Purpose: Tumor gene mutation status is becoming increasingly important in the treatment of patients with cancer. A comprehensive catalog of tumor gene–response outcomes from individual patients is needed, especially for actionable mutations and rare variants. We created a proof-of-principle database [DNA-mutation Inventory to Refine and Enhance Cancer Treatment (DIRECT)], starting with lung cancer-associated EGF receptor (EGFR) mutations, to provide a resource for clinicians to prioritize treatment decisions based on a patient's tumor mutations at the point of care.

Methods: A systematic search of literature published between June 2005 and May 2011 was conducted through PubMed to identify patient-level, mutation–drug response in patients with non–small cell lung cancer (NSCLC) with EGFR mutant tumors. Minimum inclusion criteria included patient's EGFR mutation, corresponding treatment, and an associated radiographic outcome.

Results: A total of 1,021 patients with 1,070 separate EGFR tyrosine kinase inhibitor therapy responses from 116 different publications were included. About 188 unique EGFR mutations occurring in 207 different combinations were identified: 149 different mutation combinations were associated with disease control and 42 were associated with disease progression. Four secondary mutations, in 16 different combinations, were associated with acquired resistance.

Conclusions: As tumor sequencing becomes more common in oncology, this comprehensive electronic catalog can enable genome-directed anticancer therapy. DIRECT will eventually encompass all tumor mutations associated with clinical outcomes on targeted therapies. Users can make specific queries at http://www.mycancergenome.org/about/direct to obtain clinically relevant data associated with various mutations. Clin Cancer Res; 19(7); 1894–901. ©2013 AACR.

See related commentary by Oxnard and Jänne, p. 1634

Translational Relevance

Tumor gene mutation status is becoming increasingly important in the treatment of patients with cancer. A comprehensive catalog of tumor gene–response outcomes from individual patients is needed, especially for actionable mutations and rare variants. Currently, no such catalog exists. Therefore, we created a proof-of-principle database [DNA-mutation Inventory to Refine and Enhance Cancer Treatment (DIRECT)], starting with lung cancer–associated EGF receptor mutations, to provide a resource for clinicians to prioritize treatment decisions based on a patient's tumor mutations. As tumor sequencing becomes more common in oncology, this comprehensive electronic catalog can enable genome-directed anticancer therapy. DIRECT will eventually encompass all tumor mutations associated with clinical outcomes on targeted therapies.

The treatment of patients with cancer in the 21st century has evolved into a complicated algorithm, requiring knowledge ofn individual patient's tumor mutation status before initiating therapy. Making mastery of knowledge even more difficult, mutational profiling studies have revealed that within a single type of cancer, even a single gene can harbor multiple different mutations in different individuals. EGF receptor (EGFR) mutations in non–small cell lung cancer (NSCLC) represent a prime example of the complexity of disease at the molecular level. In North America, about 17% of patients with non–small cell lung cancer (NSCLC) harbor EGFR mutations (1), of which approximately 80% to 90% are composed of exon 19 deletions and L858R point mutations (2), which inherently confer sensitivity to therapy with the EGFR tyrosine kinase inhibitors (TKI), gefitinib or erlotinib (3). However, there are rare mutations that can be associated with primary drug sensitivity, primary drug resistance, or secondary resistance, while other rarer EGFR mutations are of less clear clinical significance (4). This rapid evolution of information promises to overwhelm practicing clinicians who will need a mechanism whereby they can quickly and conveniently reference the most accurate therapeutic options for patients with known tumor mutations.

Here, we report on the development of a catalog of clinically relevant somatic mutations, named the DNA-mutation Inventory to Refine and Enhance Cancer Treatment (DIRECT). This clinical database is designed to enable a genetically informed approach to cancer medicine by providing clinicians access to tumor gene therapy–response information based on individual patient data published in the literature. As proof-of-principle, we started the catalog on EGFR mutations in lung cancer, aiming to provide a comprehensive database on rare and common mutations associated with disease control and disease progression to EGFR TKI therapy.

Searches of the PubMed database from June 2005 to May 2011 were completed to identify relevant studies that reported on EGFR mutations in patients with NSCLCs. The search strategy used the terms Receptor, Epidermal Growth Factor [Mesh] OR (EGFR protein, human [Substance Name]) AND Lung Neoplasms [Mesh]. Additional articles were identified from searching the bibliographies of retrieved articles. An English language restriction was applied. Details of the search strategies are available on request from the authors. All articles were manually reviewed to determine whether the publication met minimum criteria for inclusion by 2 independent reviewers (P. Yeh and J. Andrews). Additional articles were identified from searching the bibliographies of retrieved articles. The review of articles was conducted manually as most of the individual patient-level data, if presented, was within a table or image form, undetectable by automated, computerized methods. After the cutoff date for the PubMed search query, a retrospective search on the database was conducted to identify duplicate patient data published in multiple studies. The database was further verified by having 2 independent investigators (L. Horn and W. Pao) recheck at least every tenth patient data entry with the primary source reference. Any discrepancies were resolved through collective agreement.

All data was entered into an open source, electronic data capture program, the Research Electronic Data Capture (REDCap) project, from Vanderbilt University (Nashville, TN; refs. 5, 6). Only articles with individual patient data that provided the minimum following criterion were included: EGFR mutation, type of systemic treatment, and associated outcome including response rate (RR), progression-free survival (PFS), and/or overall survival (OS). At this time, only patients treated with the EGFR TKIs, gefitinib and erlotinib, were included. Data from studies with second-generation EGFR TKIs, such as afatinib (BIBW2992), neratinib (HKI-272), dacomitinib (PF-00299804), and XL647, will be included in future studies.

Response rate was classified into two categories, disease control and disease progression. Disease control was used to describe patients who achieved a complete response (CR), partial response (PR), or stable disease (SD), as defined by Response Evaluation Criteria in Solid Tumors (RECIST; ref. 7), following treatment with gefitinib or erlotinib. Disease progression was categorized as patients who had progressive disease (PD) as defined by the RECIST (7) following treatment with gefitinib or erlotinib. If a particular mutation had patients in both groups, that particular mutation was classified in the disease control or disease progression group based on whatever group had the higher percentage of a particular response. Additional data included if available were patient age, gender, ethnicity, smoking status, tumor histology, cancer stage at diagnosis, Eastern Cooperative Oncology Group (ECOG) performance status, and number of prior therapies. The PubMed ID, primary author name, journal name, and year of publication were recorded for each individual patient. Patients with acquired resistance to EGFR TKI therapy were catalogued separately (n = 61 from 9 different studies) from those with de novo mutations (i.e., disease previously untreated with EGFR TKIs).

Descriptive statistics, including medians, and ranges of continuous variables, as well as percentages and frequencies for categorical variables, were tabulated. The demographic analysis of the final DIRECT patient pool for this article was conducted using R Statistical Computing Program Version 14·1 (8).

The initial search identified 2085 publications including prospective studies, retrospective studies, and case reports (Fig. 1). Of the 2085 publications, 1,939 articles were excluded for not meeting minimum inclusion criteria. The most common reason for exclusion was data presented in an aggregate population form such as a clinical trial. Of the 2085 publications, 146 papers met minimum criterion for inclusion into the DIRECT database. The number of therapy–response entries exceeds the number of individual patient entries, as some patient reports included more than one line of treatment.

Figure 1.

Schematic flowchart of DIRECT database and patient cohort formation. Number of unique patients does not match the number of therapy-response entries as one patient may have been treated with EGFR TKIs in different lines of treatment and thus have multiple associated therapy-response entries.

Figure 1.

Schematic flowchart of DIRECT database and patient cohort formation. Number of unique patients does not match the number of therapy-response entries as one patient may have been treated with EGFR TKIs in different lines of treatment and thus have multiple associated therapy-response entries.

Close modal

After filtering out patients treated with only with other therapies (e.g., chemotherapy regimens), 1,149 patients with a total of 1,207 unique therapy–response entries still remained. Forty-two patients (with 47 therapy–response entries) were excluded because the exact EGFR mutation was unreported, and 79 patients (with 83 therapy–response entries) were excluded because the responses to EGFR TKI therapies were not reported. An additional 7 patients (with 7 therapy–response entries) were deemed duplicate data and deleted. About 1,070 separate therapy–response entries to gefitinib or erlotinib therapy from 1,021 patients and 116 different studies were used in our analysis. The number of unique patients does not match the number of therapy–response entries as one patient may have been treated with EGFR TKIs in different lines of treatment and thus have multiple associated therapy–response entries.

The median age of patients was 62 years (Table 1). The majority of the DIRECT patients with reported clinical features was female (42.8% vs. 28.0% male), Asian (66.6% vs. 18.5% Caucasian), never smokers (42.4% vs. 26.5% current/former smokers) with adenocarcinoma histology (65.8% vs. 24.8% other histology). About 63.6% had stage IIIB/IV or IV lung cancer (32.5% had unrecorded cancer staging). A significant portion of the patient pool had unknown demographic features (Table 1).

Table 1.

Demographic data for DIRECT patients treated with EGFR TKIs

Categoryn (% of total cohort)
Median age (range), y 62 (27–92) 
Gender 
 Female 437 (42.8) 
 Male 286 (28) 
 Unknown 298 (29.2) 
Ethnicity 
 Caucasian 189 (18.5) 
 Asian 680 (66.6) 
 African-American 2 (0.2) 
 Unknown 150 (14.7) 
Histologya 
 Adenocarcinoma 672 (65.8) 
 Adenosquamous 5 (0.5) 
 Bronchioloalveolar 23 (2.3) 
 Large cell 15 (1.5) 
 Squamous 34 (3.3) 
 Spindle 1 (0.1) 
 Unknown 96 (9.4) 
Smoking status 
 Current 55 (5.4) 
 Former smokers 215 (21.1) 
 Never smokers 433 (42.4) 
 NSCLC nosb 175 (17.5) 
 Unknown 318 (31.2) 
Stage 
 I 4 (0.4) 
 IB 1 (0.1) 
 IIA 3 (0.3) 
 IIB 5 (0.5) 
 III 2 (0.2) 
 IIIA 25 (0.25) 
 IIIB 39 (3.8) 
 IIIB/IV 401 (39.3) 
 IV 209 (20.5) 
 Unknown 332 (32.5) 
ECOG performance status 
 0 11 (1.1) 
 1 35 (3.4) 
 2 21 (2.1) 
 3 8 (0.8) 
 4 5 (0.5) 
 Unknown 941 (92.1) 
Number of prior therapies before starting EGFR TKI therapy 
 0 179 (17.5) 
 1 49 (4.8) 
 2 86 (8.4) 
 3 18 (1.8) 
 4 9 (0.9) 
 6 2 (0.2) 
 Unknown 678 (66.4) 
Total number of mutation combinations by response category 
De novo disease control 149 (72.0) 
De novo disease progression 42 (20.3) 
 Acquired resistance 16 (7.7) 
 TOTAL unique mutation combinations 207 (N/A) 
Categoryn (% of total cohort)
Median age (range), y 62 (27–92) 
Gender 
 Female 437 (42.8) 
 Male 286 (28) 
 Unknown 298 (29.2) 
Ethnicity 
 Caucasian 189 (18.5) 
 Asian 680 (66.6) 
 African-American 2 (0.2) 
 Unknown 150 (14.7) 
Histologya 
 Adenocarcinoma 672 (65.8) 
 Adenosquamous 5 (0.5) 
 Bronchioloalveolar 23 (2.3) 
 Large cell 15 (1.5) 
 Squamous 34 (3.3) 
 Spindle 1 (0.1) 
 Unknown 96 (9.4) 
Smoking status 
 Current 55 (5.4) 
 Former smokers 215 (21.1) 
 Never smokers 433 (42.4) 
 NSCLC nosb 175 (17.5) 
 Unknown 318 (31.2) 
Stage 
 I 4 (0.4) 
 IB 1 (0.1) 
 IIA 3 (0.3) 
 IIB 5 (0.5) 
 III 2 (0.2) 
 IIIA 25 (0.25) 
 IIIB 39 (3.8) 
 IIIB/IV 401 (39.3) 
 IV 209 (20.5) 
 Unknown 332 (32.5) 
ECOG performance status 
 0 11 (1.1) 
 1 35 (3.4) 
 2 21 (2.1) 
 3 8 (0.8) 
 4 5 (0.5) 
 Unknown 941 (92.1) 
Number of prior therapies before starting EGFR TKI therapy 
 0 179 (17.5) 
 1 49 (4.8) 
 2 86 (8.4) 
 3 18 (1.8) 
 4 9 (0.9) 
 6 2 (0.2) 
 Unknown 678 (66.4) 
Total number of mutation combinations by response category 
De novo disease control 149 (72.0) 
De novo disease progression 42 (20.3) 
 Acquired resistance 16 (7.7) 
 TOTAL unique mutation combinations 207 (N/A) 

Abbreviation: N/A, not applicable.

a2004 World Health Organization (WHO) Classification used.

bNSCLC not otherwise specified.

The mutations from each of the 1,021 patients were then classified into mutations associated with disease control, disease progression, or acquired resistance (Supplementary Table S1, Tables 2 and 3). Several mutations had patients that fit more than one category. For example, 2 of 3 patients with L861R mutation had disease control with EGFR TKI treatment; L861R was classified as a mutation with EGFR TKI disease control. Five mutation combinations that had a perfectly even split between 2 groups were classified as mutations associated with disease control because half of the patient sample achieved a SD, PR, or CR response, indicating the mutation seems at least somewhat sensitive to EGFR TKI therapy.

Table 2.

EGFR mutations associated with disease progression

NNumber of unique
MutationTherapy–response instancesPD responsesPatientsaStudiesbReferences
A763V 11 
A859T 12, 13 
Del19 A767-V769 11 
Del19 unspecified + ins20 unspecified 14 
Del19 unspecified + V769M 14 
E709G + G719C 14 
E711K 15 
G719A + S768I 16, 17 
G721D 18 
G729R 15 
I744M 19 
I759T 20 
Ins20 769-770insGVV 18 
Ins20 A767-V769dupASV 17 
Ins20 D770-N771insD 17 
Ins20 P772-H773insYNP + H773Y 17 
Ins20 S768-D770dupSVD 17 
Ins20 SVD768-770 21 
Ins20 unspecified 22 
K806E 17 
K860E 15 
L703F 15 
L747P 23 
L777G 18 
L838P + E868G 19 
L858R + A871E 14 
L858R + G719S 24 
L858R + L747S (de novo14 
L858R + L861P 15 
L858R + T790M (de novo17, 18, 21, 25, 26 
L858R + R776G 17 
L861Q + G719S 11 
N826Y 14 
N842S 14 
S768I + V769L 27 
S784F 14 
T847A + G863S 15 
T847I 14 
T790M (de novo25, 28 
V774M 14 
V802I 25 
V852I 29 
NNumber of unique
MutationTherapy–response instancesPD responsesPatientsaStudiesbReferences
A763V 11 
A859T 12, 13 
Del19 A767-V769 11 
Del19 unspecified + ins20 unspecified 14 
Del19 unspecified + V769M 14 
E709G + G719C 14 
E711K 15 
G719A + S768I 16, 17 
G721D 18 
G729R 15 
I744M 19 
I759T 20 
Ins20 769-770insGVV 18 
Ins20 A767-V769dupASV 17 
Ins20 D770-N771insD 17 
Ins20 P772-H773insYNP + H773Y 17 
Ins20 S768-D770dupSVD 17 
Ins20 SVD768-770 21 
Ins20 unspecified 22 
K806E 17 
K860E 15 
L703F 15 
L747P 23 
L777G 18 
L838P + E868G 19 
L858R + A871E 14 
L858R + G719S 24 
L858R + L747S (de novo14 
L858R + L861P 15 
L858R + T790M (de novo17, 18, 21, 25, 26 
L858R + R776G 17 
L861Q + G719S 11 
N826Y 14 
N842S 14 
S768I + V769L 27 
S784F 14 
T847A + G863S 15 
T847I 14 
T790M (de novo25, 28 
V774M 14 
V802I 25 
V852I 29 

NOTE: Disease progression noted as a PD response as defined by the RECIST criteria (7), following treatment with EGFR TKI (gefitinib or erlotinib therapy).

aNumber of unique patients does not match the number of therapy–response entries as one patient may have been treated with EGFR TKIs in different lines of treatment and thus may have multiple associated therapy–response entries.

bNumber of unique studies refers to the number of different studies that encompass all of the patients with a particular mutation in DIRECT.

Table 3.

EGFR mutations found in patients with acquired resistance to EGFR TKIs

Number of unique
Initial mutationAcquired resistance mutationaPatientsbStudiescReferences
Del19 E746-A750 T790M 14 30–34 
Del19 E746-T751insA T790M 32 
Del19 E746-T751insV T790M 35 
Del19 L747-A750insP T790M 31 
Del19 L747-E749;A750P T790M 34, 35 
Del19 L747-P753insQ T790M 36 
Del19 L747-P753insS T790M 32, 34 
Del19 L747-S752 T790M 31, 32 
Del19 L747-T751 T790M 35 
Del19 L747-T751;K754E T790M 34, 35 
Del19 unspecified T790M 37 
L858R D761Y 32 
L858R L747S 38 
L858R T854A 35 
L858R T790M 21 31, 32, 34, 35, 37, 39 
Wild-type T790M 31 
Number of unique
Initial mutationAcquired resistance mutationaPatientsbStudiescReferences
Del19 E746-A750 T790M 14 30–34 
Del19 E746-T751insA T790M 32 
Del19 E746-T751insV T790M 35 
Del19 L747-A750insP T790M 31 
Del19 L747-E749;A750P T790M 34, 35 
Del19 L747-P753insQ T790M 36 
Del19 L747-P753insS T790M 32, 34 
Del19 L747-S752 T790M 31, 32 
Del19 L747-T751 T790M 35 
Del19 L747-T751;K754E T790M 34, 35 
Del19 unspecified T790M 37 
L858R D761Y 32 
L858R L747S 38 
L858R T854A 35 
L858R T790M 21 31, 32, 34, 35, 37, 39 
Wild-type T790M 31 

aThese resistance mutations were acquired after treatment with EGFR TKI.

bNumber of unique patients does not match the number of therapy-response entries as one patient may have been treated with EGFR TKIs in different lines of treatment and thus may have multiple associated therapy–response entries.

cNumber of unique studies refers to the number of different studies that encompass all of the patients with a particular mutation in DIRECT.

We identified more than 188 unique EGFR mutations, occurring in 207 different combinations. For example, some patients had either L858R alone (n = 225 from 62 studies), T790M alone (n = 6 from 3 studies), or both L858R and T790M in combination (n = 27 from 11 studies), resulting in an analysis of just 2 unique mutations (L858R and T790M) but 3 different noted mutation combinations. Of the 207 EGFR mutation combinations, 149 (72.0%) met criteria for disease control, 42 (20.3%) were associated with disease progression, and 16 (7.7%) were linked to acquired resistance (Table 1).

L858R

About 210 of the 235 therapy–response instances (89%) from 225 patients with a single L858R mutation achieved disease control with either gefitinib or erlotinib. L858R mutations were also frequently found with a second de novo mutation (n = 31). Including all combinations of L858R mutation with concurrent de novo mutations, 237 of the 266 therapy–response instances (89%) from 256 unique patients achieved disease control. L858R mutations occurring in combination with either A871E, L747S, de novo G719S, de novo T790M, L861P, or R776G were associated with disease progression.

Del19

A total of 191 of the 215 therapy–response instances (89%) from 204 separate patients with a single del19 E746-A750 mutation achieved disease control. Including patients with del19 E746-A750 mutation and a concurrent de novo mutation, 201 of the >225 therapy–response instances (89%) from 213 separate patients achieved disease control. Factoring in all patients from the 53 mutation combinations involving a del19 mutation, including those with an unspecified amino acid deletion in exon 19 (del19 unspecified) and those with concurrent de novo mutation(s), 424 of the 466 therapy–response instances (91%) from 439 patients were classified as having disease control. Del19 unspecified in combination with either ins20 unspecified or V769M were noted to be associated with disease progression. Only one patient with each combination was in DIRECT, making any determination of their therapeutic significance difficult.

G719X

Fourteen of 18 therapy–responses and 18 patients (78%) with a single G719X mutation—G719A, G719C, G719D, and G719S—achieved disease control. Expanding the pool to all patients with a G719X and concurrent de novo mutations, 20 of 24 therapy–response instances and 24 patients (83%) fit disease control criteria.

L861Q

Twelve of 14 therapy–response instances (86%) from 13 patients with a single L861Q mutation achieved disease control. Including all patients with L861Q and concurrent de novo mutations, 17 of 19 therapy–response instances (89%) from 18 patients had disease control. However, only 7 of these 17 (41%) disease control therapy–responses were partial responses; the other 10 met stable disease criteria. In contrast, 177 of the 237 disease control therapy–responses (75%) for L858R, 332 of the 424 disease control therapy–responses (78%) for del19, and 16 of the 20 disease control therapy–responses (80%) for G719X were partial responses, suggesting that L858R, exon 19 deletion, and G719X mutations respond more favorably to gefitinib or erlotinib than L861Q.

T790M

Both therapy–responses from 2 patients in DIRECT with a de novo T790M mutation had progressive disease. Five of 6 therapy–response instances (83%) and 6 patients with de novo L858R + T790M had disease progression.

Acquired resistance

Fifty-seven of 60 patients (95%) with acquired resistance to EGFR TKI therapy had T790M. The other 3 patients developed a D761Y, L747S, or T854A resistance mutation, in line with known acquired resistance mutations to EGFR TKI therapy (9). Other resistance mechanisms (e.g., MET amplification, PIK3CA mutation, etc.) were outside the focus of this current report but will be included in future studies. Del19 E746-A750 was found with T790M acquired resistance mutation in 14 patients from 5 studies, whereas the L858R mutation was found with acquired resistance mutation T790M in 21 patients from 6 studies.

The DIRECT database represents the most comprehensive catalog of EGFR mutations and response to gefitinib or erlotinib therapy to date. The majority of the mutations in the DIRECT database are rare mutations with limited clinical data that may not be readily accessible to the busy clinician. As expected for more common mutations, L858R and exon 19 deletions were associated with disease control, whereas the T790M mutation was associated with disease progression on an EGFR TKI. We further confirm that G719X and L861Q mutations confer sensitivity to EGFR TKIs, although L861Q mutations were associated with a lower response rate. As an increasing number of patients have their tumors genotyped, there will be (and is already) a dire need for an interactive easily accessible educational tool that provides up-to-date information to physicians on the clinical relevance of mutations in cancers. The database is currently available for query at http://www.mycancergenome.org/about/direct, where users can complete a brief questionnaire on their mutation of interest, and a detailed report on the specific mutation is emailed back (10). Since its inception, we have averaged 5 to 6 queries per month from around the world, the majority (>80%) for rare mutations. The number of queries continues to increase monthly. With time and with proper data sharing agreements for data from large clinical trials, we intend to change this into an interactive platform that will provide immediate data at the point of care to aid in therapeutic decisions.

Currently, DIRECT catalogs only patients with NSCLCs and published individual patient data on EGFR mutations. We are working with collaborators to expand the database to include individual patient data from large randomized phase II and III trials on patients with EGFR mutations and response to EGFR TKIs and chemotherapy. We also plan to incorporate other known mutations in solid tumors (e.g. melanoma, gastrointestinal stromal tumors, etc.) for which mutation–response data exist. To truly realize the promise of personalized medicine, all publications involving genotype-driven trials should at a minimum publish Supplementary Tables listing individual patients by tumor mutation status and treatment outcome.

This study has several limitations. First, not all demographic data including gender, smoking status, tumor histology, and stage were available for inclusion. These data do not affect the reported mutation–response data but do not allow classification of specific mutations within demographic categories. Second, of our 207 mutation combinations, 67% or 138 mutation combinations (97 disease control, 33 disease progressing, and 8 acquired resistance) are rare mutations for which there are only data from one patient, making any determination of their therapeutic significance difficult. Finally, we did not record the specific method of EGFR mutation testing from each study and thus this article does not take into account the variance in EGFR mutation detection between differing studies. Nevertheless, these data may serve as an initial guide in selecting therapy. As more and more patients have their tumors genotyped, the number of patients with such “rare mutations” will grow and their clinical significance will emerge.

W. Pao has received consulting fees from MolecularMD, AstraZeneca, Bristol-Myers Squibb, Symphony Evolution, and Clovis Oncology and research funding for other projects from Enzon, Xcovery, AstraZeneca, and Symphogen. W. Pao is part of a patent regarding EGFRT790M mutation testing that was licensed by Memorial Sloan-Kettering Cancer Center to Molecular MD. L. Horn is a consultant/advisory board member of Astellas, BI, Genentech. No potential conflicts of interest were disclosed by the other authors.

Conception and design: W. Pao, L. Horn

Development of methodology: J. Andrews, W. Pao, L. Horn

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Yeh, J. Andrews, W. Pao, L. Horn

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Chen, W. Pao, L. Horn

Writing, review, and/or revision of the manuscript: P. Yeh, J. Andrews, W. Pao, L. Horn

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Yeh, R. Naser, W. Pao, L. Horn

Study supervision: W. Pao, L. Horn

The authors thank Christine Lovly, Christine Micheel, and Mia Levy for their help with the DIRECT database.

This work was supported by NIH NCI grants R01-CA121210 and P01-CA129243. Funding was also received from the Kleberg Foundation, the Martell Foundation, and an Anonymous Foundation. W. Pao received additional support from the VICC Cancer Center Core Grant (P30-CA68485).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Kris
MG
,
Johnson
BE
,
Kwiatkowski
DJ
,
Iafrate
AJ
,
Wistuba
II
,
Aronson
SL
, et al
Identification of driver mutations in tumor specimens from 1,000 patients with lung adenocarcinoma: The NCI's Lung Cancer Mutation Consortium (LCMC)
.
J Clin Oncol
2011
;
29
(
18 Suppl
):
CRA7506
.
2.
Ladanyi
M
,
Pao
W
. 
Lung adenocarcinoma: guiding EGFR-targeted therapy and beyond
.
Mod Pathol
2008
;
21
Suppl 2
:
S16
22
.
3.
Riely
GJ
,
Politi
KA
,
Miller
VA
,
Pao
W
. 
Update on epidermal growth factor receptor mutations in non-small cell lung cancer
.
Clin Cancer Res
2006
;
12
:
7232
41
.
4.
Dahabreh
IJ
,
Linardou
H
,
Siannis
F
,
Kosmidis
P
,
Bafaloukos
D
,
Murray
S
. 
Somatic EGFR mutation and gene copy gain as predictive biomarkers for response to tyrosine kinase inhibitors in non-small cell lung cancer
.
Clin Cancer Res
2010
;
16
:
291
303
.
5.
REDCAP
.
[cited 2012 May 1] Available from
: http://project-redcap.org/.
6.
Harris
PA
,
Taylor
R
,
Thielke
R
,
Payne
J
,
Gonzalez
N
,
Conde
JG
. 
Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support
.
J Biomed Inform
2009
;
42
:
377
81
.
7.
Eisenhauer
EA
,
Therasse
P
,
Bogaerts
J
,
Schwartz
LH
,
Sargent
D
,
Ford
R
, et al
New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1)
.
Eur J Cancer
2009
;
45
:
228
47
.
8.
Team
RDC
. 
R: a language and environment for statistical computing
.
Vienna, Austria
:
R Foundation for statistical computing
; 
2008
.
ISBN 3-90051-07-0. Available from
: http://www.r-project.org/.
9.
Nguyen
KS
,
Kobayashi
S
,
Costa
DB
. 
Acquired resistance to epidermal growth factor receptor tyrosine kinase inhibitors in non-small-cell lung cancers dependent on the epidermal growth factor receptor pathway
.
Clin Lung Cancer
2009
;
10
:
281
9
.
10.
Yatabe
Y
,
Pao
W
,
Jett
JR
. 
Encouragement to submit data of clinical response to EGFR-TKIs in patients with uncommon EGFR mutations
.
J Thorac Oncol
2012
;
7
:
775
6
.
11.
Chou
TY
,
Chiu
CH
,
Li
LH
,
Hsiao
CY
,
Tzen
CY
,
Chang
KT
, et al
Mutation in the tyrosine kinase domain of epidermal growth factor receptor is a predictive and prognostic factor for gefitinib treatment in patients with non-small cell lung cancer
.
Clin Cancer Res
2005
;
11
:
3750
7
.
12.
Cappuzzo
F
,
Bemis
L
,
Varella-Garcia
M
. 
HER2 mutation and response to trastuzumab therapy in non-small-cell lung cancer
.
N Engl J Med
2006
;
354
:
2619
21
.
13.
Taron
M
,
Ichinose
Y
,
Rosell
R
,
Mok
T
,
Massuti
B
,
Zamora
L
, et al
Activating mutations in the tyrosine kinase domain of the epidermal growth factor receptor are associated with improved survival in gefitinib-treated chemorefractory lung adenocarcinomas
.
Clin Cancer Res
2005
;
11
:
5878
85
.
14.
Wu
JY
,
Yu
CJ
,
Chang
YC
,
Yang
CH
,
Shih
JY
,
Yang
PC
. 
Effectiveness of tyrosine kinase inhibitors on “uncommon” epidermal growth factor receptor mutations of unknown clinical significance in non-small cell lung cancer
.
Clin Cancer Res
2011
;
17
:
3812
21
.
15.
Pallis
AG
,
Voutsina
A
,
Kalikaki
A
,
Souglakos
J
,
Briasoulis
E
,
Murray
S
, et al
‘Classical' but not ‘other’ mutations of EGFR kinase domain are associated with clinical outcome in gefitinib-treated patients with non-small cell lung cancer
.
Br J Cancer
2007
;
97
:
1560
6
.
16.
Wu
SG
,
Chang
YL
,
Hsu
YC
,
Wu
JY
,
Yang
CH
,
Yu
CJ
, et al
Good response to gefitinib in lung adenocarcinoma of complex epidermal growth factor receptor (EGFR) mutations with the classical mutation pattern
.
Oncologist
2008
;
13
:
1276
84
.
17.
Wu
JY
,
Wu
SG
,
Yang
CH
,
Gow
CH
,
Chang
YL
,
Yu
CJ
, et al
Lung cancer with epidermal growth factor receptor exon 20 mutations is associated with poor gefitinib treatment response
.
Clin Cancer Res
2008
;
14
:
4877
82
.
18.
Ichihara
S
,
Toyooka
S
,
Fujiwara
Y
,
Hotta
K
,
Shigematsu
H
,
Tokumo
M
, et al
The impact of epidermal growth factor receptor gene status on gefitinib-treated Japanese patients with non-small-cell lung cancer
.
Int J Cancer
2007
;
120
:
1239
47
.
19.
Hsieh
MH
,
Fang
YF
,
Chang
WC
,
Kuo
HP
,
Lin
SY
,
Liu
HP
, et al
Complex mutation patterns of epidermal growth factor receptor gene associated with variable responses to gefitinib treatment in patients with non-small cell lung cancer
.
Lung Cancer
2006
;
53
:
311
22
.
20.
Takahashi
T
,
Yamamoto
N
,
Nukiwa
T
,
Mori
K
,
Tsuboi
M
,
Horai
T
, et al
Phase II study of erlotinib in Japanese patients with advanced non-small cell lung cancer
.
Anticancer Res
2010
;
30
:
557
63
.
21.
Yang
CH
,
Yu
CJ
,
Shih
JY
,
Chang
YC
,
Hu
FC
,
Tsai
MC
, et al
Specific EGFR mutations predict treatment outcome of stage IIIB/IV patients with chemotherapy-naive non-small-cell lung cancer receiving first-line gefitinib monotherapy
.
J Clin Oncol
2008
;
26
:
2745
53
.
22.
Cappuzzo
F
,
Ligorio
C
,
Janne
PA
,
Toschi
L
,
Rossi
E
,
Trisolini
R
, et al
Prospective study of gefitinib in epidermal growth factor receptor fluorescence in situ hybridization-positive/phospho-Akt-positive or never smoker patients with advanced non-small-cell lung cancer: the ONCOBELL trial
.
J Clin Oncol
2007
;
25
:
2248
55
.
23.
Kimura
H
,
Suminoe
M
,
Kasahara
K
,
Sone
T
,
Araya
T
,
Tamori
S
, et al
Evaluation of epidermal growth factor receptor mutation status in serum DNA as a predictor of response to gefitinib (IRESSA)
.
Br J Cancer
2007
;
97
:
778
84
.
24.
Hata
A
,
Yoshioka
H
,
Fujita
S
,
Kunimasa
K
,
Kaji
R
,
Imai
Y
, et al
Complex mutations in the epidermal growth factor receptor gene in non-small cell lung cancer
.
J Thorac Oncol
2010
;
5
:
1524
8
.
25.
Jackman
DM
,
Miller
VA
,
Cioffredi
LA
,
Yeap
BY
,
Janne
PA
,
Riely
GJ
, et al
Impact of epidermal growth factor receptor and KRAS mutations on clinical outcomes in previously untreated non-small cell lung cancer patients: results of an online tumor registry of clinical trials
.
Clin Cancer Res
2009
;
15
:
5267
73
.
26.
Tokumo
M
,
Toyooka
S
,
Ichihara
S
,
Ohashi
K
,
Tsukuda
K
,
Ichimura
K
, et al
Double mutation and gene copy number of EGFR in gefitinib refractory non-small-cell lung cancer
.
Lung Cancer
2006
;
53
:
117
21
.
27.
Asahina
H
,
Yamazaki
K
,
Kinoshita
I
,
Yokouchi
H
,
Dosaka-Akita
H
,
Nishimura
M
. 
Non-responsiveness to gefitinib in a patient with lung adenocarcinoma having rare EGFR mutations S768I and V769L
.
Lung Cancer
2006
;
54
:
419
22
.
28.
Donovan
MJ
,
Kotsianti
A
,
Bayer-Zubek
V
,
Verbel
D
,
Teverovskiy
M
,
Cordon-Cardo
C
, et al
A systems pathology model for predicting overall survival in patients with refractory, advanced non-small-cell lung cancer treated with gefitinib
.
Eur J Cancer
2009
;
45
:
1518
26
.
29.
Cappuzzo
F
,
Hirsch
FR
,
Rossi
E
,
Bartolini
S
,
Ceresoli
GL
,
Bemis
L
, et al
Epidermal growth factor receptor gene and protein and gefitinib sensitivity in non-small-cell lung cancer
.
J Natl Cancer Inst
2005
;
97
:
643
55
.
30.
Costa
DB
,
Nguyen
KS
,
Cho
BC
,
Sequist
LV
,
Jackman
DM
,
Riely
GJ
, et al
Effects of erlotinib in EGFR mutated non-small cell lung cancers with resistance to gefitinib
.
Clin Cancer Res
2008
;
14
:
7060
7
.
31.
Chen
HJ
,
Mok
TS
,
Chen
ZH
,
Guo
AL
,
Zhang
XC
,
Su
J
, et al
Clinicopathologic and molecular features of epidermal growth factor receptor T790M mutation and c-MET amplification in tyrosine kinase inhibitor-resistant Chinese non-small cell lung cancer
.
Pathol Oncol Res
2009
;
15
:
651
8
.
32.
Balak
MN
,
Gong
Y
,
Riely
GJ
,
Somwar
R
,
Li
AR
,
Zakowski
MF
, et al
Novel D761Y and common secondary T790M mutations in epidermal growth factor receptor-mutant lung adenocarcinomas with acquired resistance to kinase inhibitors
.
Clin Cancer Res
2006
;
12
:
6494
501
.
33.
Bean
J
,
Brennan
C
,
Shih
JY
,
Riely
G
,
Viale
A
,
Wang
L
, et al
MET amplification occurs with or without T790M mutations in EGFR mutant lung tumors with acquired resistance to gefitinib or erlotinib
.
Proc Natl Acad Sci U S A
2007
;
104
:
20932
7
.
34.
Engelman
JA
,
Zejnullahu
K
,
Mitsudomi
T
,
Song
Y
,
Hyland
C
,
Park
JO
, et al
MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling
.
Science
2007
;
316
:
1039
43
.
35.
Bean
J
,
Riely
GJ
,
Balak
M
,
Marks
JL
,
Ladanyi
M
,
Miller
VA
, et al
Acquired resistance to epidermal growth factor receptor kinase inhibitors associated with a novel T854A mutation in a patient with EGFR-mutant lung adenocarcinoma
.
Clin Cancer Res
2008
;
14
:
7519
25
.
36.
Ruppert
AM
,
Beau-Faller
M
,
Neuville
A
,
Guerin
E
,
Voegeli
AC
,
Mennecier
B
, et al
EGFR-TKI and lung adenocarcinoma with CNS relapse: interest of molecular follow-up
.
Eur Respir J
2009
;
33
:
436
40
.
37.
Onitsuka
T
,
Uramoto
H
,
Nose
N
,
Takenoyama
M
,
Hanagiri
T
,
Sugio
K
, et al
Acquired resistance to gefitinib: the contribution of mechanisms other than the T790M, MET, and HGF status
.
Lung Cancer
2010
;
68
:
198
203
.
38.
Costa
DB
,
Schumer
ST
,
Tenen
DG
,
Kobayashi
S
. 
Differential responses to erlotinib in epidermal growth factor receptor (EGFR)-mutated lung cancers with acquired resistance to gefitinib carrying the L747S or T790M secondary mutations
.
J Clin Oncol
2008
;
26
:
1182
4
.
39.
Inukai
M
,
Toyooka
S
,
Ito
S
,
Asano
H
,
Ichihara
S
,
Soh
J
, et al
Presence of epidermal growth factor receptor gene T790M mutation as a minor clone in non-small cell lung cancer
.
Cancer Res
2006
;
66
:
7854
8
.