Lung Cancer Screening Project

For family medicine, we had to do an evidence based medicine project. The residency director publishes an online newsletter "FP Revolution" and included this article based on what I found. Some of the wording is a little stronger than what I originally had, but here it is.

Does Screening for Lung Cancer Make Any Kind of Sense?
Craig Chen
General Background:
Approximately 160,000 Americans die each year; these are more persons than who die from colon+breast+prostate+melanoma cancers all put together.
Thus the disease burden is substantial and treatment for lung cancer that is clinically diagnosed is usually ineffective. Of patients who present with symptoms due to advanced local or metastatic disease, 75% are not amenable to cure; their 5-year survival rates are <> 8mm, with less than 3.3 mSv dose of radiation (slightly more than annual exposure from natural sources)). In the recent and surprising ELCAP Study [Early Lung Cancer Action Project] baseline screening with low-dose spiral CT (n=1000) in smokers with >10 pack-years and age > 60, spiralCT detected malignant nodules in 2.7% vs. 0.7% with chest x-ray only.
In the International ELCAP Study with 31,567 subjects; CT scans found cancer in 13% of the positive CT scans (484 of 4186 abnormal CTs) including 412 stage I cancers. Estimated 10 year survival rate was 88% for stage I. If resection is performed within 1 month of diagnosis, the estimated 10 year survival rate was 92%. All lung cancers were estimated to have over 80% 10-year survival rate. The cost of low dose CT was < $200. The investigators concluded that lung cancers detected through CT screening are early stage and resectable; routine CT in asymptomatic smokers yields high prevalence of lung cancer. These optimistic comments must be tempered by certain obvious problems including observational cohort study type introducing lead-time bias and problems of overdiagnosis and absence of actual mortality data available from long-term follow-up. Also, extremely noteworthy, are the objections of Welch et al. who a year later wrote in the NEJM: “Last year, the New England Journal of Medicine ran a lead article reporting that patients with lung cancer had a 10-year survival approaching 90% if detected by screening spiral computed tomography. The publication garnered considerable media attention, and some felt that its findings provided a persuasive case for the immediate initiation of lung cancer screening. We strongly disagree.” Furthermore, other critics uncovered the following sources of systematic bias--undisclosed economic interest in the products being tested. The authors of the ELCAP study, Henschke and Yankelevitz, have since acknowledged that their widely cited 2006 article in The New England Journal of Medicine, for one, should have disclosed that they received royalties from their patented “methods to assess tumor growth and regression in imaging tests”—inventions that have been licensed to General Electric (GE), a maker of CT machines. 27 patents, royalties from 1. -In addition, they acknowledged that “virtually all” of the money from a foundation listed as a sponsor of their research actually came from an “unrestricted gift by the Vector Group, the parent company of Liggett Tobacco, which manufactures cigarettes.” $3.6 million improperly disclosed. These putative results must be thrown out completely because the authors are basically crooks, aided and abetted by the lax standards of the NEJM. Finally, another Mayo Clinic CT Study (5-years) with 520 current or former smokers over age 50 with > 20 pack years of smoking, found that 51% of these subjects at baseline had non-calcified nodules, and 1.7% had primary lung cancer. After 3 years of annual CT, 73.5% had non-calcified nodules, 95% were benign (15 underwent surgeries), 68 had primary cancers, of which 61% stage I. There was, however, no difference in mortality relative to historic benchmarks (2.8% vs 2.0%).
Crestanello et. al studied thoracic surgeries done on participants of Mayo Clinic CT screening study (1999, n = 1520, age > 50, >20 pack-year hx); they found 3130 nodules found in 1112 participants (73%). 55 participants underwent surgery (3.6%). Benign disease was found in 10 (18.1%) and lung cancer in 45 (81.9%), of which 28 were stage IA. 27% experienced complications, including 1.7% operative mortality. No reduction in mortality has yet been demonstrated.

Cost-Effectiveness Considerations:
Computer simulated model of annual helical CT vs. no screening for current, quitting, and former heavy smokers, age >60, 55% men. They used a parameter of 50% stage shift with screening leading to 50% fewer advanced-stage cancers and more early stage localized cancers. The model takes into account biases (overdiagnosis, lead-time, etc) The positive outcomes include absolute and relative difference in lung cancer specific deaths; harms include false positive tests and surgeries.
Result: Over 20 years, 13% lower disease-specific mortality, and 1186/100,000 false positive tests. The cost per quality-adjusted life year (QALY) for current smokers is $116,300 [probibitively expensive even though this is the group with the highest yield], $558,600 for quitting smokers [also prohibitively expensive], $2,322,700 for former smokers [also very, very prohibitively expensive even though this is the group that we might most be interested in screening in compassion for their having changed their lifestyle.] If all parameters occurred as maximally favorable (adherence, bias, quality of life, cost of CT, anxiety), the best-case cost would be $42,500/QALY, which is just at (and slightly over) the upper bound of what most would consider cost-effective. [Craig Chen, MS III, UCSF]
COMMENT: There simply is NO reasonable case to make to screen smokers for lung cancer. All smokers should be told to quit. All screening should stop. All industry participation in further research should be banned. Drs. Henschke and Yankelevitz should go to jail.