The (non)smoking gun: Experts identify alarming increase in lung cancer among never-smokers

Lung cancer once was considered a disease for smokers.

However, several studies released in the past year — including two presented in September at the World Conference on Lung Cancer — suggest that stereotype no longer applies.

Over the last decade, the proportion of non–small cell lung cancer cases diagnosed among never-smokers more than doubled. An increasing percentage of these cases are diagnosed in younger patients — including those who are physically fit and otherwise healthy — and in women, particularly those of Asian descent.

Because a majority of these patients are asymptomatic, their cancers are detected at later stages, making treatment especially challenging.

Efforts to identify and understand the explanations for this trend have become a research priority.

“We are starting to learn more about family histories and potential linkages, as well as about the genetics of lung cancer,” Karen Reckamp, MD, MS, co-director of the lung cancer and thoracic oncology program and associate professor in the department of medical oncology and therapeutics research at City of Hope in Duarte, California, told HemOnc Today. “[This] helps us to understand that cancers of younger, nonsmoking patients tend to be very different than smoking lung cancers.”

Smoking cessation efforts have led to dramatic reductions in the number of people who smoke, contributing to an overall decline in lung cancer incidence. However, experts contend that alone does not explain the increased proportion of lung cancer among never-smokers, as research has uncovered potential roles of environmental factors and genetics.

However, a key question — if or how screening can be used in this population — remains unanswered.

“We are learning more and more about these factors — genetic and germline mutations — to help us better determine how to screen people and who to screen,” Reckamp said. “We have to find a different way to do it than we do for smokers. We are heading in the direction of blood signatures and finding the genetics that are potentially inheritable to understand how best to look for lung cancer early in this population.”

HemOnc Today spoke with researchers and thoracic oncologists about the search for these genetic risk factors, the early data that are emerging on incidence and treatment, the influence of potential environmental factors, and how diagnostic tools and screening practices can be altered to reverse these trends.

Increased incidence

Smoking cessation efforts have led to dramatic decreases in the smoking rate over the past 5 decades.

An estimated 16.8% of Americans reported actively smoking in 2014, compared with 42.4% in 1965, according to CDC statistics.

This trend, in turn, has reduced the raw number of smokers who have lung cancer.

“Overall, this is good news,” Frederic J. Kaye, MD, professor of medicine and medical co-director of the thoracic oncology program at University of Florida, told HemOnc Today. “Smoking rates have dropped considerably in the U.S. over the last 60 or 70 years. As a consequence, age-adjusted death rates for lung cancer — certainly for men and now for women — also are falling quite dramatically, leading to reductions in overall cancer mortality.

“In the background of that good news, we have a better molecular understanding of these diseases, so we are recognizing an increasingly important proportion of lung cancer that has arisen in lifelong nonsmokers,” Kaye added.

Thus, whether lung cancer actually is increasing in never-smokers — or whether the nonsmoking population simply now represents a larger proportion of cases — has been debated and has not yet been answered.

Eric Lim, MB, ChB, MD, MSc, FRCS (C-Th), consultant thoracic surgeon at Royal Brompton Hospital in London, and colleagues conducted a retrospective study of prospective data from 2,170 patients who underwent surgery for lung cancer at their institution between March 2008 and November 2014.

Results — presented last fall at the World Conference on Lung Cancer — showed 436 (20%) patients were never-smokers. Sixty-seven percent of this group were female and 54% had adenocarcinoma.

Further, their data showed the rise of actual never-smoking patients with lung cancer increased from 60 in 2010 to 97 in 2014, and the proportion of NSCLC cases involving never-smokers rose from 13% in 2008 to 28% by 2014.

“We showed that the numbers of smoking-related lung cancer are decreasing, but also that the numbers of nonsmokers are increasing — not just in percentage, but in raw numbers,” Lim told HemOnc Today.

In another study presented at the conference, Lorraine Pelosof, MD, PhD, and colleagues analyzed data from three cancer registries in the southern United States — University of Texas Southwestern, Parkland Hospital in Dallas, and Vanderbilt University in Nashville, Tennessee.

Data from all these registries showed increased incidence of NSCLC among self-reported never-smokers. The increase was greatest at UT Southwestern, where the rate rose from 8.9% between 1990 and 1995 to 19.5% from 2011 to 2013 (P < .0001).

“It looks like the trend for non–small cell lung cancer among never-smokers is increasing,” Pelosof, who served as assistant professor at UT Southwestern during the conduct of the study but since joined the FDA as medical officer for the Center for Drug Evaluation and Research’s Office of New Drugs and Office of Hematology and Oncology Products, told HemOnc Today. “However, in small cell lung cancer, the rate is stable, and this is most likely because this type is almost 100% attributed to tobacco,”

Still, the fact there are fewer smokers may influence the interpretation of these data, Kaye said.

“There is reason to think that a certain number of people are destined to develop nonsmoker lung cancer,” Kaye said. “That subgroup could have been masked in the past if those people who were destined to get nonsmoker lung cancer happened to be smokers, because so much of the adult population in the U.S. were smokers at the time.”

The apparent increased risk for NSCLC among never-smokers also could be attributed to previous misidentification of cancers as “unknown primaries,” which were common prior to advances in molecular genotyping; the prolonged lifespan of the population, thus creating a larger pool of patients; and genetic factors linked to certain mutations — such as EGFR — and to ancestry, Kaye said.

Although these confounders potentially could change the interpretation of these early data, research into the nonsmoking population is addressing an unmet need, he added.

Genetics and the environment

The increasing incidence of NSCLC in nonsmokers partially may be explained by data interpretation and changes over time; however, genetics and environmental factors likely also play a role.

For instance, radon is estimated to cause 10% of lung cancers, whereas workplace exposures to carcinogens such as asbestos and diesel exhaust cause 9% to 15%, and air pollution causes 1% to 2%, according to data from the American Lung Association.

Research also has shown that never-smokers may achieve better outcomes with EGFR inhibitors, suggesting genetic underpinnings to their disease.

Saraiva and colleagues investigated differences between smoking and nonsmoking NSCLC. Results — presented at the European Respiratory Society International Congress 2015 — showed nonsmokers were more likely to be women and have the adenocarcinoma subtype, and they were less likely to have chronic obstructive pulmonary disease.

Further, 9% of the nonsmoking cohort had been exposed to carcinogens in a professional setting, 5% had a family history of lung cancer and 6% had a previous cancer diagnosis.

A main concern prompting researchers to investigate genetic and environmental causes is the fact that NSCLC among nonsmokers is more common in women than men.

Lim and colleagues showed 67% of nonsmokers with NSCLC were women, and Pelosof and colleagues calculated a 65% rate.

Photo by James Hall Photography

“It simply may be that women are more susceptible to genetic mutations and environmental exposures,” Barbara J. Gitlitz, MD, associate professor of clinical medicine at the University of Southern California Keck School of Medicine in Los Angeles, told HemOnc Today.

A study from Baik and colleagues, published in Cancer, Epidemiology, Biomarkers & Prevention, showed hormonal factors also may influence lung cancer risk in women.

Their results showed menopause onset before age 44 years (HR = 1.39; 95% CI, 1.14-1.7) and past oral contraceptive use for longer than 5 years (HR = 1.22; 95% CI, 1.05-1.42) appeared associated with increased lung cancer risk. Specifically among never-smokers, having had more offspring appeared associated with decreased risk among women who had given birth (P = 0.03).

These trends are particularly true for Southeast Asian women. Although nonsmoking lung cancer accounts for approximately 20% of lung cancer cases in the U.S., upwards of 80% of Southeast Asian women with lung cancer are nonsmokers.

Environmental factors of excessive exposure to pollution, secondhand smoke and certain cooking oils are thought to contribute to this trend.

These women also may be more susceptible to genetic mutations.

About 60% of nonsmoking Asian women diagnosed with lung cancer harbor an EGFR mutation, Edward S. Kim, MD, chair of solid tumor oncology and investigational therapeutics and Donald S. Kim distinguished chair for cancer research at Levine Cancer Institute at Carolinas HealthCare System, as well as a HemOnc Today Editorial Board member, said in an interview.

Researchers from the NCI partnered with other countries to form the Female Lung Cancer Consortium in Asia to study reasons for this trend.

Based on their efforts, Lan and colleagues identified three genetic regions — two on chromosome 6 and one on chromosome 10 — that may predispose nonsmoking Asian women to lung cancer. These gene variations may make women more susceptible to the effects of environmental secondhand smoke.

Overall, identification of genetic factors — specifically related to EGFR biology — may be useful for identifying a subpopulation at risk for nonsmoking lung cancer, Kaye said. This group may contain Hispanics in addition to Asians.

“The more native to America you are, the more Asian ancestry you are going to have,” Kaye said “It is pretty striking that there appears to be an increase in EGFR mutation rates in populations in Central America, South America and Mexico based on recent reports. My sense is that the more Asian ancestry in the Latino population there is, the higher rate of EGFR mutations, which speaks loudly and convincingly that there is a strong genetic component underlying EGFR mutation-driven lung cancer.”

A combination of environmental and genetic factors could lead researchers to define a subset of patients at risk for lung cancer, as they have done previously for melanoma.

“The bronchial epithelium is basically just skin on the inside,” Kim told HemOnc Today. “So, when those with a higher susceptibility inhale carcinogens — such as secondhand smoke or industrial pollution — they have a greater chance of what they inhale to cause changes or damages.

“Look at George Burns,” Kim added. “He didn’t die of lung cancer but smoked his entire life. He was just one of those people who could withstand all that chronic exposure. That exists out there, but we just have not found a way to identify the equivalent of fair-skinned Irish folks and olive-skinned Italian folks in lung cancer.”

The ‘Young Lung’

The average age at the time of a lung cancer diagnosis is about 70 years. Less than 2% of those diagnosed with the disease are aged younger than 45 years.

However, early data on lung cancer in nonsmokers has revealed more of these patients are younger — in their 20s and 30s — and they often are healthy and athletic. Lung cancer in this cohort appears very different than it does in the older, smoking population.

Wang and colleagues analyzed 41 patients with lung cancer aged 30 years or younger (mean age, 26.4 ± 3.5 years). They found a majority presented with advanced stages of disease (58.5%). Only five patients had a history of smoking.

Further, the results — published last year in PLoS One — showed that, of the 22 patients with available genomic data, 22.7% harbored EGFR mutations and 27.2% harbored EML4-ALK gene fusions. Two of 17 patients evaluated for KRAS and ROS1 mutations were positive for these variations.

The Genomics of Young Lung Cancer Study — for which Gitlitz serves as a lead investigator — also seeks to elucidate the genetic causes of lung cancer in younger patients. The study has grown into an international effort that involves academic medical centers, community hospitals and researchers from around the world, and it contains an interactive component that allows patients to participate online.

Researchers plan to enroll hundreds of patients diagnosed with lung cancer prior to age 40 years.

“We’ve had 88 patients sign up already and expect to get many more,” Gitlitz said. “The majority of them have stage IV adenocarcinoma, and more than 80% have either an EGFR, ALK, ROS1, HER2 or RET driver mutation.”

Only about 33% of patients thus far ever smoked.

Researchers also are planning a larger case–control epidemiology trial to evaluate tumor and germline genomics, elucidate genome-wide associations, and determine why driver mutations are so prevalent in this population.

“It is going to be pretty novel,” Gitlitz said. “It’s not like the old-school epidemiology studies where you open up a phone book or are using DMV records to find controls. We are going to do this entirely on social media to find at least 1,000 controls under age 40 who are healthy to match about 500 cases of young lung cancer.”

The analysis will rely heavily on the willingness of healthy younger adults to take the time to answer a questionnaire and provide blood and urine samples for analysis.

“It’s going to be a big thing to ask,” Gitlitz said. “But as Bonnie J. Addario — founder of the Addario Lung Cancer Medical Institute and the Bonnie J. Addario Lung Cancer Foundation — always says, ‘Anyone with lungs can get lung cancer.’ Although we’re asking a lot, we know the reward could be great and that we may very well find an at-risk population for young lung cancer.”

The screening conundrum

A major concern regarding the increase in lung cancers among never-smokers is that most patients do not present with symptoms. Consequently, many are diagnosed with advanced disease.

Within their cohort, Pelosof and colleagues found the incidence of nonmetastastic cancers (stage I-III) in never-smoking patients did not change significantly over the time studied except within the Parkland database, where stage IV cancers became more prevalent (P = .0071).

These data suggest earlier detection of lung cancer will be crucial to more effectively treat the nonsmoking population.

“The normal everyday person needs to understand this is an emerging problem and we need to tackle it from a medical perspective, and that means funding more research into this area,” Lim said.

In their analysis, Lim and colleagues found that 36% of nonsmokers diagnosed with NSCLC did not display any symptoms. Thirty percent of NSCLCs were detected on an incidental CT scan, 14% were detected on incidental chest film, 7% were detected on incident PET/CT and 1% were detected on incidental MRI.

Further, 52% of patients presented with a cough (34%) or chest infection (18%), neither of which were specific enough individually to warrant a potential NSCLC concern.

“When patients come in with a persistent cough, for example, the first thing on your mind as their physician is not lung cancer,” Reckamp said. “We’re talking about younger, otherwise healthy individuals.”

Because there are no concrete signals to indicate cancer risk among nonsmokers, widespread screening would not be feasible. The cost would be astronomical and yield only a very small benefit.

“We see a high prevalence of EGFR in the lung cancer of Asian females who are never-smokers, but that is for those who are already diagnosed,” Kim said. “What is the incidence of cancer in an Asian female who doesn’t smoke? It’s really, really low. This is the problem with screening — you cannot take characteristics of a diagnosed population and apply it to screening.”

Further, it took decades to identify a population of smokers who warrant yearly screening with low-dose CT, Kim said. The U.S. Preventive Services Task Force defines this cohort as individuals aged 55 to 80 years with a 30 pack-year smoking history who are current smokers or who quit within the past 15 years.

The task force established those criteria based on results of the National Lung Screening Trial, which showed screening with low-dose CT rather than chest X-ray conferred a 20% reduction in lung cancer deaths among current or former heavy smokers.

It may be possible to develop a non–CT-based screening test to detect cancer in nonsmokers, experts said.

“Down the road, it would have to be something we find in the blood like PSA for prostate cancer or CA-125 for ovarian cancer,” Kim said. “It has to be a blood test that is easy and cheap like a hemoglobin A1C test that can give you a quick pass-through.”

Lim and colleagues also presented a poster at the World Conference on Lung Cancer that uses a blood test to identify a three-gene signature for the detection of lung cancer.

“We look at the mutations in the bloodstream from cancer cells that have died and released DNA into the blood and, within the blood itself, we identify mutations that are present in cancer but not normally found in normal tissue to use as a clue to help us identify patients who may harbor cancers,” Lim said.

The investigators found that, in a cohort of more than 200 patients, 68% had evidence of tumor DNA in the blood. The diagnostic accuracy was 91%.

“This is a much more realistic way to screen patients,” Lim said. “Our aim is to make the test very specific so, if you test positive, we have a high degree of certainty that you have cancer. It does not work the other way around, though. If you are negative, it does not provide reassurance.”– by Anthony SanFilippo; additional reporting by Jennifer Southall

References:

American Lung Association. Lung Cancer Fact Sheet. Available at: www.lung.org/lung-health-and-diseases/lung-disease-lookup/lung-cancer/learn-about-lung-cancer/lung-cancer-fact-sheet.html. Accessed Feb. 25, 2016.

Alberg AJ, et al. Chest. 2013;doi:10.1378/chest.12-2345.

Baik CS, et al. Cancer Epidemiol Biomarkers Prev. 2010;doi: 10.1158/1055-9965.EPI-10-0450.

CDC. MMWR. 2015;4:1233-1240.

Lan Q, et al. Nat Genet. 2012;doi:10.1038/ng.2456.

Saraiva C, et al. Differences in epidemiological and clinical features in non-small cell lung cancer (NSCLC) in never and ever smokers. Presented at: European Respiratory Society International Congress; Sept. 26-20, 2015; Amsterdam.

Wang Y, et al. PLoS One. 2015;doi:10.1371/journal.pone.0136659.

The following were presented at: 16th World Conference on Lung Cancer; Sept. 6-9, 2015; Denver, Colorado.

Gitlitz B, et al. The Genomics of Young Lung Cancer Study.

Lim E, et al. Increasing incidence of non-smoking lung cancer: Presentation of patients with early disease to a tertiary institution in the UK.

Pelosof L, et al. Abstract ORAL22.01.

Viola P, et al. P3.04-011.

For more information:

Barbara J. Gitlitz, MD, can be reached at bgitlitz@med.usc.edu.

Frederic J. Kaye, MD, can be reached at fkaye@ufl.edu.

Edward S. Kim, MD, can be reached at edward.kim@carolinashealthcare.org.

Eric Lim, MB, ChB, MD, MSc, FRCS (C-Th), can be reached at e.lim@rbht.nhs.uk.

Lorraine Pelosof, MD, can be reached at lorraine.pelosof@fda.hhs.gov.

Karen Reckamp, MD, MS, can be reached at kreckamp@coh.org.

Disclosure: Gitlitz, Kaye, Kim, Lim, Pelosof and Reckamp report no relevant financial disclosures.

 

 

Does smoking history impact the development of genetically predisposed lung cancer?

Smoking and tobacco exposure are still two of the key villains, even in the gene-mutated lung cancers.

There are a few issues here. We have not fully characterized what the risk for lung cancer is in people with favorable gene mutations such as ALK, EGFR or ROS1. The biggest risk factor for lung cancer is still smoking. That said, the favorable gene mutations are seen more frequently in light or never-smokers. Those patients do respond better to therapy, specifically to tyrosine kinase inhibitors, and ultimately have more favorable prognoses.

The data from trials evaluating TKIs — such as erlotinib (Tarceva; Genentech, Astellas) and gefitinib (Iressa, AstraZeneca) — have revolutionized the treatment of non–small cell lung cancer. However, what has not yet been fully elucidated is whether there are differences in outcomes for smokers and nonsmokers among patients with these genetic mutations.

A meta-analysis of all prospective trials that randomly assigned EGFR-mutated patients to TKI or platinum doublet chemotherapy (n = 1,649) revealed that, although all patients benefited from TKI, the benefit seen in never-smokers was greater (P < .007). The pooled HR for PFS for never-smokers was 0.29 (95% CI, 0.21-0.39), whereas for smokers, the HR was 0.54 (95% CI, 0.38-0.76). Several retrospective studies of patients with gene-mutated lung cancers have reported that smoking independently impacts response to therapy, and that patients who were never-smokers or light smokers have improved PFS and OS compared with smokers. In vitro studies demonstrated that hyporesponsiveness in smokers with EGFR-mutated lung cancers was accompanied by persistent activation of EGFR-downstream signal molecules ERK1/2 and AKT, and this activation could not be inhibited by gefitinib. Multiple cell line studies have documented decreased gefitinib sensitivity due to cigarette smoke extract exposure, mediated by various epithelial to mesenchymal transition pathways.

Although these conclusions are subject to the limitations and biases inherent in meta-analyses, retrospective studies and subset analyses — such as heterogeneity, selection bias and multiple testing — these data strongly suggest that cigarette smoking has an independent effect on outcome for patients with gene-mutated lung cancer. Although further investigation will shed more light on these issues, smoking cessation should be recommended for all patients — particularly those with lung cancer — regardless of genetic mutation status.

 

References:

Hasegawa Y, et al. Oncologist. 2015;doi:10.1634/theoncologist.2014-0285.

Kim MH, et al. Lung Cancer. 2014;doi:10.1016/j.lungcan.2014.01.022.

Li D, et al. Lung Cancer. 2016;doi:10.1016/j.lungcan.2015.12.007.

Liu M, et al. Am J Transl Res. 2015;7:2026-2035.

Lynch TJ, et al. N Engl J Med. 2004;350:2129-2139.

Andrea Wolf, MD, is a thoracic surgeon at Icahn School of Medicine at Mount Sinai. She can be reached at andreawolf@mountsinai.org. Disclosure: Wolf reports no relevant financial disclosures.

Although the prospect of cooperation between carcinogen exposure and genomic aberration is intriguing, the current body of literature contains little supporting evidence.

In recent years, there has been significant progress in understanding the genetic origins of lung cancer. Through large-scale sequencing efforts, common genes that drive the development of lung cancer, particularly adenocarcinoma, have been catalogued in detail.

Based on these findings, therapies targeting these mutations have become an important part of clinical practice. Still, the influence of inheritable polymorphisms and the interaction between these aberrations and environmental carcinogens remains unclear.

When considering whether smoking impacts genetically predispositioned lung cancer, we must first be specific about how we interpret the term “genetically predispositioned.”

In the context of tumors with an identifiable driver mutation in genes such as EGFR or BRAF, or translocations involving genes such as ALK or ROS1, there is evidence to suggest that these aberrations arise independently of environmental tobacco exposure. Whereas mutational profiles of smoking-related lung cancer suggest a preponderance of C:G or A:T transversions (eg, KRAS G12C), tumors with low transversion burden more frequently harbor point mutations or insertions in EGFR or fusion proteins such as ALK, ROS1 and RET.

This would suggest that mutational events in these genes appear to arise independently of tobacco exposure and may be more associated with stochastic events in the DNA. In addition, epidemiologic evidence would suggest that these types of mutations occur much more commonly in never-smokers.

Although some might argue that even in never-smokers, environmental factors such as exposure to secondhand smoke, occupational fumes or radon may play a role, the measurable risk of these carcinogens on lung cancer formation is poorly defined. Therefore, the existing genomic and epidemiologic data argue against a causal relationship between tobacco smoke exposure and tumor formation in this genetic context.

A more literal interpretation of “genetic predisposition” would suggest the possibility of inheritable forms of these mutations. This has been investigated in several studies, but virtually all fail to provide evidence for their existence. Germline mutations involving EGFR T790M have been described; however, inherited activating mutations in the EGFR kinase domain have not been identified.

One particular variant in EGFR — V843I — has been identified in a family with lung cancer in multiple generations, but the functional significance of this mutation is uncertain. Taking a more unbiased approach, other investigators have used genome-wide association studies to identify loci that are strongly associated with the development of lung cancer. These studies have highlighted the link between the chromosome 15q25 locus — in particular, the nicotinic acetylcholine receptor subunits CHRNA5, CHRNA3 and CHRNB4 — and the formation of lung cancer.

However, there appears to be conflicting evidence whether these variants directly contribute to tumorigenesis or instead affect an individual’s propensity to smoke tobacco. Additional studies have shown a similar association with chromosome 5p15.33. This region is found to contain genes TERT (a telomerase enzyme) and CLPTM1L (a gene of relatively unknown function) but, once again, a direct causal relationship between these genes and lung tumorigenesis is lacking. Similarly, studies investigating a link between mutations in enzymes that mediate the metabolism of tobacco-related carcinogens, particularly within the cytochrome P450 system (CYP1A1), and the formation of lung cancer have yielded conflicting results.

Further investigation into the loci proposed in these linkage studies may provide evidence for a more causal relationship and may allow for prevention initiatives targeted at these populations.

 

References:

Amos CI, et al. Nat Genet. 2008;doi:10.1038/ng.109.

Bell DW, et al. Nat Genet. 2005;doi:10.1038/ng1671.

Bergethon K, et al. J Clin Oncol. 2012;doi:10.1200/JCO.2011.35.6345.

Govindan R, et al. Cell. 2012;doi:10.1016/j.cell.2012.08.024.

Hung RJ, et al. Nature. 2008;doi:10.1038/nature06885.

Krewski D, et al. J Toxicol Environ Health A. 2006;doi:10.1080/15287390500260945.

Kwak EL, et al. N Engl J Med. 2010;doi:10.1056/NEJMoa1006448.

McKay JD, et al. Nat Genet. 2008;doi:10.1038/ng.254.

Nyberg F, et al. Cancer Causes Control. 1998;9:173-182.

Ohtsuka K, et al. J Clin Oncol. 2011;doi:10.1200/JCO.2010.31.4492.

Subramanian J and Govindan R. Lancet Oncol. 2008;doi:10.1016/S1470-2045(08)70174-8.

Taioli E, et al. Int J Epidemiol. 2003;doi:10.1093/ije/dyg001.

The Cancer Genome Atlas Research Network. Nature. 2014;doi:10.1038/nature13385.

Weir BA, et al. Nature. 2007;doi:10.1038/nature06358.

Wenzlaff AS, et al. Carcinogenesis. 2005;doi:10.1093/carcin/bgi191.

Matthew Bott, MD, is a thoracic surgeon at Memorial Sloan Kettering Cancer Center. He can be reached at bottm@mskcc.org. Disclosure: Bott reports no relevant financial disclosures.