-naive advanced NSCLC) provided diagnostic tissue/cytologic and plasma samples. DNA extracted from tissue/cytologic samples was subjected to EGFR mutation testing using local practices; designated laboratories performed DNA extraction/mutation testing of blood samples. The primary end point was level of concordance of EGFR mutation status between matched tissue/cytologic and plasma samples.


Of 1311 patients enrolled, 1288 were eligible. Concordance of mutation status in 1162 matched samples was 89% (sensitivity 46%, specificity 97%, positive predictive value 78%, and negative predictive value 90%). A group of 25 patients with apparent false-positive plasma results was overrepresented for cytologic samples, use of less sensitive tissue testing methodologies, and smoking habits associated with high EGFR mutation frequency, indicative of false-negative tumor results. In cases in which plasma and tumor samples were tested with identical highly sensitive methods, positive predictive value/sensitivity were generally improved.


These real-world data suggest that ctDNA is a feasible sample for EGFR mutation analysis. It is important to conduct mutation testing of both tumor and plasma samples in specialized laboratories, using robust/sensitive methods to ensure that patients receive appropriate treatments that target the molecular features of their disease.

Keywords: NSCLC, EGFR mutation, ctDNA, Plasma, Real-world.


NSCLC accounts for approximately 83% of all lung cancers, and most patients present with advanced disease at diagnosis.1 Adenocarcinoma (ADC), which is among the most common of the NSCLC histological subtypes, has been shown to be associated with activating mutations in the EGFR gene in 13% of European and 47% of Japanese patients (corresponding rates for non-ADC 5% and 7%, respectively).2 Response to EGFR tyrosine kinase inhibitors (TKIs) is acknowledged to be greater in patients with tumors harboring activating mutations in the tyrosine kinase domain of the EGFR gene versus wild-type EGFR.3 Furthermore, EGFR TKIs have demonstrated efficacy superior to that of doublet chemotherapy in patients with EGFR mutation–positive advanced NSCLC.4, 5, 6, 7, and 8 Clinical guidelines9 and 10 and several working groups11 and 12 thus advocate mutation testing of tumor sample DNA from patients with nonsquamous advanced NSCLC to confirm suitability for EGFR TKI treatment. However, tissue/cytologic samples are not always available or evaluable for diagnosis and mutation testing, leaving some patients unable to have the EGFR mutation status of their tumors determined.13

Several studies have demonstrated that it is feasible to assess EGFR mutation status by using circulating free tumor-derived DNA (ctDNA), which can be isolated from the plasma or serum of patients with NSCLC.4 and 14 Indeed, nucleic acid released from tumor cells into the circulation represents an alternative source of tumor-derived DNA, an approach referred to as liquid biopsy.13 Indicative of the clinical relevance of ctDNA in mutation analysis, the presence of EGFR mutations in ctDNA has been found to predict response to EGFR TKIs7, 13, and 15; similar objective response rates and progression-free survival have also been observed between patients with EGFR mutation–positive NSCLC detected by testing of tissue/cytologic samples and patients with EGFR mutations detected in their ctDNA.4 Indeed, an amendment to the summary of product characteristics for the EGFR TKI gefitinib stating that ctDNA may be used for mutation analysis of NSCLC when tumor samples are unavailable16 was adopted by the Committee for Medicinal Products for Human Use in September 2014.

As ctDNA analysis is technically challenging, it is important that the accuracy, suitability, and feasibility of use of ctDNA for mutation analysis in clinical practice be established. In the open-label IFUM study of white patients with EGFR mutation–positive NSCLC, between 652 evaluable matched tissue/cytologic and plasma (ctDNA) samples, mutation status concordance was 94%, demonstrating a high accuracy of ctDNA-based mutation testing with respect to tissue-based testing.4 However, there is a need to confirm such findings outside of the controlled clinical trial environment to improve local diagnostic practice, enable wider access to ctDNA mutation testing, and provide more patients with the opportunity to receive therapies personalized to the mutation status of their tumors. The large, multicenter, noninterventional, diagnostic ASSESS study (NCT0178588) investigated the utility of ctDNA for EGFR mutation testing in a real-world diagnostic setting.

Materials and Methods

Study Design and Participants

Eligible patients were those at least 18 years of age (≥20 years in Japan) with histologically/cytologically confirmed, systemic treatment-naive, locally advanced NSCLC (stage IIIA/B according to the American Joint Committee on Cancer staging system) unsuitable for curative treatment or chemoradiotherapy or with metastatic disease (stage IV) and those with recurrent disease after previous curative treatment (resection and/or adjuvant chemotherapy). Provision of a diagnostic tissue/cytologic sample and a fresh blood (plasma) sample at enrollment was mandatory. The blood sample was taken before initiation of any treatment.

The primary end point of ASSESS was the level of concordance between EGFR mutation status obtained by tissue/cytologic testing and that obtained by blood (plasma)-based testing. Secondary end points included EGFR mutation testing practices and mutation frequency and correlations between mutation status and demographic data/disease status and treatment decisions after mutation testing (not reported).

All patients provided written, informed consent. Study approval was obtained from the independent ethics committees at each institution. The study was conducted in accordance with the Declaration of Helsinki, the International Conference on Harmonization/Good Clinical Practice, applicable regulatory requirements for noninterventional studies, and AstraZeneca’s policy on bioethics and human biological samples.


EGFR mutation testing and results data for tumor samples obtained before enrollment in ASSESS were used where available (see Fig. 1A). For tests conducted in ASSESS, diagnostic tissue/cytologic samples were subjected to EGFR mutation testing as per local practices after histopathological review (WHO classification17) to ensure that samples were adequate for use. Each patient provided a 10-mL blood sample, which was processed to plasma, frozen, and transported to designated laboratories for mutation testing. In Europe, 43 academic, hospital, and commercial laboratories performed the tissue/cytologic testing and five central/regional expert laboratories were used for blood-based testing. In Japan, two commercial laboratories were used for both tissue/cytologic and blood-based testing.


Figure 1

Mutation testing methods used for tissue/cytologic (A) and plasma (B) samples in Europe and Japan (enrolled population). Other refers to Fragment Length Analysis/PNA-LNA PCR Clamp. PCR, polymerase chain reaction; PNA-LNA, peptide nucleic acid-locked nucleic acid; RGQ, Rotor-Gene Q.



Testing methodologies, sample types and availability, testing turnaround time, and testing success rate/mutation detection rate were captured to assess EGFR mutation testing practices. Concordance of EGFR mutation status between matched tissue/cytologic and plasma samples was assessed by using the following: concordance rate; sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) (Supplementary Table 1); and exact two-sided 95% confidence interval (CI).

Statistical Analyses

Concordance of EGFR mutation status between matched tissue/cytologic and plasma samples was calculated for the evaluable population (all eligible patients with known tumor [tissue/cytologic] and plasma sample mutation status). The sample size of ASSESS was calculated to allow accurate determination of sensitivity between tumor- and blood-based testing. For 100 patients with EGFR mutation-positive NSCLC, if the sensitivity was 50%, the 95% CI would be expected to be 40% to 60%. Assuming a 10% prevalence of EGFR mutations in Europe, it was estimated that 1000 patients should be tested to obtain 100 with EGFR mutation–positive NSCLC. As an independent replication, sensitivity was obtained in patients from Japan; assuming a 30% prevalence of EGFR mutations in Japan, it was estimated that 300 patients should be tested to obtain 100 with EGFR mutation–positive NSCLC.

EGFR mutation testing practices (enrolled population) and EGFR mutation frequency (evaluable tumor/plasma populations) were summarized by using appropriate descriptive statistics.



From April 11, 2013, to April 17, 2014, a total of 1311 patients were enrolled (Table 1), 1288 of whom were eligible (Fig. 2). A total of 23 patients were enrolled in error (they did not fulfil the inclusion/exclusion criteria) and were not included in the final analysis (Supplementary Table 2).

Table 1

Patient Demographics (Enrolled Population)


Characteristic Europe Japan Overall
(n = 997) (n = 291) (n = 1288)
Mean age ± SD 65.4 ± 9.74 70.2 ± 9.0 66.5 ± 9.8
Men, n/N (%) 675 of 997 (68) 192 of 291 (66) 867 of 1288 (67)
Race, n/N (%)
 White 976 of 997 (98) 0 of 291 (0) 976 of 1288 (76)
 Asian 5 of 997 (1) 291 of 291 (100) 296 of 1288 (23)
WHO performance status, n/N (%)
 0 388 of 997 (39) 114 of 291 (39) 502 of 1288 (39)
 1 453 of 997 (45) 117 of 291 (40) 570 of 1288 (44)
 2 136 of 997 (14) 37 of 291 (13) 173 of 1311 (13)
 3 17 of 997 (2) 18 of 291 (6) 35 of 1288 (3)
 4 3 of 997 (0) 5 of 291 (2) 8 of 1288 (1)
Smoking status
 Never-smoker, n/N (%) 174 of 996 (17) 78 of 291 (27) 252 of 1287 (20)
 Pack-years, median 40.0 45.0 40.0

Note: Pack-years = (number of cigarettes smoked per day × number of years smoked)/20.


Figure 2

Patient flow diagram.


Sampling Methodologies and EGFR Mutation Testing Practices

Tissue/cytologic samples were mostly collected during current diagnosis (Supplementary Fig. 1A) and most often sampled from the primary tumor (Supplementary Fig. 1B). The most frequent sampling sites were the lungs/lymph nodes (Supplementary Fig. 1C). Bronchoscopy was the most frequently used sample collection method (Supplementary Fig. 1D). A range of mutation testing methods for tissue/cytologic samples (>10) were observed in Europe versus only two methods in Japan (Fig. 1A). The mutation testing methods observed for plasma were limited and differed between Europe and Japan (Fig. 1B).

The median EGFR test turnaround time for tissue/cytologic samples was 11 days for Europe (95% CI: 14.0–17.3) and 8 days for Japan (95% CI: 8.2–14.1); mutation testing success rates were high in both Europe (98%) and Japan (<100%). Mutation tests were not performed on the tissue/cytologic samples of 110 patients and results were not obtained from the tested samples of 17 patients. The most frequent reason for not testing was insufficient material provided (60% [47 of 78 responses to the question] in Europe and 56% [five of nine] in Japan). There were no obvious differences in sampling methods between evaluable and nonevaluable tissue/cytologic samples.

EGFR Mutation Frequency

The overall EGFR mutation frequency was 16% for evaluable tissue/cytologic samples and 9% for evaluable plasma samples (Table 2). Consistent with previous reports,2 in Europe, 14% of ADC tissue/cytologic samples (99 of 712) and 3% of non-ADC samples (six of 180) were EGFR mutation–positive. In Japan, 40% of ADC tissue/cytologic samples (78 of 195) and 8% (six of 77) of non-ADC samples were EGFR mutation–positive.

Table 2

EGFR Mutation Frequency by Sample Type, Region, and Histologic Subtype (Population with Evaluable Tissue/Cytologic and/or Plasma Samples)


Characteristic EGFR mutation frequency
Tissue/Cytologic Samples Plasma
n/n (%) n/n (%)
Overall 191 of 1184 (16) 119 of 1263 (9)
 Europe 105 of 903 (12) 82 of 972 (8)
 Japan 86 of 281 (31) 37 of 291 (13)
Histological subtype
 ADC 177 of 907 (20) 109 of 952 (11)
 Non-ADC 12 of 257 (5) 9 of 288 (3)
TNM stage
 IIIA 5 of 66 (8) 3 of 75 (4)
 IIIB 3 of 105 (3) 2 of 119 (2)
 IV 183 of 1006 (18) 114 of 1063 (11)
TNM stage IV
 M1a 57 of 235 (24) 17 of 252 (7)
 M1b 82 of 490 (17) 67 of 528 (13)
EGFR mutation subtype
 Exon 19 deletions 97 of 191 (51) 68 of 119 (57)
 Exon 19 deletions + T790M 0 of 191 (0) 0 of 119 (0)
 L858R 73 of 191 (38) 38 of 119 (32)
 L858R + T790M 0 of 191 (0) 2 of 119 (2)
 T790M only 0 of 191 (0) 3 of 119 (3)
 T790M + othera 1 of 191 (1) 1 of 119 (1)
 Otherb 20 of 191 (10) 7 of 119 (6)

a Any other mutation that occurred in combination with T790M that is not L858R or Exon 19 deletion.

b This category included double mutations not specified.

ADC, adenocarcinoma.

Concordance of EGFR Mutation Status between Matched Tissue/Cytologic and Plasma Samples

Overall concordance of mutation status was 89% (sensitivity 46%, specificity 97%, PPV 78%, and NPV 90%) (Table 3). An EGFR mutation–positive result in plasma and mutation-negative result in tissue/cytologic testing were observed in samples from 25 patients (false-positive rate 22%); the demographics and sampling/mutation testing methodologies associated with these patients are summarized in Supplementary Table 3. False-negative results were obtained for 102 patients (10%).

Table 3

Concordance of EGFR Mutation Status between Matched Tissue/Cytologic and Plasma Samples (Population with Evaluable Tissue/Cytologic and Plasma Samples)


Characteristic Concordance Rate Sensitivity Specificity PPV NPV
n/n (%) 95% CI n/n (%) 95% CI n/n (%) 95% CI n/n (%) 95% CI n/n (%) 95% CI
Overall (n = 1162) 1035 of 1162 (89) 87.1–90.8 87 of 189 (46) 38.8–53.4 948 of 973 (97) 96.2–98.3 87 of 112 (78) 68.8–85.0 948 of 1050 (90) 88.3–92.0
 Japan (n = 281) 227 of 281 (81) 75.7–85.2 34 of 86 (40) 29.2–50.7 193 of 195 (99) 96.3–99.9 34 of 36 (94) 81.3–99.3 193 of 245 (79) 73.1–83.7
 Europe (n = 881) 808 of 881 (92) 89.7–93.4 53 of 103 (51) 41.4–61.4 755 of 778 (97) 95.6–98.1 53 of 76 (70) 58.1–79.8 755 of 805 (94) 91.9–95.4
Qiagen therascreen EGFR RGQ PCR Kit (n = 138) 131 of 138 (95) 89.8–97.9 16 of 22 (73) 49.8–89.3 115 of 116 (99) 95.3–100 16 of 17 (94) 71.3–99.9 115 of 121 (95) 89.5–98.2
Roche cobas EGFR Mutation Test (n = 23) 22 of 23 (96) 78.1–99.9 3 of 4 (75) 19.4–99.4 19 of 19 (100) 82.4–100 3 of 3 (100) 29.2–100 19 of 20 (95) 75.1–100
Cycleave (n = 190) 161 of 190 (85) 78.8–89.5 29 of 57 (51) 37.3–64.4 132 of 133 (99) 95.9–100 29 of 30 (97) 82.8–99.9 132 of 160 (83) 75.5–88.0
PNA-LNA PCR Clampa (n = 91) 76 of 91 (84) 74.3–90.5 15 of 29 (52) 32.5–70.6 61 of 62 (98) 91.3–100 15 of 16 (94) 69.8–99.8 61 of 75 (81) 70.7–89.4

a PNA-LNA PCR Clamp concordance data using optimized ctDNA extraction procedure.

PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; RGQ, Rotor-Gene Q; PCR, polymerase chain reaction; PNA-LNA, peptide nucleic acid-locked nucleic acid; ctDNA, circulating free tumor-derived DNA.

When tissue/cytologic and plasma samples were tested with identical, sensitive methods (Qiagen therascreen EGFR Rotor-Gene Q Polymerase Chain Reaction (PCR) Kit, Peptide Nucleic Acid–Locked Nucleic Acid PCR Clamp [Qiagen, Manchester, UK], Roche cobas EGFR Mutation Test [Roche Molecular Diagnostics, Pleasanton, CA], or Cycleave [Takara Bio Inc., Kusatsu, Japan]), the PPV and sensitivity were generally improved (see Table 3).

Furthermore, in a subset of 94 Japanese patients, sensitivity increased from an initial result of 17% (five of 29) to 52% (15 of 29) when samples that had originally had ctDNA extracted using the Qiagen QIAamp MinElute Virus Spin Kit for DNA (400 μL plasma) had extraction repeated using the Qiagen QIAamp Circulating Nucleic Acid Kit (3 mL plasma) specifically designed to isolate smaller, fragmented DNA, even though the same detection method (peptide nucleic acid–locked nucleic acid PCR Clamp) had been used. Including the newer Japanese subset data in the overall data increased the concordance rate (90% [1036 of 1158]), sensitivity (50% [97 of 194]), PPV (80% [97 of 122]), and NPV (91% [939 of 1036]).


To our knowledge, ASSESS is the most comprehensive study of real-world EGFR mutation testing practices in Europe and Japan. ASSESS was designed to evaluate how data obtained using local practices for tissue/cytologic and plasma sample testing compare with data obtained from well-controlled clinical testing that is generally performed in a single expert laboratory.

The mutation status concordance observed between 1162 matched tissue/cytologic and plasma samples (89%) suggests that ctDNA is a feasible sample for real-world EGFR mutation analysis if robust/sensitive DNA extraction and mutation analysis methodologies are used. ctDNA comprises less than 1% of total circulating free DNA18; including data from the subset of patients who had DNA reextracted with an optimized method increased the overall concordance rate and sensitivity.

The low PPV (78%) raised a concern that some of the plasma EGFR mutation results might be false positives, and 25 patients for whom testing yielded apparent false-positive plasma results were identified (Supplementary Table 3). This group was overrepresented for cytologic samples, which may not be representative of the disease, and/or the use of less sensitive DNA sequencing methodologies (56% of these 25 tumor samples were tested by DNA sequencing/pyrosequencing versus 25% of samples from the overall population [305 of 1201]). In addition, 80% of these patients were never-, former, or light smokers, a subgroup in which a higher EGFR mutation frequency is expected. Furthermore, in Europe, the EGFR mutation frequency by DNA sequencing/pyrosequencing of tissue/cytologic samples was 10% (37 of 370) versus 15% (68 of 465) for all other methods combined, suggesting that DNA sequencing/pyrosequencing may have missed some of the EGFR mutations present. This finding is in agreement with results of external quality assessment studies that have revealed a high rate of genotyping errors in Europe, where less sensitive techniques are still frequently used.19 and 20 Overall, these data suggest that the low PPV was due to false-negative results in tumor rather than to false-positive results in plasma. False-negative results can occur because of tumor heterogeneity, inexperience, or use of methods with low sensitivity.

In contrast, it is clear that a negative result in plasma may be incorrect (false-negative), as ctDNA mutation analysis may not pick up EGFR mutations in all patients who carry them in their tumors. The sensitivity of plasma tests, which is indicative of how likely they are to accurately detect EGFR mutations, varied (36%–100%) between countries. ASSESS revealed substantial differences within/between Europe and Japan in the mutation testing methodologies used, reflecting differences in available and/or accessible equipment and knowledge, which may contribute to this variation. However, overall sensitivity (46%) was markedly lower than the pooled sensitivities reported in two recent meta-analyses of concordance of EGFR mutation status between matched tissue/cytologic and plasma samples (62%–67%).21 and 22

Caution is advised when comparing data from clinical trials, in which testing is generally well controlled and performed in a single central laboratory, with the real-world ASSESS data. Of note, the concordance rate of 94% (sensitivity 66%, specificity 100%, PPV 99%, and NPV 94%) reported for IFUM was higher than in ASSESS.4 Likewise, the sensitivity and PPV were higher in FASTACT-223 (75% and 94%, respectively) and sensitivity in the EURTAC study24 was 78%. IFUM utilized the Qiagen therascreen EGFR Rotor-Gene Q PCR Kit for mutation analysis of matched tissue/cytologic and plasma samples. Interestingly, the ASSESS concordance data for matched tissue/cytologic and plasma samples tested with the therascreen kit were highly consistent with that of IFUM (concordance 95% [131 of 138], sensitivity 73% [16 of 22], specificity 99% [115 of 116], PPV 94% [16 of 17], and NPV 95% [115 of 121]). This confirms the findings of the IFUM study,4 which indicated that when an EGFR mutation is detected in ctDNA, the result can be trusted and a prescriber can conclude that the patient is likely to benefit from TKI therapy.

In real-world practice, test sensitivity may be further complicated by the different stage/differentiation of the disease,25 which likely influence release of ctDNA into blood. In ASSESS, a higher frequency of EGFR mutations was observed in the plasma of patients with M1b disease (13%) versus M1a disease (7%) despite the EGFR mutation frequency in tumor being similar in these two groups, which is consistent with the hypothesis of greater release of ctDNA in patients with distant metastases. It must be noted that analysis of ctDNA may allow identification of mutations in heterogeneous tumors that could carry a specific variant in only a minor tumor clone (i.e., beyond the “biopsy zone”). However, although plasma-based testing requires standardization as a technique under relatively early investigation, the ASSESS results suggest that further improvements in mutation analysis of tissue/cytologic samples are still required to reduce the occurrence of false-negative results.

In IFUM, 19% of patients had nonevaluable tissue/cytologic samples for reasons including low tumor content, insufficient sample quality/quantity, poor/inappropriate fixation, and no DNA.4 In this respect, accurate and accessible ctDNA mutation testing to address the unmet need in patients without an available/evaluable tumor sample will be important to enable more patients to receive therapies personalized to the mutation status of their tumor.


These ASSESS real-world data suggest that ctDNA is a feasible sample for EGFR mutation analysis when tumor samples are unavailable. It is important to conduct mutation testing of both tumor and plasma samples in specialized laboratories, using robust and sensitive mutation testing methods to ensure accuracy of results.


This work was supported by AstraZeneca and coordinated by Worldwide Clinical Trials, which also managed the database and performed the primary analyses. In collaboration with AstraZeneca, the study steering committee interpreted the study results. Dr. Normanno had full access to the study data and final responsibility for the decision to submit for publication. The authors thank the patients and investigators for their participation in this study and Rosemary Taylor, Jukka Montonen, Mandy Garratt, and Sylvie van Hoorde for various aspects of study data analysis and delivery. Data collection and delivery were coordinated by Worldwide Clinical Trials and funded by AstraZeneca. Iressa is a trademark of the AstraZeneca group of companies. We thank Louise Brown from Complete Medical Communications, who provided medical writing support funded by AstraZeneca. Dr. Normanno was supported by a grant from the Italian Association of Cancer Research (grant IG17135). Drs. Reck, Hagiwara, Han, Tjulandin, and Normanno acted as study investigators and contributed to the protocol design and interpretation of data. In addition, Drs. Reck, Hagiwara, and Normanno contributed to enrollment and management of patients. Drs. Grohé, Yokoi, Morabito, Novello, Arriola, and Molinier acted as study investigators and contributed to the enrollment and management of patients, and interpretation of data. Drs. McCormack and Ratcliffe contributed to the protocol design and interpretation of data. All authors participated in the writing and review of the manuscript and accept full responsibility for the overall content. Rosemary Taylor contributed to the statistical analysis. A list of the ASSESS study principal investigators and study centers is included in Supplementary Table 4.

Supplementary Data

Download file

Legend for Supplementary Figure 1 and Supplementary Tables 1–4

Download file

Supplementary Figure 1


  • 1 National Cancer Institute. SEER cancer statistics review, 1975-2012. Accessed July 8, 2016.
  • 2 S. Dearden, J. Stevens, Y.L. Wu, et al. Mutation incidence and coincidence in non-small-cell lung cancer: meta-analyses by ethnicity and histology (mutMap). Ann Oncol. 2013;24:2371-2376 Crossref
  • 3 G.J. Riely, K.A. Politi, V.A. Miller, et al. Update on epidermal growth factor receptor mutations in non-small cell lung cancer. Clin Cancer Res. 2006;12:7232-7241 Crossref
  • 4 J.Y. Douillard, G. Ostoros, M. Cobo, et al. Gefitinib treatment in EGFR mutated Caucasian NSCLC: circulating-free tumor DNA as a surrogate for determination of EGFR status. J Thorac Oncol. 2014;9:1345-1353 Crossref
  • 5 T. Mitsudomi, S. Morita, Y. Yatabe, et al. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. Lancet Oncol. 2010;11:121-128 Crossref
  • 6 T.S. Mok, Y.-L. Wu, S. Thongprasert, et al. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med. 2009;361:947-957 Crossref
  • 7 R. Rosell, E. Carcereny, R. Gervais, et al. Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial. Lancet Oncol. 2012;13:239-246 Crossref
  • 8 C. Zhou, Y.-L. Wu, G. Chen, et al. Erlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): a multicentre, open-label, randomised, phase 3 study. Lancet Oncol. 2011;12:735-742 Crossref
  • 9 National Comprehensive Cancer Network. Practice guidelines in oncology—version V.3.2012 (non-small-cell lung cancer). Accessed May 19, 2016.
  • 10 NICE. EGFR-TK mutation testing in adults with locally advanced or metastatic non-small-cell lung cancer. Accessed July 8, 2016.
  • 11 A. Marchetti, N. Normanno. Recommendations for mutational analysis of EGFR in lung carcinoma. Pathologica. 2010;102:119-126
  • 12 R. Pirker, F.J. Herth, K.M. Kerr, et al. Consensus for EGFR mutation testing in non-small cell lung cancer: results from a European workshop. J Thorac Oncol. 2010;5:1706-1713 Crossref
  • 13 F. Fenizia, A. De Luca, R. Pasquale, et al. EGFR mutations in lung cancer: from tissue testing to liquid biopsy. Future Oncol. 2015;11:1611-1623 Crossref
  • 14 R. Rosell, T. Moran, C. Queralt, et al. Screening for epidermal growth factor receptor mutations in lung cancer. N Engl J Med. 2009;361:958-967 Crossref
  • 15 J.-Y. Douillard, G. Ostoros, M. Cobo, et al. First-line gefitinib in Caucasian EGFR mutation-positive NSCLC patients: a phase-IV, open-label, single-arm study. Br J Cancer. 2014;110:55-62 Crossref
  • 16 AstraZeneca UK Ltd. IRESSA Summary of Product Characteristics. Available at: Accessed January 21, 2016.
  • 17 E. Brambilla, W.D. Travis, T.V. Colby, et al. The new World Health Organization classification of lung tumours. Eur Respir J. 2001;18:1059-1068 Crossref
  • 18 L.A. Diaz Jr., A. Bardelli. Liquid biopsies: genotyping circulating tumor DNA. J Clin Oncol. 2014;32:579-586 Crossref
  • 19 Z.C. Deans, N. Bilbe, B. O'Sullivan, et al. Improvement in the quality of molecular analysis of EGFR in non-small-cell lung cancer detected by three rounds of external quality assessment. J Clin Pathol. 2013;66:319-325 Crossref
  • 20 S. Patton, N. Normanno, F. Blackhall, et al. Assessing standardization of molecular testing for non-small-cell lung cancer: results of a worldwide external quality assessment (EQA) scheme for EGFR mutation testing. Br J Cancer. 2014;111:413-420 Crossref
  • 21 M. Qiu, J. Wang, Y. Xu, et al. Circulating tumor DNA is effective for the detection of EGFR mutation in non-small cell lung cancer: a meta-analysis. Cancer Epidemiol Biomarkers Prev. 2015;24:206-212 Crossref
  • 22 J. Luo, L. Shen, D. Zheng. Diagnostic value of circulating free DNA for the detection of EGFR mutation status in NSCLC: a systematic review and meta-analysis. Sci Rep. 2014;4:6269 Crossref
  • 23 T. Mok, Y.L. Wu, J.S. Lee, et al. Detection and dynamic changes of EGFR mutations from circulating tumor DNA as a predictor of survival outcomes in NSCLC patients treated with first-line intercalated erlotinib and chemotherapy. Clin Cancer Res. 2015;21:3196-3203 Crossref
  • 24 N. Karachaliou, C.C. Mayo-de las, C. Queralt, et al. Association of EGFR L858R mutation in circulating free DNA with survival in the EURTAC trial. JAMA Oncol. 2015;1:149-157 Crossref
  • 25 X. Zhao, R.B. Han, J. Zhao, et al. Comparison of epidermal growth factor receptor mutation statuses in tissue and plasma in stage I-IV non-small cell lung cancer patients. Respiration. 2013;85:119-125 Crossref


a Department of Thoracic Oncology, Lung Clinic Grosshansdorf GmbH, Airway Research Center North (ARCN), German Centre for Lung Research (DZL), Grosshansdorf, Germany

b Comprehensive Medicine 1, Saitama Medical Center, Jichi Medical University, Saitama Prefecture, Japan

c Department of Respiratory Medicine, Shanghai Chest Hospital, Jiao Tong University, Shanghai, People's Republic of China

d Department of Clinical Pharmacology and Chemotherapy, Russian Cancer Research Center, Moscow, Russian Federation

e Pulmonary Clinic, Evangelical Lung Clinic, Berlin, Germany

f Department of Thoracic Oncology, Kansai Medical University Hirakata Hospital, Osaka, Japan

g Thoracic Oncology Unit, National Cancer Institute “Pascale Foundation,” Institute of Hospitalization and Scientific Care (IRCCS), Naples, Italy

h Thoracic Oncology Unit, Department of Oncology, University of Turin, Turin, Italy

i Oncology Department, Hospital of the Sea, Barcelona, Spain

j Pneumology, Le Mans General Hospital, Le Mans, France

k Personalised Healthcare and Biomarkers, AstraZeneca, Macclesfield, United Kingdom

l Cell Biology and Biotherapy Unit, National Cancer Institute “Pascale Foundation,” Institute of Hospitalization and Scientific Care (IRCCS), Naples, Italy

Corresponding author. Address for correspondence: Nicola Normanno, MD, Cell Biology and Biotherapy Unit, Istituto Nazionale Tumori “Fondazione Pascale,” IRCCS, Via M. Semmola, 80131 Naples, Italy.

Disclosure: Dr. Reck discloses being a member of the speakers bureaus of AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Daiichi-Sankyo, Hoffmann-La Roche, Lilly, and Pfizer and a consultant for AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Daiichi-Sankyo, Hoffmann-La Roche, Lilly, Merck Sharp and Dohme, and Pfizer. Dr. Hagiwara discloses being a member of the speakers bureaus of AstraZeneca, Chugai Pharmaceuticals, and Pfizer and holding a patent with LSI Medience. Dr. Han discloses being a member of the speakers bureaus of AstraZeneca and Roche and a consultant for AstraZeneca and Pfizer. Dr. Tjulandin discloses being a member of the speakers bureaus of AstraZeneca, Pfizer, and Sanofi-Aventis. Dr. Grohé discloses being a member of the speakers bureaus of AstraZeneca and Roche and a consultant for AstraZeneca, Pfizer, and Roche. Dr. Morabito discloses being a member of the speakers bureaus of AstraZeneca, Boehringer Ingelheim, and Roche and a consultant for Daiichi-Sankyo and Pfizer. Dr. Novello discloses being a member of the speakers bureaus of AstraZeneca, Boehringer Ingelheim, Eli Lilly, Merck Sharp and Dohme, and Roche. Dr. Arriola discloses being a member of the speakers bureaus of Boehringer Ingelheim, Lilly, and Pfizer and receiving grants/research support from Pfizer. Dr. Molinier discloses being a consultant for Boehringer Ingelheim and Roche. Drs. McCormack and Ratcliffe disclose being employees of and holding shares in AstraZeneca. Dr. Normanno discloses receiving grants/research support from and being a consultant for AstraZeneca, Qiagen, and Roche Diagnostics. The remaining author declares no conflict of interest.