Women and lung cancer: Epidemiology, tumor biology, and emerging trends in clinical research

1. Introduction 

Women's healthcare issues related to breast cancer and gynecological malignancies have received substantial attention and research funding in the past two decades. However, differences between men and women in the development and prognosis of common tumor types, such as melanoma [1], colorectal [2], or lung cancer [3], indicate the need for broader sex-specific research to optimize treatment options for men and women. These differences in tumor etiology may be influenced by physiological factors, such as the role of sex hormones, or by behavioral and environmental factors, such as levels of sun exposure or smoking habits. Overall, the variability in tumor development is based on a complex interaction between genetic, hormonal, and behavioral factors.

Lung cancer is the leading cause of cancer-related death in women. An estimated 73,000 women will die from progressive lung cancer in 2005, accounting for 27% of all cancer death in US women (compared to 15 and 6% of cancer deaths due to breast and ovarian cancer, respectively; ref. [4]). Increasing evidence suggests that the relative risk for lung cancer, the relationship between smoking and lung cancer, and response to therapy may be different for women and men. In February 2006, the Society for Women's Health Research hosted a thought leaders’ roundtable to review issues and advances related to sex differences in the development and treatment of lung cancer. Recent data on the epidemiological aspects of lung cancer in women, lung tumor biology, and emerging trends in clinical research were presented. The discussions, summarized in this review, clearly highlight the need for basic and clinical research to translate information on differences in tumor biology and risk factors between men and women into preventative, diagnostic, and therapeutic practice to improve women's healthcare.

2. Men, women, and health 

Historically, women have been underrepresented in clinical research. In 1977, the US Food and Drug Administration (FDA) restricted research on women of childbearing potential in reaction to the birth defects resulting from thalidomide and diethylstilbestrol administered during pregnancy. As a result, findings from clinical studies on men were extrapolated to female patients, not taking into account potential differences in pharmacokinetics, pharmacodynamics, and safety between the sexes [5], [6]. In addition, drugs that are potentially beneficial to women may have been eliminated in early phases of clinical testing when the test group did not include women and no benefits were manifest in male subjects [7]. In 1993, the FDA published new guidelines encouraging participation of women in clinical studies and also requiring study sponsors to collect and analyze data separately by sex to identify any differences in drug response and toxicity. However, women are still underrepresented in cancer trials [8], [9], and for most drugs currently on the market, prospective evaluation of efficacy and toxicity by sex is not available.

Sex clearly affects health as men and women have different patterns and presentations of an illness, as well as statistically different life-spans [10], [11]. The phenotypic differences between males and females are determined, initially, by genes on the sex chromosomes. Male sex is initiated and determined by the presence of the SRY-gene on the Y-chromosome. The expression of these differences in the context of physiology and endocrine function will vary throughout childhood, adolescence, adulthood, and post-menopause. On average, women, relative to men, have a higher percentage of body fat, smaller muscle mass, lower blood pressure, higher levels of estrogens and progestins, and lower levels of androgen (reviewed in ref. [5]). In addition, women, but not men, undergo hormonal fluctuations associated with reproduction (such as the ovarian cycle and pregnancy) that influence numerous body systems. During menopause, reduced estrogen levels affect not only the ovaries, uterus, and breast, but also other tissues such as the brain, skin, cardiovascular tissue, and bone [12]. Consequently, sex should be considered an important variable in research design, both in basic and clinical research as well as epidemiology.

Clinical research should also take into consideration that differences between men and women may vary with age and endocrine status. When reporting research findings in the medical literature, analyses by sex and endocrine status should be included whenever possible, even in the absence of significant differences in outcome between the sexes. A clearer understanding of the function of sex hormones in body systems other than the reproductive organs will allow for more individualized options for the prevention, diagnosis, and treatment of disease. A growing awareness and knowledge of biological sex differences will have an impact on healthcare and health policy, and improve outcome for both men and women.

3. Lung cancer: epidemiology and risk factors 

In 2005, 172,570 patients (93,010 men and 79,560 women) were diagnosed with lung cancer, and 163,510 patients (90,490 men and 73,020 women) were expected to die of this disease [4]. Because lung cancer is asymptomatic in its early stages, only 18% of cases are still confined to the primary site at diagnosis (local disease); 38% are diagnosed after the cancer has spread to regional lymph-nodes or directly beyond the primary site (regional disease), and 37% are diagnosed after the cancer has already spread to other parts of the body (metastatic disease). The corresponding 5-year relative survival rates are 51.3% for localized, 17.1% for regional, and 2.1% for metastatic disease [13]. Women have a better relative survival compared to men for each stage of the disease (Table 1A; ref. [14]), and male sex is a consistently unfavorable prognostic indicator in advanced disease (Table 1B; refs. [15], [16], [17], [18], [19], [20]). In women, smoking status prior to lung cancer diagnosis and duration of smoking abstinence are predictive of survival: risk of death was higher for current or past female smokers compared to never-smokers, a trend that was not observed in men [21], [22]. Overall, women diagnosed with lung cancer are more likely to be non-smokers than male patients, and women are also overrepresented among younger patients [23], [24], [25], [26].

Table 1A. 5-year survival rates in men and women by stage in NSCLC and SCLC [13]

		NSCLC		SCLC
		Men (%)	Women (%)	Men (%)	Women (%)
	All stages	14.9	19.6	5.6	7.2
	Local disease	46.6	56.0	20.7	21.6
	Regional disease	15.9	18.5	9.9	11.7
	Metastatic disease	2.0	2.3	1.8	2.2

Table 1B. Male sex is a poor prognostic factor in advanced lung cancer

		N	%Female	Hazard ratio	P-value
	O’Connell et al. [15]	378	30	0.71	0.001
	Ferguson et al. [16]	299	45	0.75	0.044
	Shinkai et al. [17]	192	29	0.62	0.02
	Albain et al. [18]	2531	23	0.77	<0.0001
	Paesmans et al. [19]	1052	10	0.70	0.03
	Johnson et al. [20]	378	28	0.77	0.002

Lung cancer historically has been more prevalent in men than women; however, the male/female incidence ratio has narrowed dramatically, from 3.65 in 1975 to 1.65 in 1999, as the incidence rate in men declined while the rate in women continued to rise slowly [14]. This rising incidence of lung cancer in women is primarily due to an increase in tobacco use which started in the 1940s. The lung cancer death rate in women has subsequently increased rapidly since the 1960s, from about 5 cases per 100,000 women to an estimated 40 per 100,000 in 2000. Today, lung cancer is the most common cause of cancer death in women (27%), claiming more lives than breast and colorectal cancer combined (15 and 10%, respectively; ref. [4]).

Reducing the incidence of lung cancer death in women requires renewed efforts towards smoking cessation and prevention. Although smoking rates in women have dropped since the 1960s, 19% of women ≥18 years old were current smokers in 2004, compared to 23% of men [27]. Moreover, a near equal prevalence of cigarette smoking has been reported in male and female adolescents [27], [28]. Factors associated with the initiation of cigarette smoking in young adults vary between the sexes; particularly, in young women, vulnerability to depression and weight concerns are associated with an increased risk of smoking initiation [29], [30]. Future intervention programs may be able to specifically target young women at increased risk for smoking initiation or continuation.

Lung cancer histologies include squamous cell carcinoma, adenocarcinoma, small cell carcinoma, and large cell carcinoma. Non-small cell lung cancers (NSCLC) account for approximately 85% of lung cancer, and include the histological subtypes squamous cell carcinoma and large cell carcinoma, which arise from epithelial cells lining the bronchi, and adenocarcinoma, which develops from glandular tissue in the peripheral regions of the lung. The proportional occurrence of these histological subtypes differs significantly between men and women [14]. Adenocarcinoma is currently the most common histological subtype in both men and women, and women have proportionally more adenocarcinoma and less squamous cell carcinoma compared to men (Table 2A). Cigarette smoking has been linked to all four histological subtypes; however, the proportion of non-smokers is highest in those who develop adenocarcinoma (Table 2B; ref. [33]).

Table 2A. Proportion of patients diagnosed with either squamous cell carcinoma or adenocarcinoma by sex

		Adenocarcinoma		Squamous cell carcinoma
		Men (%)	Women (%)	Men (%)	Women (%)
	Moore et al. [31]	40	50	30	20
	Fu et al. [14]	33	45	30	21
	Visbal et al. [24]	48	59	32	22
	Chen et al. [32]	42	73	41	16
	Ringer et al. [26]	38	47	38	24

Table 2B. Proportion of patients diagnosed with lung cancer by sex, histological subtype, and smoking status (based on Muscat and Wynder; ref. [33])

		Men			Women
		Current smokers (%)	Former smokers (%)	Never-smokers (%)	Current smokers (%)	Former smokers (%)	Never-smokers (%)
	Adenocarcinoma	32	34	58	42	44	59
	Squamous cell carcinoma	35	37	19	20	20	12
	Small cell carcinoma	15	11	0	19	12	3
	Other	18	18	23	19	24	26

The incidence rates for the various histological subtypes have changed over time and reflect changes in smoking habits. For example, incidence rates for squamous cell carcinoma decreased in men between 1975 and 1999 while increasing slightly in women. On the other hand, the incidence rate of adenocarcinoma increased in both in men and women during that same time period, with a greater increase observed in women [14], [34]. The increase in adenocarcinoma has been associated with the introduction of low-tar cigarettes that enhance delivery of smoke to peripheral regions of the lungs [35].

While smoking is major risk factor, only a minority of smokers develop lung cancer. In addition, approximately 10–15% of patients diagnosed are non-smokers. The ratio of women to men in patients with lung cancer who have never smoked is approximately 3:1 [24], [36]. The risk for developing lung cancer in both smokers and non-smokers is modulated by determinants that may vary between individuals. For example, data from the Environmental Protection Agency showed that environmental tobacco smoke accounts for approximately 3000 lung cancer deaths in the US each year among non-smokers [37]. The development of lung cancer despite the low level of exposure suggests that a subset of the general population may be more susceptible to the carcinogenic effects of tobacco smoke.

Susceptibility is likely to be determined by each individual's capacity to activate and detoxify carcinogens in tobacco smoke. Certain genetic polymorphisms in genes involved in the metabolism of carcinogens have been associated with an increased risk of lung cancer in smokers and never-smokers [38], [39], [40]. Further evidence for genetic susceptibility comes from studies on family history of lung cancer. A positive family history has been defined as a risk factor in non-smokers; particularly in the development of adenocarcinoma, in women, and in cases with an earlier age at onset [41], [42], [43], [44]. Finally, an epidemiologic link between estrogen exposure and incidence of NSCLC has been suggested in Japanese non-smoking women [45]. Younger, presumably pre-menopausal, women appear to have shorter survival than older women: in an analysis of patients with advanced NSCLC enrolled in SWOG trials, women over 70 had a 34% 1-year survival compared with 11% for those under 45 [18]. Clearly, the multifactorial nature of the disease requires an improved understanding of the interplay between sex, environmental factors, and genetics [46], [47].

4. Sex differences in lung tumor biology 

The available epidemiological data clearly demonstrate sex-specific differences in lung cancer susceptibility and prognosis. Women appear to have an increased susceptibility to tobacco carcinogens, but have a lower rate of fatal outcome of lung cancer compared to men [48]. A biological explanation for these epidemiological findings is beginning to emerge as the comparison of lung tumors from male and female patients has identified various tumor characteristics that may affect tumor etiology.

Smoking-related lung cancer is induced by the formation of DNA adducts in lung epithelial cells [49]. Processing of these adducts by the cellular DNA repair machinery may lead to mutations in genes that initiate or facilitate tumor growth, such as the tumor suppressor gene p53. The formation of DNA adducts is dependent on the enzymatic activation of several tobacco-related pro-carcinogens, including polycyclic aromatic hydrocarbon (PAH; ref. [50]). PAH activation is catalyzed by the cytochrome P450 enzymes CYP1A1 and CYP1B1, and inactivated by glutathione S-transferases (GSTs). The expression of CYP1A1 and CYP1B1 in lung tissue is significantly higher in current smokers compared to former and never-smokers [51]. Levels of lung DNA adducts correlate with the level of CYP1A1 expression. Among smokers, female patients had a 3.9-fold higher median level of CYP1A1 compared to males. In non-smokers, certain CYP1A1 polymorphisms are associated with an increased lung cancer risk, and this association can be modulated by ETS and genetic variations of GSTs [38], [52], [53]. Independent of smoking history, the combined variant genotype of CYP1A1 and GSTM1 contribute to an increased risk of lung cancer in women compared to men (odds ratio: 6.54 versus 2.36; ref. [54]).

The carcinogens present in tobacco smoke are associated with specific mutations at a relatively small number of codons of tumor suppressor p53; active cigarette smoking is positively correlated with the total frequency of p53 mutations [55]. The most common p53 mutations in lung cancer are G:C to T:A transversions, a type of mutation rarely found in other tumor types. In contrast, transversions from G:C to A:T are more common in lung tumors from non-smokers and other types of cancer [56]. The prevalence of G:C to T:A transversions is approximately 30% in smokers, compared to 10% of non-smokers. The prevalence of G:C to A:T transversion is approximately 28% in smokers and 50% in non-smokers [55], [56], [57]. Analysis by sex shows that the tobacco-related p53 mutations are more common in women than men, evidence that women may be more susceptible to the carcinogenic effects of tobacco smoke [56], [57], [58], [59]. In addition, the G:C to T:A transversion rate is significantly lower, and the G:C to A:T transversion significantly higher, in female non-smokers compared to female smokers [57]. The dramatic shift in the p53 mutational spectrum between women who are non-smokers and those who smoke suggests that alternate tumorigenic pathways may be activated depending on exposure to tobacco smoke. If so, tumors from non-smoking women may be biologically different than those from male or female smokers.

The heterogeneous nature of lung cancer first became evident when clinical responses to gefitinib, a tyrosine kinase inhibitor that targets the epidermal growth factor receptor (EGFR), were noted in patients with a mutated form of EGFR. Mutations in the tyrosine kinase domain of EGFR constitutively activate downstream signaling in the absence of ligand binding, and are associated with cellular proliferation and resistance to apoptosis. In lung cancers, these activating mutations are more common in women and non-smokers [60], [61], [62], [63]. Mutations in K-RAS, a downstream signaling molecule in the EGFR pathway, may also be more common in women, but are significantly associated with smoking and levels of DNA adduct formation [64], [65].

In hormone-dependent tissues, such as breast and ovary, estrogen regulates the transcription of target genes through interaction with estrogen receptors (ERs) alpha or beta. While important for normal tissue physiology, estrogen receptor-mediated signaling also promotes tumorigenesis, either in response to estrogen-binding or independent of estrogen through the expression of growth factors, particularly EGFR [66], [67], [68], [69], [70]. Similarly, estrogen plays a role in both normal pulmonary physiology and in the biology of non-small cell lung carcinoma (NSCLC) [71], [72], [73], [74]. In vitro studies confirm that NSCLC cells respond to estrogens and anti-estrogens by altering endogenous gene expression [75]. ER-beta, and to a lesser extent ER-alpha, are expressed in lung tumors from both men and women. Expression of ER-beta is associated with improved survival, while expression of ER-alpha is a poor prognostic factor [68], [73], [76], [77]. In a large series of surgically resected NSCLC tumors, ER-beta was detected in 45.8% of cases. Overexpression of ER-beta was significantly more common in tumors from non-smokers (53.5%) than smokers (36.6%, P<.004). Among non-smokers, higher ER-beta expression was observed significantly more frequently in female patients (58.3%) than in male patients (40.9%) [73].

Current data related to tumor biology have started to provide an explanation for the risk factors associated with the development of lung cancer. Particularly, lung tumorigenesis may progress via different pathways in men and women depending on the metabolisms of tobacco-related carcinogen, molecular abnormalities, and hormonal status. A better understanding of the biological and molecular aspects contributing to these differences may allow the development of more effective therapy specifically tailored for men or women.

5. Current and experimental treatment options 

Initial treatment for patients with localized, early stage NSCLC typically includes surgical resection; patients with limited pulmonary reserve are candidates for radiotherapy. The treatment modality for early stage disease differs significantly between men and women (Table 3). In general, a higher proportion of women undergo surgery as part of a treatment compared to men, while radiotherapy is more frequently administered to men. Although surgery is the most effective treatment option in early stage disease, the lower frequency of surgical resection in men does not explain the observed survival benefit for women as a comparison of patients who received surgery as their initial treatment showed that women have better survival than men [14], [31], [78].

Table 3. Differences in treatment modality by sex

		Women (%)	Men (%)	P-value	Survival
	Surgery only
	Moore et al. [31]	27.4	24.7	.031	Median: 55.4 months vs. 37.2 months (P<.001)
	Radzikowska et al. [101]	23.3	18.8	.0001	NA
	Fu et al. [14]
	Local	63.7	56.4	<.0001	2-year: 74.3% vs. 66.0% (P<.0001)
	Regional	23.4	22.2
	Radiotherapy only
	Moore et al. [31]	17.7	19.6	.031	Median: 10.1 months vs. 8.0 months (P=.004)
	Radzikowska et al. [101]	14.5	14.6	NS	NA
	Fu et al. [14]
	Local	18.2	23.4	<.0001	2-year: 12.2% vs. 15.5% (P<.0001)
	Regional	39.4	42.5

A 1995 meta-analysis reported that in advanced metastatic NSCLC, platinum-based chemotherapy provides a survival benefit compared with best supportive care [79]. Platinum combinations including paclitaxel, docetaxel, vinorelbine, or gemcitabine, all offer similar outcomes for the first-line treatment of patients with advanced disease. Non-platinum combinations, such as paclitaxel or docetaxel plus gemcitabine, also result in similar survival [80], [81], [82], [83], [84]. Clinical factors that influence the outcome of advanced NSCLC patients treated with standard first-line chemotherapy include skin or liver metastases, loss of appetite, low performance status, four or more metastatic sites, and no prior surgery [85]. Clearly, novel treatment approaches are required to improve survival in advanced NSCLC.

The role of angiogenesis is well established in lung cancer tumor progression. High microvessel density and expression of the pro-angiogenic factor vascular endothelial growth factor (VEGF) are associated with metastasis and poor survival [86], [87]. Targeted inhibition of VEGF prevents the formation of new tumor vasculature and may also produce vascular normalization, allowing enhanced delivery of cytotoxic therapies to the tumor [88]. Bevacizumab is a humanized monoclonal antibody that acts by binding and neutralizing all VEGF isoforms. The efficacy and safety of bevacizumab in combination with carboplatin and paclitaxel has been assessed in patients with advanced or recurrent NSCLC (non-squamous only) in the Eastern Cooperative Oncology Group (ECOG) E4599 study. The results demonstrated a significantly increased response rate in favor of the addition of bevacizumab (10% versus 27%, P<0.0001) and increased progression-free survival and overall survival [89]. A subset analysis of survival by sex, however, showed that the survival advantage associated with the addition of bevacizumab was male-specific (8.7 months versus 11.7 months, P=0.001); women had similar survival in both arms, despite a significant improvement of response rate and time to progression in the bevacizumab-arm [90]. Possibly, survival may have been confounded by second- or third-line treatment as the EGFR-inhibitors are particularly active in certain patient subsets as discussed below.

The small molecule inhibitors gefitinib and erlotinib prevent activation of EGFR signaling pathways by binding to the EGFR tyrosine kinase domain. As a single agent, these tyrosine kinase inhibitors induce responses in approximately 10% of relapsed patients with advanced NSCLC. In the case of gefitinib, single agent treatment in relapsed patients or combination treatment with standard chemotherapy in the first-line setting were not associated with improved survival [91], [92], [93]. Erlotinib improved survival as a single agent compared to placebo in previously treated patients, but neither erlotinib or gefitinib improved survival in combination treatment with standard chemotherapy [94], [95]. While gefitinib and erlotinib fail to show activity in most patients, a dramatic and durable response is observed in a subset of patients, particularly women, Japanese, non-smokers, and patients with adenocarcinoma [96]. Correlative studies established a relationship between the presence of EGFR-mutations and sensitivity to gefitinib; EGFR-mutations are present in 20% of female patients compared to only 9% of male patients [62], [97]. The activity of gefitinib in selected patients is illustrative of the heterogeneous nature of lung cancer and the need to define treatment options by both clinical characteristics and tumor biology.

Paclitaxel poliglumex (PPX) is a macromolecular drug conjugate that links paclitaxel with a biodegradable polymer, poly-l-glutamic acid [98]. PPX is stable in systemic circulation and takes advantage of hyperpermeable tumor vasculature and suppressed lymphatic clearance to passively accumulate in tumor tissue (enhanced permeation and retention (EPR) effect). The efficacy and safety of PPX as a single agent or in combination with carboplatin was evaluated in chemotherapy-naïve patients with advanced NSCLC and ECOG performance status 2 (PS2) in two phase III studies, enrolling over 800 patients [99], [100]. The primary efficacy endpoint of both studies was survival. While patients receiving PPX had similar overall survival compared with the control arms, a trend towards improved survival was noted for female patients receiving PPX compared to female patients in the control arm. In contrast, survival was similar for male patients, regardless of treatment. Estrogen could potentially modulate the efficacy of PPX through the enhanced activity of cathepsin B, a rate-limiting enzyme in PPX metabolism. Alternatively, estrogen may enhance the distribution of PPX to ER-beta expressing tissues, such as lung. As a follow-up on these observations, a female-only phase III trial is underway to compare the efficacy of PPX to that of paclitaxel in women with advanced NSCLC and PS2 (PIONEER). This is the first study to take advantage of sex-based differences in lung tumor biology to optimize treatment for a subset of patients, in this case women.

6. Conclusion 

The reviewed epidemiological and biological lung cancer data highlight the impact of sex as a variable on various aspects of the disease:

To further improve survival in this disease, the panel made the following recommendations:

Conflict of interest statement 

The roundtable was hosted by the Society for Women's Health Research and supported by an unrestricted educational grant provided by CTI Inc., Seattle. No conflicts of interest were declared by the authors.