Former smokers with non‐small‐cell lung cancers: a comprehensive investigation of clinicopathologic characteristics, oncogenic drivers, and prognosis

Abstract The aim of this present investigation was to evaluate the clinicopathologic characteristics, oncogenic drivers, and prognosis of former smokers with non‐small‐cell lung cancer (NSCLC), and to compare them with those of the current and never smokers. This investigation was a single‐institution retrospective study of 2289 NSCLC patients, who were classified as former, current, or never smokers. A collection was made of the clinicopathological characteristics, spectra of well‐identified driver genes and survival rates. The survival rates were compared using log‐rank test, and independent prognostic factors, identified using Cox regression analysis. Of 2289 NSCLC patients, 257 (11.2%) were former smokers; 868 (37.9%), current smokers; and 1164 (50.9%), never smokers. Compared with the current, the former were characterized by older age at diagnosis (64.3y vs. 59.9y; P < 0.001), earlier TNM stage (stage I, 47.9% vs. 39.5%; P = 0.017), fewer solid predominance in adenocarcinomas (16.2% vs. 29.5%; P = 0.005), and more EGFR mutation (33.2% vs. 20.7%; P < 0.001) but less KRAS mutation (6.7% vs. 11.9%, P = 0.041). No statistically significant survival differences were observed between the former and current. However, the light former smokers presented favorable overall survival when compared with the light current and heavy former or current (the light former vs. the heavy former, P = 0.028; the light former vs. the light current, P = 0.048; and the light former vs. the heavy current, P = 0.048). Our findings suggest that the former smokers with NSCLCs can have distinctive clinicopathologic characteristics, oncogenic drivers, and prognosis, and they, especially the light former, can benefit from smoking cessation.


Introduction
Lung cancer is the leading cause of cancer-related death worldwide, with more than one million people dying from it each year [1]. Decades of epidemiologic studies have established cigarette smoking as the most important risk factor for the development of lung cancer [2]. These studies indicated that 85-90% of

ORIGINAL RESEARCH
Former smokers with non-small-cell lung cancers: a comprehensive investigation of clinicopathologic characteristics, oncogenic drivers, and prognosis lung cancer cases are associated with cigarette smoking [3].
The association of smoking with lung cancer and subsequent death is irrefutable. Tobacco smoke contains over 60 known carcinogens [4]. The relative risk of lung cancer development is estimated to be 10-to 30-fold in longterm smokers when compared with never smokers [5]. For lung cancer patients, smoking history is associated with diagnosis at a later stage, increased operative mortality, higher rate of local recurrence, and worse long-term survival [6,7]. Furthermore, those who quit smoking develop fewer radiation pneumonitis and infections during radiotherapy, and better response to chemotherapy and targeted therapy [8,9]. Smoking cessation is associated with improved pulmonary function, weight gain, and better quality of life [10].
Although previous studies have already demonstrated the benefits of smoking cessation for lung cancer patients, the reports are limited on associations between detailed smoking statuses and clinical outcomes in former smokers with non-small-cell lung cancer (NSCLC). Moreover, few studies have described the clinical and pathologic features and well-identified driver mutations in the former with NSCLC in comparison with the current and never with NSCLC. In this study, we undertook a comprehensive analysis to further investigate the clinicopathologic features, oncogenic drivers of former smokers and associations between detailed smoking statuses and clinical outcomes.

Patients
We consecutively enrolled the patients with newly diagnosed non-small-cell lung cancer between October 2007 and May 2013 at the Department of Thoracic Surgery of Fudan University Shanghai Cancer Center, Shanghai, China. The history of cigarette smoking were obtained at the patient interview by professional doctors, and the patients were categorized into never smokers, former smokers, and current smokers according to their smoking statuses. The never were defined as those who had smoked less than 100 cigarettes; of the smokers, the former, as those who had quitted smoking 1 year ago or more before diagnosis; and the current, as those who continued their smoking habit at diagnosis or had quitted smoking less than 1 year ago. Pack-years of smoking were calculated by multiplying the number of packs (one pack containing 20 cigarettes) smoked per day by the number of smoking years. This study was approved by the Institutional Review Board of Fudan University Shanghai Cancer Center, with written informed consent obtained from all patients.
Disease recurrence and survival were observed in the follow-up clinic or obtained by telephone.

Clinical and pathological variables
The clinical and pathological data collected for analyses covered sex, age at diagnosis, detailed information regarding cigarette smoking such as daily cigarette use, duration of smoking history and quit date if applicable, pathological TNM stage, tumor differentiation, and histological types according to the new IASLC/ATS/ERS multidisciplinary classification [11]. Pathological TNM stages were evaluated in accordance with the seventh edition of the lung cancer staging classification system [12].

Statistical analysis
Difference in proportions was analyzed using Pearson's chi-squared test, when no cell of a contingency table had an expected count less than five, or Fisher's exact test, when any cell of a contingency table had an expected count less than five. Relapse-free survival (RFS) and overall survival (OS) of patients with smoking statuses was estimated through the Kaplan-Meier method. The survival differences between groups were determined via the logrank test. Independent prognostic factors were identified through the Cox proportional hazards regression (forward likelihood ratio model). All tests were two tailed, and statistical significance was set as P < 0.05. All data were
Differences in clinicopathological features and driver mutations between former, never, and current smokers A collection was made of the clinicopathological features of the former smokers, which were compared with those of the current and never. The former were diagnosed at an older age than the current and never (64.3year vs. 59.9year, P < 0.001; 64.3year vs. 59.1year, P < 0.001, respectively). When compared with the current, fewer male patients were defined as the former (96.1% as former vs. 98.3% as current; P = 0.039). The former produced an earlier TNM stage than the current (stage I: 47.9% as former vs. 39.5% as current; P = 0.017). The former were more likely to develop squamous cell carcinoma than the never, as in the same case of the current.
An analysis was made of the associations between smoking states and histologic subtypes in adenocarcinoma. Fewer solid predominant adenocarcinomas were detected in the former than in the current (16.2% as former vs. 29.5% as current, P = 0.005; 16.2% as former vs. 11.1% as never, P < 0.104). No differences were observed in other histologic subtypes between the former and the never or the current.
Mutational analyses were performed in approximately 1800 patients (Fig. 1). Of the former, 33.2% had EGFR kinase domain mutations (including 32 exon 19 deletions, 31 L858R, and two other mutations); 6.7%, KRAS (including two G12A, three G12C, one G12D, three G12V, one L19F, one R41M, and two Q61H); 0.5%, HER2 (one exon 20 insertion mutation); and 0.5%, BRAF mutations (one L485S). ALK rearrangement was detected in the patients by 2.6% (five EML4-ALK fusions), and FGFR fusion, in those by 2.1% (four FGFR3-TACC3 fusions), whereas RET and ROS1 fusion were not found in the former. Interestingly for EGFR and KRAS, the most common mutations in NSCLC, the mutation profile of the former fell just between the never and the current (EGFR: 33.2% as former vs. 20.7% as current, P < 0.001; 33.2% as former vs. 67.9% as never, P < 0.001; KRAS: 6.7% as former vs. 11.9% as current, P = 0.041; 6.7% as former vs. 3.3% as never, P = 0.024). As in the same case of the current, the former carried fewer HER2 mutations, but more FGFR fusions than the never. As to BRAF, ALK, RET, and ROS1, no significant differences were found between the former and the current or the never.

Univariate analysis of the clinicopathologic variables on survival outcomes
The median follow-up period was 29.5 months (ranging 1-103 months). The log-rank test on the Kaplan-Meier survival analysis demonstrated that although there were differences between the former and current in terms of the median RFS and OS, they were not statistically significant ( Fig. 2A and B, RFS; median survival: 40 vs. 38 months; log rank: P = 0.348; OS: median survival not reached, log rank: P = 0.168). Additionally, the never were found to have significantly longer RFS and OS than the current ( Fig. 2A  To determine what patients could benefit more from smoking cessation, we focused on detailed smoking information of the former, intending to identify the risk factors for recurrence and mortality. According to the age of smoking cessation before or after 57 as the median, the years of smoking cessation lasting longer or shorter than 5 years as the median duration, and the smoking dosage over or less than 40 pack-years as the median dosage, the former were divided into different subgroups. The univariate analysis identified that the more advanced initial stage was associated with recurrence, and that smoking dosage over 40 pack-years, advanced initial stage, EGFR wildtype and FGFR fusion were associated with mortality ( Table 2).

Multivariate analysis of the clinicopathologic variables on survival outcomes in former smokers
As indicated by the Cox proportional hazards models (Table 3), the multivariate analysis was adjusted for age, initial stage, age of smoking cessation, years of smoking cessation, smoking dosage, EGFR, and FGFR. It was found that the late initial stage was a significant and independent risk factor for relapse, while such a stage and EGFR wild-type status were significant and independent risk factors for worse overall survival.

Survival outcomes of cigarette smoking status and dosage
As smoking dosage was an important prognosis factor in the former, we further analyzed the association between smoking dosage and outcomes in all patients, who were divided into three groups: never smokers (0 pack-years), light smokers (≤40 pack-years), and heavy smokers (>40 pack-years). It was found that the never presented more significant RFS and OS than the light and heavy (Fig. 2C & D; RFS: never vs. light, P < 0.001; never vs. heavy, P < 0.001; OS: never vs. light, P < 0.001; never vs. heavy, P < 0.001). However, the light did not show significant differences when compared with the heavy (RFS: light vs. heavy, P = 0.991; OS: light vs. heavy, P = 0.229).
Combining the two important groups of smoking information: smoking status and smoking dosage, we found out that the light former had favorable overall survival than the heavy former, the light current, and the heavy current ( Fig. 2F; light former vs. heavy former, P = 0.028; light former vs. light current, P = 0.048; light former vs. heavy current, P = 0.048).

Discussion
Over the past decades, studies have proved that patients with lung cancer can benefit from smoking cessation; however, there is a death of literature on the clinicopathological characteristics of the former smokers with NSCLC. Previous studies have identified a number of molecularly distinct subsets of lung cancer characterized by different oncogenes [17][18][19]. Researchers have found tumor mutational frequencies and spectra differences between smokers and nonsmokers, as indicated by the reported differences in the frequency of somatic mutations of EGFR and KRAS observed between smoking and nonsmoking lung cancer patients [20]. However, the mutational spectra of the former smokers with NSCLC have not been reported.
To our knowledge, this study pioneered the comprehensive analysis of well-identified driver mutations, clinical characteristics, and survival analysis in an Asian cohort of non-small-cell lung cancer patients who had quitted smoking. While previous studies have examined the characteristics and outcomes between never and ever smokers, our investigations have focused on NSCLC in former smokers. In this study, we demonstrated that the former  smokers were older at diagnosis, showing an earlier TNM stage and harboring more EGFR but less KRAS mutations than the current. Additionally, we proved that the light former presented favorable overall survival when compared with the light current and the heavy former or the current.
The former with NSCLC were found to benefit from cessation according to the analyses of clinicopathological characteristics, as indicated by the mean age of four years older than that of the current, suggesting a longer lifetime. The findings were in agreement with the data from the NCCN network that former smokers had older age than current smokers [21]. When the former had an earlier TNM stage than the current, this was an independent prognosis factor for RFS and OS. In adenocarcinomas, former smokers have been reported to be associated with a lower proportion of solid predominant histology subtype, which is indicative of more aggressive biological features and predicts worse outcomes [22][23][24]. These clinicopathological characteristics of the former may partly explain their clinical outcomes.
We now understand that NSCLCs are subdivided into different molecular subtypes according to the primary driver genes identified [25]. Driver mutations in smoking NSCLCs are different from those in nonsmoking NSCLCs [26,27]. We performed a comprehensive analysis of wellidentified driver genes in the former smokers with NSCLCs, and compared the results with those of the current and never with NSCLCs, demonstrating that the former harbored more EGFR but less KRAS mutations than the current. It has been reported that NSCLCs with EGFR mutations are associated with better outcomes, whereas KRAS mutations, with worse outcome [28,29]. Moreover, lung cancer of EGFR mutation can be treated with EGFR-TKI and prolong PFS overall than treated with chemotherapy, especially in those with exon 19 deletions, never smokers and women [30]. Our results supported the conclusion that former smokers with NSCLCs may benefit from their specific mutational profiles.
A number of studies have investigated the relationship between smoking and outcome of lung cancer patients, but the results are diverse [7,20,21,[31][32][33][34]. The present results clearly supported the conclusion that smoking history can exert a negative influence on RFS and OS. Furthermore, the never had favorable RFS and OS rates when compared with those of the current. As to the former, those who had smoked no more than 40 pack-years showed more significant OS than those who had continued smoking with the same dosage, although smoking dosage is not an independent prognostic factor. Such results emphasized the importance of smoking cessation, even for light smokers.
The present retrospective study had several limitations in the analyses. Smoking extent was reported by the patients themselves without biochemical confirmation; therefore, this could have bias. Additionally, second-hand tobacco smoke, which has been established as an inducement of lung cancer, was too difficult to quantify exactly and include it in the analyses. Nonetheless, the large number of the patients enrolled for investigation may reduce error in different groups.
In conclusion, NSCLC patients who underwent surgical resection with smoking history can benefit from smoking cessation; former smokers with NSCLCs can age older at diagnosis, presenting an earlier TNM stage and more EGFR but less KRAS mutations than current ones with NSCLCs; and light former ones can have better long-term overall survival than heavy former ones as well as light current ones.