CT Features of Epidermal Growth Factor Receptor–Mutated Adenocarcinoma of the Lung: Comparison with Nonmutated Adenocarcinoma

Accepted Manuscript CT features of epidermal growth factor receptor mutated adenocarcinoma of the lung: Comparison with non mutated adenocarcinoma Mizue Hasegawa, MD, PhD, Fumikazu Sakai, MD, PhD, Rinako Ishikawa, MD, PhD, Fumiko Kimura, MD, PhD, Hironori Ishida, MD, PhD, Kunihiko Kobayashi, MD, PhD PII:

S1556-0864(16)00429-9

DOI:

10.1016/j.jtho.2016.02.010

Reference:

JTHO 119

To appear in:

Journal of Thoracic Oncology

Received Date: 1 December 2015 Revised Date:

17 February 2016

Accepted Date: 17 February 2016

Please cite this article as: Hasegawa M, Sakai F, Ishikawa R, Kimura F, Ishida H, Kobayashi K, CT features of epidermal growth factor receptor mutated adenocarcinoma of the lung: Comparison with non mutated adenocarcinoma, Journal of Thoracic Oncology (2016), doi: 10.1016/j.jtho.2016.02.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

(a) Title:

Comparison with non mutated adenocarcinoma.

SC

(b) Author`s names and affiliations:

Rinako Ishikawa, MD, PhD 2) Fumiko Kimura, MD, PhD 1)

TE D

Hironori Ishida, MD, PhD 3)

M AN U

Mizue Hasegawa, MD, PhD 1)4) Fumikazu Sakai, MD, PhD 1)

RI PT

CT features of epidermal growth factor receptor mutated adenocarcinoma of the lung:

EP

Kunihiko Kobayashi, MD, PhD 2)

AC C

1) Saitama International Medical Center, Saitama Medical University, department of diagnostic radiology

2) Saitama International Medical Center, Saitama Medical University, department of respiratory medicine 3) Saitama International Medical Center, Saitama Medical University, department of thoracic surgery

ACCEPTED MANUSCRIPT

4) Tokyo Women’s Medical University, Yachiyo Medical Center, department of

RI PT

respiratory medicine

(c) Corresponding author:

SC

Name: Mizue Hasegawa

Address: Department of Diagnostic Radiology, Saitama International Medical Center,

Post code: 350-1298

Fax; +81-042-984-4220

TE D

Phone; +81-042-984-4111

Email: hasemizue@yahoo/co.jp

AC C

EP

Disclosure of funding: none

(d) Address for reprints: Same as (c)

(e) All sources of support: None

M AN U

Saitama Medical University, 1397-1 Hidaka, Saitama, Japan

ACCEPTED MANUSCRIPT

Disclosures of funding from NIH; Wellcome Trust; HHMI; and others:

AC C

EP

TE D

M AN U

SC

RI PT

None

ACCEPTED MANUSCRIPT

Introduction: The purpose of this study was to analyze the high resolution computed tomography (HRCT) features of lung carcinoma based on epidermal growth factor

RI PT

receptor (EGFR) mutation status.

Methods: Two hundred and sixty-three consecutive cases of lung adenocarcinoma

SC

diagnosed at our institution between January 2010 and December 2011 were enrolled in the study. All patients underwent HRCT and analysis of EGFR mutation status. The

M AN U

HRCT findings were retrospectively analyzed for tumor size, multiple bilateral lung metastases, convergence of surrounding structures, surrounding ground-glass opacity, prominent peribronchovascular extension, air bronchogram, notch, pleural indentation,

TE D

spiculation, cavity, and pleural effusions.

Results: EGFR mutations were demonstrated in 103 patients (39.2%), the remaining

EP

160 patients (60.8%) had non-mutated type. Compared with the non-mutated group, the

AC C

mutated group had significantly higher frequencies of multiple bilateral lung metastases (p = 0.0152), convergence of surrounding structures (p < 0.0001), ground glass opacity (p = 0.0011), and notch (p = 0.0428); but significantly lower frequencies of cavitation (p = 0.0004) and pleural effusions (p = 0.0064). The frequencies of the other CT findings were similar between the two groups. The devised prediction HRCT score for EGFR mutation was 78.4% sensitive and 70.4% specific.

ACCEPTED MANUSCRIPT

Conclusion: EGFR-mutated adenocarcinoma showed significantly higher frequencies of multiple bilateral lung metastases, convergence of surrounding structures, surrounding

RI PT

ground glass opacity, and notch at HRCT compared with EGFR non-mutated type. Conversely, EGFR-mutated adenocarcinoma showed cavity and pleural effusions less

Key words: epidermal

growth

factor

M AN U

SC

frequently than the non-mutated type.

receptor

AC C

EP

TE D

tomography

mutation;

lung

adenocarcinoma;

computed

ACCEPTED MANUSCRIPT

Introduction Lung cancer is one of the leading causes of death worldwide. Adenocarcinoma is one

RI PT

of the most common histologic subtypes of lung cancer. Since the discovery of somatic epidermal growth factor receptor (EGFR) mutations, targeted therapies have

SC

dramatically improved survival rates for patients with lung cancer, especially

M AN U

adenocarcinoma (1-4). Adenocarcinoma, non-smoking status, female sex, and Asian ethnicity were described as predictors of EGFR mutation (5-7) that enable efficient plans for earlier diagnosis and treatment.

Several studies have attempted to describe the relationship between computed

TE D

tomography (CT) findings and EGFR mutation status. Lee et al have reported that the volume of ground glass opacity (GGO) attenuation was significantly higher in

EP

adenocarcinoma with EGFR mutation of exon 21 missense (8). Togashi et al reported

AC C

that EGFR-mutated adenocarcinoma tends to develop diffuse and random pulmonary metastases compared with the wild type (9). However, the relationship between EGFR-mutated adenocarcinoma and CT scan features has not been fully elucidated. In the field of lung cancer, there has been a recent increase in the identification of new gene mutations for targeted therapy, including anaplastic lymphoma kinase (ALK), ROS-1, and RET (10-12). Yamamoto et al have reported that non-small cell lung cancer

1

ACCEPTED MANUSCRIPT

(NSCLC) with ALK rearrangement was associated with CT scan characteristics of central tumor location, absence of pleural tail, and large pleural effusion (13). With

RI PT

regard to cost effectiveness and early planning of treatment, prediction of gene mutation from CT findings may be valuable. In this study, we aimed to analyze the differences in

Materials and Methods

M AN U

SC

CT findings between EGFR-mutated and non-mutated lung adenocarcinoma.

This retrospective study was approved by our institutional review board, who waived

of patient anonymity.

EP

Patient selection

TE D

the need for informed consent due to the non-invasive nature of the study and because

AC C

Consecutive patients (median age, 66 years; range, 38–90 years; N = 263, 156 men, 107 women) who had pathologically-confirmed diagnoses of primary lung adenocarcinoma at Saitama International Medical Center between January 2010 and December 2011 were enrolled in this study. All patients were Japanese and had no prior diagnosis or treatment for lung cancer. Patients with active malignancies in other organs, including apparent recurrent or residual tumor; patients with recurrent lung cancer;

2

ACCEPTED MANUSCRIPT

patients without pathologic confirmation of adenocarcinoma, including adenosquamous carcinoma; and patients with inadequate samples for sequencing were excluded from

RI PT

this study. Non-smoking status was defined as lifetime exposure to less than 100

SC

cigarettes. All patients underwent chest CT and analysis of EGFR mutation status.

M AN U

Analysis of EGFR status

All patients were analyzed for EGFR mutation status using histology or cytology specimens. The mutation status of EGFR exons 18, 19, 20, and 21 was examined with

PCR) clamp (14).

EP

CT imaging protocol

TE D

the peptide nucleic acid-locked nucleic acid polymerase chain reaction (PNA-LNA

AC C

All patients underwent chest CT scan, either at our institution or at a previous institution, within 2 months before histology or cytology examination. At our institution, CT was performed using scanners with 64-detector rows (Light Speed VCT; GE Medical Systems, Milwaukee, USA) or 16-detector rows (Bright speed; GE Medical Systems), with the following parameters: detector collimation, 1–5 mm; beam pitch, 1–5; rotation time, 0.5–1.0 second; tube voltage, 120 kVp; and tube current, 150–200

3

ACCEPTED MANUSCRIPT

mA. Contiguous images of 5-mm thickness were constructed from the lung apex to base, followed by reconstruction with high-resolution kernel covering of contiguous 1-mm

RI PT

slices of the primary tumor. Contrast study was done on 256 cases at our institution.

Seven cases underwent CT scan at previous institutions using a similar image protocol

SC

with different kinds of commercially available and state-of-the-art scanners. The

M AN U

frequency of positive EGFR mutation status was not statistically different between patients who underwent CT at our institution (38.3%, 98/256) and those who underwent

CT interpretation

TE D

CT at previous institutions (71.4%, 5/7).

The CT images were independently analyzed by one chest radiologist (F.S.) and one

EP

pulmonologist (M.H.), with experience of 38 years and 14 years, respectively. The

AC C

radiologist (F.S.) was blinded to the results of EGFR mutation status; whereas the pulmonologist (M.H.) had access to the data on EGFR mutation status of all enrolled patients. In order to avoid bias, at least 6 months interval was provided before CT interpretation. Interpretation of CT images was done on mediastinal window (M: 20 to 40. W: 300) and lung window (M: -500 to -700 W: 1500 to 2000) using picture archiving and communication system. The images were analyzed for the following

4

ACCEPTED MANUSCRIPT

points: longest and shortest diameters on the largest axial surface of the primary tumor, multiple bilateral lung metastases, convergence of surrounding structures, surrounding

indentation, spiculation, cavity, and pleural effusion (Fig. 1).

RI PT

GGO, prominent peribronchovascular extension, air bronchogram, notch, pleural

SC

Multiple bilateral lung metastases were defined as more than 10 lesions in both lung

M AN U

fields. Convergence of surrounding structures pertained to vessels, bronchi, bronchioles, or pleura around the primary tumor. Surrounding GGO was delineated based on a clearly defined border with the normal lung parenchyma and was classified on a 4-point scale, depending on the percentage of the GGO component on the largest axial surface

TE D

of the tumor (1: 1%–25%; 2: 25%–50%; 3: 50%–75%; 4: >75%). Prominent peribronchovascular extension of the primary lesion was defined as tumor involvement

EP

of the lymphatic channels, which was recognized on CT as longitudinal extensions of

AC C

the tumor; thickening of the bronchovascular bundle; or perilymphatic satellite lesions. Air bronchogram was defined as a branching or tubular air density surrounded by consolidation without prominent bronchial dilation. Cavity was indicated by the presence of round or oval air density in the tumor with relatively thick wall; markedly dilated tubular lucencies with extensive bronchial dilation within the tumor were considered as cavity. Tumors arising on a cyst wall or air densities caused by

5

ACCEPTED MANUSCRIPT

pre-existing structures, such as bullae or blebs, were excluded from this category. When

consensus.

SC

Comparison of EGFR mutation status and CT findings

RI PT

interpretations of two observers were different, a discussion was made to reach a final

M AN U

The CT findings of EGFR-mutated and non-mutated groups were compared. We had to exclude 19 patients from the CT analysis because of passive atelectasis from massive pleural effusion that obscured the primary tumor. Therefore, 263 patients were evaluated

points.

EP

Statistical analyses

TE D

for the presence of pleural effusion and 244 patients were evaluated for the remaining

AC C

Patient and CT characteristics of the study population were expressed as median (range) for non-parametric variables, and as frequency and percentage for categorical variables. Interobserver agreement was assessed by the kappa coefficient. The CT scan variables were compared between the two groups using univariate (Fisher’s exact test or Chi-square test) and multivariate (multiple logistic regression) analyses. Before performing multiple logistic regression analysis, variables were selected by a stepwise

6

ACCEPTED MANUSCRIPT

method. A prediction tool for positive EGFR mutation was devised from the total scores of each

RI PT

CT finding on multiple logistic regression analysis; a cut-off value was determined by receiver operating characteristic (ROC) analysis. Values of p < 0.05 were considered

SC

significant. All statistical analyses were performed using commercially available

M AN U

software programs (JMP 9; SAS Institute Inc., Cary, NC, USA or StatMate 4; Atoms, Tokyo, Japan).

Results

TE D

Clinical characteristics of the study population The clinical characteristics of the patients are shown in Table 1. The 263 patients had a

EP

median age of 66 years (range, 38–90 years); 156 were men, 107 were women. EGFR

AC C

mutation was positive in 103 (39.0%) patients (44 men, 59 women), who had a median age of 68 years (range, 38–86 years). Specifically, the EGFR mutations were on exon 21 in 51 patients (49.5%), exon 19 in 50 patients (48.5%), and exon 18 in 2 patients (1.9%). There was no EGFR exon 20 mutation observed in this study. Non-mutated type adenocarcinoma was seen in 160 (61%) patients (112 men, 48 women) who had a median age of 69 years (range, 38–90 years). The percentage of smokers was 45.0% in

7

ACCEPTED MANUSCRIPT

the EGFR-mutated group and 74.4% in the non-mutated group. The proportions of

compared with those in the non-mutated group.

SC

Interobserver agreement of CT interpretation

RI PT

women and non-smoking status were significantly higher in the EGFR-mutated group

M AN U

Concordance rates between the two observers were almost perfect, with kappa coefficients ranging between 0.89 and 1.0.

1) Pleural effusion

TE D

Relevance of EGFR mutation and CT findings

Pleural effusion was observed in 45 patients, but was negative in 218 patients. Among

EP

45 patients with pleural effusion, 9 (20%) were positive for EGFR mutation, whereas 36

AC C

(80%) had non-mutated adenocarcinoma. Among 218 patients without pleural effusion, 94 (43.1%) were positive for EGFR mutation, whereas 124 (56.9%) had non-mutated adenocarcinoma. Overall, the frequency of pleural effusion was significantly higher in the non-mutated group than the EGFR-mutated group (p = 0.0064). 2) Tumor size The average size of primary lesions was similar between the EGFR-mutated group and

8

ACCEPTED MANUSCRIPT

the non-mutated group (29.2 mm vs. 34.7 mm, p = 0.1953) (Table 1). 3) CT findings of the primary lesions

RI PT

The CT findings of the primary lesions were evaluated in 244 patients without passive atelectasis from massive pleural effusion (Table 2). Compared with the non-mutated

SC

group, the mutated group had significantly higher frequencies of multiple bilateral lung

M AN U

metastases (univariate, p = 0.0258; multivariate, p = 0.0152); convergence of surrounding vessels (univariate, p <0.0001; multivariate, p < 0.0001); GGO (univariate, p = 0.0007; multivariate, p = 0.0011); and notch (multivariate, p = 0.0428), but significantly lower frequencies of cavitation (univariate, p = 0.0007; multivariate, p =

TE D

0.0004). The percentage of GGO component was similar between the EGFR-mutated and the non-mutated groups (p = 0.1097). The frequencies of the other CT findings were

EP

not statistically significant between the two groups. This study combined multiple

AC C

disease stages into a single study, but this did not affect the outcomes of the multivariate model.

Prediction of EGFR mutation from CT findings (Figure 2) Prediction score was calculated by the product sum of CT scan variables and the logarithm of the odds ratio; -0.79 + spiculation × (-0.49) + notch × (0.70) + pleural

9

ACCEPTED MANUSCRIPT

effusion × (-0.86) + multiple bilateral metastases × (2.28) + prominent peribronchovascular extension × (-0.43) + air bronchogram × (-0.74) + convergence of

RI PT

surrounding structures × (2.01) + cavity × (-1.61) + surrounding ground glass opacity × (1.21). At a set cut-off value of ≥0.23 by ROC curve analysis, the sensitivity and

M AN U

SC

specificity for predicting a positive EGFR status were 78.4% and 70.4%, respectively.

Differences in CT features between exon 19 and exon 21 EGFR mutations (Table 3) Patients with exon 19 and exon 21 EGFR mutations had statistically similar frequencies of multiple bilateral lung metastases, convergence of surrounding vessels, surrounding

EP

Discussion

TE D

GGO, notch and cavity, as well as percentage of GGO component, on CT scan.

AC C

This study indicated that CT findings of multiple bilateral lung metastases, convergence of surrounding structures, surrounding GGO, and notch were associated with EGFR-mutated adenocarcinoma; whereas cavity and pleural effusion were associated with non-mutated adenocarcinoma. The CT prediction score that we devised for EGFR mutation was 78.4% sensitive and 70.4% specific. Since EGFR-TKI can be used in all patients with positive EGFR mutation status, this prediction score may be

10

ACCEPTED MANUSCRIPT

useful in helping decide whether invasive examination is suitable in some high risk patients, such as those with old age or severe illness. Also, the prediction score may

RI PT

help decide whether re-biopsy is worth trying in patients with discordant clinical

features and EGFR mutation status. In the future, the utility of this prediction score

SC

would be more evident if the CT findings of other gene mutations become clear.

M AN U

Several studies have reported that EGFR-mutated lung cancer was associated with GGO on CT (15-18), findings that coincided with our result. Lee et al have reported that the frequency of GGO was significantly higher in adenocarcinoma with EGFR exon 21 missense mutation (8). In contrast with our findings, their pathologic evaluation of

TE D

surgically resected specimens showed that the percentage of GGO component on CT scan was significantly higher in lepidic-predominant adenocarcinoma (LPA), which

EP

contained a higher frequency of exon 21 missense mutations compared with exon 19

AC C

mutations. This might be due to differences in ethnicity of the study population and diagnostic procedures that were studied. Although pathologic correlation with GGO was not assessed in our study, several studies pointed out that adenomatous hyperplasia, atypical adenomatous hyperplasia, adenocarcinoma in situ, and LPA had relatively frequent EGFR mutations (8, 19-24). Sun et al reported that lepidic growth pattern was seen in 59% of cases with EGFR

11

ACCEPTED MANUSCRIPT

mutation and in only 33.3% of cases with non-mutated adenocarcinoma (22). This rate was similar to the frequency of surrounding GGO in our patients with EGFR-mutated

RI PT

adenocarcinoma.

For adenocarcinoma, the relationship between EGFR mutation status and CT findings

SC

can be related with the pathologic features. Noguchi et al have previously classified

M AN U

adenocarcinoma with lepidic growth pattern replacing normal alveolar cells with varying degrees of fibrosis into subtypes A, B, and C (25). Type A referred to localized lepidic pattern adenocarcinoma; type B referred to localized lepidic pattern with foci showing collapse of the alveolar structure; and type C referred to lepidic pattern with

TE D

foci of active fibroblastic proliferation associated with architectural destruction. Collapse or destruction of Noguchi subtypes B and C can be related to convergence of

EP

surrounding structures on CT. On the other hand, other subtypes with destructive acinar

AC C

filling adenocarcinoma without lepidic component, which were classified as Noguchi D, E, and F, occasionally showed less convergence of surrounding structures (26). This may explain our results on a high frequency of convergence of surrounding structures in EGFR-mutated adenocarcinoma. Notch sign usually indicates expansive growth of tumor that is often observed in Noguchi D, E, and F. Our study suggested that the presence of notch on CT was weakly

12

ACCEPTED MANUSCRIPT

associated with EGFR mutation. This result seemed to be inconsistent with the fact these Noguchi types do not contain lepidic pattern and that they may be more associated

RI PT

with non-mutated adenocarcinoma. The reason for the association between the presence of notch and EGFR-mutated adenocarcinoma in our study needs to be further explained

SC

in future studies.

M AN U

The presence of cavitation in lung cancer is associated with adverse prognosis; it was reported to be detected by CT in 22% of stage 1 primary lung cancers, most frequently in squamous cell carcinomas (37.9%) and less in adenocarcinoma (8.7%) (27). A cavity was pathologically defined in literature as “air-filled spaces within a zone of pulmonary

TE D

consolidation or within a mass or nodule that is produced by the expulsion of a necrotic part of the lesion via the bronchial tree” (28). According to our study, cavity in the

EP

primary tumor was associated with EGFR non-mutated adenocarcinoma. Zaibo et al

AC C

reported that necrosis was associated more with ALK rearrangements and KRAS mutations than with EGFR mutations (29). The high frequency of cavitation in the non-mutated group in our study supports this previous report. Onn et al have reported that Stage 1 NSCLCs with cavitary lesions were likely to have EGFR overexpression (27). On the other hand, Lee et al have reported that in Stage 1 NSCLC, there was no difference in the frequency of cavitation between patients

13

ACCEPTED MANUSCRIPT

with overexpression of EGFR and those without overexpression of EGFR (30). Although Sun et al have reported that EGFR overexpression was seen more frequently

RI PT

in tumors with EGFR mutation than in tumors without EGFR mutations (22), other

studies described that EGFR overexpression, negative EGFR mutation, and cavitation

SC

were factors associated with poor prognosis (27,30-32). In this regard, we suppose that

M AN U

mutation and overexpression may be independent factors for cavitation. Taken together, cavitation of lung adenocarcinoma may be associated with EGFR overexpression and negative EGFR mutation, but further study is essential to evaluate this hypothesis. Togashi et al have reported that diffuse, random pulmonary metastasis is

TE D

associated with EGFR mutation in lung adenocarcinoma. In a previous report, 2 of 4 patients who had diffuse, random pulmonary metastasis without EGFR mutation were

EP

reported to achieve partial response to gefitinib therapy (33). It was also reported that

AC C

the presence of multiple lung metastases was associated with improved response to gefitinib, regardless of EGFR mutation status (34). In the present study, the frequency of multiple bilateral lung metastases was significantly higher in the EGFR mutated group than the non-mutated group. In our study, pleural effusion was associated more with non-mutated adenocarcinoma than with EGFR-positive adenocarcinoma. However, it is not clear

14

ACCEPTED MANUSCRIPT

whether this association was due to the gene mutation itself or secondary to delayed diagnosis due to rapid progression of adenocarcinoma without EGFR mutation.

RI PT

Recently, Rizzo et al have reported the association of EGFR mutation with CT

findings in European patients (35). Specifically, the CT findings of air bronchogram,

SC

pleural retraction, small size, and absence of fibrosis were all related to EGFR-positive

M AN U

adenocarcinoma. In our study, air bronchogram, pleural indentation, and size of primary tumor had no association with EGFR-positive adenocarcinoma. In contrast to the study by Rizzo et al, our study showed that the presence of fibrosis, which may be similar to convergence of surrounding structures, was associated with EGFR-positive

TE D

adenocarcinoma. Interestingly, the CT findings that were considered in our study as characteristic of EGFR-positive adenocarcinoma (ground glass opacity, nodules on

EP

non-tumor lobes, lobulation, absence of pleural effusion or cavity, etc.) were also

AC C

assessed in their study; however, they did not report any association between these CT features and EGFR-positive adenocarcinoma. Their result was totally discordant with our study. One explanation may be differences in patient selection. Although our study was a retrospective study, we included consecutive patients who were diagnosed as adenocarcinoma within 2 years in a single institution. In contrast to our study, the study by Rizzo et al have included only patients who have completed both CT scan and EGFR

15

ACCEPTED MANUSCRIPT

analysis between 2006 and 2014; this period was relatively long and may have been affected by the recent changes and developments in sequence techniques. Another

RI PT

possible reason for the discordance may be differences in ethnicity and EGFR profiles between Asian and non-Asian individuals with lung adenocarcinoma (36). Further

SC

studies that account for these differences would be needed. To the best of our

M AN U

knowledge, our study is essentially the first report on the CT features of EGFR-mutated adenocarcinoma on consecutive patients.

Our study has several limitations. The retrospective design had inevitable selection bias. Analyses of CT findings were only judged by qualitative evaluation and

TE D

might have lacked accuracy and repeatability; additionally, differentiation between cavity and air bronchogram with minimal bronchiectasis might have been confusing in

EP

some cases. The prediction score developed in this study requires validation in a

AC C

different cohort and should be planned as a future research.

Conclusion

EGFR mutation in lung adenocarcinoma may be predicted by CT scan findings of multiple bilateral lung metastases, convergence of surrounding structures, GGO, and notch; whereas the presence of cavity and pleural effusion may indicate non-mutated

16

ACCEPTED MANUSCRIPT

adenocarcinoma. In accordance with discovery of new targeted therapies, CT prediction

RI PT

scores for EGFR or new gene mutations may be valuable.

Acknowledgements

SC

We would like to thank Drs. Tomoyuki Akita and Junko Tanaka (Department of

M AN U

Epidemiology, Infectious Disease Control and Prevention, Institute of Biomedical and Health Science, Hiroshima University, Japan) for their valuable advice and assistance

Conflicts of interest

AC C

EP

None declared.

TE D

on the statistical analysis of this study.

17

ACCEPTED MANUSCRIPT

References (1) Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth

RI PT

factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 2004;350:2129-2139.

SC

(2) Paez JG, Janne PA, Lee JC, et al. EGFR mutations in lung cancer: correlation with

M AN U

clinical response to gefitinib therapy. Science 2004;304:1497-1500.

(3) Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med 2009;361:947-957.

(4) Tsao MS, Sakurada A, Cutz JC, et al. Erlotinib in lung cancer – Molecular and

TE D

clinical predictors of outcome. N Engl J Med 2005;353:133-144. (5) Kosaka T, Yamabe Y, Endoh H, et al. Mutations of the epidermal growth factor

EP

receptor gene in lung cancer: Biological and clinical implications. Cancer Res

AC C

2004;64:8919-8923.

(6) Shigematsu H, Lin L, Takahashi T, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst 2005;97(5):339-346. (7) Sakaki H, Endo K, Takada M, et al. L858R EGFR mutation status correlated with clinic-pathological features of Japanese lung cancer. Lung cancer 2006;54:103-108.

18

ACCEPTED MANUSCRIPT

(8) Lee HJ, Kim YT, Kang CH, et al. Epidermal growth factor receptor mutation in lung adenocarcinomas: relationship with CT characteristics and histologic subtypes.

RI PT

Radiology 2013;268:254-264.

(9) Togashi Y, Masago K, Kubo T, et al. Association of diffuse, random pulmonary

SC

metastases, including miliary metastases, with epidermal growth factor receptor

M AN U

mutations in lung adenocarcinoma. Cancer 2011;117(4):819-825.

(10) Gandhi L, Jänne P. Crizotinib for AKL-rearranged non-small cell lung cancer: A new targeted therapy for a new target. Clin Cancer Res 2012;18(14):3737-3742. (11) Yasuda H, Figueiredo-Pontes LL, Kobayashi S, et al. Preclinical rationale for use of

TE D

the clinically-available multi-targeted tyrosine kinase inhibitor crizotinib in ROS1 translocated lung cancer. J Thorac Oncol 2012;7(7):1086-1090.

EP

(12) Kohno T, Tsuta K, Tsuchihara K, et al. RET fusion gene: translation to personalized

AC C

lung cancer therapy. Cancer Sci 2013;104(11):1396-1400. (13) Yamamoto S, Korn R, Oklu R, et al. ALK molecular phenotype in non-small cell lung cancer: CT radiogenomic characterization. Radiology 2014;272:568-576. (14) Nagai Y, Miyazawa H, Huqun, et al. Genetic heterogeneity of the epidermal growth factor receptor in non-small cell lung cancer cell lines revealed by a rapid and sensitive detection system, the peptide nucleic acid-locked nucleic acid PCR clamp. Cancer Res

19

ACCEPTED MANUSCRIPT

2005;65(16):7276-7282. (15) Usuda K, Sagawa M, Motono N, et al. Relationship between EGFR mutation status

RI PT

of lung cancer and preoperative factors - Are they predictive? Asian Pac J Cancer 2014;15:657-662.

SC

(16) Yano M, Sasaki H, Kobayashi Y, et al. Epidermal growth factor receptor gene

J Thorac Oncol 2006;1(5):413-416.

M AN U

mutation and computed tomographic findings in peripheral pulmonary adenocarcinoma.

(17) Sugano M, Shimizu K, Nakano T, at al. Correlation between computed tomography findings and epidermal growth factor receptor and KRAS gene mutations in patients

TE D

with pulmonary adenocarcinoma. Oncol Report 2011;26:1205-1211. (18) Aoki T, Hanamiya M, Uramoto H, et al. Adenocarcinoma with predominant

EP

ground-glass opacity: correlation of morphology and molecular biomarkers. Radiology

AC C

2012;264:590-596.

(19) Sakuma Y, Matsukuma S, Yoshihara M, et al. Distinctive evaluation of non-mucinous and mucinous subtypes of bronchoalveolar carcinomas in EGFR and K-ras gene-mutation analysis for Japanese lung adenocarcinomas. Am J Clin Pathol 2007;128:100-108. (20) Sakuma Y, Matsukuma S, Yoshihara M, et al. Epidermal growth factor receptor

20

ACCEPTED MANUSCRIPT

gene mutations in atypical adenomatous hyperplasia of the lung. Modern Pathol 2007;20:967-973.

RI PT

(21) Yanagawa N, Shiono S, Abiko M, et al. The correlation of the international association for the study of lung cancer (IASLC)/American thoracic society

SC

(ATS)/European respiratory society (ERS) classification with prognosis and EGFR

M AN U

mutation in lung adenocarcinoma. Ann Thorac Surg 2014;98:453-458.

(22) Sun PL, Seol H, Lee HJ, et al. High incidence of EGFR mutations in Korean men smokers with no intratumoral heterogeneity of lung adenocarcinomas. Correlation with

2012;7(2):323-330.

TE D

histologic subtypes, EGFR/TTF-1 expressions, and clinical features. J Thorac Oncol

(23) Yoshizawa A, Sumiyoshi S, Sonobe M, et al. Validation of the IASLC/ATS/ERS

EP

lung adenocarcinoma classification for prognosis and association with EGFR and

AC C

KRAS gene mutations: analysis of 440 Japanese patients. J Thorac Oncol 2013;8(1):52-61.

(24) Song Z, Zhu H, Guo Z, et al. Correlation of EGFR mutation and predominant histologic subtype according to the new lung adenocarcinoma classification in Chinese patients. Med Oncol 2013;30:645. (25) Noguchi M, Morikawa A, Kawasaki M et al. Small adenocarcinoma of the lung:

21

ACCEPTED MANUSCRIPT

histologic characteristics and prognosis. Cancer 1995;75:191-202.

Japan: Medical Science International; 2011:178-197.

RI PT

(26) Murata K, T Jyokoh, S Murayama. Ikezoe`s CT of the chest 3rd Edition. Tokyo,

(27) Onn A, Choe DH, Herbst RS, et al. Tumor cavitation in stage 1 non-small cell lung

SC

cancer: Epidermal growth factor receptor expression and prediction of poor outcome.

M AN U

Radiology 2005;237:342-347.

(28) Tuddenham W J. Glossary of terms for thoracic radiology: recommendations of the Nomenclature Committee of the Fleischner Society. Am J Roentgenol 1984;143:509-517.

TE D

(29) Zaibo L, Dacic S, Pantanowitz L, et al. Correlation of cytomorphology and molecular findings in EGFR+, KRAS+, and ALK+ lung adenocarcinomas. Am J Clin

EP

Pathol 2014;141:420-428.

AC C

(30) Lee Y, Lee HJ, Kim YT, et al. Imaging characteristics of stage 1 non-small cell lung cancer on CT and FDG-PET: Relationship with epidermal growth factor receptor protein expression status and survival. Korean J Radiol 2013;14(2):375-383. (31) Selvaggi G, Novello S, Torri V, et al. Epidermal growth factor receptor overexpression correlates with a poor prognosis in completely resected non-small-cell lung cancer. Ann Oncol 2004;15:28-32.

22

ACCEPTED MANUSCRIPT

(32) Sasaki H, Shimizu S, Okuda K, et al. Epidermal growth factor receptor gene amplification in surgical resected Japanese lung cancer. Lung cancer

RI PT

2009;64(3):295-300.

(33) Togashi Y, Masago K, Kubo T, et al. Association of diffuse, random pulmonary

SC

metastasis, including miliary metastases, with epidermal growth factor receptor

M AN U

mutations in lung adenocarcinoma. Cancer 2011;15:819-825.

(34) Park K, Goto K. A review of the benefit-risk profile of gefitinib in Asian patients with advanced non-small-cell lung cancer. Curr Med Res Opin 2006; 22(3):561-573. (35) Rizzo S, Petrella F, Buscarino V, et al. CT radiogenomic characterization of EGFR,

TE D

K-ras, and ALK mutations in non-small cell lung cancer. Eur Radiol 2016; 26:32-42. (36) Soh J, Toyooka S, Matsuo K, et al. Ethnicity affects EGFR and KRAS gene

AC C

EP

alterations of lung adenocarcinoma. Oncol Lett 2015; 10:1775-1782.

23

ACCEPTED MANUSCRIPT

Figure legends

RI PT

Figure 1. CT findings that were analyzed on patients with lung adenocarcinoma

A. Multiple bilateral lung metastases and prominent peribronchovascular extension of

SC

the primary lesion; B. convergence of surrounding structures and air bronchogram; C.

M AN U

surrounding ground glass opacity; and D. cavity in the primary lesion

Figure 2. Score for prediction of EGFR mutation by ROC curve analysis Score = -0.79 + spiculation × (-0.49) + notch × (0.70) + pleural effusion × (-0.86) +

TE D

multiple bilateral metastases × (2.28) + prominent peribronchovascular extension × (-0.43) + air bronchogram × (-0.74) + convergence of surrounding structures × (2.01) +

AC C

EP

cavity × (-1.61) + surrounding ground glass opacity × (1.21)

24

ACCEPTED MANUSCRIPT

Table 1. Baseline characteristics of patients with lung adenocarcinoma (N = 263)

P-value

Negative

Number of patients

103 (39%)

160 (61%)

Male-to-female ratio

44 (42.7%):59 (57.3%)

112 (70%):48(30%)

<0.001

68 (range, 38–86)

69 (range, 38–90)

N.S.

45 (43.7%)

119 (74.4%)*2

<0.001

Heavy smoker*3

31

106

Light smoker*3

11

Never smoker

58

N/A*4

3

Median patient age (years)

Stage

40 2

1

55 (53.4%)

59 (36.9%)

2

7 (6.8%)

19 (11.9%)

3

11 (10.7%)

25 (15.6%)

4

30 (29.1%)

57 (35.6%)

29.2mm

34.7mm

*1 Non-smoking

N.S.

N.S.

TE D

Tumor size (average)*5

status was defined as those with lifetime exposure to 100 cigarettes or

less *2 Smoking

11

M AN U

Smoking history (yes)*1

RI PT

Positive

SC

EGFR mutation

history was not available in one patient in the non-mutated group

*3

Heavy smoking was defined as 20 or more pack-year, and light smoking as less than

AC C

20 pack-year. *4

EP

Fisher’s exact test and Mann-Whitney were used for statistical analysis

Information about pack-year was not available in 3 patients in the mutated group and

2 patients in the non-mutated group *5

Size of the major axis on axial image. Nineteen patients without visible primary

lesions because of massive pleural effusion were excluded. Tumor size was not available in one patient in the non-mutated group. EGFR, epidermal growth factor recepter N.S., not significant

ACCEPTED MANUSCRIPT

Table 2. Difference in CT features between EGFR mutated and non-mutated adenocarcinoma (N=244)

Univariateanalysis

+

-

OR

95%CI

p-value

OR

95%CI

p-value

1.05-25.37

0.0258

9.76

1.82-82.46

0.0152

3.24-18.42

<0.0001

1.64-7.10

0.0011

0.31-1.29

0.2256

0.4721

0.0505

1.04-4.04

0.0428

+

9

7 (77.8%)

2 (22.2%)

5.16

-

235

95 (40.4%)

140(59.6%)

1.00

Convergenceofsurrounding

+

70

44 (62.9%)

26 (37.1%)

3.38

structures

-

174

58 (33.3%)

116(66.7%)

1.00

Surrounding ground-glass

+

103

56 (54.4%)

47 (45.6%)

2.46

opacity

-

141

46 (32.6%)

95 (67.4%)

1.00

Prominent peribronchovascular

+

91

43 (47.3%)

48 (52.8%)

1.43

extension

-

153

59 (38.6%)

94 (61.4%)

1.00

+

113

50 (44.2%)

63 (55.8%)

1..21

-

131

52 (39.7%)

79 (60.3%)

1.00

+

148

62 (41.9%)

86 (58.1%)

1.01

-

96

40 (41.7%)

56 (58.3%)

1.00

+

184

81 (44.0%)

103(56.0)

1.46

-

60

21 (35%)

39 (65.0%)

1.00

+

166

66 (39.8%)

100(60.2%)

0.77

-

78

36 (46.2%)

42 (53.9%)

1.00

+

46

9 (19.6%)

37 (80.4%)

0.27

105(53.0%)

1.00

Multiplebilateralmetastases

Spiculation

EP

Cavity

TE D

Notch

-

198

93 (47.0%)

1.00

1.90-6.03

<0.0001

7.48

SC

1.00

1.46-4.15

0.0007

3.36 1.00

M AN U

Air-bronchogram

Pleuralindentation

Multivariate analysis

RI PT

EGFRstatus N

0.84-2.41

0.72-2.01

0.60-1.70

0.1832

0.4721

0.9722

0.65 1.00 0.48 1.00 3.36 1.00

0.80-2.67

0.2186

-

-

-

0.45-1.33

0.3449

0.61

0.31-1.21

0.1592

0.13-0.60

0.0007

0.20

0.08-0.47

0.0004

1.00

AC C

CT: Computed tomography, EGFR: Epidermal growth factor recepter, OR: Odds ratio, CI: Confidence interval

ACCEPTED MANUSCRIPT

Table 3. Comparison of EGFR exon mutations based on CT findings

Exon 21 (n = 51)

Multiple bilateral lung metastases

4 (8.2%)

2 (3.9%)

17

26

(34.7%)

(51.0%)

24

31

Convergence of surrounding structures GGO

(60.8%)

M AN U

(49.0%)

P value

RI PT

Exon 19 (n = 49)

SC

EGFR mutation

1.0000

0.3003

0.7067

25 (51.0%)

20 (30.2%)

≤ 25%

10 (20.4%)

10 (%)

25% to ≤ 50%

4 (8.2%)

9 (%)

50% to ≤ 75%

2 (4.1%)

6 (%)

75% to 100%

8 (16.3%)

6 (%)

32 (65.3%)

30 (58.8%)

0.7815

2 (4.1%)

7 (13.7%)

0.2762

Cavity

EP

Notch

TE D

GGO-negative

0.3365

AC C

EGFR: Epidermal growth factor receptor, CT: Computed tomography, GGO: Ground glass opacity

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

RI PT

M AN US C

Figure 2. Score for prediction of EGFR mutation and ROC curve Scores for prediction of EGFR mutation (+)

(-)

multiple bilateral metastases

2.28

0

convergence of surrounding structures

2.01

0

surrounding ground glass opacity

1.21

0

prominent peribronchovascular extension

-0.43

0

air-bronchogram

-0.74

0

notch

0.70

0

spiculation

-0.49

0

cavity

-1.61

0

pleural effusions

-0.86

0

AC C

EP

TE

D

Sensitivity

CT findings

1-Specificity

Cut-off value 0.23, AUC 0.78897, sensitivity 78.4%, specificity 70.4%, accuracy 73.8%