SECONDARY PREVENTION OF LUNG CANCER

Over 210,000 new cases of lung cancer will be diagnosed in 2008.
Of these, only 20% are surgically resectable. 1 Of this subset, only
15% are stage I lesions, that is, lesions that have not progressed
to involve the lymph nodes. While detecting early stage NSCLC
is considered secondary lung cancer prevention (or prevention
of metastatic cancer), it is important because the patient still has
the possibility of being cured, typically with a surgical approach.
In addition to being associated with an improved survival, this
is crucial to the prevention of the presentation of the typical patient
with lung cancer with advanced or metastatic lung cancer.
Currently, the most promising approach to early lung cancer
d etection is with an imaging study, namely, a chest CT scan.
Radiology Screening
CT Screening Screening with chest CT has the potential to
detect small lesions at an earlier stage than would occur with
a conventional approach. Seventy-five percent of cancers diagnosed
following the onset of symptoms are stage III or IV. In
one recent study, 85% of lung cancers diagnosed in the context
of screening current or former smokers were stage I, and only
5 of the 484 lung cancer diagnoses were made following the
development and evaluation of symptoms. Five-year survival
was estimated at 84%, 116 compared to 14% for lung cancer
diagnosed conventionally. 117 Further details on the CT screening
of lung cancer are seen in Chapter 16.
In 2003 based on early CT trial reports, the U.S. Preventive
Services Task Force did not find the evidence strong enough to
support a benefit with CT-based screening but neither did they
find evidence of harm. This inconclusive recommendation is
the same as their current recommendation for prostate cancer
screening and surprisingly prostate cancer screening has been
federally reimbursed for many years. 118
Clinical Management for Screened Lesions One of the
most challenging areas of lung cancer screening is the rapidly
evolving ability of higher-resolution CT scanners to find a large
number of lung nodules in the lungs of smokers. CT screening
identifies nodules that have baseline features suggestive of lung
cancer such a size, shape, or pattern of calcification. Newer approaches
to CT-based lung cancer detection involve evaluation
of features such as nodule growth. 97 In the most recent series,
about 13% of individuals undergoing an initial CT screening
will have findings that require further action. Of those with
negative or “non-actionable” initial scans, only about 5% will
have findings requiring further intervention on their annual
follow-up screening CT. In both cases, this intervention is
most often a repeat CT scan, with only approximately 2% to
4% of individuals being screened requiring an invasive test to
achieve a definitive diagnosis. 119,120 This may involve various
approaches. In the screening setting, it is critical to optimize the
efficiency of this process especially through the parsimonious
use of invasive procedures. At highly qualified institutions, the
diagnostic workup typically involves using an aspiration needle
biopsy to confirm that a growing nodule is cancer. When a
biopsy is performed in the context of the specific protocol, it
yields a diagnosis of cancer at least 90% of the time. 119 The frequency
of complications such as pneumothorax requiring tube
thoracostomy are low enough (5%) to justify its use. 121 This
approach is associated with an extremely low frequency of futile
thoracotomy (thoracic resection for a nonmalignant d iagnosis).
Other institutions include positron emission tomography (PET)
scan to evaluate questionable nodules. The investigators in their
screening trial reported an efficient process with an accuracy
rate of CT/PET being 91%, but they did report a futile thoracotomy
rate of approximately 5% (8/157). 122
Surgical Management Issues In the large prospective
experience of the I-ELCAP, postoperative mortality for
earlier stage lung cancers removed by anatomic resection has
been 0.5% (2/411). This mortality rate within the context of
a lung cancer screening trial is similar to the mortality rate of
0.6% reported by Pastorino et al.99 associated with lobectomy
for stage IA NSCLCs identified by current conventional methods.
Interestingly, these same investigators also participated in
a nother lung cancer screening trial in the same city and reported
a surgical mortality rate of 0% for early stage lung cancers.
Through the efforts of population-based screening protocols,
the detection of smaller nodules and ultimately earlier
stage lung cancers is occurring. 123,124 Commensurate with
the lesser size of early stage lung cancers has been the advent
and popularization of less invasive procedures. This has ultimately
translated into a lower mortality and morbidity rates
for the thoracic surgeon. In the largest video-assisted thoracoscopic
surgery (VATS) lobectomy series reported from the
United States, McKenna et al.125 reported a morbidity rate of
approximately 15% and a mortality rate of less than 1%. Most
of the complications were non–life threatening. In the recently
published ACOSOG Z0030 trial that prospectively collected
data to compare outcomes of patients undergoing mediastinal
lymph node dissection versus lymph node sampling, the morbidity
rate after lobectomy was 37% and 1%, respectively. The
majority of these tumors were early stage lesions. 126
In addition to minimally invasive approaches to the very
early stage I lung cancer, lung cancer screening with modern
day CT scanners has resurrected interest in parenchymal sparing
operations for small and peripheral NSCLCs. In 1995, the
results of the only randomized, controlled trial comparing limited
resections, including wedge resection and anatomic segmentectomy,
to the formal anatomic lobectomy were reported.
This study found that, while not statistically significant, there
was an increased rate of recurrence and mortality in the limited
resection population. Based on these findings, the investigators
advised against lesser resections for known stage I NSCLCs 127 .
The applicability of this study to current practice may be limited
by a number of differences. The tumors included in the
aforementioned study were those that were up to 3 cm in size.
Asamura et al. 128 argued that the current staging system that
defines T 1 tumors as less than 3 cm may be too heterogeneous.
Their review of the 1994 Japanese lung cancer registry found
that patient with lesions less than 2 cm actually fared better
than those between 2.1 and 3.0 cm, which would still be classified
as a T 1 lesion in the current staging system. Their findings
suggested, and they argue, that subclassification of T 1 lesions
into those that are 2.0 cm would be of benefit for prognostic
purposes. In addition to this potential difference in tumor biology,
current CT scanner technology is more advanced than
when the limited resection study was performed. Furthermore,
since lung cancer screening at this time was not in practice,
none of the enrolled patients were those identified by lung
cancer screening protocols.
A metaanalysis evaluating limited resection versus lobectomy
for small ( 3 cm) peripheral NSCLCs found that survival was not
significantly different. Interestingly, this finding occurred despite
the fact that the most common reason for a limited resection was
poor cardiopulmonary function, a condition that would suggest a
more fragile patient. 129 In the context of lung cancer screening, the
role of a limited resection has resurfaced as a question. Asamura
et al.123 reviewed their experience with subcentimeter lung cancers,
some of which were identified by lung cancer screening protocols.
The subcentimeter nodules identified were categorized into three
groups based on their appearance, nonsolid ground-glass opacity
(GGO), part-solid GGO, and solid-type tumor. Of the 28 GGO
lesions, 15 underwent a limited resection, versus only 4 of the 20
solid-type lesions. No difference in 5-year survival differences were
observed between the GGO, part-solid GGO, and solid-type tumors.
Similar outstanding survival rates have been observed in
studies evaluating sublobar resections for slightly larger GGOs
( 2 cm) identified by high-resolution CT scanners. 130 The findings
of these studies might suggest that small nodules identified
by screening CT scans, particularly GGOs, may not require an
anatomic lobectomy and may be equally served with a limited
resection. The question of whether a limited resection suffices
for stage IA tumors, of which many will be identified by screening
protocols, may not be answerable until the mature data from
randomized comparisons of standard compared to more limited
surgical approaches, such as with Cancer and Leukemia Group B
(CALGB) Study 140503, are reported. 131
Pennathur et al.132 and Dupuy et al.133 have furthered the
role of minimally invasive surgery to include percutaneous ablative
techniques without the use of the conventional thoracotomy
or newer thoracoscopy. 134 Currently, this modality of therapy
has not replaced the established role of surgery in the medically
fit patient. However, it is a reasonable alternative in the medically
inoperable patient that is found to have stage I d isease. 132,133
With increased experience in this subset of patients, the indications
for percutaneous ablation may expand to include those
with subcentimeter stage I lung cancers that are identified by
screening protocols. This, in fact, may be o ccurring as some
pulmonary radiofrequency ablation studies have i ncorporated
medically fit patients who have simply r efused surgery. 134
From a survival standpoint, it has been suggested that by
virtue of detecting earlier stage lesions, 5-year survival may be
improved over conventionally cited survival. 119 Some investigator
argue that an increased survival following surgery for any
screened lung cancer may reflect lead-time bias. Henschke et al.135
have suggested that a lung cancer mortality reduction benefit will
not ultimately be found unless there is an improved lead time.
Furthermore, true lead-time bias is thought to occur when relatively
short-term survivals (such as 5-year interval) are compared
between groups. This may not be the case with current screening
regimens that have improved long-term (10 year) survival. 136
Other investigators have suggested that the apparent benefit
of lung cancer screening relates to finding more indolent lung
cancers. This is referred to as overdiagnosis bias and has been
suggested as a reason why prior efforts at lung cancer screening
with chest radiographs and sputum cytopathology did not result
in significantly improved lung cancer–related mortality. 137,138
This has been questioned recently by Raz et al. 136 who reviewed
the long-term survival of patients with completely untreated
stage I NSCLC. They discovered that the median survival
among patients not treated by any modality including surgery,
chemotherapy, or radiation was 9 months overall and 13 months
in the subset with T 1 disease. This led the authors to conclude
that treatment of identified stage I NSCLC is imperative.
The medically stable patient with a lung cancer detected by
screening techniques deserves an opportunity for intervention.
Despite the criticisms leveled against screening for lung cancer,
recent mathematical modeling has suggested benefit in the
form of mortality reduction associated with screening. 139,140
Furthermore, the decreased morbidity and mortality associated
with minimally invasive techniques in the proper hands makes
surgical resection of the suspicious lesions that are identified by
screening an attractive and effective treatment option.
While the goal of surgical resection or ablation includes
the eradication of ever smaller early stage NSCLC, invariably
there will be lung nodules that are removed, which will
ultimately prove to be benign. The decreased mortality and
morbidity associated with newer surgical approaches and other
experimental intervention techniques may improve the therapeutic
index with surgical intervention in screening settings.
Special Considerations
Gene Therapy in the Prevention of Primary Lung
Cancer As mentioned earlier, the development of lung cancer
is thought to follow a stepwise progression. The underlying
mechanism of this progression is thought to be dependent
on genetic alterations that occur at key points. Naturally, the
earlier genetic changes that manifest as the early or precancerous
changes could possibly be targets of therapy. Despite this
knowledge, genetic therapy for lung cancer is well short of this
desired goal. Most genetic therapy centers on the role of the
tumor suppressor gene p53. 62,141 Despite its purported role
in the progression of disease from either bronchial atypia or
squamous metaplasia to low-grade dysplasia, much of the work
on the p53 tumor suppressor gene appears to center on promoting
its activity in patients with advanced malignancy. 62,141
At present, including the work regarding the p53 tumor
suppressor gene, no genetic therapy is available that definitively
halts the progression of normal bronchial epithelial cells
down the sequence of hyperplasia to dysplasia and ultimately
to carcinoma. This remains a conceptually exciting area for ongoing
research.
Biomarkers in the Early Detection and Prevention of
Primary Lung Cancer In terms of early detection, serum
and plasma analysis of circulating DNA has been thought to
be a novel biomarker. In particular, identifying free DNA that
has undergone changes such as methylation, (Refer to Laird
Offinga Chapter) loss of heterozygosity, allele shifts, microsatellite
instability, or other mutations may allow for identification
of early lung cancers. 142 These genetic changes may add
to the variety of proteomic biomarkers that are forthcoming.
From a population-based lung cancer screening perspective,
there are currently no established biomarkers routinely targeted
in the prevention of lung cancer or used in the early detection
of lung cancer. However, much work is being pursued in the
field of proteomics to arrive at a point where proteins expressed
in the sputum or serum of patients may be able to complement
or direct image-guided screening protocols. Presently, the
American Association of Cancer Research has established a task
force to evaluate the possibility of identifying biomarkers of precancerous
lesion along the multistep carcinogenesis pathway.
Identifying and intervening at these surrogate end point biomarker
milestones may potentially ultimately reduce the rates
of lung c ancer. 143 A detailed description of all the researched
biomarkers is beyond the scope of this chapter, but presently
a concerted effort in proteomic analysis is being undertaken to
identify targets for lung cancer detection and/or prevention.
Within the context of chemoprevention, two pathways that
have received a substantial amount of attention for their potential
role in biomarker protein detection and cancer prevention include
the COX pathway and epidermal growth factor receptor (EGFR)
pathway. The EGFR pathway involves a tyrosine kinase receptor
that results in autophosphorylation and activation of downstream
pathways. One of the significant downstream pathways involves
the COX-2 pathway. The products of both pathways may work
to potentiate the effectiveness of the other. 144,145 Simultaneous
inhibition of the COX-2 and EGFR pathways has demonstrated
an augmented antiproliferative and pro-apoptotic effect in cancer
cell lines in vitro. This presumed potentiation of the EGFR
pathway inhibition by COX-2 is thought to be representative of
a more complete blockade of both pathways. 144 Chemoprevention
efforts aimed toward inhibiting the pathway or products of
COX-2 and identifying and blocking the EGFR pathway are attractive
targets for chemoprevention, but in both instances approaches
to manage side effects of this therapy may be needed
prior to large-scale application, if sufficient efficacy data emerges.
Identification of circulating DNA released from tumor
cells, genetic alterations such as microsatellite instability in addition
to actual mutations, epigenetic changes such as tumor
suppressor gene promoter methylation, as well as an entire host
of other expressed proteins represent potential biomarkers that
may be available in the future. 146 Unfortunately, these and other
serum, plasma, sputum, and tissue assays are only in their investigational
stages and have yet to definitively be associated with
lung cancer and much less play a role in chemoprevention.