Lung Cancer Screening Test
Search for the Holy Grail
By Barry B. Sheppard, M.D.
Lung
Cancer is the deadliest cancer in the United States. It is deadlier than
the three runners-up (breast, colon, and prostate) combined. Every year
lung cancer kills more American women than breast cancer. Is this an
unstoppable cancer? No. Stage IA lung cancer patients treated with surgery
alone can expect an 80 percent chance of long-term survival. Admittedly,
this does not quite match the surgical success reached in breast cancer,
where Stage IA disease can be treated with a greater than 90 percent
five-year survival. It is, however, a respectable result. The chasm in
care separating breast and lung cancer does not derive from that small
inequality; and although the clinical arenas of lung and breast cancer
treatment are as different as the night compared with the day, that
difference arises because of the deficiency of a single element. Let me
elaborate.
Pop
Quiz:
The findings of the Mayo Lung Project randomized controlled trial (1983) best support which of the following clinical practices?
A. Quarterly screening with plain CXR for smokers
B. Quarterly sputum cytology combined with CXR
for smokers
C. Annual screening with plain CXR for smokers
No surveillance for lung cancer in
smokers
Hint: What defined the Control Group in the Mayo Lung Project?
In the arena of breast
cancer care, Sutter’s oversight committee has taken quality indicators,
initially generated by the American College of Surgeons, and further, has
targeted goals for its nine accredited Cancer Centers associated with each
of them. Sutter, therefore, anticipates that a top-notch breast-cancer
program, applying the tried-and-true screening test, mammography, in
concert with fine needle, stereotactic, and wire-localized biopsy methods
to the population at greatest risk, i.e. women of a certain age, will
diagnose at least 70 percent of its tumors at a size no greater than 2cm.
Additionally, at least 35 percent of newly diagnosed breast cancers are
expected to be less than or equal to 1cm, and 21 percent of all new breast
cancer diagnoses should be Stage 0, i.e. in situ disease. These
results can be targeted and reasonably expected if the various specialists
and primary care docs work in concert with each other as well as with an
educated and responsive community. The presence of a proven screening test
with widely accepted guidelines for its application provides an
environment where the quality of a breast cancer program will affect
directly the cancer stage at diagnosis and, through this mechanism alone,
have an enormous impact on cancer survival.
Contrast that with the
current environment surrounding lung cancer care. Upon meeting the
infrequent individual with early-stage lung cancer, the pulmonologist,
surgeon, or oncologist shouts “Hallelujah,” which, it turns out, is an
appropriate response since we are reliant on divine intervention or
serendipity to coordinate the events necessary for an early diagnosis of
lung cancer. Lung cancer diagnosis prompted by patient symptoms is
synonymous with late-stage disease, the only exception being the
occasional post-obstructive pneumonia, which clues a clinician in to the
presence of a resectable tumor. The vast majority of Stage I lung cancers
are discovered incidentally on CXRs performed for another purpose.
Occasionally early lung cancers are diagnosed on CXRs performed
specifically as screening exams by the rare physicians who actually read
the Mayo Lung Project study, circa 1983, and remember its findings. They
persist in using an imperfect screen in the absence of any alternative
tool. These medieval methods of cancer detection result in a situation
where, at the time of diagnosis, 75 percent of lung cancers already have
spread beyond the lungs as regional (Stage III) or distant (Stage IV)
metastases.
The population at risk
for lung cancer is easily discernible. It includes current or prior
smokers with moderate to heavy history of tobacco abuse, especially those
greater than 60 years of age or with evidence of at least mild emphysema.
Without a reliable screening tool, however, lung cancer care is left
floundering in a haphazard milieu with a complete disconnect between the
desire to improve a patient’s chances and any real ability to positively
impact the stage at diagnosis.
Progress has been made
in the clinical arena of lung cancer treatment. New chemotherapeutic
agents have impressive response rates to the point that pathologic
complete responses are now not only possible, but occur regularly.
Advances have been made adjusting the timing of adjuvant chemotherapy and
radiation treatment both with regard to surgical intervention as well as
in relation to each other. I once had high hopes for concomitant
neoadjuvant chemo-irradiation followed by surgery, but at the end of the
day, that is a very long haul for, at best, a moderate improvement in a
subgroup of patients with abysmal survival rates to start with.
The hard reality in
thoracic oncology is that the possibilities for significant improvement in
patient outcomes are twofold: smoking prevention/cessation and detection
of cancer at an earlier stage. These two possibilities are huge, however.
Since 85 percent of lung cancers are directly attributable to smoking,
there are approximately 135,000 lives a year that we could salvage from
lung cancer alone, and twice that many from atherosclerotic cardiovascular
disease, emphysema, and premature infant deaths, if we were able to
completely eradicate tobacco abuse.
But for those that
already have quit but not before amassing significant risk for cancer, as
well as those who refuse to quit and undoubtedly those who will begin
smoking in the years to come, we need to find an effective screening tool
for lung cancer. The quest for that Grail has eluded valiant attempts to
locate it. Several large crusades were launched in the ‘70s and ‘80s
and their combined result revealed that neither sputum cytology nor plain
chest radiograph was the sought-after treasure. In fact, the
disappointment over the shortcomings of CXR as a screening tool led to an
unfortunate backlash that resulted in a marked decline in usage of a tool
that—although not the Chalice of lore—still served as a crude tool to
screen for lung cancers and has the ability to diagnose some early lung
cancers and thereby improve curability for many patients.
The Mayo Lung Project
study was widely interpreted to indicate that screening smokers with CXR
was of no benefit. In actuality, the Mayo study compared the use of
quarterly, regimented CXR with a control group receiving the standard of
care at that time, i.e., a recommendation for annual CXR, the obtaining of
which was left to the discretion of the control patients in the study.
Approximately half of the control group did undergo annual CXR as a
screening test. The ultimate finding of the study was that the quarterly
application of CXR as a screening tool showed no difference in lung cancer
mortality compared with the control group, and this should have been
interpreted as evidence against doing more than an annual CXR. The study
does not prove that CXR has no efficacy because half of the control group
also had screening CXRs, only on a less frequent basis. Enigmatically,
despite the finding of identical lung-cancer mortality rates, a large
number of early lung cancers were detected in the experimental group,
subsequently completely resected and constituted the basis of the much
higher lung cancer survival seen in that arm of the study.
This unfortunate
misinterpretation of the Mayo study fueled much of the drastic reduction
in CXRs ordered during the past two decades and has undoubtedly worsened
the shift of lung cancer to late-stage disease at the time of diagnosis.
The Mayo Lung Project
closed in 1983 with the final report being published in 1984. So for the
rest of that decade and through the ‘90s there was effectively
little-to-no screening for lung cancer, while breast cancer experienced
gratifying advancements, culminating in the expectations mentioned
earlier, where targets could actually be set regarding what percentage of
patients were diagnosed in early stages.
In 1999 I ran across an
announcement that the Mayo was starting a new lung cancer screening trial
utilizing Low Dose CT (LDCT) scans. These were the newest iteration of CT
scans, which not only had a very low associated radiation exposure
(approximately equal to one and a half regular chest radiographs) but
could also be done extremely quickly during the single breath hold of a
patient. This immediately caught my attention since I was reminded of
several early Stage I lung cancers that we diagnosed during my thoracic
surgery residency at Davis, when we were performing CT scans on emphysema
patients as a work-up for lung-volume reduction surgery. The incidental
lung cancers discovered in this fashion were the smallest of any that I
operated on during my entire training period.
In the same year the
first Early Lung Cancer Action Project (ELCAP) paper was published out of
Cornell Medical Center by Dr. Claudia Henschke. In that study 1,000
patients with a 10-pack-year smoking history or greater were screened with
low-dose CT scans, and 233 patients had 1-6 non-calcified nodules imaged.
CXRs performed in the same patients only detected nodules in 68 patients.
Utilizing a conservative algorithm based on nodule growth over time, and
persistence after treating with antibiotics, 28 biopsies were recommended
and 27 of those turned out to be cancer.
Twenty-three of the 27 cancers were Stage I. It appeared that the hunt for
the Grail was back on and that LDCT might be the real thing.
Suspecting that this
had the potential to be the biggest advancement in lung cancer since the
first successful lung cancer resection, I went to the head of the oncology
program at Mills-Peninsula, Dr. Gale Katterhagen, with my idea to get
involved in this groundbreaking research. We submitted a grant proposal to
the Mills-Peninsula Foundation and were subsequently awarded full funding
for a study to investigate low-dose CT scanning by scanning 400 patients
for two years. We intended to marry our study to one of the larger
investigations to statistically strengthen any findings. We succeeded
beyond our most optimistic dreams by being accepted as one of the 30
international centers involved in the I-ELCAP, headed up by Dr. Claudia
Henschke, herself, the woman who started the push in the United States to
evaluate CT as a screening modality for lung cancer.
To date, the I-ELCAP
has jointly performed more than 35,000 screening CTs. In addition there is
an ongoing randomized trial supported by the NIH, the National Lung
Screening Trial (NLST), which has enrolled 50,000 subjects at 50 U.S.
centers and will provide the definitive answer regarding low-dose CT scans
as the long-awaited screening test for lung cancer, but not until 2013.
Within the next year I-ELCAP hopes to provide statistically irrefutable
evidence that LDCT markedly increases curability of lung cancer and
improves survival when applied as a screening test to high-risk patients.
Mills-Peninsula
Hospital and the Dorothy E. Schneider Cancer Center have the honor of
being an integral part of the successful quest for the Holy Grail of lung
cancer. Even more important, with a still-open protocol utilizing LDCT,
smokers and ex-smokers residing in San Mateo County with a greater than
30-pack-year history of smoking have free access to the soon-to-be-proven
screening test, thereby exercising the best course of action to survive
any potential lung cancer.
Dr. Sheppard, a cardiovascular and thoracic
surgeon in Burlingame, is principal investigator for the Mills-Peninsula
Spiral CT Lung Cancer Screening study. For more information, contact Jaime
Brown 650-696-44