Lung Cancer Screening

1
Lung Cancer Screening

• Caryn Gee Morse, MD
• March 20, 2001

2
Lung Cancer the problem

• In the United States, the second most common
  cancer in men and women
• Leading cause of cancer mortality
• Accounts for more deaths from than cancer than
  breast, prostate and colorectal cancer combined

3
Common and Lethal

• The American Cancer Society estimates 169,500 new
  cases of lung cancer will be diagnosed in 2001
• Overall 5-year survival remains poor,
  approximately 15
• ACS anticipates 157,400 deaths from lung cancer
  in 2001

4
Estimated U.S. Cancer Deaths10 Leading Sites, by
Gender 2001
5
Age-adjusted cancer death rates by siteUS males,
1930-1997
6
Age-adjusted cancer death rates by siteUS
females, 1930-1997
7
Prognosis of Non-small Cell Lung Cancer, By Stage
8
Table 4. Current guidelines for lung cancer
screening
Adapted from Mandel, J, Weinberger, S.
Screening for lung cancer. UpToDate 2000 81-2.
Mandel, J, Weinberger, S. Screening for lung
cancer. UpToDate 2000 81-2.
9
Presentation overview

• History of lung cancer screening
• Trials of the 1950-60s
• National Cancer Institute Cooperative Early Lung
  Cancer Project
• Mayo Lung Project
• Memorial Sloan-Kettering
• Johns Hopkins
• Czechoslovakian Trial
• Lost interest

10
Presentation Overview Cont.

• Renewed interest and new directions
• Chest radiographs
• Computed tomography
• PET
• Biomarkers
• Fluorescence bronchoscopy
• Conclusions and recommendations

11
(No Transcript)
12
(No Transcript)
13
Screening Principles

• Successful cancer screening program
• needs to detect disease in the preclinical stage
• when it is amenable to curative treatment
• reduce mortality by preventing progression of
  disease

14
Lung cancer screening

• A successful randomized trial of screening for
  lung cancer
• Enhances detection of lung cancers, particularly
  asymptomatic, early stage cancers, in the study
  group when compared to a control group during the
  screening phase
• As the trial progresses, the number of lung
  cancers in the two groups should equalize as
  asymptomatic, early stage cancers undetected in
  the control group grow, spread, and present as
  symptomatic, advanced stage cancers
• If treatment is more effective for asymptomatic,
  early stage cancers compared with symptomatic,
  more advance lung cancers, fewer deaths would be
  expected in the screened group compared to
  non-screened controls
• Model assumes that the bulk of early stage lung
  tumors progress to lethal disease without
  detection and early treatment and assumes that
  early detection reduces mortality

15
Screening mortality and survival

• No randomized, controlled trial has demonstrated
  that lung cancer screening leads to a reduction
  in disease-specific mortality
• Mortality vs. survival
• Mortality(death rate) cancer deaths /
  patients screened, expressed as deaths per 1000
  persons screened per year
• Survival patients alive following cancer
  diagnosis / cancers detected, expressed as a
  percentage over time

16
Screening

• Mortality can be influenced by selection bias in
  nonrandomized trials, however, in a RCT, a
  statistically significant mortality reduction is
  considered proof of screening effectiveness or at
  least best evidence for efficacy

17
Screening bias

• All other measures of outcome can be affected
  bias, including
• Selection bias error in patient assignment
  between groups that permits a confounding
  variable to arise from study design rather than
  by chance alone usually eliminated by
  randomization
• Lead-time bias mistakenly attributing increased
  survival of patients to a screening intervention
  when longer survival is only a reflection of
  earlier detection in the preclinical phase of
  disease
• Length-sampling bias Slow growing tumors are
  detectable longer than fast growing ones and will
  be preferentially identified by any early
  diagnosis strategy. Fast growing tumors, with
  shorter survival, will be left for routine
  diagnosis.
• Over-diagnosis a portion of detected cancers may
  have remained indolent and undetected because of
  patient death from other causes

18
Early, early trials

• In the 1950s and 1960s a number of uncontrolled
  and nonrandomized controlled studies were
  performed to evaluate combinations of chest x-ray
  and sputum cytology screening at various time
  intervals, ranging up to once every 6 months
• Three nonrandomized, uncontrolled trials
• Philadelphia Pulmonary Neoplasm Research Project
• Veterans Administration Lung Cancer Screening
  Study
• South London Lung Cancer Study
• Two randomized controlled trials
• North London Cancer Study
• Kaiser Foundation Health Plan Study

19
Early, early trials cont.

• All of these studies failed to demonstrate a
  statistically significant mortality benefit from
  lung cancer screening.
• Small cohorts with limited follow-up periods,
  limiting demonstration of small-moderate
  improvements or longer-term benefits in mortality

20
NCI Cooperative Early Lung Cancer Group

• In 1971, the National Cancer Institute (NCI)
  sponsored three large-scale, long-term,
  randomized controlled trials created to determine
  
• whether a program of lung cancer screening
  might lead to earlier detection, that is, to
  finding a larger proportion of lung cancers at a
  localized, potentially curable stage, and whether
  with appropriate treatment this would result in a
  substantial reduction of lung cancer deaths in
  the screened group.

21
NCI Cooperative Early Lung Cancer Group, cont.

• Specifically, the trials sought to establish if
  detection of early lung cancer could be improved
  by the addition of sputum cytology to routine
  chest x-ray and if lung cancer mortality could be
  reduced by this screening and appropriate
  therapy.
• The trials, completed in 1984, were conducted at
  Mayo Clinic, Johns Hopkins Medical Center and the
  Memorial-Sloan Kettering Cancer Center and the
  participating institutions were designated the
  Cooperative Early Lung Cancer Group.

22
Mayo Lung Project

• From late 1971 to mid-1976, enrolled
• 10,933 male volunteers
• Age 45 years or older
• At least one pack per day cigarette use in the
  previous year
• Referred for participation by their primary care
  physician during routine physical examination
• All participants were offered an initial
  prevalence screen including chest x-ray and
  sputum cytology.

23
Mayo Lung Project Prevalence
24
Mayo Lung Project

• 91 prevalent cancer patients removed from the
  initial volunteers
• 978 patients ruled ineligible because of serious
  underlying medical problems and predicted life
  expectancy of less than 5 years
• 653 volunteers refused participation
• Remaining 9211 participants were randomized to
  two groups, screening and control
• Screening group participants received chest x-ray
  and sputum cytology examination every 4 months
  for 6 years
• Control group participants were advised to seek
  annual chest x-ray and sputum cytology, standard
  Mayo advice at the time and no reminders were
  sent
• Study group was followed for a total of 9 years,
  6 years of screening and 3 years of follow-up
  observation.

25
MLP Incidence and Mortality
26
MLP Staging and 5-year survival
27
MLP Late stage cancers, nonresectable cases and
number of deaths
28
MLP Extended follow-up
29
Memorial Sloan-Kettering

• From 1974 to 1978, enrolled
• 10,040 male volunteers
• Age 45 years or older
• At least 1 pack per day cigarette use currently
  or in the preceding year
• On initial intake all participants received PA
  and lateral chest films and pooled sputum cytology

30
Memorial Sloan-Kettering

• Following the initial prevalence screen
• 5072 men randomized to the chest x-ray only
  group
• 4985 men randomized to the dual-screened group
• Both groups received annual chest x-rays.
• The dual-screened group additionally received
  3-day pooled sputum cytology every 4 months.

31
Memorial Sloan-KetteringIncidence
32
M S-K Conclusions

• There was no statistically significant difference
  in early stage lung cancers identified, 5-year
  survival or mortality between the dual-screen
  group and the chest x-ray only group
• Sputum cytology, even as often as q4 months, does
  not improve mortality compared to CXR alone

33
Johns Hopkins

• Uncanny-ly similar to Memorial Sloan-Kettering
• From 1973 to1977 enrolled
• 10,387 male volunteers
• Age 45 years or older
• At least one pack per day smoking history in
  preceding year
• Volunteers were randomly allocated to two groups
• Control, or single-screen group, received annual
  radiographic screening only
• Dual-screen group received annual radiography
  plus annual sputum cytologic examination

34
Johns Hopkins Incidence

• Lung cancer detected in 396 participants
• 194 in the dual-screen group
• 202 in the control group
• Over half (51) of the cancers identified were
  detected incidentally by chest x-ray or sputum
  cytology performed outside of the screening
  protocol
• Compared with clinical diagnosis by symptoms,
  screening by both chest x-ray and sputum cytology
  identified a greater proportion of the lung
  cancer cases at an earlier stage
• Addition of sputum cytology improved detection of
  squamous cell lung cancer but did not effect
  disease-specific mortality

35
NCI Cooperative Lung Conclusions

• Demonstrated improvements in stage distribution,
  resectability and survival in screened groups
• No improvement in disease-specific mortality with
  screening
• Cooperative authors recognize the potential
  effects of bias, lead-time, length-sampling and
  over-diagnosis
• Role of control contamination, screening
  non-compliance?
• However, they hedge,
• It is probable that some patients who had
  lung cancers detected by screening would have
  died of their malignancies had they not been
  detected at earlier stages.

36
NCI Cooperative LungRecommendations, 1984

• 1. If screening for lung cancer is to be carried
  out, it should be done within the framework of
  general health care that is, in the private
  practitioner's office, HMO, or general medicine
  clinic.
• 2. The chest x-ray is the most sensitive method
  for detection of lung cancer currently available.
  
• 3. Sputum cytology is the most effective and
  specific method of detecting early squamous cell
  lung carcinoma. Patients with positive sputum
  cytology in the setting of radiologically occult
  malignancy have good 5 year survival.
• 4. Data do not indicate if prolonged survival in
  prevalent cases of lung cancer represented
  decreased mortality from disease or reflects one
  or more screening artifacts.

37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
Czechoslovakian Lung Cancer Study

• Began in Czechoslovakia in 1976
• Designed to evaluate semi-annual screening by
  chest x-ray and sputum cytology
• 6364 males, ages 40-64, with a greater than 20
  pack year history of tobacco abuse, were
  screened with PA chest x-ray and 24-hour sputum
  cytology to identify prevalent cases

41
Czech study cont.

• After the prevalence screen, remaining
  participants were randomized to a screening group
  or a control group
• 3172 randomized to screening received PA chest
  x-ray and sputum cytology every 6 months for 3
  years
• 3174 randomized to control received a single
  screening chest x-ray at the end of the 3 year
  trial
• All participants received annual chest x-ray for
  an additional 3 years following the screening
  period

42
Czechoslovakian Lung Cancer Study

• Prevalence
• Initial screen identified 19 cancers, 9 squamous
  cell carcinomas and 7 small cell lung cancers
• Overall prevalence was 3/1000 examinations
• 5-year survival for prevalent cases was 25

43
Czechoslovakian Lung Cancer Study

• During the three year screening period, 55
  confirmed lung cancers were identified
• 36 cases were identified in the screening group.
  26(75) cancers found in asymptomatic
  participants
• 19 cases were identified in the control group. 4
  (25)found incidentally or at autopsy
• Following the screening period, annual CXR
  surveillance revealed an additional 35 cases of
  lung cancer in the screening group and 38 cases
  in the control group
• Overall mortality after nine years was 3.6/1000
  person/years in both the screened and control
  groups

44
Randomized control trials, summary
45
RCT, summary cont.

• Four RCTs collectively screened 37,724
  participants
• All studies demonstrated improvements in stage
  distribution, resectability and survival in
  screened groups
• No improvement in disease-specific mortality with
  screening

46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
Interest lost, found

• With failure to demonstrate mortality benefits
  from screening, all major advisory organizations
  adopted recommendations against screening for
  lung cancer
• Research funding waned
• But,with emergence of new diagnostic and
  therapeutic modalities, new interest in lung
  cancer screening arose in the mid-1990s

52
New directions in early lung cancer detection

• Chest x-ray reexamined
• PLCO
• Low-dose spiral CT
• ELCAP
• PET
• Biomarkers
• hnRNP expression
• Fluorescence bronchoscopy

53
Chest x-ray reexamined PLCO

• The National Cancer Institute is readdressing the
  use of periodic chest x-ray screening in the
  Prostate, Lung, Colorectal and Ovarian Cancers
  Trial (PLCO)
• Currently in progress, PLCO seeks to enroll
  150,000 Americans, age 55-75, for randomization
  to various screening or no screening strategies
• 14 years follow-up planned

54
Computed Tomography

• In the 1990s a number of population-based trials
  demonstrated increased sensitivity of CT scan for
  the detection of resectable lung cancer vs. CXR
• Majority of reports come from Japan where a
  government sponsored screening program with CXR
  and sputum cytology has been in place since
  mid-1980s. The addition of low-dose spiral CT
  to screening programs there lead to higher
  detection rates of early stage non-small cell
  lung cancers.
• A number of studies are currently underway to
  evaluate the use of low-dose spiral CT for early
  lung cancer detection.

55
What is low-dose spiral CT?

• Takes 15-30 seconds to perform
• Allows complete chest imaging in one breath hold
  using wide slices
• Radiation exposure equivalent to mammography
• Can detect lesions as small as 2mm

56
Computed tomography ELCAP

• ELCAP Early Lung Cancer Action Project
• On-going study begun in 1992
• Single cohort, non-comparative design
• Aims to establish a cure rate based on lung tumor
  size to be used subsequently to assess other
  screening protocols or novel tests
• Compared with a randomized controlled trial, this
  design is less costly and allows more rapid
  acquistion of data

57
ELCAP

• Enrolled 1000 symptom-free volunteers from New
  York Hospital-Cornell University and NYU Medical
  Center hospitals, associated physicians' offices
• 60 years of age or older
• Both men and women!!
• 10 pack-year history of cigarette abuse or
  greater
• No prior history of malignancy, no
  contraindications to thoracic surgery.

58
ELCAP

• All participants underwent PA and lateral chest
  radiographs and low-dose helical CT
• Positive results were defined as CT evidence of 1
  or greater non-calcified nodules.
• When nodules were identified parameters were
  recorded including number of nodules, size,
  location (lobe and distance from pleura,) shape
  (round, non-round,) edge (smooth, non-smooth,)
  and benign calcification (present or absent.)

59
ELCAP

• Patients with CT evidence of non-calcified
  pulmonary nodules underwent standard-dose high
  resolution diagnostic CT scan of the chest
• If standard CT demonstrated benign calcifications
  in a nodule with smooth edges, nodule was classified as benign
• Suspicious nodules were evaluated according to
  ELCAP protocol
• months. If no growth noted _at_ 24 months the
  lesion was considered benign.
• 6-10mm biopsy recommended. If contraindications
  to biopsy exist, follow-up serial CTs as
  described above.
• 11mm biopsy strongly recommended

60
ELCAP RESULTS (brief)
61
ELCAP Conclusions

• Compared with CXR, low-dose CT greatly increases
  the likelihood of detection of small
  non-calcified nodules, and, thus, of lung cancer
  at earlier, more curable stages
• CT vs CXR
• detected 3x as many non-calcified nodules
• 4x as many malignant tumors
• 6x as many stage I cancers

62
ELCAP Conclusions

• High false positive rate of 233 suspicious
  nodules only 27 malignant tumors were confirmed.
  Only 4 patients underwent biopsy for benign
  disease and no biopsy complications were noted.
• Risks of cumulative radiation likely low
• Cost-effectiveness ???

63
(No Transcript)
64
PETPositron Emission Tomography

• Utilizes metabolic activity of a pulmonary lesion
  to provide information about the malignant
  potential of a pulmonary nodule
• Allows imaging of structures by virtue of their
  ability to concentrate specific molecules that
  have been labeled with a positron-emitting
  isotope
• In evaluation of solitary pulmonary nodules some
  studies suggest a 95 sensitivity and 70
  specificity for the detection of malignancy

65
(No Transcript)
66
(No Transcript)
67
Biomarkers Introduction

• Screening with conventional sputum cytology
  failed to decrease mortality as demonstrated in
  the Johns Hopkins and Memorial Sloan-Kettering
  trials.
• However, new techniques and immunostaining for
  biomarkers promises much greater sensitivity for
  sputum cytology evaluation.

68
Biomarkers

• Lung cancer results from an interaction between
  genetic predisposition and environmental causes.
• Exposure of the respiratory epithelium to
  carcinogens triggers mutation in specific genes,
  proto-oncogenes, and tumor suppressor genes.
  Once epithelial cells undergo malignant
  transformation, cell proliferation depends on
  tumor promoters, cellular growth factors.
• Several genetic abnormalities have been
  identified in association with lung cancer
  (table13.) Mutations and molecular products can
  be identified from sputum and tissue samples
  using polymerase chain reaction (PCR) and other
  techniques in molecular biology.

69
Biomarkers cont.
70
Prospective Detection of Preclinical Lung Cancer
Results from studies of heterogeneous nuclear
ribonucleoprotein A2/B1 overexpression

• Lung Cancer Early Detection Working Group, a
  cooperative NCI-sponsored group
• Earlier studies identified potentially useful
  lung CA biomarkers expressed in archieved sputum
• Initiated to address
• Does hnRNP A2/B1 overexpression correctly detect
  preclinical lung cancer?
• Does overexpression precede dysplastic
  morphological changes in sputum epithelial cells?

71
hnRNP overexpression study, cont.

• Prospective, case-cohort design
• Monitored Chinese tin workers over age 40 with
  annual sputum cytology and immunocytochemistry
• Blinded investigators at Johns Hopkins analyzed
  sputum samples from participants with clinically
  detected and confirmed non-small lung cancers and
  from age-matched controls

72
Immunodetection of preclinical primary lung
cancer by hnRNP overexpression hnRNP test
characteristics

• Primary lung cancer risk 56/62850.9
• Sensitivity82 (CI 68-92)
• Specificity65(CI 50-78)
• Relative risk1480/432, 3.4

73
Biomarkers conclusions

• May be possible to use biomarkers to identify an
  early clonal phase of progression of lung cancer
  in high-risk populations, enabling cancer
  detection earlier than visualization by spiral
  CT.
• May be used to complement spiral CT to ?
  sensitivity and specificity for malignancy.
• May be used to identify targets for treatment
  allowing chemical, radiation or pharmacological
  targeting of minute primary lung cancers.

74
Contemporary Screening for Lung Cancer,
Pre-malignancy and Malignancy Project

• Recruitment underway now
• Multicenter, randomized controlled trial
• Aims to address whether a screening program using
  lung cancer associated molecular markers in
  sputum combined with low-dose helical CT can
  improve lung-cancer specific survival in
  individuals at high risk for lung cancer.

75
Contemporary Screening for Lung Cancer,
Pre-malignancy and Malignancy Project cont.

• With enrollment, participants will undergo chest
  x-ray and sputum cytologic evaluation to select
  out prevalent cases of lung cancer prior to
  randomization. Enrollees without evidence of
  lung cancer will be randomized to screening and
  control groups.
• Screening group will undergo sputum (cytologic
  and biomarker evaluation) and radiographic (chest
  x-ray and low dose spiral CT) at 6 month
  intervals. The control group will undergo no
  screening but will complete annual health
  questionnaires.
• NC Baptist/Wake Forest University Medical Center
  is one of thirteen participating institutions for
  this study.

76
Fluorescence Bronchoscopy

• Undergoing early multi-center trials for
  screening in smokers with established obstructive
  lung disease and abnormal sputum cytology
• Utilizes differences in the fluorescence
  properties of normal and abnormal bronchial
  epithelium to identify dysplasia and metaplasia

77
Fluorescence bronchoscopy

• Demonstrated ability to enhance detection of
  severe dysplasia and carcinoma in situ over
  white-light bronchoscopy
• May be most useful in localization of sputum
  cytology positive, CXR/CT (-) malignancies and in
  determination of endobronchial involvement by
  malignancy
• Technology and technique remain under development
  
• Anticipate improved sensitivity and specificity
  with future systems

78
Conclusions

• No randomized controlled trial to date has
  demonstrated reduced mortality from a lung cancer
  screening program
• Trials demonstrate improvement in stage at
  diagnosis, resectability and survival with
  screening programs
• No evidence to support the use of chest x-ray or
  sputum cytology for routine lung cancer screening
  in asymptomatic patients

79
Conclusions, cont.

• New diagnostic techniques may prove promising for
  use alone or together in early lung cancer
  detection
• Low-dose spiral CT looks promising for detection
  of early stage lung cancer however high false
  positive rate could lead to unnecessary morbidity
  and mortality in disease-free patients

80
Conclusions, conclude

• PET and biomarkers may improve sensitivity and
  specificity of other diagnostic tests especially
  CT
• Not sure what to expect from fluorescence
  bronchoscopy, sounds neat

81
Should you screen?

• 1. Does the burden of disability from the disease
  warrant action?
• 2. Are at-risk populations well-defined?
• 3. Does early diagnosis really lead to improved
  clinical outcomes (in terms of survival, function
  and quality of life)?
• 4. Are the cost and accuracy of the screening
  test acceptable?
• 5. Can you manage the additional clinical time
  required to confirm diagnosis and provide
  long-term care to those who screen positive?
• 6. Will patients in whom an early diagnosis is
  achieved comply with subsequent recommendations
  and treatment regimens?
• Adapted from Sackett et al. Clinical
  Epidemiology. A Basic Science for Medicine.
  1991 391.

82
Recommendations

• At this time, would hold on screening of
  asymptomatic high-risk patients
• Could consider lowering your threshold for
  screening in patients with documented airflow
  obstruction
• Lung Health Study found a 1 cancer mortality at
  5 years in patients with COPD
• Eagerly await the results of pending trials (that
  do indeed include women)

83
Recommendations, cont.

• Anticipate screening will have greater mortality
  benefit as lung cancer therapy advances
• Above all, emphasize to patients that the
  greatest reduction in lung cancer mortality comes
  from smoking cessation