Summary of Key Points
- •
The known or suspected etiologic factors for lung cancer arising in never-smokers are weak carcinogens or rare factors, which cannot explain the relatively high frequency of cancer in never-smokers. This also applies to environmental tobacco smoke.
- •
Genetic factors play an increasing role in the etiology of lung cancer in never-smokers. These include rare high penetrance mutations in crucial genes such as the T790M mutation in the EGFR gene. However, high-frequency, low-penetrance variations in susceptibility genes are playing an increasingly prominent role. These include loci that predispose to smoking as well as those that may contribute directly to cancers arising in smokers and never-smokers.
- •
The molecular alterations in lung cancers arising in smokers and never-smokers are very different. Smoke related tumors are associated with high numbers of mutations, especially C:G>A:T transversions, while never-smoker tumors are associated with low numbers of mutations targeting C:G>T:A transitions.
- •
The specific mutational targets are also different in smoker and never-smoker tumors. Thus KRAS mutations are more frequent in ever-smoker tumors, while EGFR mutations and ALK translocations are more frequent in never-smokers. Paradoxically, the number of therapeutic actionable mutations is more frequent in never-smoker tumors.
- •
Lung cancers arising in never-smokers show major differences based on ethnicity, gender, and histology. The ethnic differences point out the importance of genetic susceptibility loci in the development of lung cancers.
- •
The major clinical, ethnic, gender, and histology differences between lung cancers arising from smokers and never-smokers, coupled with their different etiologic factors and major molecular differences, indicate that they represent very different tumor types, confirming that lung cancers in never-smokers represent a different form of cancer.
Key Terms
Ever-smoker
An individual who has smoked 100 or more cigarettes during his or her lifetime.
Never-smoker
An individual who has smoked fewer than 100 cigarettes during his or her lifetime.
Current smoker
An individual who is currently smoking or who has quit smoking during the past 12 months.
Former smoker
An ever-smoker who quit more than 12 months earlier.
Epidemiology of Nonsmoking-Related Lung Cancer
Although numerous articles on lung cancer in never-smokers in Asia have been published, some data are inconsistent and other data are suspect, as the definitions of never-smokers are not uniform, and the quality of some of the data is questionable. Also, smoking incidence rates differ among women even within a single country. For example, the smoking incidence rates among women in northeastern China are considerably higher than the rates among women in southern China. For these reasons, we extensively cite reviews or meta-analyses that combine data from multiple published reports and from cancer registries. By doing so, we can avoid some of the biases from small, individual studies, and we can place ethnic, gender, and geographic differences in their proper context. Findings from a case–control study on epidemiologic risk factors for lung cancer in never-smokers are described in a 2010 article by Brenner et al.
A review of published studies on the epidemiology of lung cancer (18 studies, comprising 82,037 people) showed a marked gender bias that lung cancer among never-smokers appears to affect women more frequently than men, irrespective of geography ( p < 0.0001). The proportion of women with lung cancer who reported never having smoked regularly is particularly high in East Asia (61%) and South Asia (83%), whereas only 15% of women with lung cancer in the United States are never-smokers. In contrast, only 11% of men with lung cancer in East Asia are never-smokers.
Thun et al. published an analysis of 13 cohorts and 22 cancer registry studies with data from nearly 2.5 million never-smokers and from cancer registries in 10 countries covering several decades. Some of the key findings from this comprehensive analysis regarding lung cancer in never-smokers include the following:
- •
death rates from lung cancer were higher among men than women across all age and racial groups
- •
incidence rates among men and women were similar, with some variation by age
- •
death rates were higher among East Asian individuals (but not among those living in the United States) and black Americans than among white Americans
- •
no temporal trends were seen for American women
- •
lung cancer incidence rates were higher and more variable among East Asian women.
Known or Suspected Etiologic Factors for Lung Cancer in Never-Smokers
Because tobacco use is a powerful carcinogen and the major cause of lung cancer, most attention has focused on environmental tobacco exposure as the major cause of lung cancer in lifetime never-smokers. Although environmental tobacco exposure has been identified as a contributing factor for lung cancer in never-smokers since 1986, the Surgeon General of the United States 2006 report confirmed that environmental tobacco exposure modestly increased the risk of lung cancer. However, the odds ratios for the development of lung cancer in the United States indicated that such exposure is a very weak carcinogen compared with active smoking, as cited in the Surgeon General’s report. According to that report, the odds ratio is 1.0 for never-smokers who do not have environmental tobacco exposure and 1.2 for never-smokers who do have such exposure; in contrast, the odds ratio is 40.4 for ever-smokers. Thus, if environmental tobacco exposure is a weak carcinogen and cannot be the major cause of lung cancer in never-smokers, other known or suspected factors should be considered, such as indoor air pollution, environmental and occupational toxins (e.g., arsenic, radon, asbestos), and human papillomavirus (HPV) infection (and possibly other infections). Genetic factors should also be considered, and these are discussed later.
Indoor Air Pollution
The relatively high burden of lung cancer among women in China who have no history of regular smoking is attributed to indoor air pollution from coal smoke generated by unventilated coal-fueled stoves, volatilization of oils from cooking at high temperatures in open woks, and secondhand smoke. A meta-analysis of seven studies from China and Taiwan of never-smokers found that cooking oil vapors are the risk factor associated with lung cancer for women, and indoor coal and wood burning is a risk factor for both women and men. Indeed, a retrospective study on the association of household stove improvement and risk of lung cancer in rural China indicated that changing from unvented fire pits to stoves with chimneys was associated with a subsequent reduction in the lung cancer incidence rate. Other factors thought to contribute to higher lung cancer incidence among rural Chinese women who are never-smokers include a higher prevalence of nonsmoking women in Asian countries and viral factors of HPV infection.
Environmental and Occupational Toxins
Exposure to some environmental and occupational toxins has been shown to increase the risk of lung cancer for smokers and, to a lesser extent, for never-smokers. These toxins include arsenic, radon, asbestos, chromium, organic dust, and others.
As summarized in a meta-analysis, several studies indicate that high levels of arsenic in the major source of drinking water of highly defined geographic regions (southwestern Taiwan, the Niigata Prefecture, Japan, and northern Chile) were associated with increased incidence of lung cancer, both for smokers and never-smokers. The authors of the meta-analysis concluded: “Despite methodologic limitations, the consistent observation of strong, statistically significant associations from different study designs carried out in different regions provide[s] support for a causal association between ingesting drinking water with high concentrations of arsenic and lung cancer.”
Present in soil and groundwater, radon is a gaseous decay product of uranium-238 and radium-226, which is capable of damaging respiratory epithelium by emitting alpha particles. The increased risk of lung cancer among uranium miners has been more clearly established and is thought to be caused by radiation from radon, although most miners are ever-smokers. The role of radon in the home is more difficult to assess.
An analysis of occupational asbestos exposure in the Netherlands found a relative risk of lung cancer of 3.5 after controlling for age, smoking, and other factors. In a French study of 1493 cases, some occupational exposure was identified in 9.4% and 48.6% of female and male never-smokers, respectively, in whom lung cancer developed. In a Canadian case–control study, the odds ratio for lung cancer risk from occupational exposures in never-smokers was 2.1 (95% confidence interval [CI], 1.3–3.3) but was higher for exposure to solvents, paints, or thinners (odds ratio, 2.8; 95% CI, 1.6–5.0). A meta-analysis that focused on lung cancer risk for painters demonstrated a relative risk of lung cancer for all painters of 1.35 (95% CI, 1.29–1.41), but 2.0 (95% CI, 10.9–3.67) among never-smokers.
Human Papillomavirus
Several studies have found that HPV infection is associated with lung cancer, particularly in China and Taiwan. Cheng et al. reported a high incidence of HPV infection among never-smoking women in Taiwan. Results of a case–control study (141 cases and 60 controls) in Taiwan showed that the prevalence of HPV16 and HPV18 infection was significantly higher among never-smoking women with lung cancer who were older than 60 years; HPV 16 and HPV 18 infection was thought to be associated with the high lung cancer incidence and death rates among never-smoking women in Taiwan. Results of a similar study in Wuhan, China, indicated that no association with clinical-pathologic features was noted. However, the role of HPV infection in lung cancer pathogenesis in never-smokers might be restricted to certain geographic areas because the incidence rate of lung cancer associated with HPV infection varies widely based on geographic location and is reported to be low in Australia, Europe, and North America.
Clinical-Pathologic Features of Lung Cancer in Never-Smokers
Adenocarcinoma is the most common form of nonsmall cell lung cancer (NSCLC) in most parts of the world and the predominant form of lung cancer in never-smokers worldwide, followed by large cell carcinoma, which may represent an undifferentiated form of adenocarcinoma. Squamous cell carcinoma is rare among never-smokers with lung cancer, and small cell lung cancer almost never occurs. However, another neuroendocrine tumor, the bronchial carcinoid, may be slightly more common among never-smokers, although no relationship to smoking status has been shown.
The age at which lung cancer is diagnosed varies according to geographic location and smoking status. Studies from East Asian countries, such as Singapore and Japan, as well as Hong Kong, demonstrate an earlier age at the time of diagnosis among never-smokers compared with smokers, whereas the same or older age at the time of diagnosis among never-smokers has been found in studies from the United States and Europe. The possible reasons for this geographic variation include the greater contribution of risk factors other than active smoking in East Asian countries, much later age of initiation of smoking among East Asians with a smoking history compared with individuals from Western countries, and different degrees of detection bias between countries.
A retrospective study in Singapore comparing differences in the epidemiologic characteristics and survival outcomes between never-smokers and former and current smokers showed that never-smoker status was associated with a significantly better performance status, younger age at the time of diagnosis (10 years and 5 years earlier, respectively), higher proportion of women (68.5% vs.12% to 13%), and more advanced stage at the time of diagnosis. The variation in disease stage at the time of diagnosis might be explained by late presentation of symptoms and delayed diagnosis by physicians. The survival outcome for never-smokers was significantly better than that for smokers, with a 5-year overall survival rate of 10.8% and 7.7%, respectively ( p = 0.0003). Differences in treatment response and survival outcome between never-smokers and smokers with lung cancer may be attributed to differences in pathogenesis and tumor biology.
The Genetics of Lung Cancer
Inherited cancer syndromes are associated with rare and highly penetrant single-gene mutations, but genetic factors also play a role in sporadic cancers, as reported in numerous family-based studies. About 100 genes with mendelian inheritance cause an even smaller number of cancer syndromes, but these syndromes provide an explanation for only a minor part of the familial clustering of common cancers. Linkage analyses of high-risk families may identify other rare high-penetrance genes, and such studies have identified a lung cancer susceptibility locus on chromosome 6q. Smoking appeared to increase the susceptibility. Further studies indicated that the regulator of G-protein signaling 17 ( RGS17 ) gene at this location was a major candidate for lung cancer susceptibility.
The major mechanism of acquired resistance to tyrosine kinase inhibitors in lung cancers with epidermal growth factor receptor ( EGFR ) gene mutations is the appearance of a second activating mutation, T790M (substitution of threonine 790 with methionine). However, T790M may be inherited as a rare familial mutation. Our recently published study of a large family with an inherited T790M mutation and lung cancer, combined with analysis of published cases, indicates that inherited T790M predisposed lifetime never-smokers (and women) to lung cancer. These findings have, in part, been independently confirmed. Of interest, although EGFR mutations occur more frequently among East Asians, no case of an inherited T790M mutation has been described among East Asians. However, V843I, an even rarer inherited EGFR gene mutation that predisposes to lung cancer, has been reported in both Asian and non-Asian families. Another recent report described a Japanese family with an autosomal inherited germline mutation in human epidermal growth factor receptor 2 ( HER2 ) associated with lung cancer risk, which may also target women and light or never-smokers.
It is now believed that alleles with high frequency (typically greater than 10%) and low penetrance (typically less than a twofold increased lifetime risk) contribute substantially to susceptibility to many diseases, including lung cancer. Genome-wide association studies (GWAS) using population-based designs have identified many genetic loci associated with risk of a range of complex diseases, including lung cancer; however, each locus exerts a very small effect, and combinations of genes are required to exert a significant effect on risk. GWAS are often based on large microchip analyses of single nucleotide polymorphisms (SNPs), and more than one million SNPs can be analyzed on a single microchip. These studies of weak associations often consist of many thousands of cases and controls, and meta-analyses may be required for confirmation. In lung cancer, more than 150 GWAS have been published. Although some findings are widely accepted, others are controversial or require confirmation. In 2008, three studies identified three potential susceptibility loci for lung cancer. Two of these loci, on chromosomes 15q25 and 5p15.33—the site of the telomerase reverse transcriptase ( TERT ) gene, essential for telomerase activation—have been confirmed, but the cancer-associated role of the locus on 6p21-6p22 remained more controversial; however, it may be histology related. Additional studies, including a meta-analysis, confirmed that the major susceptibility locus was on 15q25, encoding several genes, including the nicotinic acetylcholine receptor (nAChR) genes: cholinergic receptor, nicotinic, beta 4 (neuronal) ( CHRNB4 ), alpha 5 (neuronal) ( CHRNA5 ), and alpha 3 (neuronal) ( CHRNA3 ). Because the variants at 15q25 are also associated with nicotine dependence, they may influence lung cancer risk at least in part through an effect on smoking behavior rather than a direct effect on lung carcinogenesis. A large meta-analysis of lung cancer in female never-smokers in six Asian countries showed no evidence of association for lung cancer at 15q25 in that population, which the authors said provided “strong evidence that this locus is not associated with lung cancer independent of smoking.” Other studies, including meta-analyses, have identified additional variants associated with increased risk, such as smoking, ethnicity, gender, and histology. Thus, although genetic variation of the TERT locus appears to be involved in susceptibility to all lung cancers, the 15q25 locus predisposes to smoking, the cyclin-dependent kinase inhibitor 2A ( CDKN2 ) locus at 9p21 may influence susceptibility to squamous cell carcinomas, and the tumor protein p63 (TP63) locus may influence susceptibility to lung adenocarcinoma in East Asian populations.
As confirmed by the GWAS cited previously, nicotine and its derivatives, by binding to nAChR on bronchial epithelial cells, can regulate cellular proliferation and apoptosis by activating the protein kinase B (PKB/Akt) pathway. Lam et al. found different nAChR subunit gene expression patterns between NSCLCs from smokers and nonsmokers, and a 65-gene expression signature was associated with nonsmoking nAChR alpha-6 beta-3 expression.
Molecular Characteristics of Nonsmoking East Asian Individuals With Lung Cancer
With the development of molecular genetic therapies for lung cancer, the molecular profile of East Asian individuals with lung cancer was found to differ from that of white individuals with lung cancer. Mutations in Kirsten rat sarcoma viral oncogene homolog ( KRAS ) and EGFR genes are mutually exclusive and demonstrate striking frequency differences related to ethnicity. EGFR mutation is the first specific molecular alteration associated with lung cancers arising among never-smokers. A relatively high incidence of somatic mutations in EGFR has been found in a specific subpopulation: women, never-smokers, patients with adenocarcinoma, and Asians. In the First Line Iressa versus Carboplatin/Paclitaxel in Asia (Iressa Pan-Asia Study [IPASS]) study, with 1214 (99.8%) of 1217 patients of East Asian origin and 1140 (93.7%) of 1217 never-smokers, the incidence rate of EGFR mutation was 59.7% in the 437 patients evaluable for EGFR mutation. A recent multinational study demonstrated variations in the EGFR gene mutation rates in Asian countries, with the lowest frequencies from India.
Nevertheless, a review of nine published studies showed that the frequency of EGFR mutation among US never-smokers with NSCLC was substantially lower (20%). In addition, even in an unselected population, the frequency of EGFR mutation among East Asian individuals with NSCLC was also considerably higher than that for white individuals. In the review, which included an analysis of data on 2347 patients for whom ethnicity was noted, the frequency of EGFR mutations among East Asian patients was significantly higher compared with non-Asian patients (33% vs. 6%; p < 0.001). Unlike EGFR mutations, KRAS mutations occur less commonly in lung cancers among individuals from East Asia and more frequently in lung cancers among smokers.
In pooled data summarizing three published studies comprising 1536 patients with NSCLC, EGFR and KRAS were shown to be mutually exclusive in the same tumors. KRAS mutations were detected in 20% of patients with NSCLC, particularly patients who smoked or who had adenocarcinoma. A study investigating the EGFR and KRAS status of 519 unselected patients with NSCLC showed that KRAS mutations were present more frequently in smokers than never-smokers (10% vs. 4%; p = 0.01), among non–East Asians than East Asians (12% vs. 5%; p = 0.001), and among patients with adenocarcinoma than patients with nonadenocarcinoma histologies (12% vs. 2%; p < 0.001). Several studies found that KRAS mutations were present in 20% to 30% of white patients with lung adenocarcinoma but only 5% of patients with lung adenocarcinoma from East Asia. In addition, in previous studies from Hong Kong and Taiwan, KRAS mutations were found in 13% to 19% of men with adenocarcinoma but in none of the women studied. A potential explanation for the distinction between genders may be that the vast majority of Chinese female patients were never-smokers.
A Japanese case–control study assessing the impact of smoking and gender on the risk of NSCLC with or without EGFR mutation demonstrated that ever-smoking was a substantial risk factor for NSCLC without EGFR mutation but not for NSCLC with EGFR mutation. Cumulative exposure to smoking was associated with a linear increased risk of NSCLC without EGFR mutation only. This finding was consistent for both men and women. Age at the start of smoking among ever-smokers and years since quitting smoking among former smokers also showed a strong correlation between NSCLC without EGFR mutation and smoking. EGFR mutation was present more frequently among patients who smoked no more than 20 pack-years. Similarly, in another Japanese study, EGFR mutation was found more frequently among patients who quit smoking at least 20 years before the date of lung cancer diagnosis. These findings suggest an inverse correlation between EGFR mutation and exposure dose of cigarette smoking.
Smoking status is a risk factor affecting not only EGFR mutations but other somatic mutations as well. The authors of a Korean study screened genetic tests for EGFR mutations, KRAS mutations, and enchinoderm-microtubule-associated protein-like 4-anaplastic lymphoma kinase ( EML4-ALK ) fusions in 200 fresh surgical specimens of primary lung adenocarcinoma by polymerase chain reaction (PCR), Sanger sequencing, and fluorescence in situ hybridization. They then performed high-throughput RNA sequencing in 87 lung adenocarcinoma specimens that were negative for the three known driver mutations (three samples with insufficient RNA quality were excluded). The results showed that people who had a smoking history of at least 40 pack-years harbored significantly more somatic point mutations than did people who had a smoking history of fewer than 40 pack-years or of never-smoking. In addition, important differences in mutation patterns exist between lung cancer in never-smokers and ever-smokers.
Given the difference in the incidence rate of EGFR mutations between East Asian and white populations, several studies have investigated ethnic differences in ALK , c-ros oncogene 1 receptor tyrosine kinase ( ROS1 ), and ret proto-oncogene ( RET ) fusions after these three novel driver fusions were identified in NSCLC. Most studies showed that ALK fusions occurred in 2.4% to 5.6% of NSCLC cases, and no differences in incidence rate between Asian and non-Asian populations have been identified to date. However, a Chinese study screening ALK fusions by rapid amplification of complementary DNA ends (RACE)-coupled PCR sequencing found that ALK fusions existed in 12 (11.6%) of 103 individuals with NSCLC, 10 (16.13%) of 62 individuals with adenocarcinomas, and 10 (19.23%) of 52 never-smokers. This high incidence of ALK fusions in a selected East Asian population may be explained by the relatively small sample size and use of a different screening method. Unlike ethnicity, smoking status is regarded as an important factor affecting the incidence rate of fusion genes. Similar to ALK fusions, ROS1 and RET fusions appear to occur more frequently among never-smokers. Given a very low frequency of ROS1 as well as RET fusions identified in NSCLC, a large sample study is warranted to prove the role of smoking status in the occurrence of fusion genes.
Several Chinese studies have demonstrated the previously described differences in the molecular profile of lung cancer between never-smokers and smokers. An et al. screened for candidate driver genes in 524 Chinese patients with NSCLC with the use of several methods, including sequencing, high-resolution melt analysis, quantitative PCR, or multiplex PCR and RACE, and analyzed the differences in driver gene alterations among a subgroup based on histology and smoking status ( Table 4.1 ). The findings demonstrated that the driver gene alterations in nonsmokers differ completely from driver gene alterations in smokers, irrespective of histologic type. In adenocarcinoma, EGFR , phosphatase and tensin homolog ( PTEN ), and phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha ( PIK3CA ) mutations and ALK fusions were present more frequently among never-smokers, whereas KRAS and serine/threonine kinase 11 ( STK11 ) mutations were present more frequently among smokers. The met proto-oncogene methylated CpG (mCpG) sequences ( MET ) and v-raf murine sarcoma viral oncogene homolog B ( BRAF ) mutations did not differ substantially by smoking status. As expected, fewer squamous cell carcinomas were present, and discoidin domain receptor tyrosine kinase 2 ( DDR2 ) and fibroblast growth factor receptor 2 ( FGFR2 ) mutations, although infrequent, were present only in tumors from smokers.
| Adenocarcinoma ( n = 347) | ||
|---|---|---|
| Gene | Never-Smokers (66%) | Ever-Smokers (34%) |
| EGFR | 49.8 | 22.0 |
| PTEN | 9.9 | 2.6 |
| ALK | 9.3 | 4.5 |
| PIK3CA | 5.2 | 2.1 |
| STK11 | 2.7 | 11 |
| KRAS | 4.5 | 12 |
| C-MET | 4.8 | 4 |
| BRAF | 1.9 | 3.1 |
| Squamous Cell Carcinoma ( n = 144) | ||
|---|---|---|
| Gene | Never-Smokers (35%) | Ever-Smokers (65%) |
| DDR2 | 0 | 4.4 |
| FGFR2 | 0 | 2.2 |
In another Chinese study limited to lung cancers from never-smokers, Li et al. identified driver mutations in 89% of the tumors ( Table 4.2 ). Of interest, these mutations were mutually exclusive, consistent with their driver status. Although the mutation figures may be lower among lung cancers in never-smokers in Western countries, most of these tumors contain potentially actionable driver mutations. In conclusion, driver gene alterations in NSCLC are shown to be associated with smoking status rather than gender.
| Driver Mutation a | Percentage (%) of Patients ( n = 202) |
|---|---|
| EGFR | 75 |
| HER2 | 6 |
| ALK fusion | 5 |
| KRAS | 2 |
| ROS1 fusion | 1 |
| BRAF | 0 |
| No mutation | 11 |
| Any mutation | 89 |
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
