- Email: [email protected]
ERS histologic subtype in surgically resected lung adenocarcinomas
Accepted Manuscript Title: Added prognostic value of CT characteristics and IASLC/ATS/ERS histologic subtype in surgically resected lung adenocarcinomas Authors: Young Joo Suh, Hyun-Ju Lee, Young Tae Kim, Chang Hyun Kang, In Kyu Park, Yoon Kyung Jeon, Doo Hyun Chung PII: DOI: Reference:
S0169-5002(18)30334-9 https://doi.org/10.1016/j.lungcan.2018.04.007 LUNG 5620
To appear in:
Lung Cancer
Received date: Revised date: Accepted date:
11-1-2018 2-4-2018 7-4-2018
Please cite this article as: Suh Young Joo, Lee Hyun-Ju, Kim Young Tae, Kang Chang Hyun, Park In Kyu, Jeon Yoon Kyung, Chung Doo Hyun.Added prognostic value of CT characteristics and IASLC/ATS/ERS histologic subtype in surgically resected lung adenocarcinomas.Lung Cancer https://doi.org/10.1016/j.lungcan.2018.04.007 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.
Added prognostic value of CT characteristics and IASLC/ATS/ERS histologic subtype in surgically resected lung adenocarcinomas
In Kyu Park, MDc, Yoon Kyung Jeon, MDd, Doo Hyun Chung, MDd
a
IP T
Young Joo Suh, MDa,b,1, Hyun-Ju Lee, MDa,b, Young Tae Kim, MDc, Chang Hyun Kang, MDc,
SC R
Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu,
Seoul, 03080, Korea,
Department of Radiology, Seoul National University College of Medicine, 103 Daehak-ro,
U
b
N
Jongnogu, Seoul, 03080, Korea, c
A
Department of Thoracic and Cardiovascular Surgery, Seoul National University Hospital,
M
Cancer Research Institute, Seoul National University College of Medicine, Seoul, Republic
ED
of Korea,
Department of Pathology, Seoul National University Hospital, Seoul, Republic of Korea
1
Present address: Department of Radiology, Severance Hospital, Yonsei University College of
PT
d
A
CC E
Medicine, Korea
*Corresponding
author : Hyun-Ju Lee, MD.
Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul, 03080, Korea Department of Radiology, Seoul National University College of Medicine, 103 Daehak-ro,
IP T
Jongnogu, Seoul, 03080, Korea Tel: 82-2-2072-1861, Fax: 82-2-743-6385
A
CC E
PT
ED
M
A
N
U
SC R
E-mail: [email protected]
Highlights In surgically resected adenocarcinomas, prognostic factors were investigated. EGFR mutation were not associated with tumor recurrence. CT characteristics and histologic subtype were independent predictors of recurrence.
SC R
IP T
Both had added prognostic values over tumor size and stage.
Abstract
U
Objectives: Our study investigates the added value of computed tomography (CT)
N
characteristics, histologic subtype classification of the International Association for the Study
A
of Lung Cancer (IASLC)/the American Thoracic Society (ATS)/the European Respiratory
M
Society (ERS), and genetic mutation for predicting postoperative prognoses of patients who
ED
received curative surgical resections for lung adenocarcinoma. Materials and Methods: We retrospectively enrolled 988 patients who underwent curative
PT
resection for invasive lung adenocarcinoma between October 2007 and December 2013. Cox’s proportional hazard model was used to explore the risk of recurrence-free survival,
CC E
based on the combination of conventional prognostic factors, CT characteristics, IASLC/ATS/ERS histologic subtype, and epidermal growth factor receptor (EGFR)
A
mutations. Incremental prognostic values of CT characteristics, histologic subtype, and EGFR mutations over conventional risk factors were measured by C-statistics. Results: During median follow-up period of 44.7 months (25th to 75th percentile 24.6-59.7 months), postoperative recurrence occurred in 248 patients (25.1%). In univariate Cox proportion hazard model, female sex, tumor size and stage, CT characteristics, and
predominant histologic subtype were associated with tumor recurrence (P<0.05). In multivariate Cox regression model adjusted for tumor size and stage, both CT characteristics and histologic subtype were independent tumor recurrence predictors (P<0.05). Cox proportion hazard models combining CT characteristics or histologic subtype with size and
IP T
tumor stage showed higher C-indices (0.763 and 0.767, respectively) than size and stage-only models (C-index 0.759, P>0.05).
SC R
Conclusion: CT characteristics and histologic subtype have relatively limited added
U
prognostic values over tumor size and stage in surgically resected lung adenocarcinomas.
N
Keywords
A
CC E
PT
ED
M
A
Lung adenocarcinoma; prognosis; CT characteristics; histologic subtype; EGFR mutation
1. Introduction Lung cancer is the leading cause of cancer death worldwide, and adenocarcinoma is the most common histologic type of lung cancer. Tumor stage is a well-established prognostic factor in lung adenocarcinomas, which enables discrimination and prediction of recurrence-
IP T
free survival (RFS) and overall survival (OS) [1]. Recently, several nonconventional indicators have been suggested to be prognosticators, including histologic subtype proposed
SC R
by the International Association for the Study of Lung Cancer (IASLC)/the American
Thoracic Society (ATS)/the European Respiratory Society (ERS), and the presence of genetic mutations such as those within the epidermal growth factor receptor (EGFR).
U
Prognostic value of the predominant histology subtype classification, proposed by
N
IASLC/ATS/ERS in 2011, has been established across various stages of surgically resected
A
adenocarcinomas [2-7]. In lung adenocarcinomas, RFS and OR have been discriminated
M
according to the predominant histologic subtype. The presence of EGFR mutation, which is
ED
frequently observed in Asians, females, and non-smokers, is associated with good response to EGFR tyrosine kinase inhibitor (TKI) therapy [8-10], and impact of EGFR mutation on
PT
survival remains controversial [8,11-18].
CC E
Meanwhile, a few studies have reported the prognostic utility of lesion characteristics on computed tomography (CT), demonstrating that adenocarcinomas presenting with pure ground-glass nodule (GGN) correlate with better survival than those with part-solid nodules
A
(PSNs) or solid nodules; and at identical tumor stage, solid nodules correspond to worse prognoses than PSN [19-21]. Moreover, ground-glass opacity (GGO) ratio, size, and solid portion within tumors presenting PSN impact postoperative prognosis [22,23]. Although these histologic, genetic, and radiologic markers can be potentially prognostic in surgically resected lung adenocarcinomas, their incremental prognostic values
over traditional prognosticators, such as tumor stage and size, have yet to be adequately established. The aim of this study was to investigate the added prognostic value of CT characteristics, IASLC/ATS/ERS histologic subtype, and genetic mutations in predicting postoperative prognoses in a large cohort of patients who received curative surgical resection
IP T
for lung adenocarcinomas.
SC R
2. Materials and Methods 2.1. Patients
U
This study was approved by the Institutional Review Board of our institution, and
N
informed consent was waived for this retrospective study. Using a surgical database of our
A
institution, we searched 1,310 patients who received curative resections for lung
M
adenocarcinoma between October 2007 and December 2013. We excluded 177 patients with IASLC/ATS/ERS histologic subtypes, as well as those with adenocarcinoma in situ or
ED
minimally invasive adenocarcinoma, which are known to have a good prognosis of almost 100% RFS rate 5 years after surgery [4,5,24,25]. We also excluded 129 patients with
PT
unavailable histologic subtype classifications, 13 patients who received second operations
CC E
due to recurrent or multiple lesions, two with enteric subtypes, and one without available preoperative CT examination. Our final study population consisted of 988 patients. 2.2. Clinical Characteristics
A
Preoperative, intraoperative, and postoperative clinical records were reviewed, and
baseline clinical characteristics such as sex, age, and smoking history were recorded. Whether patients received neoadjuvant or adjuvant chemotherapy was identified from their medical records. In our institution, platinum- and taxane-based neoadjuvant treatment were indicated for patients with bulky N2 diseases. Adjuvant platinum-based doublet
chemotherapy was indicated for patients with pathological stage IIA or greater. 2.3. CT image Analysis Two thoracic radiologists reviewed preoperative CT images. Lesion characteristics were assessed and classified into four categories (pure GGN, ground-glass opacity [GGO]-
IP T
dominant PSN, solid-dominant PSN, and solid nodule; Supplemental Figure 1). GGO portion was defined as increased opacity that did not obscure adjacent airway and pulmonary
SC R
vascular structures. The proportion of GGO was calculated according to the ratio of
maximum GGO diameter to that of the total tumor across the largest cross section, and classified as follows: pure GGN; GGO 100%, GGO-dominant PSN; 50%≤ GGO<100%,
U
solid-dominant PSN; 0%< GGO<50%, solid nodule; GGO 0% [26]. If disagreement occurred
N
between the two readers, a consensus reading was performed to reach the final conclusion.
A
2.4. Histologic Evaluation and Molecular Analysis
M
One board-certificated thoracic pathologist reviewed the pathologic slides and
ED
assessed predominant subtype among IASLC/ATS/ERS classification of lung adenocarcinoma [27]. According to the predominant histologic subtype, tumors were graded
PT
using a modified version of system by Sica et al [28]. In addition to the previous 3-tier grading system, we classified invasive mucinous adenocarcinoma as a separate category,
CC E
since tumors with invasive mucinous adenocarcinoma patterns show a different clinical course [29]. Therefore, the final grading system was defined as follows: grade 1, histologic
A
subtypes of lepidic predominant invasive adenocarcinoma; grade 2, acinar or papillary predominant; grade 3, micropapillary or solid predominant type; and grade 4, invasive mucinous adenocarcinoma. Tumor stage was assessed by the 7th edition of TNM classification for lung cancer [30]. Mutation status of EGFR exons 18, 19, 20, and 21 was examined with polymerase chain reaction–based assay and confirmed through direct sequencing, as previously
described [31,32]. 2.5. Outcomes The primary outcome was time to recurrence. Recurrence after surgery was assessed according to follow-up medical records and imaging study results. Recurrence was defined
IP T
when disease appeared at anatomically contiguous sites, or regional lymphatics of the primary tumor, lung parenchyma, or extrapulmonary distant site, after a disease-free interval
SC R
of more than three months after curative resection. Date of recurrence was defined as the date of examination in which recurrence was documented. 2.6. Statistical Analysis
U
The relationship between clinical variables and EGFR mutation was analyzed using
N
chi-square or student t-tests. For analysis, we selected the following clinical variables that are
A
known to affect long-term survival of lung adenocarcinoma: sex, age, smoking status, tumor
M
size, and stage. Interobserver agreement for the classification of CT characteristics was analyzed using weighted kappa statistics. RFS was investigated by the Kaplan–Meier method,
ED
and differences between groups, determined by risk factors, were tested using the log-rank
PT
test. Cox’s proportional hazard models were used to identify predictors of RFS. Hazard ratios (HRs) were used to assess as risk estimations, with 95% confidence intervals (CIs). For HR
CC E
assessment of CT characteristics, Firth logistic regression was performed to solve a separation issue, as none of the patients with pure GGN showed postoperative recurrence [33]. For assessment of the incremental prognostic values of CT characteristics, IASLC/ATS/ERS
A
histologic subtypes, and EGFR mutation status over conventional risk factors, C-statistics were performed and C-indices were compared between models [34]. P-values less than 0.05 were considered statistically significant. All statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC) and STATA (version 12.1, Stata Corp, College Station, TX).
3. Results 3.1. Patients The study population consisted of 988 patients (463 men, median age 64 years [25th
IP T
to 75th percentile, 56-70 years]), and their clinical characteristics are shown in Table 1.
Maximal tumor size by pathology specimen was median 2.45 cm (25th to 75th percentile 1.7-
SC R
3.2 cm). Pathologic stages were stage I in 693 patients (70.1%), stage II in 125 patients
(12.7%), stage III in 148 patients (15.0%), and stage IV in 22 patients (2.2%). Nine (0.9%) and 182 patients (18.4%) received neoadjuvant and adjuvant chemotherapy, respectively.
U
3.2. CT image Analysis
N
Among the four kinds of CT characteristics, solid-dominant PSN was the most
A
common feature (372 patients, 37.7%), and we achieved good interobserver variability for the
M
classification of CT characteristics (weighed kappa 0.744, 95% CI 0.708-0.781).
ED
3.3. Histologic Subtypes and Genetic Mutations The most common histologic subtype was acinar predominant (53%), followed by
PT
lepidic predominant (14%), papillary predominant (13.9%), and solid predominant (12.2%).
CC E
Invasive mucinous adenocarcinoma was observed in 44 (4.4%) out of 988 patients. Thirtyfive patients (3.5%) lacked EGFR mutation status information. Among the 553 patients (56%) with recorded EGFR mutations, gene 19 deletion was the most common type. CT
A
characteristics of lesions significantly differed according to predominant histologic subtype (Table 2, P<0.0001). Patients with EGFR mutations were more often females, never smokers, early stages (stage I), diagnosed with solid-dominant PSN on CT, and histologic grades 1 or 2 (lepidic, acinar, or papillary predominant subtypes), whereas patients with wild-type EGFR were more
often males, smokers, and diagnosed with solid nodule on CT (Table 3, P<0.0001). 3.4. Prognostic Models for Recurrence-free Survival Twenty patients lacked follow-up records of recurrence. During the median followup period of 44.7 months (25th to 75th percentile 24.6-59.7 months), postoperative recurrences
IP T
occurred in 248 out of 988 patients (25.1%). Kaplan-Meier survival curves demonstrated differences in RFS according to the histologic grading system (Figure 1). According to the
SC R
histologic grading system, the 5-year RFS rate was 92.1% for grade 1, 71.6% for grade 2, 53.5% for grade 3, and 72.3% for grade 4 (P<0.0001), showing discrimination of RFS
U
between different histologic grades.
N
In univariate Cox proportion hazard models, female sex, tumor size and stage, CT
A
characteristics, and predominant IASLC/ATS/ERS histologic subtype (grade) were associated
M
with tumor recurrence (P<0.05, Table 4). Age, smoking status, and presence of EGFR mutation were not associated with increasing HRs for tumor recurrence (P>0.05).
ED
We constructed multiple prognosis prediction models that combined tumor size and stage with CT characteristics or IASLC/ATS/ERS histologic type. In multivariate Cox
PT
regression model adjusted for tumor size and stage, both CT characteristics and histologic
CC E
subtype were independent predictors of tumor recurrence (P<0.05, Models 2 and 3 in Table 5). Among factors that were significant in univariate model, female sex was not associated with tumor recurrence. In Model 2, which combined CT characteristics with tumor size and stage,
A
patients with solid-dominant PSNs and solid nodules showed increased HRs for recurrence (HR 7.254, 95% CI 1.047-915.721, P=0.0428 for solid-dominant PSN; HR 11.119, 95% CI 1.612-1402.475, P=0.0067 for solid nodule) compared to patients who displayed CT characteristics of pure GGN. In Model 3, which combined histologic subtype with tumor size and stage, tumors with grades 2 and 3 showed increased HR for shorter RFS (HR 2.287, 95%
CI 1.264-4.138, P=0.0063; HR 2.773, 95% CI 1.461-5.264, P=0.0018) compared to grade 1 tumors. C-indices of Model 2 (size + stage + CT characteristics, C-index 0.763) and Model 3 (size + stage + IASLC/ATS/ERS histologic subtype, c-index 0.767) were higher than Model 1 (size + stage, c-index 0.759); however, these differences were not statistically significant
IP T
(p=0.647 for Model 1 vs. Model 2; p=0.158 for Model 1 vs. Model 3).
SC R
4. Discussion
Our study demonstrates that CT characteristics and IASLC/ATS/ERS histologic
U
subtypes have added prognostic value, with tumor stage and size, for predicting RFS in
N
patients who receive curative surgical resections for lung adenocarcinomas, while the
A
presence of EGFR mutations does not. Compared to lung adenocarcinomas that present as
M
pure GGNs on CT scan, adenocarcinomas that present as solid-dominant PSNs and solid nodules are associated with increased HRs for recurrence. Compared to lepidic predominant
ED
type tumors, we found that acinar or papillary predominant type adenocarcinomas and micropapillary or solid predominant types are associated with increased HR for tumor
PT
recurrence. Prognosis models that include CT characteristics or histologic subtype, in
CC E
combination with stage and size, show higher C-indices than models constructed with stage and tumor size only, although the difference does not show statistical significance. The prognostic utility of IASLC/ATS/ERS histologic classification of lung
A
adenocarcinomas has been investigated in many studies, which have found that the predominant histologic type of adenocarcinomas is an independent prognosticator that can properly differentiate prognoses, after adjusting for tumor stage and other clinical variables [2-7]. For example, adenocarcinoma in situ and minimally invasive adenocarcinoma have 100% 5-year disease free survival rates. Adenocarcinomas with lepidic predominant
histologic subtypes have better prognoses than other histologic subtypes, whereas those with solid or micropapillary subtypes have worse prognoses. Our study results were consistent with those of previous studies in terms of achieving a clear discrimination of prognoses according to histologic type, and identifying independent prognostic values of predominant
IP T
histologic types. Additionally, our study demonstrated the incremental prognostic value of IASLC/ATS/ERS histologic classification for predicting RFS over conventional prognostic
SC R
indicators; however, our results should be carefully interpreted as no statistical significance
was observed between the two models, despite the higher c-index of prognostic model including histologic classification.
particularly
within
a
specific
stage
N
adenocarcinomas,
U
Previous studies have investigated the utility of CT parameters for lung of
surgically
resected
A
adenocarcinomas [19-23]. Most of these studies reported that CT parameters, such as nodule
M
type, size of solid portion, and GGO ratio, within lesions are associated with the
ED
postoperative prognoses; however, no existing study has yet examined the incremental value of CT characteristics in addition to tumor stage, especially across a large patient population.
PT
Our study included a large cohort of more than 900 patients with various tumor stages, and found that CT characteristics in combination with tumor stage and size have better prognostic
CC E
values in comparison with stage and size only, which is similar to histologic subtype results. This observation can be explained by the positive correlation between CT characteristics and
A
predominant histologic subtype, as shown in both our study and previous publications. Generally, GGO in lung adenocarcinomas manifesting as pure GGN or PSN on CT reflect the lepidic component of tumor, whereas solid portion in CT well correlates with non-lepidic, invasive component. Therefore, lesions with grade 1 subtype (lepidic predominant), which are associated with better prognoses, mostly present as pure GGN or GGO-predominant PSN, while lesions with grade 2 (acinar or papillary predominant) or grade 3 (micropapillary or
solid predominant) subtypes present as solid-predominant PSN or pure solid nodules [19,3538]. Invasive mucinous adenocarcinomas have distinct radiologic characteristics, such as airspace consolidation or pure solid nodules [29]. Although the prognosis model that includes CT characteristics shows a slightly lower C-index compared to the model with histologic
IP T
subtypes, the model with CT characteristics has a higher C-index than the one that includes tumor stage and size only. Therefore, CT characteristics are useful as a surrogate marker in
SC R
preoperative predictions of postoperative prognoses.
Interestingly, grade 4 subtype (invasive mucinous adenocarcinoma) showed distinguishable survival pattern between grade 1 and grade 2 tumors in our study. Invasive
U
mucinous adenocarcinomas have been known to have unique clinical, histologic, genetic, and
N
radiologic characteristics from non-mucinous lung adenocarcinomas. However, the prognosis
A
of invasive mucinous adenocarcinomas has been under debate [2,24,39]. A recent study
M
demonstrated that disease-free survival of invasive mucinous adenocarcinomas was between
ED
that of patients with lepidic-predominant non-mucinous adenocarcinomas (grade 1 in our study) and patients with acinar and papillary predominant adenocarcinomas (grade 2 in our
PT
study), which could be explained by aggressiveness of mucinous tumor and high rate of tumor spread in invasive mucinous adenocarcinomas [29]. Our study result was consistent
CC E
with that of previous study, and may support other recent studies which reported that invasive mucinous adenocarcinomas could be classified into intermediate histologic grade groups of
A
lung adenocarcinomas [4,40]. EGFR mutations are present in 41.1–60.5% of surgically resected lung
adenocarcinomas [8,9,14,18]; and in recurrent lung adenocarcinomas, EGFR mutations have a predictive value in patient response to TKI therapy. However, the prognostic value of EGFR mutations in survival predictions remain controversial. Some studies have argued that the
presence of EGFR mutations is an independent predictor of OS and RFS in surgically resected adenocarcinomas [11-15], whereas others have insisted that EGFR mutations are not significant predictors for postoperative survival and are predictive, rather than prognostic, markers [8,16-18]. We observed EGFR mutations in 58.1% of our study population, and our
IP T
results support the latter opinion. We hypothesize that such differences among studies may have resulted for the following reasons. Distributions of EGFR mutations significantly differ
SC R
according to CT characteristics and histologic subtypes, and the variability of characteristics
between study populations, including distributions of tumor stage, histologic subtypes, and prevalence of EGFR mutations, can affect outcomes. For instance, the solid nodule type
U
displayed a lower prevalence of EGFR mutations (47.2%, 188/399) than other CT
N
characteristics (63.6-73.1%) in our study, which was consistent with previous
A
findings [38,41-46]. Moreover, a higher prevalence of EGFR mutation was seen in grades 1
ED
observations [7,25,36,47,48].
M
and 2 histologic subtypes than in grades 3 and 4, which also agreed with previous
Our study had several limitations. First, we used the 7th edition lung cancer staging
PT
system, as our study was conducted in a retrospective manner; therefore, out database included patients who underwent surgery and received treatment based on the 7th edition
CC E
staging system, and our inclusion period occurred before the new 8th edition staging system was proposed. Second, CT images were analyzed by semi-quantitative methods, rather than
A
fully quantitative methods, to classify CT characteristics. However, we anticipate that the methods used in our study will assist patient prognosis estimations readily in clinical situations.
5. Conclusions
CT characteristics and histologic subtype have relatively limited added prognostic values over tumor stage and size in surgically resected lung adenocarcinomas. Preoperative CT characteristics represent feasible surrogate markers in preoperative predictions of postoperative prognoses.
IP T
Conflicts of interest.
SC R
The authors declare no conflicts of interest.
N
U
Acknowledgements
A
This research was supported by Basic Science Research Program through the
M
National Research Foundation of Korea (NRF) funded by the Ministry of Education (2017R1D1A1B04032467). We would like to thank the Medical Research Collaboration
ED
Center of Seoul National University Hospital for providing statistics consultation and statistical support, and Sang Min Lee, MD, Youkyung Lee, MD, and Mi-Jin Kang, MD for
A
CC E
PT
their expert support and contributions to the preparation and review of CT image data.
References
[1]
R. Rami-Porta, V. Bolejack, P. Goldstraw, The new tumor, node, and metastasis staging system, Seminars in respiratory and critical care medicine 32 (2011) 44-51.
[2]
A. Warth, T. Muley, M. Meister, et al., The novel histologic International Association for the Study
of
Lung
Cancer/American
Thoracic
Society/European
Respiratory
Society
IP T
classification system of lung adenocarcinoma is a stage-independent predictor of survival, Journal of clinical oncology : official journal of the American Society of Clinical Oncology 30 (2012) 1438-1446.
K. Tsuta, M. Kawago, E. Inoue, et al., The utility of the proposed IASLC/ATS/ERS lung
SC R
[3]
adenocarcinoma subtypes for disease prognosis and correlation of driver gene alterations, Lung cancer (Amsterdam, Netherlands) 81 (2013) 371-376. [4]
A. Yoshizawa, S. Sumiyoshi, M. Sonobe, et al., Validation of the IASLC/ATS/ERS lung adenocarcinoma classification for prognosis and association with EGFR and KRAS gene
U
mutations: analysis of 440 Japanese patients, Journal of thoracic oncology : official [5]
N
publication of the International Association for the Study of Lung Cancer 8 (2013) 52-61. J. Gu, C. Lu, J. Guo, et al., Prognostic significance of the IASLC/ATS/ERS classification in
A
Chinese patients-A single institution retrospective study of 292 lung adenocarcinoma, [6]
M
Journal of surgical oncology 107 (2013) 474-480.
H.Y. Lee, J.Y. Jeong, K.S. Lee, et al., Histopathology of lung adenocarcinoma based on new IASLC/ATS/ERS classification: prognostic stratification with functional and metabolic
[7]
ED
imaging biomarkers, Journal of magnetic resonance imaging : JMRI 38 (2013) 905-913. A. Mansuet-Lupo, A. Bobbio, H. Blons, et al., The new histologic classification of lung primary adenocarcinoma subtypes is a reliable prognostic marker and identifies tumors 643.
Y.T. Kim, Y.W. Seong, Y.J. Jung, et al., The presence of mutations in epidermal growth factor
CC E
[8]
PT
with different mutation status: the experience of a French cohort, Chest 146 (2014) 633-
receptor gene is not a prognostic factor for long-term outcome after surgical resection of non-small-cell lung cancer, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 8 (2013) 171-178.
A
[9]
[10]
M.W. Lin, C.T. Wu, J.Y. Shih, et al., Clinicopathologic characteristics and prognostic significance of EGFR and p53 mutations in surgically resected lung adenocarcinomas
[11]
T.Y. Chou, C.H. Chiu, L.H. Li, 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, Clinical cancer research : an official journal of the American Association for Cancer Research 11 (2005) 3750-3757. [12]
M. Sonobe, M. Kobayashi, M. Ishikawa, et al., Impact of KRAS and EGFR gene mutations on recurrence and survival in patients with surgically resected lung adenocarcinomas, Annals of surgical oncology 19 Suppl 3 (2012) S347-354.
[13]
J. Cadranel, A. Mauguen, M. Faller, et al., Impact of systematic EGFR and KRAS mutation
IP T
evaluation on progression-free survival and overall survival in patients with advanced nonsmall-cell lung cancer treated by erlotinib in a French prospective cohort (ERMETIC
project--part 2), Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 7 (2012) 1490-1502.
W.S. Liu, L.J. Zhao, Q.S. Pang, et al., Prognostic value of epidermal growth factor receptor
SC R
[14]
mutations in resected lung adenocarcinomas, Medical oncology (Northwood, London, England) 31 (2014) 771. [15]
V.N. Maturu, N. Singh, A. Bal, et al., Relationship of epidermal growth factor receptor
U
activating mutations with histologic subtyping according to International Association for
N
the Study of Lung Cancer/American Thoracic Society/European Respiratory Society 2011 adenocarcinoma classification and their impact on overall survival, Lung India : official V.H. Lee, V.P. Tin, T.S. Choy, et al., Association of exon 19 and 21 EGFR mutation patterns
M
[16]
A
organ of Indian Chest Society 33 (2016) 257-266.
with treatment outcome after first-line tyrosine kinase inhibitor in metastatic non-smallcell lung cancer, Journal of thoracic oncology : official publication of the International [17]
ED
Association for the Study of Lung Cancer 8 (2013) 1148-1155. H. Hu, Y. Pan, Y. Li, et al., Oncogenic mutations are associated with histological subtypes but do not have an independent prognostic value in lung adenocarcinoma, OncoTargets [18]
PT
and therapy 7 (2014) 1423-1437.
T. Ohba, G. Toyokawa, T. Kometani, et al., Mutations of the EGFR and K-ras genes in resected stage I lung adenocarcinoma and their clinical significance, Surgery today 44
CC E
(2014) 478-486.
[19]
Y. Kudo, J. Matsubayashi, H. Saji, et al., Association between high-resolution computed tomography findings and the IASLC/ATS/ERS classification of small lung adenocarcinomas in Japanese patients, Lung cancer (Amsterdam, Netherlands) 90 (2015) 47-54.
A
[20]
E.J. Hwang, C.M. Park, Y. Ryu, et al., Pulmonary adenocarcinomas appearing as part-solid ground-glass nodules: is measuring solid component size a better prognostic indicator?, European radiology 25 (2015) 558-567.
[21]
Y. Tsutani, Y. Miyata, T. Yamanaka, et al., Solid tumors versus mixed tumors with a groundglass opacity component in patients with clinical stage IA lung adenocarcinoma: prognostic comparison using high-resolution computed tomography findings, The Journal of thoracic and cardiovascular surgery 146 (2013) 17-23.
[22]
H. Uehara, Y. Tsutani, S. Okumura, et al., Prognostic role of positron emission tomography and high-resolution computed tomography in clinical stage IA lung adenocarcinoma, The Annals of thoracic surgery 96 (2013) 1958-1965.
[23]
S.H. Song, J.H. Ahn, H.Y. Lee, et al., Prognostic impact of nomogram based on whole tumour size, tumour disappearance ratio on CT and SUVmax on PET in lung adenocarcinoma, European radiology 26 (2016) 1538-1546.
[24]
A. Yoshizawa, N. Motoi, G.J. Riely, et al., Impact of proposed IASLC/ATS/ERS classification of
IP T
lung adenocarcinoma: prognostic subgroups and implications for further revision of staging based on analysis of 514 stage I cases, Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc 24 (2011) 653-664. [25]
N. Yanagawa, S. Shiono, M. Abiko, et al., The correlation of the International Association
SC R
for the Study of Lung Cancer (IASLC)/American Thoracic Society (ATS)/European Respiratory Society (ERS) classification with prognosis and EGFR mutation in lung adenocarcinoma, The Annals of thoracic surgery 98 (2014) 453-458. [26]
E.A. Park, H.J. Lee, Y.T. Kim, et al., EGFR gene copy number in adenocarcinoma of the lung
U
by FISH analysis: investigation of significantly related factors on CT, FDG-PET, and W.D. Travis, E. Brambilla, M. Noguchi, et al., International association for the study of lung cancer/american
thoracic
society/european
A
[27]
N
histopathology, Lung cancer (Amsterdam, Netherlands) 64 (2009) 179-186. respiratory
society
international
M
multidisciplinary classification of lung adenocarcinoma, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 6 (2011) 244-285.
G. Sica, A. Yoshizawa, C.S. Sima, et al., A grading system of lung adenocarcinomas based
ED
[28]
on histologic pattern is predictive of disease recurrence in stage I tumors, The American journal of surgical pathology 34 (2010) 1155-1162. H.Y. Lee,
M.J. Cha, K.S. Lee, et
PT
[29]
Adenocarcinomas
of
the
Lung:
al., Prognosis
Related
Factors
in Resected Invasive Mucinous and
Comparison
with
Resected
Nonmucinous Adenocarcinomas, Journal of thoracic oncology : official publication of the
CC E
International Association for the Study of Lung Cancer 11 (2016) 1064-1073.
[30]
P. Goldstraw, J. Crowley, K. Chansky, et al., The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh)
A
edition of the TNM Classification of malignant tumours, Journal of thoracic oncology :
[31]
official publication of the International Association for the Study of Lung Cancer 2 (2007) 706-714. T.J. Lynch, D.W. Bell, R. Sordella, et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib, N Engl J Med 350 (2004) 2129-2139.
[32]
H.G. Yi, H.J. Kim, Y.J. Kim, et al., Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) are effective for leptomeningeal metastasis from non-small cell lung
cancer patients with sensitive EGFR mutation or other predictive factors of good response for EGFR TKI, Lung cancer (Amsterdam, Netherlands) 65 (2009) 80-84. [33]
D. Firth, Bias Reduction of Maximum Likelihood Estimates, Biometrika 80 (1993) 27-38.
[34]
F.E. Harrell, Jr., K.L. Lee, D.B. Mark, Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors, Statistics in medicine 15 (1996) 361-387.
[35]
M.J. Cha, H.Y. Lee, K.S. Lee, et al., Micropapillary and solid subtypes of invasive lung thoracic and cardiovascular surgery 147 (2014) 921-928.e922.
[36]
IP T
adenocarcinoma: clinical predictors of histopathology and outcome, The Journal of H.J. Lee, Y.T. Kim, C.H. Kang, et al., Epidermal growth factor receptor mutation in lung
adenocarcinomas: relationship with CT characteristics and histologic subtypes, Radiology [37]
SC R
268 (2013) 254-264.
J.P. Ko, J. Suh, O. Ibidapo, et al., Lung Adenocarcinoma: Correlation of Quantitative CT Findings with Pathologic Findings, Radiology 280 (2016) 931-939.
[38]
Y. Yang, Y. Yang, X. Zhou, et al., EGFR L858R mutation is associated with lung
U
adenocarcinoma patients with dominant ground-glass opacity, Lung cancer (Amsterdam, [39]
N
Netherlands) 87 (2015) 272-277.
P.A. Russell, Z. Wainer, G.M. Wright, et al., Does lung adenocarcinoma subtype predict
A
patient survival?: A clinicopathologic study based on the new International Association for
M
the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary lung adenocarcinoma classification, Journal of thoracic oncology : official publication of the International Association for the Study of Lung [40]
ED
Cancer 6 (2011) 1496-1504.
H. Nakamura, H. Saji, T. Shinmyo, et al., Close association of IASLC/ATS/ERS lung adenocarcinoma subtypes with glucose-uptake in positron emission tomography, Lung
[41]
PT
cancer (Amsterdam, Netherlands) 87 (2015) 28-33. M. Hasegawa, F. Sakai, R. Ishikawa, et al., CT Features of Epidermal Growth Factor Receptor-Mutated
Adenocarcinoma
of
the
Lung:
Comparison
with
Nonmutated
CC E
Adenocarcinoma, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 11 (2016) 819-826.
[42]
S.J. Hong, T.J. Kim, Y.W. Choi, et al., Radiogenomic correlation in lung adenocarcinoma with
A
epidermal growth factor receptor mutations: Imaging features and histological subtypes,
[43]
European radiology 26 (2016) 3660-3668. O. Gevaert, S. Echegaray, A. Khuong, et al., Predictive radiogenomics modeling of EGFR mutation status in lung cancer, Scientific reports 7 (2017) 41674.
[44]
Y. Liu, J. Kim, F. Qu, et al., CT Features Associated with Epidermal Growth Factor Receptor Mutation Status in Patients with Lung Adenocarcinoma, Radiology 280 (2016) 271-280.
[45]
T.J. Kim, C.T. Lee, S.H. Jheon, et al., Radiologic Characteristics of Surgically Resected NonSmall Cell Lung Cancer With ALK Rearrangement or EGFR Mutations, The Annals of
thoracic surgery 101 (2016) 473-480. [46]
K. Usuda, M. Sagawa, N. Motono, et al., Relationships between EGFR mutation status of lung cancer and preoperative factors - are they predictive?, Asian Pacific journal of cancer prevention : APJCP 15 (2014) 657-662.
[47]
C. Villa, P.T. Cagle, M. Johnson, et al., Correlation of EGFR mutation status with predominant histologic subtype of adenocarcinoma according to the new lung adenocarcinoma classification of the International Association for the Study of Lung
IP T
Cancer/American Thoracic Society/European Respiratory Society, Archives of pathology & laboratory medicine 138 (2014) 1353-1357. [48]
H. Nakamura, H. Saji, T. Shinmyo, et al., Association of IASLC/ATS/ERS Histologic Subtypes
A
CC E
PT
ED
M
A
N
U
Resected Cases, Clinical lung cancer 16 (2015) 209-215.
SC R
of Lung Adenocarcinoma With Epidermal Growth Factor Receptor Mutations in 320
Figure Captions Figure 1. Kaplan-Meier survival curves for recurrence-free survival according to histologic
A
CC E
PT
ED
M
A
N
U
SC R
IP T
grading system.
Tables
SC R U N
A
CC E
PT
ED
M
A
n (%) Clinical variables Age (years)* 64 (56-70) Male sex 463 (46.9) Smoking history Never smoker 596 (60.3) Former smoker 390 (39.5) Current smoker 2 (0.2) Pathologic stage 1 693 (70.1) 2 125 (12.7) 3 148 (15) 4 22 (2.2) Size (cm)* 2.5 (1.7-3.2) Location Upper 499 (50.5) Middle 82 (8.3) Lower 341 (34.5) More than one lobe 66 (6.7) Type of surgery Lobectomy 900 (91.1) Segmentectomy 28 (2.8) Wedge resection 53 (5.4) Pneumonectomy 7 (0.7) CT characteristics Pure ground-glass nodule 27 (2.7) GGO-dominant PSN 180 (18.2) Solid-dominant PSN 372 (37.7) Solid nodule 409 (41.4) Missing information 1 (0.1) IASLC/ATS/ERS classification Lepidic predominant 138 (14) Acinar predominant 524 (53) Papillary predominant 137 (13.9) Micropapillary predominant 24 (2.4) Solid predominant 121 (12.2) Invasive mucinous adenocarcinoma 44 (4.4) EGFR mutation Wild type 400 (40.5) Mutation 553 (56) Gene 19 deletion 284 (51.4) Gene 21 missense 218 (39.45) Miscellaneous 51 (9.2) Missing information 35 (3.5) * Data are presented with median (25th to 75th percentile).
IP T
Table 1. Clinical characteristics of the study population (n=988).
CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society; EGFR, epidermal growth factor receptor
Table 2. Distribution of CT characteristics according to the histologic subtype. GGOdominant PSN
Soliddominant PSN
Solid nodule
20 (14.5) 6 (1.1) 1 (0.7) 0 (0) 0 (0) 0 (0)
55 (39.9) 91 (17.4) 25 (18.2) 0 (0) 4 (3.3) 5 (11.4)
48 (34.8) 231 (44.1) 45 (32.8) 6 (25) 26 (21.5) 16 (36.4)
15 (10.9) 196 (37.4) 66 (48.2) 18 (75) 91 (75.2) 23 (52.3)
20 (14.5) 7 (1.1) 0 (0) 0 (0)
55 (39.9) 116 (17.5) 4 (2.8) 5 (11.4)
48 (34.8) 276 (41.8) 32 (22.1) 16 (36.4)
P value
IP T
<0.0001
<0.0001
15 (10.9) 262 (39.6) 109 (75.2) 23 (52.3)
SC R
IASLC/ATS/ERS classification Lepidic predominant Acinar predominant Papillary predominant Micropapillary predominant Solid predominant Invasive mucinous adenocarcinoma Grading Grade 1 (lepidic predominant) Grade 2 (acinar or papillary predominant) Grade 3 (micropapillary or solid predominant) Grade 4 (invasive mucinous adenocarcinoma) .
Pure groundglass nodule
U
CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International
N
Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory
A
CC E
PT
ED
M
A
Society
Table 3. Comparison of clinical, histologic, radiologic, and genetic information according to the presence of EGFR mutations. EGFR mutation Absent Present (n=400) (n=553)
<0.0001
0.1684 <0.0001
385 (69.7) 166 (30.3) 0.1439
SC R
401 (72.4) 61 (11.2) 78 (14.1) 13 (2.3)
IP T
220 (39.9) 333 (60.1) 63 (62-65)
<0.0001
U
19 (3.4) 120 (21.7) 226 (40.9) 188 (34)
<0.0001
97 (17.5) 398 (72) 54 (9.7) 4 (0.7)
M
A
N
Sex Male 231 (57.7) Female 169 (42.2) Age 65 (64-66) Smoking Never 193 (48.3) Ever 207 (51.7) Staging 1 264 (66) 2 60 (15) 3 68 (17) 4 8 (2) CT characteristics (n=953) Pure ground-glass nodule 7 (1.8) GGO-dominant PSN 53 (13.3) Solid-dominant PSN 129 (32.2) Solid nodule 211 (52.7) IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) 40 (10) Grade 2 (acinar or papillary predominant) 237 (59.2) Grade 3 (micropapillary or solid predominant) 84 (21) Grade 4 (invasive mucinous adenocarcinoma) 39 (9.8) Data are numbers, with percentages in parentheses.
P value
A
CC E
PT
ED
EGFR, epidermal growth factor receptor; CT, computed tomography; GGO, ground-glass opacity; PSN, partsolid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society
Table 4. Univariate Cox hazard regression models for predictions of recurrence-free survival. Recurrence-free survival (n=968) HR 95% CI P value 0.764 0.595-0.980 0.0342 1 0.987-1.013 0.9825 1.415 1.328-1.507 <0.0001 1.223 0.951-1.573 0.1175 <0.0001 1 N/A N/A 3.193 2.271-4.49 <0.0001 6.521 4.898-8.68 <0.0001 3.234-13.655 .<0.000 6.645 1 0.828 0.642-1.068 0.1465
IP T
Sex (female) Age (per 1 year increase) Size (per mm increase) Smoking (smoker) Stage 1 (reference) 2 3 4
A
CC E
PT
ED
M
A
N
U
SC R
EGFR mutation (n=954) CT characteristics Pure ground-glass nodule (reference) 1 N/A N/A GGO-dominant PSN 5.569 0.76-709.919 0.1078 Solid-dominant PSN 13.661 1.989-1722.327 0.002 Solid nodule 25.295 3.712-3184.733 <0.0001 <0.0001 IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) 1 N/A N/A Grade 2 (acinar or papillary predominant) 3.397 1.89-6.103 <0.0001 Grade 3 (micropapillary or solid predominant) 6.432 3.456-11.97 <0.0001 Grade 4 (invasive mucinous adenocarcinoma) 2.634 1.11-6.252 0.0281 HR, hazard ratio; CI, confidence interval; N/A, non-applicable; EGFR, epidermal growth factor receptor; CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society
Table 5. Multivariate Cox hazard models for predictions of recurrence-free survival. Recurrence-free survival (n=968) HR 95% CI P value Model 1 (size + stage) Size Stage 1 (reference) 2 3 4 Model 2 (size + stage + CT characteristics) Size Stage 1 (reference) 2 3 4 CT characteristics Pure ground-glass nodule (reference) GGO-dominant PSN
1.212
1.128-1.302
<0.0001
1 2.261 5.099 5.437
N/A 1.561-3.275 3.745-6.942 2.63-11.24
N/A <.0001 <.0001 <.0001
1.21
1.124-1.299
<0.0001
1 1.891 4.056 4.311
N/A 1.296-2.728 2.948-5.567 1.956-8.31
N/A 0.0011 <.0001 0.0008
0.626586.125 1.047-915.721 1.612-1402.475
0.1679
1.215
0.0428 0.0067
1.13-1.306
<0.0001
1 2.026 4.364 5.162
N/A 1.389-2.956 3.171-6.005 2.478-10.751
N/A 0.0002 <.0001 <.0001
1 2.283 2.782 1.604
N/A 1.262-4.132 1.465-5.281 0.665-3.869
N/A 0.0064 0.0018 0.2931
A
M
ED
SC R
N
7.254 11.119
U
1 4.598
IP T
0.763 (0.733-0.792)
Solid-dominant PSN Solid nodule Model 3 (size + stage + IASLC/ATS/ERS classification) Size Stage 1 (reference) 2 3 4 IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) Grade 2 (acinar or papillary predominant) Grade 3 (micropapillary or solid predominant) Grade 4 (invasive mucinous adenocarcinoma)
PT
c-index (95% CI) 0.759 (0.729-0.789)
0.765 (0.738-0.795)
A
CC E
HR, hazard ratio; CI, confidence interval; N/A, non-applicable; CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society
S0169-5002(18)30334-9 https://doi.org/10.1016/j.lungcan.2018.04.007 LUNG 5620
To appear in:
Lung Cancer
Received date: Revised date: Accepted date:
11-1-2018 2-4-2018 7-4-2018
Please cite this article as: Suh Young Joo, Lee Hyun-Ju, Kim Young Tae, Kang Chang Hyun, Park In Kyu, Jeon Yoon Kyung, Chung Doo Hyun.Added prognostic value of CT characteristics and IASLC/ATS/ERS histologic subtype in surgically resected lung adenocarcinomas.Lung Cancer https://doi.org/10.1016/j.lungcan.2018.04.007 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.
Added prognostic value of CT characteristics and IASLC/ATS/ERS histologic subtype in surgically resected lung adenocarcinomas
In Kyu Park, MDc, Yoon Kyung Jeon, MDd, Doo Hyun Chung, MDd
a
IP T
Young Joo Suh, MDa,b,1, Hyun-Ju Lee, MDa,b, Young Tae Kim, MDc, Chang Hyun Kang, MDc,
SC R
Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu,
Seoul, 03080, Korea,
Department of Radiology, Seoul National University College of Medicine, 103 Daehak-ro,
U
b
N
Jongnogu, Seoul, 03080, Korea, c
A
Department of Thoracic and Cardiovascular Surgery, Seoul National University Hospital,
M
Cancer Research Institute, Seoul National University College of Medicine, Seoul, Republic
ED
of Korea,
Department of Pathology, Seoul National University Hospital, Seoul, Republic of Korea
1
Present address: Department of Radiology, Severance Hospital, Yonsei University College of
PT
d
A
CC E
Medicine, Korea
*Corresponding
author : Hyun-Ju Lee, MD.
Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul, 03080, Korea Department of Radiology, Seoul National University College of Medicine, 103 Daehak-ro,
IP T
Jongnogu, Seoul, 03080, Korea Tel: 82-2-2072-1861, Fax: 82-2-743-6385
A
CC E
PT
ED
M
A
N
U
SC R
E-mail: [email protected]
Highlights In surgically resected adenocarcinomas, prognostic factors were investigated. EGFR mutation were not associated with tumor recurrence. CT characteristics and histologic subtype were independent predictors of recurrence.
SC R
IP T
Both had added prognostic values over tumor size and stage.
Abstract
U
Objectives: Our study investigates the added value of computed tomography (CT)
N
characteristics, histologic subtype classification of the International Association for the Study
A
of Lung Cancer (IASLC)/the American Thoracic Society (ATS)/the European Respiratory
M
Society (ERS), and genetic mutation for predicting postoperative prognoses of patients who
ED
received curative surgical resections for lung adenocarcinoma. Materials and Methods: We retrospectively enrolled 988 patients who underwent curative
PT
resection for invasive lung adenocarcinoma between October 2007 and December 2013. Cox’s proportional hazard model was used to explore the risk of recurrence-free survival,
CC E
based on the combination of conventional prognostic factors, CT characteristics, IASLC/ATS/ERS histologic subtype, and epidermal growth factor receptor (EGFR)
A
mutations. Incremental prognostic values of CT characteristics, histologic subtype, and EGFR mutations over conventional risk factors were measured by C-statistics. Results: During median follow-up period of 44.7 months (25th to 75th percentile 24.6-59.7 months), postoperative recurrence occurred in 248 patients (25.1%). In univariate Cox proportion hazard model, female sex, tumor size and stage, CT characteristics, and
predominant histologic subtype were associated with tumor recurrence (P<0.05). In multivariate Cox regression model adjusted for tumor size and stage, both CT characteristics and histologic subtype were independent tumor recurrence predictors (P<0.05). Cox proportion hazard models combining CT characteristics or histologic subtype with size and
IP T
tumor stage showed higher C-indices (0.763 and 0.767, respectively) than size and stage-only models (C-index 0.759, P>0.05).
SC R
Conclusion: CT characteristics and histologic subtype have relatively limited added
U
prognostic values over tumor size and stage in surgically resected lung adenocarcinomas.
N
Keywords
A
CC E
PT
ED
M
A
Lung adenocarcinoma; prognosis; CT characteristics; histologic subtype; EGFR mutation
1. Introduction Lung cancer is the leading cause of cancer death worldwide, and adenocarcinoma is the most common histologic type of lung cancer. Tumor stage is a well-established prognostic factor in lung adenocarcinomas, which enables discrimination and prediction of recurrence-
IP T
free survival (RFS) and overall survival (OS) [1]. Recently, several nonconventional indicators have been suggested to be prognosticators, including histologic subtype proposed
SC R
by the International Association for the Study of Lung Cancer (IASLC)/the American
Thoracic Society (ATS)/the European Respiratory Society (ERS), and the presence of genetic mutations such as those within the epidermal growth factor receptor (EGFR).
U
Prognostic value of the predominant histology subtype classification, proposed by
N
IASLC/ATS/ERS in 2011, has been established across various stages of surgically resected
A
adenocarcinomas [2-7]. In lung adenocarcinomas, RFS and OR have been discriminated
M
according to the predominant histologic subtype. The presence of EGFR mutation, which is
ED
frequently observed in Asians, females, and non-smokers, is associated with good response to EGFR tyrosine kinase inhibitor (TKI) therapy [8-10], and impact of EGFR mutation on
PT
survival remains controversial [8,11-18].
CC E
Meanwhile, a few studies have reported the prognostic utility of lesion characteristics on computed tomography (CT), demonstrating that adenocarcinomas presenting with pure ground-glass nodule (GGN) correlate with better survival than those with part-solid nodules
A
(PSNs) or solid nodules; and at identical tumor stage, solid nodules correspond to worse prognoses than PSN [19-21]. Moreover, ground-glass opacity (GGO) ratio, size, and solid portion within tumors presenting PSN impact postoperative prognosis [22,23]. Although these histologic, genetic, and radiologic markers can be potentially prognostic in surgically resected lung adenocarcinomas, their incremental prognostic values
over traditional prognosticators, such as tumor stage and size, have yet to be adequately established. The aim of this study was to investigate the added prognostic value of CT characteristics, IASLC/ATS/ERS histologic subtype, and genetic mutations in predicting postoperative prognoses in a large cohort of patients who received curative surgical resection
IP T
for lung adenocarcinomas.
SC R
2. Materials and Methods 2.1. Patients
U
This study was approved by the Institutional Review Board of our institution, and
N
informed consent was waived for this retrospective study. Using a surgical database of our
A
institution, we searched 1,310 patients who received curative resections for lung
M
adenocarcinoma between October 2007 and December 2013. We excluded 177 patients with IASLC/ATS/ERS histologic subtypes, as well as those with adenocarcinoma in situ or
ED
minimally invasive adenocarcinoma, which are known to have a good prognosis of almost 100% RFS rate 5 years after surgery [4,5,24,25]. We also excluded 129 patients with
PT
unavailable histologic subtype classifications, 13 patients who received second operations
CC E
due to recurrent or multiple lesions, two with enteric subtypes, and one without available preoperative CT examination. Our final study population consisted of 988 patients. 2.2. Clinical Characteristics
A
Preoperative, intraoperative, and postoperative clinical records were reviewed, and
baseline clinical characteristics such as sex, age, and smoking history were recorded. Whether patients received neoadjuvant or adjuvant chemotherapy was identified from their medical records. In our institution, platinum- and taxane-based neoadjuvant treatment were indicated for patients with bulky N2 diseases. Adjuvant platinum-based doublet
chemotherapy was indicated for patients with pathological stage IIA or greater. 2.3. CT image Analysis Two thoracic radiologists reviewed preoperative CT images. Lesion characteristics were assessed and classified into four categories (pure GGN, ground-glass opacity [GGO]-
IP T
dominant PSN, solid-dominant PSN, and solid nodule; Supplemental Figure 1). GGO portion was defined as increased opacity that did not obscure adjacent airway and pulmonary
SC R
vascular structures. The proportion of GGO was calculated according to the ratio of
maximum GGO diameter to that of the total tumor across the largest cross section, and classified as follows: pure GGN; GGO 100%, GGO-dominant PSN; 50%≤ GGO<100%,
U
solid-dominant PSN; 0%< GGO<50%, solid nodule; GGO 0% [26]. If disagreement occurred
N
between the two readers, a consensus reading was performed to reach the final conclusion.
A
2.4. Histologic Evaluation and Molecular Analysis
M
One board-certificated thoracic pathologist reviewed the pathologic slides and
ED
assessed predominant subtype among IASLC/ATS/ERS classification of lung adenocarcinoma [27]. According to the predominant histologic subtype, tumors were graded
PT
using a modified version of system by Sica et al [28]. In addition to the previous 3-tier grading system, we classified invasive mucinous adenocarcinoma as a separate category,
CC E
since tumors with invasive mucinous adenocarcinoma patterns show a different clinical course [29]. Therefore, the final grading system was defined as follows: grade 1, histologic
A
subtypes of lepidic predominant invasive adenocarcinoma; grade 2, acinar or papillary predominant; grade 3, micropapillary or solid predominant type; and grade 4, invasive mucinous adenocarcinoma. Tumor stage was assessed by the 7th edition of TNM classification for lung cancer [30]. Mutation status of EGFR exons 18, 19, 20, and 21 was examined with polymerase chain reaction–based assay and confirmed through direct sequencing, as previously
described [31,32]. 2.5. Outcomes The primary outcome was time to recurrence. Recurrence after surgery was assessed according to follow-up medical records and imaging study results. Recurrence was defined
IP T
when disease appeared at anatomically contiguous sites, or regional lymphatics of the primary tumor, lung parenchyma, or extrapulmonary distant site, after a disease-free interval
SC R
of more than three months after curative resection. Date of recurrence was defined as the date of examination in which recurrence was documented. 2.6. Statistical Analysis
U
The relationship between clinical variables and EGFR mutation was analyzed using
N
chi-square or student t-tests. For analysis, we selected the following clinical variables that are
A
known to affect long-term survival of lung adenocarcinoma: sex, age, smoking status, tumor
M
size, and stage. Interobserver agreement for the classification of CT characteristics was analyzed using weighted kappa statistics. RFS was investigated by the Kaplan–Meier method,
ED
and differences between groups, determined by risk factors, were tested using the log-rank
PT
test. Cox’s proportional hazard models were used to identify predictors of RFS. Hazard ratios (HRs) were used to assess as risk estimations, with 95% confidence intervals (CIs). For HR
CC E
assessment of CT characteristics, Firth logistic regression was performed to solve a separation issue, as none of the patients with pure GGN showed postoperative recurrence [33]. For assessment of the incremental prognostic values of CT characteristics, IASLC/ATS/ERS
A
histologic subtypes, and EGFR mutation status over conventional risk factors, C-statistics were performed and C-indices were compared between models [34]. P-values less than 0.05 were considered statistically significant. All statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC) and STATA (version 12.1, Stata Corp, College Station, TX).
3. Results 3.1. Patients The study population consisted of 988 patients (463 men, median age 64 years [25th
IP T
to 75th percentile, 56-70 years]), and their clinical characteristics are shown in Table 1.
Maximal tumor size by pathology specimen was median 2.45 cm (25th to 75th percentile 1.7-
SC R
3.2 cm). Pathologic stages were stage I in 693 patients (70.1%), stage II in 125 patients
(12.7%), stage III in 148 patients (15.0%), and stage IV in 22 patients (2.2%). Nine (0.9%) and 182 patients (18.4%) received neoadjuvant and adjuvant chemotherapy, respectively.
U
3.2. CT image Analysis
N
Among the four kinds of CT characteristics, solid-dominant PSN was the most
A
common feature (372 patients, 37.7%), and we achieved good interobserver variability for the
M
classification of CT characteristics (weighed kappa 0.744, 95% CI 0.708-0.781).
ED
3.3. Histologic Subtypes and Genetic Mutations The most common histologic subtype was acinar predominant (53%), followed by
PT
lepidic predominant (14%), papillary predominant (13.9%), and solid predominant (12.2%).
CC E
Invasive mucinous adenocarcinoma was observed in 44 (4.4%) out of 988 patients. Thirtyfive patients (3.5%) lacked EGFR mutation status information. Among the 553 patients (56%) with recorded EGFR mutations, gene 19 deletion was the most common type. CT
A
characteristics of lesions significantly differed according to predominant histologic subtype (Table 2, P<0.0001). Patients with EGFR mutations were more often females, never smokers, early stages (stage I), diagnosed with solid-dominant PSN on CT, and histologic grades 1 or 2 (lepidic, acinar, or papillary predominant subtypes), whereas patients with wild-type EGFR were more
often males, smokers, and diagnosed with solid nodule on CT (Table 3, P<0.0001). 3.4. Prognostic Models for Recurrence-free Survival Twenty patients lacked follow-up records of recurrence. During the median followup period of 44.7 months (25th to 75th percentile 24.6-59.7 months), postoperative recurrences
IP T
occurred in 248 out of 988 patients (25.1%). Kaplan-Meier survival curves demonstrated differences in RFS according to the histologic grading system (Figure 1). According to the
SC R
histologic grading system, the 5-year RFS rate was 92.1% for grade 1, 71.6% for grade 2, 53.5% for grade 3, and 72.3% for grade 4 (P<0.0001), showing discrimination of RFS
U
between different histologic grades.
N
In univariate Cox proportion hazard models, female sex, tumor size and stage, CT
A
characteristics, and predominant IASLC/ATS/ERS histologic subtype (grade) were associated
M
with tumor recurrence (P<0.05, Table 4). Age, smoking status, and presence of EGFR mutation were not associated with increasing HRs for tumor recurrence (P>0.05).
ED
We constructed multiple prognosis prediction models that combined tumor size and stage with CT characteristics or IASLC/ATS/ERS histologic type. In multivariate Cox
PT
regression model adjusted for tumor size and stage, both CT characteristics and histologic
CC E
subtype were independent predictors of tumor recurrence (P<0.05, Models 2 and 3 in Table 5). Among factors that were significant in univariate model, female sex was not associated with tumor recurrence. In Model 2, which combined CT characteristics with tumor size and stage,
A
patients with solid-dominant PSNs and solid nodules showed increased HRs for recurrence (HR 7.254, 95% CI 1.047-915.721, P=0.0428 for solid-dominant PSN; HR 11.119, 95% CI 1.612-1402.475, P=0.0067 for solid nodule) compared to patients who displayed CT characteristics of pure GGN. In Model 3, which combined histologic subtype with tumor size and stage, tumors with grades 2 and 3 showed increased HR for shorter RFS (HR 2.287, 95%
CI 1.264-4.138, P=0.0063; HR 2.773, 95% CI 1.461-5.264, P=0.0018) compared to grade 1 tumors. C-indices of Model 2 (size + stage + CT characteristics, C-index 0.763) and Model 3 (size + stage + IASLC/ATS/ERS histologic subtype, c-index 0.767) were higher than Model 1 (size + stage, c-index 0.759); however, these differences were not statistically significant
IP T
(p=0.647 for Model 1 vs. Model 2; p=0.158 for Model 1 vs. Model 3).
SC R
4. Discussion
Our study demonstrates that CT characteristics and IASLC/ATS/ERS histologic
U
subtypes have added prognostic value, with tumor stage and size, for predicting RFS in
N
patients who receive curative surgical resections for lung adenocarcinomas, while the
A
presence of EGFR mutations does not. Compared to lung adenocarcinomas that present as
M
pure GGNs on CT scan, adenocarcinomas that present as solid-dominant PSNs and solid nodules are associated with increased HRs for recurrence. Compared to lepidic predominant
ED
type tumors, we found that acinar or papillary predominant type adenocarcinomas and micropapillary or solid predominant types are associated with increased HR for tumor
PT
recurrence. Prognosis models that include CT characteristics or histologic subtype, in
CC E
combination with stage and size, show higher C-indices than models constructed with stage and tumor size only, although the difference does not show statistical significance. The prognostic utility of IASLC/ATS/ERS histologic classification of lung
A
adenocarcinomas has been investigated in many studies, which have found that the predominant histologic type of adenocarcinomas is an independent prognosticator that can properly differentiate prognoses, after adjusting for tumor stage and other clinical variables [2-7]. For example, adenocarcinoma in situ and minimally invasive adenocarcinoma have 100% 5-year disease free survival rates. Adenocarcinomas with lepidic predominant
histologic subtypes have better prognoses than other histologic subtypes, whereas those with solid or micropapillary subtypes have worse prognoses. Our study results were consistent with those of previous studies in terms of achieving a clear discrimination of prognoses according to histologic type, and identifying independent prognostic values of predominant
IP T
histologic types. Additionally, our study demonstrated the incremental prognostic value of IASLC/ATS/ERS histologic classification for predicting RFS over conventional prognostic
SC R
indicators; however, our results should be carefully interpreted as no statistical significance
was observed between the two models, despite the higher c-index of prognostic model including histologic classification.
particularly
within
a
specific
stage
N
adenocarcinomas,
U
Previous studies have investigated the utility of CT parameters for lung of
surgically
resected
A
adenocarcinomas [19-23]. Most of these studies reported that CT parameters, such as nodule
M
type, size of solid portion, and GGO ratio, within lesions are associated with the
ED
postoperative prognoses; however, no existing study has yet examined the incremental value of CT characteristics in addition to tumor stage, especially across a large patient population.
PT
Our study included a large cohort of more than 900 patients with various tumor stages, and found that CT characteristics in combination with tumor stage and size have better prognostic
CC E
values in comparison with stage and size only, which is similar to histologic subtype results. This observation can be explained by the positive correlation between CT characteristics and
A
predominant histologic subtype, as shown in both our study and previous publications. Generally, GGO in lung adenocarcinomas manifesting as pure GGN or PSN on CT reflect the lepidic component of tumor, whereas solid portion in CT well correlates with non-lepidic, invasive component. Therefore, lesions with grade 1 subtype (lepidic predominant), which are associated with better prognoses, mostly present as pure GGN or GGO-predominant PSN, while lesions with grade 2 (acinar or papillary predominant) or grade 3 (micropapillary or
solid predominant) subtypes present as solid-predominant PSN or pure solid nodules [19,3538]. Invasive mucinous adenocarcinomas have distinct radiologic characteristics, such as airspace consolidation or pure solid nodules [29]. Although the prognosis model that includes CT characteristics shows a slightly lower C-index compared to the model with histologic
IP T
subtypes, the model with CT characteristics has a higher C-index than the one that includes tumor stage and size only. Therefore, CT characteristics are useful as a surrogate marker in
SC R
preoperative predictions of postoperative prognoses.
Interestingly, grade 4 subtype (invasive mucinous adenocarcinoma) showed distinguishable survival pattern between grade 1 and grade 2 tumors in our study. Invasive
U
mucinous adenocarcinomas have been known to have unique clinical, histologic, genetic, and
N
radiologic characteristics from non-mucinous lung adenocarcinomas. However, the prognosis
A
of invasive mucinous adenocarcinomas has been under debate [2,24,39]. A recent study
M
demonstrated that disease-free survival of invasive mucinous adenocarcinomas was between
ED
that of patients with lepidic-predominant non-mucinous adenocarcinomas (grade 1 in our study) and patients with acinar and papillary predominant adenocarcinomas (grade 2 in our
PT
study), which could be explained by aggressiveness of mucinous tumor and high rate of tumor spread in invasive mucinous adenocarcinomas [29]. Our study result was consistent
CC E
with that of previous study, and may support other recent studies which reported that invasive mucinous adenocarcinomas could be classified into intermediate histologic grade groups of
A
lung adenocarcinomas [4,40]. EGFR mutations are present in 41.1–60.5% of surgically resected lung
adenocarcinomas [8,9,14,18]; and in recurrent lung adenocarcinomas, EGFR mutations have a predictive value in patient response to TKI therapy. However, the prognostic value of EGFR mutations in survival predictions remain controversial. Some studies have argued that the
presence of EGFR mutations is an independent predictor of OS and RFS in surgically resected adenocarcinomas [11-15], whereas others have insisted that EGFR mutations are not significant predictors for postoperative survival and are predictive, rather than prognostic, markers [8,16-18]. We observed EGFR mutations in 58.1% of our study population, and our
IP T
results support the latter opinion. We hypothesize that such differences among studies may have resulted for the following reasons. Distributions of EGFR mutations significantly differ
SC R
according to CT characteristics and histologic subtypes, and the variability of characteristics
between study populations, including distributions of tumor stage, histologic subtypes, and prevalence of EGFR mutations, can affect outcomes. For instance, the solid nodule type
U
displayed a lower prevalence of EGFR mutations (47.2%, 188/399) than other CT
N
characteristics (63.6-73.1%) in our study, which was consistent with previous
A
findings [38,41-46]. Moreover, a higher prevalence of EGFR mutation was seen in grades 1
ED
observations [7,25,36,47,48].
M
and 2 histologic subtypes than in grades 3 and 4, which also agreed with previous
Our study had several limitations. First, we used the 7th edition lung cancer staging
PT
system, as our study was conducted in a retrospective manner; therefore, out database included patients who underwent surgery and received treatment based on the 7th edition
CC E
staging system, and our inclusion period occurred before the new 8th edition staging system was proposed. Second, CT images were analyzed by semi-quantitative methods, rather than
A
fully quantitative methods, to classify CT characteristics. However, we anticipate that the methods used in our study will assist patient prognosis estimations readily in clinical situations.
5. Conclusions
CT characteristics and histologic subtype have relatively limited added prognostic values over tumor stage and size in surgically resected lung adenocarcinomas. Preoperative CT characteristics represent feasible surrogate markers in preoperative predictions of postoperative prognoses.
IP T
Conflicts of interest.
SC R
The authors declare no conflicts of interest.
N
U
Acknowledgements
A
This research was supported by Basic Science Research Program through the
M
National Research Foundation of Korea (NRF) funded by the Ministry of Education (2017R1D1A1B04032467). We would like to thank the Medical Research Collaboration
ED
Center of Seoul National University Hospital for providing statistics consultation and statistical support, and Sang Min Lee, MD, Youkyung Lee, MD, and Mi-Jin Kang, MD for
A
CC E
PT
their expert support and contributions to the preparation and review of CT image data.
References
[1]
R. Rami-Porta, V. Bolejack, P. Goldstraw, The new tumor, node, and metastasis staging system, Seminars in respiratory and critical care medicine 32 (2011) 44-51.
[2]
A. Warth, T. Muley, M. Meister, et al., The novel histologic International Association for the Study
of
Lung
Cancer/American
Thoracic
Society/European
Respiratory
Society
IP T
classification system of lung adenocarcinoma is a stage-independent predictor of survival, Journal of clinical oncology : official journal of the American Society of Clinical Oncology 30 (2012) 1438-1446.
K. Tsuta, M. Kawago, E. Inoue, et al., The utility of the proposed IASLC/ATS/ERS lung
SC R
[3]
adenocarcinoma subtypes for disease prognosis and correlation of driver gene alterations, Lung cancer (Amsterdam, Netherlands) 81 (2013) 371-376. [4]
A. Yoshizawa, S. Sumiyoshi, M. Sonobe, et al., Validation of the IASLC/ATS/ERS lung adenocarcinoma classification for prognosis and association with EGFR and KRAS gene
U
mutations: analysis of 440 Japanese patients, Journal of thoracic oncology : official [5]
N
publication of the International Association for the Study of Lung Cancer 8 (2013) 52-61. J. Gu, C. Lu, J. Guo, et al., Prognostic significance of the IASLC/ATS/ERS classification in
A
Chinese patients-A single institution retrospective study of 292 lung adenocarcinoma, [6]
M
Journal of surgical oncology 107 (2013) 474-480.
H.Y. Lee, J.Y. Jeong, K.S. Lee, et al., Histopathology of lung adenocarcinoma based on new IASLC/ATS/ERS classification: prognostic stratification with functional and metabolic
[7]
ED
imaging biomarkers, Journal of magnetic resonance imaging : JMRI 38 (2013) 905-913. A. Mansuet-Lupo, A. Bobbio, H. Blons, et al., The new histologic classification of lung primary adenocarcinoma subtypes is a reliable prognostic marker and identifies tumors 643.
Y.T. Kim, Y.W. Seong, Y.J. Jung, et al., The presence of mutations in epidermal growth factor
CC E
[8]
PT
with different mutation status: the experience of a French cohort, Chest 146 (2014) 633-
receptor gene is not a prognostic factor for long-term outcome after surgical resection of non-small-cell lung cancer, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 8 (2013) 171-178.
A
[9]
[10]
M.W. Lin, C.T. Wu, J.Y. Shih, et al., Clinicopathologic characteristics and prognostic significance of EGFR and p53 mutations in surgically resected lung adenocarcinomas
[11]
T.Y. Chou, C.H. Chiu, L.H. Li, 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, Clinical cancer research : an official journal of the American Association for Cancer Research 11 (2005) 3750-3757. [12]
M. Sonobe, M. Kobayashi, M. Ishikawa, et al., Impact of KRAS and EGFR gene mutations on recurrence and survival in patients with surgically resected lung adenocarcinomas, Annals of surgical oncology 19 Suppl 3 (2012) S347-354.
[13]
J. Cadranel, A. Mauguen, M. Faller, et al., Impact of systematic EGFR and KRAS mutation
IP T
evaluation on progression-free survival and overall survival in patients with advanced nonsmall-cell lung cancer treated by erlotinib in a French prospective cohort (ERMETIC
project--part 2), Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 7 (2012) 1490-1502.
W.S. Liu, L.J. Zhao, Q.S. Pang, et al., Prognostic value of epidermal growth factor receptor
SC R
[14]
mutations in resected lung adenocarcinomas, Medical oncology (Northwood, London, England) 31 (2014) 771. [15]
V.N. Maturu, N. Singh, A. Bal, et al., Relationship of epidermal growth factor receptor
U
activating mutations with histologic subtyping according to International Association for
N
the Study of Lung Cancer/American Thoracic Society/European Respiratory Society 2011 adenocarcinoma classification and their impact on overall survival, Lung India : official V.H. Lee, V.P. Tin, T.S. Choy, et al., Association of exon 19 and 21 EGFR mutation patterns
M
[16]
A
organ of Indian Chest Society 33 (2016) 257-266.
with treatment outcome after first-line tyrosine kinase inhibitor in metastatic non-smallcell lung cancer, Journal of thoracic oncology : official publication of the International [17]
ED
Association for the Study of Lung Cancer 8 (2013) 1148-1155. H. Hu, Y. Pan, Y. Li, et al., Oncogenic mutations are associated with histological subtypes but do not have an independent prognostic value in lung adenocarcinoma, OncoTargets [18]
PT
and therapy 7 (2014) 1423-1437.
T. Ohba, G. Toyokawa, T. Kometani, et al., Mutations of the EGFR and K-ras genes in resected stage I lung adenocarcinoma and their clinical significance, Surgery today 44
CC E
(2014) 478-486.
[19]
Y. Kudo, J. Matsubayashi, H. Saji, et al., Association between high-resolution computed tomography findings and the IASLC/ATS/ERS classification of small lung adenocarcinomas in Japanese patients, Lung cancer (Amsterdam, Netherlands) 90 (2015) 47-54.
A
[20]
E.J. Hwang, C.M. Park, Y. Ryu, et al., Pulmonary adenocarcinomas appearing as part-solid ground-glass nodules: is measuring solid component size a better prognostic indicator?, European radiology 25 (2015) 558-567.
[21]
Y. Tsutani, Y. Miyata, T. Yamanaka, et al., Solid tumors versus mixed tumors with a groundglass opacity component in patients with clinical stage IA lung adenocarcinoma: prognostic comparison using high-resolution computed tomography findings, The Journal of thoracic and cardiovascular surgery 146 (2013) 17-23.
[22]
H. Uehara, Y. Tsutani, S. Okumura, et al., Prognostic role of positron emission tomography and high-resolution computed tomography in clinical stage IA lung adenocarcinoma, The Annals of thoracic surgery 96 (2013) 1958-1965.
[23]
S.H. Song, J.H. Ahn, H.Y. Lee, et al., Prognostic impact of nomogram based on whole tumour size, tumour disappearance ratio on CT and SUVmax on PET in lung adenocarcinoma, European radiology 26 (2016) 1538-1546.
[24]
A. Yoshizawa, N. Motoi, G.J. Riely, et al., Impact of proposed IASLC/ATS/ERS classification of
IP T
lung adenocarcinoma: prognostic subgroups and implications for further revision of staging based on analysis of 514 stage I cases, Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc 24 (2011) 653-664. [25]
N. Yanagawa, S. Shiono, M. Abiko, et al., The correlation of the International Association
SC R
for the Study of Lung Cancer (IASLC)/American Thoracic Society (ATS)/European Respiratory Society (ERS) classification with prognosis and EGFR mutation in lung adenocarcinoma, The Annals of thoracic surgery 98 (2014) 453-458. [26]
E.A. Park, H.J. Lee, Y.T. Kim, et al., EGFR gene copy number in adenocarcinoma of the lung
U
by FISH analysis: investigation of significantly related factors on CT, FDG-PET, and W.D. Travis, E. Brambilla, M. Noguchi, et al., International association for the study of lung cancer/american
thoracic
society/european
A
[27]
N
histopathology, Lung cancer (Amsterdam, Netherlands) 64 (2009) 179-186. respiratory
society
international
M
multidisciplinary classification of lung adenocarcinoma, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 6 (2011) 244-285.
G. Sica, A. Yoshizawa, C.S. Sima, et al., A grading system of lung adenocarcinomas based
ED
[28]
on histologic pattern is predictive of disease recurrence in stage I tumors, The American journal of surgical pathology 34 (2010) 1155-1162. H.Y. Lee,
M.J. Cha, K.S. Lee, et
PT
[29]
Adenocarcinomas
of
the
Lung:
al., Prognosis
Related
Factors
in Resected Invasive Mucinous and
Comparison
with
Resected
Nonmucinous Adenocarcinomas, Journal of thoracic oncology : official publication of the
CC E
International Association for the Study of Lung Cancer 11 (2016) 1064-1073.
[30]
P. Goldstraw, J. Crowley, K. Chansky, et al., The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh)
A
edition of the TNM Classification of malignant tumours, Journal of thoracic oncology :
[31]
official publication of the International Association for the Study of Lung Cancer 2 (2007) 706-714. T.J. Lynch, D.W. Bell, R. Sordella, et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib, N Engl J Med 350 (2004) 2129-2139.
[32]
H.G. Yi, H.J. Kim, Y.J. Kim, et al., Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) are effective for leptomeningeal metastasis from non-small cell lung
cancer patients with sensitive EGFR mutation or other predictive factors of good response for EGFR TKI, Lung cancer (Amsterdam, Netherlands) 65 (2009) 80-84. [33]
D. Firth, Bias Reduction of Maximum Likelihood Estimates, Biometrika 80 (1993) 27-38.
[34]
F.E. Harrell, Jr., K.L. Lee, D.B. Mark, Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors, Statistics in medicine 15 (1996) 361-387.
[35]
M.J. Cha, H.Y. Lee, K.S. Lee, et al., Micropapillary and solid subtypes of invasive lung thoracic and cardiovascular surgery 147 (2014) 921-928.e922.
[36]
IP T
adenocarcinoma: clinical predictors of histopathology and outcome, The Journal of H.J. Lee, Y.T. Kim, C.H. Kang, et al., Epidermal growth factor receptor mutation in lung
adenocarcinomas: relationship with CT characteristics and histologic subtypes, Radiology [37]
SC R
268 (2013) 254-264.
J.P. Ko, J. Suh, O. Ibidapo, et al., Lung Adenocarcinoma: Correlation of Quantitative CT Findings with Pathologic Findings, Radiology 280 (2016) 931-939.
[38]
Y. Yang, Y. Yang, X. Zhou, et al., EGFR L858R mutation is associated with lung
U
adenocarcinoma patients with dominant ground-glass opacity, Lung cancer (Amsterdam, [39]
N
Netherlands) 87 (2015) 272-277.
P.A. Russell, Z. Wainer, G.M. Wright, et al., Does lung adenocarcinoma subtype predict
A
patient survival?: A clinicopathologic study based on the new International Association for
M
the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary lung adenocarcinoma classification, Journal of thoracic oncology : official publication of the International Association for the Study of Lung [40]
ED
Cancer 6 (2011) 1496-1504.
H. Nakamura, H. Saji, T. Shinmyo, et al., Close association of IASLC/ATS/ERS lung adenocarcinoma subtypes with glucose-uptake in positron emission tomography, Lung
[41]
PT
cancer (Amsterdam, Netherlands) 87 (2015) 28-33. M. Hasegawa, F. Sakai, R. Ishikawa, et al., CT Features of Epidermal Growth Factor Receptor-Mutated
Adenocarcinoma
of
the
Lung:
Comparison
with
Nonmutated
CC E
Adenocarcinoma, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 11 (2016) 819-826.
[42]
S.J. Hong, T.J. Kim, Y.W. Choi, et al., Radiogenomic correlation in lung adenocarcinoma with
A
epidermal growth factor receptor mutations: Imaging features and histological subtypes,
[43]
European radiology 26 (2016) 3660-3668. O. Gevaert, S. Echegaray, A. Khuong, et al., Predictive radiogenomics modeling of EGFR mutation status in lung cancer, Scientific reports 7 (2017) 41674.
[44]
Y. Liu, J. Kim, F. Qu, et al., CT Features Associated with Epidermal Growth Factor Receptor Mutation Status in Patients with Lung Adenocarcinoma, Radiology 280 (2016) 271-280.
[45]
T.J. Kim, C.T. Lee, S.H. Jheon, et al., Radiologic Characteristics of Surgically Resected NonSmall Cell Lung Cancer With ALK Rearrangement or EGFR Mutations, The Annals of
thoracic surgery 101 (2016) 473-480. [46]
K. Usuda, M. Sagawa, N. Motono, et al., Relationships between EGFR mutation status of lung cancer and preoperative factors - are they predictive?, Asian Pacific journal of cancer prevention : APJCP 15 (2014) 657-662.
[47]
C. Villa, P.T. Cagle, M. Johnson, et al., Correlation of EGFR mutation status with predominant histologic subtype of adenocarcinoma according to the new lung adenocarcinoma classification of the International Association for the Study of Lung
IP T
Cancer/American Thoracic Society/European Respiratory Society, Archives of pathology & laboratory medicine 138 (2014) 1353-1357. [48]
H. Nakamura, H. Saji, T. Shinmyo, et al., Association of IASLC/ATS/ERS Histologic Subtypes
A
CC E
PT
ED
M
A
N
U
Resected Cases, Clinical lung cancer 16 (2015) 209-215.
SC R
of Lung Adenocarcinoma With Epidermal Growth Factor Receptor Mutations in 320
Figure Captions Figure 1. Kaplan-Meier survival curves for recurrence-free survival according to histologic
A
CC E
PT
ED
M
A
N
U
SC R
IP T
grading system.
Tables
SC R U N
A
CC E
PT
ED
M
A
n (%) Clinical variables Age (years)* 64 (56-70) Male sex 463 (46.9) Smoking history Never smoker 596 (60.3) Former smoker 390 (39.5) Current smoker 2 (0.2) Pathologic stage 1 693 (70.1) 2 125 (12.7) 3 148 (15) 4 22 (2.2) Size (cm)* 2.5 (1.7-3.2) Location Upper 499 (50.5) Middle 82 (8.3) Lower 341 (34.5) More than one lobe 66 (6.7) Type of surgery Lobectomy 900 (91.1) Segmentectomy 28 (2.8) Wedge resection 53 (5.4) Pneumonectomy 7 (0.7) CT characteristics Pure ground-glass nodule 27 (2.7) GGO-dominant PSN 180 (18.2) Solid-dominant PSN 372 (37.7) Solid nodule 409 (41.4) Missing information 1 (0.1) IASLC/ATS/ERS classification Lepidic predominant 138 (14) Acinar predominant 524 (53) Papillary predominant 137 (13.9) Micropapillary predominant 24 (2.4) Solid predominant 121 (12.2) Invasive mucinous adenocarcinoma 44 (4.4) EGFR mutation Wild type 400 (40.5) Mutation 553 (56) Gene 19 deletion 284 (51.4) Gene 21 missense 218 (39.45) Miscellaneous 51 (9.2) Missing information 35 (3.5) * Data are presented with median (25th to 75th percentile).
IP T
Table 1. Clinical characteristics of the study population (n=988).
CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society; EGFR, epidermal growth factor receptor
Table 2. Distribution of CT characteristics according to the histologic subtype. GGOdominant PSN
Soliddominant PSN
Solid nodule
20 (14.5) 6 (1.1) 1 (0.7) 0 (0) 0 (0) 0 (0)
55 (39.9) 91 (17.4) 25 (18.2) 0 (0) 4 (3.3) 5 (11.4)
48 (34.8) 231 (44.1) 45 (32.8) 6 (25) 26 (21.5) 16 (36.4)
15 (10.9) 196 (37.4) 66 (48.2) 18 (75) 91 (75.2) 23 (52.3)
20 (14.5) 7 (1.1) 0 (0) 0 (0)
55 (39.9) 116 (17.5) 4 (2.8) 5 (11.4)
48 (34.8) 276 (41.8) 32 (22.1) 16 (36.4)
P value
IP T
<0.0001
<0.0001
15 (10.9) 262 (39.6) 109 (75.2) 23 (52.3)
SC R
IASLC/ATS/ERS classification Lepidic predominant Acinar predominant Papillary predominant Micropapillary predominant Solid predominant Invasive mucinous adenocarcinoma Grading Grade 1 (lepidic predominant) Grade 2 (acinar or papillary predominant) Grade 3 (micropapillary or solid predominant) Grade 4 (invasive mucinous adenocarcinoma) .
Pure groundglass nodule
U
CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International
N
Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory
A
CC E
PT
ED
M
A
Society
Table 3. Comparison of clinical, histologic, radiologic, and genetic information according to the presence of EGFR mutations. EGFR mutation Absent Present (n=400) (n=553)
<0.0001
0.1684 <0.0001
385 (69.7) 166 (30.3) 0.1439
SC R
401 (72.4) 61 (11.2) 78 (14.1) 13 (2.3)
IP T
220 (39.9) 333 (60.1) 63 (62-65)
<0.0001
U
19 (3.4) 120 (21.7) 226 (40.9) 188 (34)
<0.0001
97 (17.5) 398 (72) 54 (9.7) 4 (0.7)
M
A
N
Sex Male 231 (57.7) Female 169 (42.2) Age 65 (64-66) Smoking Never 193 (48.3) Ever 207 (51.7) Staging 1 264 (66) 2 60 (15) 3 68 (17) 4 8 (2) CT characteristics (n=953) Pure ground-glass nodule 7 (1.8) GGO-dominant PSN 53 (13.3) Solid-dominant PSN 129 (32.2) Solid nodule 211 (52.7) IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) 40 (10) Grade 2 (acinar or papillary predominant) 237 (59.2) Grade 3 (micropapillary or solid predominant) 84 (21) Grade 4 (invasive mucinous adenocarcinoma) 39 (9.8) Data are numbers, with percentages in parentheses.
P value
A
CC E
PT
ED
EGFR, epidermal growth factor receptor; CT, computed tomography; GGO, ground-glass opacity; PSN, partsolid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society
Table 4. Univariate Cox hazard regression models for predictions of recurrence-free survival. Recurrence-free survival (n=968) HR 95% CI P value 0.764 0.595-0.980 0.0342 1 0.987-1.013 0.9825 1.415 1.328-1.507 <0.0001 1.223 0.951-1.573 0.1175 <0.0001 1 N/A N/A 3.193 2.271-4.49 <0.0001 6.521 4.898-8.68 <0.0001 3.234-13.655 .<0.000 6.645 1 0.828 0.642-1.068 0.1465
IP T
Sex (female) Age (per 1 year increase) Size (per mm increase) Smoking (smoker) Stage 1 (reference) 2 3 4
A
CC E
PT
ED
M
A
N
U
SC R
EGFR mutation (n=954) CT characteristics Pure ground-glass nodule (reference) 1 N/A N/A GGO-dominant PSN 5.569 0.76-709.919 0.1078 Solid-dominant PSN 13.661 1.989-1722.327 0.002 Solid nodule 25.295 3.712-3184.733 <0.0001 <0.0001 IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) 1 N/A N/A Grade 2 (acinar or papillary predominant) 3.397 1.89-6.103 <0.0001 Grade 3 (micropapillary or solid predominant) 6.432 3.456-11.97 <0.0001 Grade 4 (invasive mucinous adenocarcinoma) 2.634 1.11-6.252 0.0281 HR, hazard ratio; CI, confidence interval; N/A, non-applicable; EGFR, epidermal growth factor receptor; CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society
Table 5. Multivariate Cox hazard models for predictions of recurrence-free survival. Recurrence-free survival (n=968) HR 95% CI P value Model 1 (size + stage) Size Stage 1 (reference) 2 3 4 Model 2 (size + stage + CT characteristics) Size Stage 1 (reference) 2 3 4 CT characteristics Pure ground-glass nodule (reference) GGO-dominant PSN
1.212
1.128-1.302
<0.0001
1 2.261 5.099 5.437
N/A 1.561-3.275 3.745-6.942 2.63-11.24
N/A <.0001 <.0001 <.0001
1.21
1.124-1.299
<0.0001
1 1.891 4.056 4.311
N/A 1.296-2.728 2.948-5.567 1.956-8.31
N/A 0.0011 <.0001 0.0008
0.626586.125 1.047-915.721 1.612-1402.475
0.1679
1.215
0.0428 0.0067
1.13-1.306
<0.0001
1 2.026 4.364 5.162
N/A 1.389-2.956 3.171-6.005 2.478-10.751
N/A 0.0002 <.0001 <.0001
1 2.283 2.782 1.604
N/A 1.262-4.132 1.465-5.281 0.665-3.869
N/A 0.0064 0.0018 0.2931
A
M
ED
SC R
N
7.254 11.119
U
1 4.598
IP T
0.763 (0.733-0.792)
Solid-dominant PSN Solid nodule Model 3 (size + stage + IASLC/ATS/ERS classification) Size Stage 1 (reference) 2 3 4 IASLC/ATS/ERS classification Grade 1 (lepidic predominant, reference) Grade 2 (acinar or papillary predominant) Grade 3 (micropapillary or solid predominant) Grade 4 (invasive mucinous adenocarcinoma)
PT
c-index (95% CI) 0.759 (0.729-0.789)
0.765 (0.738-0.795)
A
CC E
HR, hazard ratio; CI, confidence interval; N/A, non-applicable; CT, computed tomography; GGO, ground-glass opacity; PSN, part-solid nodule; IASLC, International Association for the Study of Lung Cancer; ATS, the American Thoracic Society; ERS, the European Respiratory Society