LUNG FIELD CARCINOGENESIS

The original description of field carcinogenesis , or field cancerization
as it was originally called, dates back to Slaughter and
coworkers, 5 who showed that multiple foci of epithelial hyperplasia,
hyperkeratinization, atypia, dysplasia, and carcinoma
in situ occurred in otherwise normal-appearing epithelium
adjacent to cancers of the oral pharynx. This empiric evidence
strongly suggested that carcinogen exposure has widespread
effects throughout the entire carcinogen-damaged epithelial
field. These diffuse histological changes suggested that the development
of malignancy in the proximal upper aerodigestive
tract is a result of the dose of carcinogen impacting exposed
cell populations spread throughout this area. Auerbach and
colleagues 6 observed and demonstrated the same pattern of
heterogeneous, multifocal histological changes in the bronchial
epithelium of smokers, many of whom have lung cancer.
Thus, the concept of field cancerization or carcinogenesis was
coined to describe the diffuse damage done by chronic exposure
to carcinogens. Field carcinogenesis also forms the basis
for the observation that individuals who survive a first cancer
are often likely to develop a second malignancy in the region
of the condemned epithelium. 7–9
As these findings have been validated over the last several
decades, increasing evidence has come to bear explaining
the molecular genetic abnormalities that lead ultimately to the
stepwise development of clonal invasion-competent cancer of
the airway. Among the earliest findings in the aerodigestive
tract of chronic smokers are loss of heterozygosity in chromosomes
3p, 9p, and 17p 2,10–14 (see Chapter 6). Clonal genetic
abnormalities seen in normal histologic regions of the airway
and not necessarily in areas that had detectable changes were
further evidence that damage was scattered throughout the entire
field (see Chapter 19). It also brought to bear the limitations
of modern histologic techniques in detecting the extent
of the damage induced by tobacco smoke in the airway. The
importance of these losses of hererozygosity are crystallized in
the fact that, for example, the short arm of chromosome 3,
which is often the earliest loss and is demonstrated on occasion
even in early hyperplasia, 10 results in the loss of a region rich
in tumor suppressor genes. There is also evidence for losses of
the short arms of chromosomes 9 and 17, resulting in loss of
tumor suppressor genes in p16 and p53, both of which are
important to the cell’s ability to repair DNA damage done by
the tobacco carcinogen. p53 in particular, acts as a transcription
factor in the control of G1 arrest and apoptosis, or programmed
cell death, thereby allowing the cell to repair existing
DNA damage or make the determinant decision to push a cell
into the apoptotic process once it is too far damaged for repair.
p16, found on chromosome 9p, negatively controls cyclindependent
kinases (CDK)–cyclin activity by overexpression of
cyclin-D1. By inhibiting cyclin-D1/CDK, the damaged cell
is also prevented from entering into mitosis and proliferating
with damaged DNA (see also Chapter 14).
Increasing evidence shows that the loss of these and other
tumor suppressor genes in lung cancer is augmented by the activation
of several critical proto-oncogenes. For example, RAS,
c-MYC, and the epidermal growth factor receptors (EGFRs),
are all tumor-promoting genes that are activated progressively
during lung carcinogenesis. K- ras mutations are particularly
common in adenocarcinomas of the lung, with some studies
showing 30% of lung adenocarcinomas in smokers containing
K- ras mutations. 15 The EGFR is frequently amplified in
non–small cell lung cancers (NSCLCs) and in fact in lung
carcinogenesis 16 (see Chapter 49). Mutations have been detected
in the EGFR tyrosine kinase binding domain conferring
sensitivity to small molecule EGFR-tyrosine kinase inhibitors
(EGFR-TKIs). 17,18 Interestingly, both K- ras mutations and
EGFR mutations have been found in the airways of smokers
and nonsmokers in regions distant from the primary tumor,
suggesting that both play important roles in field carcinogenesis.
Thus, modern molecular tools can help us better understand
the process of field carcinogenesis and, as we will discuss,
implement the development of targeted agents that can arrest
or inhibit this otherwise inexorable progression.
Epigenetic Abnormalities in the Airway Loss
of function of tumor suppressor genes and greater function
of oncogenes through mutations or gene activations are not
the only mechanisms by which carcinogenesis is augmented
in lung cancer patients. A growing body of evidence suggests
that gene silencing through epigenetic means can be crucial in
lung carcinogenesis (see Chapter 7). For example, in NSCLC,
some tumor suppressor genes such as RASSF1A, also located
in the tumor suppressor gene–rich chromosome 3p, encodes a
protein that heterodimerizes with Nore-1, an important RAS
effector with pro-apoptotic effect. 19 Evidence from several
groups suggests that in NSCLC, RASSF1A can be frequently
inactivated by hypermethylation. 20,21 Another important gene
that can be inactivated by epigenetic means is the retinoic acid
r eceptor- (RAR- ), which also maps to chromosome 3p.
RAR- is a nuclear retinoid receptor with vitamin A– dependent
transcriptional activity. 22 RAR- is a gene that is shown to be
gradually lost in lung carcinogenesis, 23 and this can be regulated
either by loss or by hypermethylation. 24,25 Interestingly,
the maintenance of RAR- in mature tumors is a factor for
poor prognosis 26 and it is differentially regulated in current
versus former smokers. For example, a recent retrospective
study on methylation status and occurrence of second primary
tumors (SPTs) in completely resected NSCLC patients
revealed a strong association between RAR- 2 hypermethylation
in the development of SPTs only in former smokers. 27
In current smokers, hypermethylation was associated with a
protective effect, pointing to the critical value of context with
regard to smoking status in understanding the biological effects
of retinoid receptors in lung cancer. 27,28
Lung Cancer Susceptibility and Risk Stratification
Although there is little doubt that exposure to tobacco
smoke is the major risk factor for the development of lung cancer,
there is also progressive evidence that the risk of developing lung
cancer, even among the heaviest smokers, varies widely according
to genetic and perhaps even dietary factors. This is reinforced
by the fact that 85% of heavy smokers will not develop lung
cancer, implicating important differences in lung cancer susceptibility
across the population in the likelihood of developing overt
disease. However, there are significant competing risks for death
of heavy smokers, including cardiovascular disease, chronic obstructive
pulmonary disease, and other malignancies including
those of the upper airway, bladder, cervix, kidney, and pancreas.
Evidence exists, however, to show that every other smoker dies of
a smoking-related cause. 29 Despite this, accumulating evidence
suggests that genetic and epigenetic factors are critical in modulating
individual susceptibility to lung cancer or other consequences
of tobacco exposure. 30 The various carcinogens from cigarette
smoke, benzpyrine and 4-(methylnitrosoamino)-1-(3-pyridil)-1
butanone (NNK) require metabolic activation before they can
exert their full carcinogenic effects. Various activation pathways
compete with detoxification pathways and the balance between
the two is critical in modulating cancer risk. Cytochrome P450s
serve as carcinogen-metabolizing enzymes, whereas glutathione
transferases serve as detoxification enzymes. Both sets of these
important genes are known to have significant polymorphisms
correlating with variations in lung cancer risk (see Chapter 4).
Other important modulators of risk in lung cancer include
diet and gender. Dietary factors play an important role
as epigenetic modulators of lung cancer susceptibility. In several
case-controlled studies, defective detoxification and defective
repair of genetic damage have been shown to be associated
with increased individual susceptibility to lung cancer. 31–34
However, certain food constituents appear to afford a significant
degree of protection to individuals with limitations to their
detoxification capacity. 35 This relationship between diet and
lung cancer has been extensively explored and used as the basis
for the development of novel approaches to the prevention of
lung cancer. For example, large studies suggest that diets that
have a high intake of fruits and vegetables are associated with
a reduced risk of lung cancer. 36 Further evidence suggests that
a Mediterranean diet is associated with a substantial reduction
in several cancers, including lung cancer. 37
Basis of Clinical Strategies in Lung Cancer Prevention
As several dietary adjustments and micronutrients
have been associated with lower risks of lung cancer,
including vitamin E, selenium, isothiocyanates, and BMTpolyphenols,
several large randomized trials have sought to
delineate which compounds in complex human diets are
most responsible for the protective effects seen. 36,38 Initially,
the greatest degree of evidence existed around the potency
of carotenoids and retinoids and the broad epidemiologic
indications implicating them in reduction of cancer risk in
general, and lung cancer risk in particular. 39,40 In fact, the
original definition of chemoprevention coined by Michael
Sporn as “attempts to reverse, suppress, and prevent carcinogenic
progression to overt cancer” 41 is based on his and
other’s experimental work showing that vitamin A analogues
are capable of reversing or preventing epithelial carcinogenesis
in mouse models of cancer.
Consequently, the identification of populations at risk for
development of this disease followed suit. For example, target
populations in many of the early prevention studies included
broad populations of individuals at risk because of their exposure
to tobacco and/or asbestos. Others sought to develop
strategies for uranium miners, also a population believed to
be at enhanced risk for developing lung cancer. Thus, with
several decades of broad epidemiologic and dietary investigation,
augmented by important experimental systems showing
several classes of compounds, especially the retinoids and carotenoids,
were capable of reversing cell damage, a strong drive
to develop this approach in human populations was born. The
trials focused on broad patient populations, including those
of individuals at even higher risk, in other words, individuals
with known premalignancy of the airway, or those patients
who had already developed a primary tobacco-related cancer,
and thus were at the greatest risk for developing a second primary
tumor.
Clinical Chemoprevention Trials Laudable as smoking
cessation efforts have been, it is clear that lung cancers are
developing in former smokers. 42,43 Thus, the design of lung
cancer trials has taken the approach to develop strategies based
on the different populations considered for each intervention.
Risk categories were different across these populations, and
thus the categories of chemoprevention were defined as primary,
secondary, and tertiary prevention (Table 21.1). Primary
prevention involved intervening in patient populations at increased
risk but with no histological or clinical evidence of
cancer. The end point here was reduction in incidence of lung
cancer, and reduction in death from lung cancer. Secondary
prevention involved attempts in individuals with evidence of
lung premalignancy to prevent progression of that premalignancy
or, ideally, to reverse it to an earlier stage of carcinogenesis.
Finally, studies in the highest risk population were
undertaken in those individuals who had already developed
tobacco-related malignancy. Here, the goal was to prevent the
development of SPTs. As expected, many of the early trials focused
on using retinoids and carotenoids as the chemoprevention
agents of choice, given that the weight of epidemiological
and experimental evidence for these compounds was significantly
greater than that for other classes of compounds.
Primary Chemoprevention Studies Several large randomized
studies were undertaken in populations deemed to be
at increased risk of lung cancer caused by exposure to tobacco,
asbestos, or their occupation as uranium miners. Three major
studies, the -tocopherol/ -carotene (ATBC), -carotene and
retinol efficacy trial (CARET), and the U.S. Physician Study,
were carried out using increasingly high doses of the carotenoid,
-carotene (Table 21.2). None of the three studies showed any
reduction of lung cancer risk utilizing these compounds, and
two of these studies, the ATBC and CARET studies, had surprising
findings, both showing significant increases in lung cancer
risk associated with supplementation with -carotene. Both
of these studies, which were 2 plus 2 factorial studies, also involved
a second agent as an intervention. In these cases neither
the second agent nor vitamin A itself was found to be associated
with an increased risk of lung cancer, but neither were they protective
against lung cancer incidence. Interestingly, both of these
studies showed that the risk was increased only in populations
that were heavy smokers, and particularly so in those individuals
who had evidence of pulmonary asbestosis, as measured by
manifestation of pleural plaques on chest radiograph.
The ATBC cancer prevention study was a randomized
2 by 2 factorial, double-blind, placebo-controlled, primary prevention
study in which 29,133 Finnish male smokers received
either -tocopherol (50 mg/day alone), -carotene (20 mg/day),
both -tocopherol and -carotene, or a placebo. The participants,
between 50 and 69 years of age, all of whom smoked
at least five cigarettes per day, were enrolled on the study and
received follow-up observation for 5 to 8 years. Although lung
cancer incidence, the primary end point, was not modified by
the addition of -tocopherol alone, both groups who received
-carotene supplementation either alone or with -tocopherol
had an 18% increase in the incidence of lung cancer and a significant
increase in lung cancer mortality. This study showed
a stronger adverse effect from -carotene in men who smoked
more than 20 cigarettes per day, and was the first to raise the
concern that pharmacologic doses of -carotene could be
harmful in active smokers. A recent update suggested that the
excess risk for -carotene recipients was no longer evident 4 to
6 years after any intervention and at that point, morbidity was
largely caused by cardiovascular diseases. 39,44
The CARET confirmed the results of the ATBC trial.
This was also a randomized, double-blind, placebo-controlled
trial testing the combination of 30 mg/day of -carotene and
25,000 IU of retinyl palmitate in 18,314 men and women ages
50 to 69 years old who were considered at increased risk for
lung cancer. The majority of participants had a smoking history
of at least 20 pack-years, and were either current smokers
or recent former smokers. Significant and even extensive occupational
exposure to asbestos was noted in 4060 men on
this trial. 45 The trial was stopped early by the Data Safety
Monitoring Committee because of evidence of possible harm,
consistent with the ATBC study. In fact, lung cancer incidence,
the primary end point, increased 28% in the active intervention
group, with an increase in overall mortality of 17% in
this group. 46
By contrast, the Physicians Health Study a randomized,
double-blind, placebo-controlled trial studied 22,071 healthy
male physicians, half of whom received 50 mg/day of -carotene
on alternate days and the other half received placebo. The use
of supplemental -carotene in this study comprising a majority
of nonsmokers showed virtually no adverse or beneficial effects
on cancer incidence or overall mortality rate during a 12-year
follow-up. 47
Subsequent subgroup analyses of the ATBC and the
CARET studies have indicated that -carotene had a harmful
effect only in the high-risk heavy smokers, or those with previous
exposure to asbestos. 48
Reversal of Premalignancy or Secondary Prevention
Although there has been much dispute over which
are the optimal premalignant markers of lung cancer to follow,
the success in developing chemoprevention agents in this
arena has been extremely limited. To date, eight randomized
trials have used various end points, including reversal of sputum
atypia, reduction in DNA micronuclei, and reversal of
dysplasia or hyperplasia (Table 21.3). Some of these trials have
used retinoids, and have shown that in the absence of smoking
cessation, retinoids are incapable of reversing premalignant lesions.
Although some biomarker-driven studies utilizing novel
pan-retinoids, such as 9- cis -retinoic acid, or atypical retinoids
such as fenretinoid, have shown that these agents can modulate
biomarkers such as RAR- or human telomerase reverse transcriptase
(hTERT) expression, respectively, there is little evidence
to suggest that any of the compounds listed in Table 21.3
are capable of consistently reversing premalignant lesions. To
date, one of the most positive studies has been one that utilized
folate and vitamin B 12 , which showed some improvement in
bronchial epithelial metaplasia in smokers. Given the difficulty
with validation of the end points, even these positive results
must be viewed with caution. Larger trials using biologic end
points are needed to confirm efficacy. 49
Although several trials examining reversal of premalignancy
in upper aerodigestive tumors with high doses of retinoids
had shown significant initial efficacy, results have yet to
be duplicated using retinoids in reversal of lung preneoplasia.
Most importantly, findings from a trial by Lee et al. 50 indicated
that retinoids could be effective in the presence of smoking
cessation, but were highly ineffectual in active smokers.
Recent studies employing novel retinoids have indicated that
they may have significant promise. One such trial by Kurie
et al. 51 reported the results of a randomized controlled trial
in former smokers who received either 9- cis -retinoic acid or
13- cis -retinoic acid with - tocopherol. The end point of this
trial was upregulation of RAR- , the loss of which is gradually
seen in pulmonary carcinogenesis. Of 177 evaluable patients,
those treated with 9- cis -retinoic acid were found to have
restoration of RAR- expression (p 0.03) and reduction of
metaplasia (p 0.01). There was no significant effect in the
13- cis -retinoic acid with -tocopherol arm, encouraging this
group of investigators to move forward with 9- cis -retinoic acid,
a pan-retinoid agonist, targeting former smokers. Similar results
were seen with fenretinide, an atypical retinoid that does
not bind to any of the cognate X retinoid receptors, which
was shown to downregulate hTERT, a critical component in
human telomerase. 52 These promising data, albeit preliminary,
suggest that continued investigation with novel retinoids may
have substantial merit.
Second Primary Tumor Prevention Patients with tobacco-
related cancers who have undergone successful treatment
remain at a substantially elevated risk for developing subsequent
tumors in the tobacco-damaged epithelium. 7,8,53–57 Although
the treatment for the initial cancer can often be successful, these
patients are at dramatically increased risk for developing and
dying from a second primary tumor 58 (Table 21.4).The concept
of SPTs initially described by Warren and Gates 59 explains
the high likelihood of multiple oral and pharyngeal premalignancies,
both synchronous and metachronous, disseminated
throughout the condemned epithelium. Although modern molecular
work by Sidransky et al. 60 has shown that the clonal
origin of some of these SPTs can be indeterminate, some are
a result of clonal evolution and the malignant clone can be
found throughout the epithelium. 61,62 Clinical data continue
to suggest that this remains a devastating long-term problem
for patients definitively treated for primary tumor. The Warren
and Gates criteria defined SPT as one that (a) is a new cancer
of a different histological type, or (b) is a cancer, regardless of
site, that occurs after more than 3 years, and (c) if it is also located
in the head and neck region, the lesion is separated from
the initial primary tumor by at least 2 cm of clinically normal
epithelium, and (d) if found in the lung, and of squamous histologic
s ubtype and developing within 3 years, it presents as a
solitary mass with no evidence in the patient of local or regional
disease, with changes consistent with dysplasia or carcinoma in
situ within the surrounding bronchial epithelium. Using these
criteria, the risk of local recurrence was shown by Vikram 58 to
decline over time, whereas the risk of SPTs is constant for approximately
the first decade following the initial head and neck
cancer. Therefore, the lifetime risk of developing a SPT in the
head and neck region is approximately 20%, with a similar incidence
in the lung. Although estimates have varied between 3%
to 7% per year, evidence remains incontrovertible that SPTs
are the major cause of death after curative surgery in head and
neck cancer, and can be a major cause of death in early stage
disease. 56–58,63–65
Because of the high likelihood of both recurrence and
SPTs in patients with advanced oral, oropharyngeal, or laryngeal
squamous cell cancers, Hong et al. 66 launched a randomized,
placebo-controlled study of 103 patients with stage I to IV
head and neck squamous cell cancer, definitively treated with
either previous surgery and/or radiation therapy. The patients
were randomized to receive either high dose 13- cis -retinoic acid
(100 mg/m 2 /day) or placebo for 1 year after definitive local
therapy. The dosage of 13-cRA was reduced to 50 mg/m 2 /day
after 13 of the first 44 patients experienced intolerable side
effects. The primary end points were primary tumor recurrence
and SPT development. In the two treatment arms, there was no
difference in local recurrence or distant metastases. However,
the patients treated with 13-cRA had a dramatically lower
incidence of SPTs. Of the 103 patients followed for a median
of 42 months, SPTs developed in 6% (3 of 49) of those in the
13-cRA arm, whereas 28% (14 of 51) of the patients on the
placebo arm developed SPTs. Consistent with the findings of
field carcinogenesis, the vast majority of the SPTs developed
in the upper aerodigestive tract, esophagus, and lung, and 14
of 17 were found to be histologically of squamous cell type.
Despite only 47% of the 13-cRA–treated patients completing
the therapy as prescribed, the reduction in SPT development
was still significant. Long-term follow-up has also revealed that
the effects of 13-cRA decreased over time. 67
A large scale follow-up to this trial using a much lower
dose of 13-cRA was completed and presented a few years ago.
In this trial, a randomized, double-blind, placebo-controlled
study, launched in 1991, more than 1382 patients were registered
and 1192 were randomized to either low dose 13-cRA
at 30 mg/day versus placebo. These patients were definitively
treated for stage I or II head and neck squamous cell cancer.
With a median follow-up of 7 years, no effect was seen of the
low dose retinoid on reducing the incidence of SPT in the lung
or aerodigestive tract. 68
Several other phase III trials have been launched in lung
cancer in an attempt to prevent SPTs. The first of these was
a trial by Pastorino et al. 69 that randomized over 300 patients
with early stage lung cancer to retinyl palmitate or placebo. This
study showed a significant reduction in the development of second
primary lung cancers on the retinyl palmitate arm. A subsequent
study by Bolla et al. 70 using a different synthetic retinoid,
etretinate, failed to show a significant reduction in SPTs.
Two large phase III trials reported in the last decade include
the EUROSCAN and the U.S. Intergroup NCI trial.
EUROSCAN, a randomized adjuvant chemoprevention study
of the European Organization for Research and Treatment of
Cancer (EORTC), head and neck and lung groups studied the
effects of vitamin A (retinyl palmitate) and N-acetylcysteine
(NAC) in patients with early stage head and neck and lung
cancer. 71 In this trial 2592 patients with cancers of the larynx,
TIS-T3 and 0-N1, oral cavity, TIS-T2 and 0-N1, and NSCLC,
T1-T2 and 0-N1, received retinyl palmitate, 300,000 IU per
day in year 1, 150,000 IU/day in year 2, NAC 600 mg/day for
2 years, both drugs, or placebo. No significant differences were
seen between the three active treatment arms and the placebo
group in terms of recurrence rates, SPT development, or survival.
More than 90% of the patient population was regular
smokers, with 43 median pack-years of tobacco exposure.
The U.S. Intergroup NCI 91-0001 trial was a randomized,
double-blind, placebo-controlled study utilizing the same
low dose of 30 mg 13-cRA after complete resection of stage I
NSCLC. 72 This trial completed accrual in 1997, having accrued
1486 participants. The study objectives were to evaluate the
e fficacy of low-dose daily 13-cRA for 3 years at 30 mg/day versus
placebo in the prevention of SPTs. Patients were required to have
complete resection of primary, stage I NSCLC (postoperative
T1 or T2 and 0) and registered between 6 weeks to 3 years after
completion of therapy. After a median follow-up of 3.5 years, no
statistically significant differences were seen between the placebo
and 13-cRA arms with respect to time to SPT development,
recurrence rate, or mortality. Multivariate analyses showed that
the rate of SPTs was not affected by any stratification factor, with
recurrence rate affected only by treatment stage, and evidence
for a treatment-by-smoking interaction (hazard ratio [HR] for
treatment-by-current versus never-smoking status 3.11, 95%
confidence interval [CI], 1.00 to 9.71). Therefore, 13-cRA was
not shown to affect overall survival rates or SPTs, recurrences,
or mortality in stage I NSCLC. Subsequent subset analyses have
indicated that 13-cRA was potentially harmful in current smokers
and beneficial in never-smokers.
Finally, researchers at Yale University conducted a randomized,
double-blind, placebo-controlled trial studying the
effects of -carotene at 50 mg/day in reducing local recurrence
and SPTs in head and neck cancer. 73 Two hundred
sixty-four patients with curatively treated early stage squamous
cell carcinoma of the oral cavity, pharynx, or larynx
were randomized to either 50 mg/day of -carotene or placebo
and were followed for 90 months for the development
of SPTs and local recurrences. After a median follow-up of
51 months, no difference was seen between the two groups
in time to failure, local recurrences or SPTs. Supplemental
-carotene had no effect on overall mortality, SPT rates, or
rates of local recurrences.