Lung cancer in never smokers

INTRODUCTION — Lung cancer is the leading cause of cancer death worldwide, causing about 1.8 million deaths per year [1,2]. Exposure to tobacco smoke is the primary etiologic factor responsible for lung cancer, and its importance is illustrated by the decline in lung cancer incidence and mortality in the United States that has accompanied the decline in smoking [3-5]. (See "Cigarette smoking and other possible risk factors for lung cancer".)

Despite the predominance of tobacco smoking as its presumed etiology, lung cancer is also a significant health problem in those with no history of smoking [6,7]. Lung cancer in people who have never smoked is distinct enough from an epidemiologic and biologic standpoint to be considered a separate entity [8-11]. Furthermore, as the number of people who have never smoked in the United States and other countries rise, the issue of lung cancer in this group is a growing public health concern.

The epidemiology, risk factors, biologic differences, and potential implications for treatment of lung cancer in people who have never smoked are reviewed here. General aspects of lung cancer biology and management are discussed elsewhere. (See "Clinical manifestations of lung cancer" and "Overview of the initial treatment and prognosis of lung cancer" and "Overview of the initial treatment of advanced non-small cell lung cancer".)

EPIDEMIOLOGY — In general, the term "never smoker" refers to individuals who have smoked fewer than 100 cigarettes in their lifetime. The incidence of small cell lung cancer in never smokers is exceedingly small [12], and this discussion will therefore focus on non-small cell lung cancer (NSCLC), predominantly adenocarcinoma (while 50 to 60 percent of cancers in never smokers are adenocarcinomas, only approximately 10 to 20 percent are squamous cell carcinomas [13]).

Firm data on the incidence of lung cancer in never smokers is difficult to ascertain because most population-based cancer registries, including the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database, do not collect patient smoking information. SEER data have been linked with population-based tobacco use information [14], but this only allows for estimates of lung cancer in never smokers within broad geographic areas.

Incidence — Worldwide, smoking accounts for approximately two-thirds of lung cancer cases, with the remaining occurring due to other causes [1]. There are major geographic and gender differences, particularly in Asia, where 60 to 80 percent of women with the disease have never smoked [6].

In one analysis in the United States, an estimated 19 percent of women with lung cancer had never smoked, compared with approximately 9 percent of men with the disease [15]. These results are based upon an analysis of data from patients in five large United States cohort studies [16-21]. The age-adjusted incidence rate for lung cancer in never smokers aged 40 to 79 years ranged from 11.2 to 13.7 per 100,000 person-years for males and from 15.2 to 20.8 per 100,000 person-years for females. These incidence rates are similar to those for myeloma in men or cervical cancer in women in this country [22]. By contrast, the age-adjusted rates of lung cancer in current smokers in the same cohorts were approximately 12 to 30 times higher.

An analysis of lung cancer patients diagnosed between 1991 and 2005 in Southern California utilized a novel text-mining programming algorithm to evaluate smoking status based on information from electronic medical record abstracts in the regional cancer registry. This study determined that 8.9 percent of over 25,000 patients with known smoking status were never smokers, including 6 percent of males and 14 percent of females [23,24].

Another analysis of 13 cohorts and 22 cancer registries, however, did not find any apparent difference in the incidence rate of lung cancer in people who had never smoked by sex overall, although females had a higher incidence rate than males among never smokers aged 40 to 59 years [25]. In this multinational analysis, the age-standardized incidence rates for lung cancer (per 100,000 using the 2000 world population age standard) for females were 12.4 for women of European descent, 15 for Asian women, and 19.4 for African American women. This compared with rates of 11.2, 12.9, and 12.3, respectively, for men of European, Asian, and African American background.

Whether the incidence of lung cancer in never smokers has been changing over time had been debated, with older studies yielding conflicting results:

●A study from the Swedish construction worker's health care program, which is linked to the national cancer registry, showed a substantial increase in the age-adjusted incidence rate of nonsmoking lung cancer comparing the period 1976 to 1980 with 1991 to 1995 (1.5 versus 5.4 per 100,000) [26].

●Two American Cancer Society prevention study cohorts found a statistically significant but smaller increase in the mortality rate in women due to nonsmoking-associated lung cancer comparing the period 1959 to 1972 with 1982 to 2000 (12.3 to 14.7 per 100,000) in 1982 to 2000 [27]. The rates did not change over time for men.

●Indirect evidence comes from a series of nearly 12,000 patients with lung cancer in Southern California [28]. Comparing the interval 1995 to 1999 with 1999 to 2003, the percentage of never smokers with the subtype formerly known as bronchioloalveolar carcinoma increased from 19 to 26 percent, while the percentage with other forms of lung cancer went from 8.6 to 9.4 percent.

●Data collected from cancer registries in Japan and Southern California with smoking information focused on lung cancer in never smokers over two time periods, 1991 to 1996 and 1997 to 2001, and did not show a substantial numerical change in the percent of lung cancer patients who were never smokers during the two time periods (28 and 31 percent versus 7.2 and 7.4 percent for the Japanese and Southern Californian registries at earlier and later time periods, respectively) [29].

●In a previous comparison of two case-control studies from the United Kingdom done at the same hospitals and community settings in 1950 and in 1990, the percentage of never smokers among male lung cancer cases did not change, although the percentage of never smokers among the controls increased from 4.5 to 19 percent [30]. The percentage of lung cancer in never smokers who were women dropped dramatically during this time period, while the rate of smoking in the controls dropped less dramatically. A subsequent analysis of early-stage lung cancer cases from a single institution in the United Kingdom, though, demonstrated an increase in the annual frequency of lung cancer in never smokers from 13 to 28 percent looking from 2008 to 2014 [31].

●A single institution evaluation from Japan found that over the 30-year period 1974 to 2004, the proportion of never smoking, early-stage NSCLC patients increased from approximately 16 to 33 percent [8]. The large pooled analysis of 13 cohorts compared contemporary rates with those in Connecticut in the 1930s when smoking was very rare in women and found no substantial change over time [25].

A pooled analysis of over 10,000 patients treated from 1990 to 2013 at three large academic hospitals in the southern United States identified an increase in the proportion of people who had never smoked with lung cancer from 8 percent in 1990 to 1995 to 14.9 percent in 2011 to 2013 (p <0.001) in a multivariable logistic regression analysis that accounted for sex, stage, and ethnicity [32].

A similar analysis looking at the proportion of newly diagnosed patients who were never smokers at a large institution in Singapore in 1999 to 2002 versus 2008 to 2011 found the proportion of lung cancer patients who were never smokers increased from 31 to 48 percent (p <0.001) [33].

Cancer registry data will need to collect information on the smoking status of patients in order to carefully examine trends in the incidence rate of lung cancer among never smokers. Self-reported smoking status can be challenging to obtain, but the validity of this method has been verified in several studies that have shown only small smoker misclassification rates [34,35].

Though the data are unclear, the increasing prevalence of lung cancer in patients without a smoking history is widely accepted. In Taiwan for example, 55 percent of the lung cancers develop in people who have never smoked, with 90 percent of women who develop lung cancer having no history of smoking. These observations have led to lung cancer screening programs in patients without a smoking history (TALENT trial) [36].

Sex — Despite the fact that a higher percentage of women with lung cancer have never smoked compared with men as reported in several analyses [6,15,37,38], the age-adjusted mortality rate from lung cancer among never smokers was higher in men than women in two American Cancer Society cancer prevention study cohorts [27]. These two cohorts formed the bulk of patients of European ancestry in a larger pooled analysis of 13 cohort studies and multiple cancer registries that showed age-standardized lung cancer mortality rates of 12 per 100,000 for never smoking men and 9.5 per 100,000 for never smoking women of European ancestry. In Asian populations, the rates were 26 and 16.1 per 100,000 for men and women, respectively [25]. Women with lung cancer live longer than men, but how much this explains the differences seen between incidence and mortality is not known [39]. (See "Females and lung cancer", section on 'Outcome'.)

By contrast, in a Japanese study, 33 percent of lung cancer deaths in men and 85 percent of lung cancer deaths in women were not due to cigarette smoking. In Japan, the percentage of women with lung cancer who are never smokers is much higher [40,41].

Ethnicity — The risk of smoking-associated lung cancer differs by race/ethnicity. Accurate data on differences in rates of lung cancer among never smokers based upon race or ethnicity are extremely limited.

●A Southern California analysis of over 20,000 patients found that for women, 46 percent of all lung cancer in Asian/Pacific Islanders occurred in never smokers compared with 25 percent in Hispanic individuals, 11 percent in African Americans, and 10 percent in non-Hispanic White individuals [24].

●A Northern California analysis of lung cancer cases diagnosed from 1998 to 2008 found that for women, of all lung cancer cases identified, 71 percent of lung cancer in Asian/Pacific Islanders occurred in never smokers compared with 35 percent of cases in Hispanic women and 40 percent in non-Hispanic White individuals [42]. This study also looked at survival differences in these groups.

●Further data on this topic come from a large pooled analysis of cohort and registry data that show wide variability in incidence and mortality from lung cancer in never smokers, even within certain Asian countries including China and Thailand, but overall increased rates in African Americans compared with other ethnic groups [25].

Age — Although it has been postulated that lung cancer occurs at a younger age in never smokers [43], this has only been validated in some [38], but not all, cohort analyses in Western populations [15]. In studies from Asia, however, a younger age at diagnosis is characteristic of lung cancer people who have never smoked [44].

RISK FACTORS — The known and potential causes of lung cancer are discussed in detail separately. (See "Cigarette smoking and other possible risk factors for lung cancer".)

The causative factors for lung cancer in never smokers are not well understood. The risk factors that are considered to be most important among never smokers are discussed here.

Secondhand smoke — Secondhand smoke is an important risk factor for lung cancer among never smokers [34,45-47]. However, the extent of this problem is not clear. (See "Secondhand smoke exposure: Effects in adults", section on 'Lung cancer'.)

Several studies suggest that approximately 15 to 35 percent of lung cancer among never smokers is due to secondhand smoke [48-52] and that risk may be increased in those with exposure prior to age 25 [53].

●A study from France of 1493 lung cancer patients found that 79 percent of women and 21 percent of men who were never smokers had secondhand smoke exposure [52].

●A population-based case-control study in Canada of 445 cases and 948 controls found no association between exposure to secondhand smoke at home or the workplace and development of lung cancer [38].

●A population-based prospective study in Japan reported a hazard ratio (HR) of 2.03 (95% CI 1.07-3.86) for development of adenocarcinoma for never smoking women who lived with a smoking husband as compared with never smoking women with a nonsmoking spouse [54].

●A United States case-control study reported an increase in lung cancer risk in never smokers with secondhand smoke exposure during childhood (odds ratio [OR] 2.5; 95% CI 1.04-4.90) [55].

●In a report based upon the largest sample, analysis from a case-control study in Europe of 520,000 people estimated that the proportion of lung cancer related to secondhand smoke was 16 to 24 percent in nonsmokers [51]. An additional 5 to 7 percent of lung cancer among never smokers was thought to be due to air pollution.

●Spousal smoking increased the risk of lung cancer in a never smoking spouse (relative risk [RR] 1.27, 95% CI 1.17-1.37) in a meta-analysis that included 55 international studies [56].

●In the European Prospective Investigation into Cancer and Nutrition study evaluating the risk of cancer in nearly 500,000 people, the risk of lung cancer in never smokers was significantly elevated in those with secondhand smoke exposure (HR 1.65, 95% CI 1.04-2.63) [57].

Subsequent data demonstrate the challenges in proving secondhand smoke as a truly causative factor in the development of lung cancer in never smokers.

●The Women's Health Initiative (WHI) observational study included over 76,000 women with a complete smoking history [58]. With a mean follow-up of 10.5 years, 901 cases of lung cancer were identified. Among the never smokers, there was no overall association with passive smoking exposure (HR 0.88, 95% CI 0.52-1.49), although there was a trend toward an increase in women never smokers with adult-home passive smoking exposure (HR 1.61, 95% CI 1.00-2.58).

The hypothesis that secondhand smoke may be responsible for higher rates of lung cancer in never smokers in Asia was questioned by whole-genome sequencing of lung cancer from Asian patients with and without a smoking history. The mutational signatures of the Asian never-smoker lung cancers were very similar to those of European never-smoker lung cancers but different from the patterns of smokers in either group [59].

Radon — Radon is a gaseous decay product of uranium-238 and radium-226, which is capable of damaging respiratory epithelium via the emission of alpha particles. The increased risk of lung cancer among uranium miners has been more clearly established and is thought to be due to radiation from radon [60]. (See "Cigarette smoking and other possible risk factors for lung cancer", section on 'Radon'.)

Radon is present in soil, rock, and groundwater, and it can accumulate in homes. Exposure to radon within the home may play a role in the development of lung cancer in never smokers [61-63], although this is controversial [64].

Environmental exposures — Other occupational exposures are known to increase lung cancer risk in smokers, and to a lesser extent, in never smokers as well.

●Commonly associated toxins include asbestos, chromium, and arsenic [65-68]. Occupational exposure to organic dust has been linked as a risk factor for lung cancer [69]. A Dutch analysis of occupational asbestos exposure found an RR of lung cancer of 3.5 after controlling for age, smoking, and other factors [68]. In a French study of 1493 cases, some occupational exposure was identified in 9.4 percent of women and 48.6 percent of men who developed lung cancer as never smokers [52]. In a Canadian case-control study, the OR for lung cancer risk from occupational exposures in never smokers was 2.1 (95% CI 1.3-3.3) but was higher for exposure to solvents, paints, or thinners (OR 2.8, 95% CI 1.6-5.0) [38]. A meta-analysis focused on lung cancer risk in painters reported an RR for lung cancer in all painters of 1.35 (95% CI 1.29-1.41) but 2.0 (95% CI 1.09-3.67) in the never smokers [70]. (See "Asbestos-related pleuropulmonary disease".)

●Arsenic has implicated as a cause of lung cancer where it contaminates drinking water, such as in some areas of Taiwan and Chile [71,72].

●Various dietary factors have been studied as possible causes of lung cancer but none are clearly implicated. Higher intake of fruits and vegetables may be protective against lung cancer [40,73,74]. A case-control study from Italy found an RR for lung cancer of 2.4 (95% CI 1.4-4.0) for the highest versus lowest quartiles of red meat consumption in never smokers [75], and another report indicated that higher consumption of fish may protect against lung cancer in never smokers [76]. A separate United States study focused exclusively on never smokers reported a reduced risk of lung cancer in those with the "healthiest" dietary patterns (ie, high consumption of fruits and vegetables and of low-fat foods) [77]. In one pooled analysis of 22 case-control and cohort studies including over 2500 never smoking lung cancer patients and over 9000 never smoking controls, the consumption of low-moderate amounts of alcohol was potentially protective against lung cancer [78].

●Indoor air pollutants such as vapors from cooking oil and the smoke from burning coal have been linked to lung cancer, particularly in Asia [79-82], although there is some controversy about the importance of this [83-85].

●Outdoor air pollution is also associated with lung cancer risk. Using mean long-term ambient fine particulate matter (PM2.5) as a measure of air pollution, there was a 15 to 27 percent increase in lung cancer mortality in never smokers for each 10 mcg/m³ increase in PM2.5 concentrations in the Cancer Prevention Study-II [86]. A meta-analysis including this and 17 other studies concluded that the meta-RR for lung cancer associated with PM2.5 was 1.09 (95% CI 1.04-1.14); this was particularly evident in never smokers (meta-analysis RR 1.18, 95% CI 1.00-1.39) [87]. A link between air pollution and lung cancer has been described. Increasing PM2.5 levels have been associated with increased rates of epidermal growth factor receptor (EGFR) mutated lung cancer, amongst other malignancies [88]. EGFR mutated non-small cell lung cancer (NSCLC) is exceedingly common in never smokers who develop lung cancer and very rare in those who develop lung cancer with a smoking history. PM has been shown to promote inflammation (partly through the IL1B pathway) that allows lung cells with driver mutations to fully develop into malignancy.

Lung disease — The risk of lung cancer among never smokers is increased in those with prior damage to the lungs from underlying pulmonary disease or exposure to radiation or chemotherapy [38,89-92]. The magnitude of this risk is unclear because the vast majority of pulmonary disease is related to smoking. When one looks at idiopathic pulmonary fibrosis, the literature is conflicting regarding the association with lung cancer, though an association does seem to be supported by the majority of studies [93,94]. (See "Cigarette smoking and other possible risk factors for lung cancer", section on 'Inflammation and benign lung disease'.)

Oncogenic viruses — A viral etiology for some lung cancer is under investigation, with a particular focus on the role of human papillomavirus (HPV) [95,96].

In a case-control study of 141 lung cancer patients and 60 noncancer controls from Taiwan, HPV serotypes 16 and 18 were significantly more common among those with lung cancer (55 versus 27 percent) [97]. This association was most prominent among older nonsmoking women. However, this high incidence of HPV positivity was not observed in a series of patients [98], and a subsequent meta-analysis of the literature was inconclusive about the potential role of HPV in the etiology of lung cancer in never smokers [99]. Detailed pathologic evaluation and HPV genotyping of 334 lung tumors identified HPV DNA in 10 percent, but viral oncogenes were not identified, thus failing to support a causal association [100]. In a detailed analysis of early-stage lung cancer in Japan, only 0.3 percent of patients were found to be HPV positive [101].

Genetic factors — Multiple studies have shown an association between lung cancer in never smokers and a family history of lung cancer, suggesting a role for genetic factors [38,46,102-106].

As an example, in a case-control study of 257 cases of lung cancer among never smokers and an equivalent number of control never smokers, lung cancer was significantly more common among those with a positive family history (OR 7.2) [105]. In another case-control study that included over 2400 relatives of 316 never smoker lung cancer cases, there was a 25 percent excess risk of any type of cancer among first-degree relatives of cases, particularly of early-onset lung cancer (<50 years) among those relatives who smoked [106]. In an analysis from the Environment And Genetics in Lung cancer Etiology study, the RR of lung cancer associated with a positive family history, adjusted for age, gender, residence, education, and smoking, was 1.57 (95% CI 1.25-1.98) [107]. Efforts to decipher the interactions between family history and environment exposure in the development of lung cancer in never smokers will be aided by an environmental exposure index developed to look at these questions in Chinese never smokers. There was an OR of 30.6 (95% CI 9.4-99.9) for men with both positive family history and high environmental exposure compared with those with neither, but the associations in women were not as strong with this model [108]. Suboptimal DNA repair capacity as measured by the host-cell reactivation assay nearly doubled the risk of lung cancer in a case-control study of never smokers (RR 1.92, 95% CI 1.3-2.9) [109].

Studies have suggested susceptibility loci for lung cancer in never smokers at various locations including chromosomes 6q [110], 5p15.33 (TERT) [111-113], 13q31.3 (GPC5) [114], 15q25, and 6p21 [115], and in a Korean population at 18p11.22 [116]. However, the results of genome-wide association studies (GWAS) have not been consistent [117]. Pooled data from 14 studies of ethnic Chinese patients (n = 5510 never smoking women with lung cancer and 4544 controls) reported three new susceptibility loci (10q25.2, 6q22.2, and 6p21.32) and confirmed the 5p15.33, 3q28, and 17q24.3 loci [118]. However, the studies in this analysis included some of the trials used initially to identify certain loci.

A large GWAS in never smokers from two large United States academic centers and a confirmation cohort from Taiwan identified two potential loci associated with decreased survival in never smokers, but these were different from loci identified by other groups, which might be related to lung cancer susceptibility [119]. In Korea, an analysis of 181 never smoking women with lung cancer and 179 controls identified colony-stimulating factor 1 receptor (CSF1R) as a candidate gene of interest for lung cancer susceptibility in this group [120].

Specific germline mutations have been identified in limited numbers of families with high rates of NSCLC in never smokers. The best understood are germline mutations in the EGFR [121-124]. Few other germline mutations have been reported in association with increased lung cancer risk in never smokers, but a novel germline mutation in human epidermal growth factor 2 (HER2) was identified in a Japanese family with nine affected family members [125].

Estrogens — The role of estrogens and other female hormones in lung cancer risk in women is uncertain, regardless of smoking status. Multiple studies have shown that the majority of NSCLCs express the estrogen receptor beta [126-128] and that the expression of this receptor is more common in never smokers compared with smokers [129].

The clinical implications of these laboratory studies are unclear. Multiple studies have looked at the relationship of lung cancer incidence to early menopause [130,131], use of hormone therapy [132,133], age at first birth [134], number of children [135], and use of tamoxifen [136]. However, the results have been inconsistent in establishing a relationship. Data from the randomized WHI trial indicated that women assigned to postmenopausal estrogen and progesterone had a higher risk of death from lung cancer without a higher incidence of the disease compared with women assigned to placebo [137]. However, women randomized to estrogen alone had no such increase in lung cancer mortality, nor in lung cancer incidence, compared with those on placebo [138]. Further analyses of the WHI trial indicated that the incidence and mortality of lung cancer did not differ by hormone use in never smoking women, but mortality from lung cancer was significantly higher in smoking women who used estrogen plus progesterone hormone therapy compared with smoking women who did not [139].

BIOLOGIC DIFFERENCES COMPARED WITH LUNG CANCER IN PEOPLE WHO HAVE SMOKED

Pathology — Adenocarcinoma is more common in never smokers, light smokers, and former smokers while squamous cell carcinoma and small cell lung cancer are seen with a higher incidence in heavy smokers [6,7]. In a review of published reports on the histology of lung cancer that included smoking data, adenocarcinoma was more common than squamous cell carcinoma among nonsmokers (62 versus 18 percent, based upon 5144 cases). By contrast, adenocarcinoma was less frequent among smokers (19 versus 53 percent in 21,853 cases) [6]. Subsequent series of lung cancer in never smokers continue to report adenocarcinoma as the most common histology [15,41,44].

Molecular biology — Contemporary advances in the understanding of the molecular biology of lung cancer have led to the identification of substantial differences between never smokers compared with smokers with lung cancer. These factors are arguably more important than any clinical differences. (See "Personalized, genotype-directed therapy for advanced non-small cell lung cancer".)

In one study performing whole-exome and RNA sequencing in patients with lung adenocarcinoma, there was a higher prevalence of clinically actionable driver alterations in never smokers compared with smokers (78 to 92 percent versus 50 percent) [140].

One of the most apparent differences between lung cancer in never smokers versus current and former smokers is in the expression and mutations of the epidermal growth factor receptor (EGFR). Mutations in the EGFR gene are more common in lung tumors from people who have never smoked compared with those with a smoking history [141-146]. Several analyses of patients with the most common activating mutations in the EGFR gene (predominantly deletions in exon 19 and the L858R mutation on exon 21) report that the frequency of the mutations, especially exon 19 and 21 mutations, is much higher in never smokers than ever smokers (p <0.001) in both men and women [147,148]. In addition, a distinct immunohistochemical profile of the EGFR pathway has also been identified in never versus current smokers, even in the absence of EGFR mutations [149]. (See "Systemic therapy for advanced non-small cell lung cancer with an activating mutation in the epidermal growth factor receptor".)

The contribution of secondhand exposure to tobacco smoke to the incidence of EGFR mutations in never smokers is unclear. In a Korean study of 179 never smokers with non-small cell lung cancer (NSCLC), EGFR mutations were significantly less frequent in those with a positive history of secondhand tobacco smoke (39 versus 61 percent), and there was a progressive decrease in the incidence with increasing secondhand tobacco smoke exposure [150]. This is in contrast to a Japanese publication focused on 126 never smokers with NSCLC, which reported an increase in EGFR-activating mutations with increasing secondhand smoke exposure [151]. (See 'Secondhand smoke' above.)

Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations have been thought to be more common in lung cancer patients who are ever smokers [146,152]. An analysis of 482 lung adenocarcinomas, however, found that the rate of KRAS mutations in codons 12 and 13 was not significantly different in never smokers (15 percent) compared with former smokers (22 percent) and current smokers (25 percent) [153]. The nature of the mutations was significantly different though, with more transition mutations (G to A) compared with transversion mutations (G to T or G to C) in tumors from patients with a smoking history.

A fusion gene comprising portions of the echinoderm microtubule-associated protein-like 4 (EML4) gene and the anaplastic lymphoma kinase (ALK) gene in NSCLC is found in 3 to 7 percent of NSCLC and appears to be nearly mutually exclusive with EGFR and KRAS mutations [154]. This frequency is higher in those who develop lung cancer as never smokers [154-156]. The rate of ALK fusion oncogene positivity was nearly 20 percent of NSCLC in one study from China and 42 percent in those patients in that analysis who were known to have wildtype EGFR and KRAS [157]. (See "Anaplastic lymphoma kinase (ALK) fusion oncogene positive non-small cell lung cancer".)

Differences have also been observed in the expression patterns and mutations for other genes when never smokers are compared with smokers. Examples of these include p53 [152,158], the nucleotide excision repair family of proteins, including ERCC1, involved in DNA repair [159], p38 (downstream of mitogen-activated protein kinase [MAPK]) [160], ataxia telangiectasia mutated (ATM) [161], nitrotyrosine (a marker of nitric oxide protein damage) [152], other chromosomal abnormalities [162,163], and methylation of p16 [164,165]. The microRNA-21 (miR-21) appears to be increased in lung cancer in never smokers, particularly in those with EGFR mutations and may play a significant role in lung carcinogenesis [166]. Additionally, mitochondrial DNA mutations in the respiratory complex-I have been reported to differ between never smokers and smokers with lung cancer [167].

The results of whole-genome and transcriptome sequencing of tumor (adenocarcinoma in 16 and large cell in 1) and adjacent tissue from 17 early-stage NSCLC patients identified average mutation frequency to be significantly higher in NSCLC patients with a smoking history compared with never smokers (n = 5 plus 1 light smoker) [168]. The number of mutations per megabase (Mb) in the cancers of the smokers (median 10.5) was significantly higher compared with the cancers in the never smokers (median 0.6 mutations per Mb) and the single former light smoker (0.3 mutations per Mb). The work also confirmed presence of EGFR and KRAS mutation in founder clones, supporting the hypotheses that these are true "driver" mutations when identified, and found novel gene changes in DNA repair and cell cycle pathways. The tumors from both smokers and never smokers were heterogeneous with multiple subclones. In another study of six Korean women never smokers, high-throughput multidimensional sequencing of lung adenocarcinoma tumor and adjacent normal tissue found multiple gene mutations, but only one patient had a recognized "driver" mutation, CCDC6-RET fusion [169].

Focused efforts to determine frequencies of the known "driver" mutations in NSCLC (ie, those mutations that appear to be the driving force in a cancer) are ongoing in multiple collaborations. The National Cancer Institute's Lung Cancer Mutation Consortium initially presented data at the American Society of Clinical Oncology 2011 annual meeting and for the first approximately 800 patients tested, reported mutation frequencies of 25 percent KRAS; 23 percent EGFR; 6 percent ALK rearrangements; and 3 percent or less each of BRAF, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA), mesenchymal-epithelial transition factor (MET) amplification, human epidermal growth factor receptor 2 (HER2), MAP/ERK kinase 1 (MEK1), and neuroblastoma RAS viral oncogene homolog (NRAS) [170]. Though those results are for adenocarcinomas in general, the identification of such mutations in over 60 percent of patients sampled and the mutually exclusive nature of the majority of those mutations indicate how far we have come in our biologic understanding of adenocarcinoma, the most frequent histology in never smokers.

Further updates have confirmed these relative percentages of mutations, which have also been looked at by The Cancer Genome Atlas effort [171]. Additional mutations, such as ROS1 and RET, are frequently more common in never smokers and have been identified at a rapid pace [172,173]. (See "Personalized, genotype-directed therapy for advanced non-small cell lung cancer", section on 'ROS1 rearrangements'.)

An analysis of 229 tumors from never smokers in Korea identified EGFR mutations, ALK gene rearrangements, and KRAS mutations in 48, 8, and 4 percent of cases, respectively [174]. A surgical series of 349 never smoker women with lung adenocarcinoma in China identified known driver mutations in nearly 90 percent; these were predominantly EGFR (76 percent) mutations, with less than 5 percent each of HER2, EML4-ALK, KRAS, and BRAF [175]. A smaller study in an area of China known for high rates of lung cancer in never smokers potentially related to coal use, found very high rates of an otherwise less common EGFR mutation, G719X, and a 15 percent rate of KRAS mutations in never smokers with lung cancer [176]. An analysis in the United States of over 800 patients with advanced-stage NSCLC, of whom one-third were never smokers, found only a 4 percent rate of KRAS mutations but a 49 percent rate of EGFR mutations in lung cancers from never smokers compared with 15 percent KRAS and approximately 30 percent EGFR mutations in those with up to 15 pack-years of smoking history and 40 percent KRAS and only 11 percent EGFR mutations in those with at least a 15 pack-year smoking history [177]. This provides further support to the assertion that targetable oncogenic mutant kinases can now be identified in the majority of lung adenocarcinomas from never smokers, especially in East Asian patients.

PROGNOSIS AND TREATMENT IMPLICATIONS — Whether or not patients with non-small cell lung cancer (NSCLC) who are never smokers have a better response to standard treatment and/or a better prognosis than those with a positive smoking history is uncertain. Some observational studies have demonstrated a better outcome [28,178-181], although this has not been true in all studies [182-184].

As an example, in an analysis of 12,000 patients from California, smokers had a shorter survival compared with never smokers (hazard ratio [HR] for death 1.09, 95% CI 1.00-1.18) [28]. Similarly, in another study of patients with lung adenocarcinoma comparing 132 never smokers with 522 current smokers, it was found that the five-year survival was significantly better among never smokers (23 versus 16 percent), a difference that was significant on multivariate analysis [178]. By contrast, in another study of 254 patients with lung cancer, the five-year survival was not significantly different in never smokers compared with smokers (27 versus 31 percent, respectively) [182]. Subsequently, a much larger lung cancer registry trial, which included over 15,000 Japanese and over 13,000 White American patients from Southern California, demonstrated a statistically significant improvement in survival for the patients with no smoking history [29]. In an expanded analysis of 26,957 NSCLC patients from Japan, the median overall survival was significantly longer for never smokers compared with ever smokers (30 versus 19 months) [185].

Other studies suggest that the cancers that develop in nonsmokers are more likely than those in smokers to harbor actionable driver mutations. In a detailed prospective molecular epidemiology study conducted in Japan in over 950 patients, half of whom were never smokers, epidermal growth factor receptor (EGFR) mutations were found in over 65 percent of the never smokers compared with <25 percent of the ever smokers; anaplastic lymphoma kinase (ALK) translocations were twice as likely in the never smoker population (5.1 percent) compared with the ever smoker cohort (2.7 percent). Tumor protein p53 (TP53) and Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations, however, were much more common in the ever smoker group [101].

A meta-analysis of 167 epidemiologic studies including over 63,000 lung cancer cases evaluated the frequency of specific driver mutations by smoking status. Lung cancer patients who were never smokers had increased odds of EGFR mutations compared with ever smokers, though this association was less apparent for Asian never smokers compared with other groups (odds ratio [OR] 3.89 versus 6.29) [186]. Similar group differences were seen for ALK translocations (OR 2.21 in Asian never smokers versus OR 4.09 in other groups).

EGFR inhibitors — Clinically important differences in treatment response are seen with epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (erlotinib, gefitinib, afatinib, osimertinib). Higher response rates and better survival have been reported in never smokers with these drugs compared with current or former smokers, though the true factor is presence or absence of an EGFR-activating mutation [143,187,188].

The iPASS and First-Signal trials comparing first-line gefitinib versus chemotherapy for never smoking Asian patients with advanced-stage NSCLC emphasize this point [189,190]. In both studies, high frequencies of EGFR mutations were found among never smokers, and gefitinib conferred an overall progression-free survival benefit only in those with activating EGFR mutations, while those without the EGFR-activating mutations did better with first-line chemotherapy. Thus, smoking status should not be used as a determining factor in selection of therapy, though the pretest probability of an actionable driver mutation is higher in those who develop lung cancer as a never smoker, and testing for these mutations is critical in that patient population. (See "Systemic therapy for advanced non-small cell lung cancer with an activating mutation in the epidermal growth factor receptor".)

Checkpoint inhibitors — Many developments in lung cancer have been in the use of immune checkpoint inhibitors. The activity of the checkpoint inhibitors can be predicted to some extent by programmed death-ligand 1 (PD-L1) expression levels, but other factors play a key role as well. One of the earliest clinical observations was that responses were more likely in patients who developed lung cancer who had a smoking history compared with never smokers. However, in one meta-analysis of five studies of second-line checkpoint inhibitor therapy versus chemotherapy, the never-smoker patient cohorts had similar survival benefits compared with the current/former smokers (HR 0.73, 95% CI 0.62-0.87 for never smokers versus HR 0.69, 95% CI 0.61-0.78 for current/former smoker groups) [191].

In a specific analysis of patients with KRAS-mutated NSCLC, investigators found that patients with a smoking history were more likely to express PD-L1 (44 percent) versus former (20 percent) or never smokers (13 percent). Though PD-L1 expression is not the only marker of programmed cell death protein 1 checkpoint inhibitor efficacy, this does add to the data supporting mechanisms of resistance to checkpoint inhibitor therapy in those who are never smokers [192].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diagnosis and management of lung cancer".)

SUMMARY

●Incidence – Although most cases of lung cancer are due to tobacco smoking, lung cancer among never smokers is an important problem, with the age-adjusted incidence rate for lung cancer in never smokers aged 40 to 79 years ranging from 11.2 to 13.7 per 100,000 person-years for men and from 15.2 to 20.8 per 100,000 person-years for women. (See 'Incidence' above.)

●Risk factors – Although the etiology for lung cancer among never smokers is not always clear, risk factors include air pollution, secondhand tobacco smoke, radon, and other environmental exposures. (See 'Risk factors' above and "Cigarette smoking and other possible risk factors for lung cancer".)

●Pathology – Adenocarcinoma is the most common pathology among never smokers and is more common than among ever smokers. (See 'Pathology' above.)

●Molecular biology – There are important differences at the molecular level between lung cancers arising in people who have never smoked and those who currently smoke or previously smoked. The best understood of these are the abnormalities in the epidermal growth factor receptor (EGFR) pathway. These EGFR pathway abnormalities seen in never smokers have been associated with a particular responsiveness to erlotinib, gefitinib, afatinib, and osimertinib and other agents that inhibit EGFR tyrosine kinase. The anaplastic lymphoma kinase (ALK) fusion oncogene is another molecular variant in non-small cell lung cancer that is seen more frequently in never smokers and for which targeted therapy is now available. Similar increases in frequency of other driver mutations are reported in lung cancer in never smokers. (See 'Molecular biology' above.)