MAIN POINTS

• ▪
  

  Lung cancer is a prevalent disease, and is the leading cause of cancer death in men and women.

  
  
• ▪
  

  There is a large difference in survival between lung cancer detected at an early stage and lung cancer detected at an advanced stage. Of the four most common cancers, the screening tests have made a big difference in survival for breast, prostate, and colon cancer, but there is no officially approved screening test for lung cancer. A screening test that can detect lung cancer at an early stage would be very desirable because the mortality rate for lung cancer is so high.

  
  
• ▪
  

  Low-dose computed tomography (LDCT) is sensitive for the detection of early-stage, curable lung cancer

  
  
• ▪
  

  However, there has been no randomized, prospective screening study to prove that LDCT will reduce the overall mortality for lung cancer. Studies are currently underway to evaluate this more comprehensively.

SCREENING—DEFINITIONS

• ▪
  

  Screening : a test used to detect cancer before symptoms develop.

  
  
• ▪
  

  To be effective, a screening test must be readily available, of low cost, and ultimately influence mortality.

LUNG CANCER – BACKGROUND INFORMATION / RATIONALE FOR SCREENING

• ▪
  

  Epidemiology:

  
  
  
  
  • ○
    

    Lung cancer is the leading cause of cancer death in men and women.

    
    
  • ○
    

    2007 estimated new cases: 213,380

    
    
  • ○
    

    2007 estimated deaths: 160,390 (compared with 40,910 for breast cancer and 52,180 for colon cancer)

    
    
  • ○
    

    Accounts for 15% of cancer diagnoses and 29% of cancer deaths

  
  
  
  
• ▪
  

  Prognosis:

  
  
  
  
  • ○
    

    Average 5-year survival rate for all patients is very poor, at 16%.

    
    
  • ○
    

    Five-year survival rate for localized disease is considerably better, at 67% for stage IA, to 39% for stage IIB.

    
    
  • ○
    

    At present, only 16% of lung cancers are diagnosed when still localized.

    
    
  • ○
    

    Small cell lung cancer accounts for 13% of all lung cancer cases.

    
    
  • ○
    

    Non–small cell lung cancer (NSCLC) constitutes 87%. Screening epidemiology is based on NSCLC data.

  
  
  
  
• ▪
  

  Treatment:

  
  
  
  
  • ○
    

    For localized cancer, surgery is most often the treatment of choice.

    
    
  • ○
    

    Chemotherapy and radiation therapy are used in metastatic disease.

    
    
  • ○
    

    For patients with localized (stage I) NSCLC who are not surgical candidates, optimal treatment includes radiofrequency ablation or localized radiation therapy, or both.

  
  

STATISTICAL ANALYSIS

• ▪
  

  For a screening test to prove effectiveness, it must ultimately improve mortality, and not just detect cancer and improve 5-year survival.

  
  
• ▪
  

  Theoretic reasons why a screening test may give a false impression of benefit:

  
  
  
  
  • ○
    

    Lead time bias—False impression of improved survival in a screened population without affecting mortality, because the cancer is diagnosed earlier in the natural history of the disease but the patient still dies of the cancer ( Fig. 20-1 ).

    
    

    
    

    
    

    

    
    

    
    

    Schematic showing concept of lead time bias. Hypothetical patients A and B both develop lung cancer (i.e., onset of pathology) at similar times. In A, screening identifies a mass on the chest radiograph that proves to be non–small cell lung cancer. The patient receives appropriate surgical therapy. For patient B, the diagnosis is only made following the development of symptoms, at a much later time compared with patient A. Both patients A and B die from cancer-related causes at similar times. We conclude from this scenario that although the apparent survival of patient A is significantly longer than that of patient B, the mortality and inherent natural history, however, are the same for both patients.

    
    
    
    
  • ○
    

    Length time bias—Screened populations are by definition asymptomatic and may therefore be more likely to have less aggressive malignancies.

    
    
  • ○
    

    Overdiagnosis bias—A screening program has a tendency to discover cancer that will not affect the life expectancy of the patient.

    
    
    
    
    • ▪
      

      For example, autopsy studies show that up to 22% of men older than 70 dying from unrelated causes still have small kidney cancers.

      
      
    • ▪
      

      There is also a 1% incidence of lung cancer in the same population, with neither cancer contributing to the patient’s demise.

    
    

  
  

SCREENING MODALITY: CHEST X-RAY STUDY AND SPUTUM CYTOLOGY

The usefulness of chest x-ray study (CXR) and sputum cytology in lung cancer screening was examined in five randomized control trials in the 1970s and 1980s. Screening with CXR (with or without sputum cytology) increased lung cancer diagnosis and an improved survival. However, no definite overall mortality reduction was found. The applicability of these findings is limited by the fact that the American studies used recommended annual CXR as the standard of care for the control group.

• ▪
  

  A randomized control trial (RCT) in London evaluated 55,034 men. The experimental group received CXR every 6 months for 3 years, whereas the control group received CXR at the beginning and end of the study. Sixty-two patients died in the semiannual screening group, whereas 59 patients died in the control group. The study was reported as showing no benefit; however, if the data are analyzed differently, interesting results and trends are observed. For example, if the prevalent cases were not included, there was a 60% survival for the incident cases in the screened group and a 0% survival for the incident cases in the control group.

  
  
• ▪
  

  An RCT in the Czech Republic evaluated 6364 male smokers. Semiannual screening CXR and sputum cytology was compared with a single CXR at the end of 3 years. Lung cancer detection was increased and prognosis improved in the screened group. However, lung cancer mortality was similar. At 15-year follow-up, 247 patients in the semiannual screening group died of lung cancer, compared with 216 in the control group.

  
  
• ▪
  

  The Mayo Lung Project evaluated 10,933 male smokers older than 45 years of age. Initial screening found 51 lung cancers by CXR, 17 by sputum cytology, and 15 by CXR and sputum cytology; these patients were excluded from the study.

  
  
  
  
  • ○
    

    The experimental group (4618 men) received CXR and sputum cytology every 4 months for 6 years

    
    
  • ○
    

    The control group (4593 men) was less intensively screened. They were recommended to get annual CXR and sputum cytology. At least half of this control group did have at least one CXR in the interim.

    
    
  • ○
    

    More cancers were diagnosed in the experimental group (206 compared with 160 in the control group).

    
    
  • ○
    

    More of the detected cancers were resectable in the experimental group (48% versus 32%).

    
    
  • ○
    

    Overall mortality was similar (3.2 per 1000 person years for the intensively screened group compared with 3.0 per 1000 person years for the less intensively screened group).

    
    
  • ○
    

    Long-term follow-up of the data from this study discounts the concept of lead time bias for lung cancer.

  
  
  
  
• ▪
  

  Studies performed at the Memorial Sloan-Kettering Cancer Center (MSKCC) in New York and at Johns Hopkins Hospital compared dual screening with CXR and sputum cytology relative to annual CXR. Dual screening did not improve survival.

SCREENING WITH LOW-DOSE COMPUTED TOMOGRAPHY

Because of the early results with LDCT, showing a four-fold increased ability of LDCT to detect lung cancer compared with CXR, and a six-fold increased ability of LDCT to detect Stage I lung cancer (Early Lung Cancer Action Program, 1999), there was renewed interest in the value of imaging to detect early stage lung cancer.

EXISTING TRIALS

Two prospective, single-arm trials demonstrate the capability of LDCT to detect early-stage lung cancer. The International Early Lung Cancer Action Program (I-ELCAP) study was a multicenter international trial that screened 31,567 patients with LDCT between 1993 and 2005. Another study, combining data from the Mayo Clinic, the Istituto Tumori in Milan, and the Moffitt Center in Tampa, screened 3246 patients, and followed them for an average of 3.9 years.

• ▪
  

  I-ELCAP Study

  
  
  
  
  • ○
    

    31,567 patients older than 40 years of age with a history of smoking or occupational/second-hand smoke exposure were screened for lung cancer with LDCT. Of those, 27,546 patients underwent repeat screening 7 to 18 months later.

    
    
  • ○
    

    The detection rate for lung cancer was 1.3% on baseline examination and 0.3% on follow-up studies. For patients older than 60 years of age who were smokers, the detection rate was 2.7% on baseline examination and 0.6% on follow-up studies.

    
    
  • ○
    

    Overall, lung cancer was detected in 484 patients. The 10-year survival rate for these patients was projected at 80%. This compares favorably with the 16% 5-year survival rate observed in the average clinically detected lung cancer.

    
    
  • ○
    

    Most of the patients (412 of 484) were diagnosed as stage I at time of detection.

    
    
  • ○
    

    In 302 patients with stage I lung cancer who had surgical resection of their tumor within 1 month of cancer detection, there was a 92% 10-year survival for this group.

  
  
  
  
• ▪
  

  Mayo Clinic/Istituto Tumori/Moffitt study

  
  
  
  
  • ○
    

    Starting in 1998, 3246 current or former smokers were screened with LDCT. Screened patients followed for an average of 3.9 years after screening were compared with expected cancer incidence and mortality based on age- and sex-matched data from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) cancer registry database from 1993 to 1998.

    
    
  • ○
    

    One hundred and forty-four lung cancers were diagnosed in the screened population, compared with an expected 44.3 cases of lung cancer.

    
    
  • ○
    

    One hundred and nine lung resections were performed, compared with 10.9 expected operations for a nonscreened population.

    
    
  • ○
    

    LDCT screening detected lung cancer at an earlier stage than expected. Ninety-six lung cancers (67%) were stage I or stage II.

    
    
  • ○
    

    Lung cancer deaths in the screened group:

    
    
    
    
    • ▪
      

      Twelve patients were diagnosed with early-stage lung cancer.

      
      
    • ▪
      

      Thirteen patients were diagnosed with stage III or stage IV lung cancer.

      
      
    • ▪
      

      Six patients died of lung cancer who had negative findings on their screening examinations.

    
    
    
    
  • ○
    

    Only 1 death from lung cancer occurred in screened patients in the first year. The investigators hypothesized that the mortality in the first year was so low because all study patients had to be asymptomatic on enrollment.

    
    
  • ○
    

    Overall, 38 screened patients died. This is not statistically different than the expected number of deaths (38.8) in a comparable nonscreened population, as predicted by SEER data.