Early lung cancer detection

Current and ongoing challenges

Authors


Lung cancer is the most common cancer in the world, and the leading cause of death from cancer. By any conventional measure, the enormity of the global problem is immense. In some countries incidence and mortality rates have peaked and are beginning to decline. In many developing nations the burden of disease is rising and will continue to rise because aggressive tobacco industry marketing is leading to a growing prevalence of cigarette smoking. The implications of these incidence and mortality trends are staggering in their own right, but perhaps even more so because of decades of confirming epidemiologic evidence linking prolonged exposure to tobacco smoke as the prime etiologic agent associated with lung cancer. The World Health Organization estimates that by the year 2025 tobacco-related deaths worldwide will exceed 10 million annually, with lung cancer expected to contribute at least 30 percent of that total.1

From a public health standpoint, lung cancer is unique among leading cancers because the underlying etiology is obvious and avoidable through various individual and societal means. However, competing economic interests have frustrated, and continue to frustrate disease control efforts throughout the world. Nearly half of U.S. adults are current or former smokers, and currently about 1 in 4 adults still smoke.2 In the U.S. and other industrialized countries, gains that had been made in reducing tobacco use among children and youth have stagnated and, in some cases, begun to reverse.3, 4 Thus, millions of individuals who are current smokers, former smokers, or eventual smokers will be at risk of lung cancer for the foreseeable future, and each is deserving of the broad spectrum of support and intervention that may help avoid a death from lung cancer.

Among current smokers, cessation of smoking has a marked impact on risk reduction.5 As Peto and colleagues showed recently, the earlier one quits smoking, the lower the absolute risk compared with continuing smokers.5 This is perhaps the soundest advice that can be given to any current smoker. However, while the risk of lung cancer declines after smoking cessation, long-term risk still is associated with the range of dose-response characteristics (age at initiation, duration, number of cigarettes per day, depth of inhalation, etc.) that are associated with lung cancer among current smokers. Further, many current smokers will fail in their attempts to quit. For individuals at elevated risk for lung cancer, the importance of early detection looms large since current disease prognosis is so poor. Five year survival in the U.S. is 14.1%, and while survival is far more favorable for early stage disease, only 15% of cases are diagnosed while still localized.6 The large number of cases and poor survival results in an estimated 2.24 million person years of life lost prematurely due to death from lung cancer.6

The International Conference on Prevention and Early Diagnosis of Lung Cancer was held in Varese, Italy in December 1998 to review the historical evidence related to lung cancer screening, and to review on-going research in new modalities for early lung cancer detection. In attendance were the early trialists, clinicians, epidemiologists, and advocates, each with considerable expertise and interest in lung cancer. It was a gathering that also could be readily be divided into three groups on the issue of whether of not there was sufficient evidence to support a recommendation for lung cancer screening, specifically, those opposed to screening, those in support of screening, and the ambivalent. However, because of attendee's willingness to avoid the rancor and impasse increasingly common during debates about screening, the conference concluded with a consensus statement (included in this supplement) that was both consistent with the evidence and a call to action. It emphasizes the uncertainty of the existing evidence to shape public health policy. The statement also acknowledges that individuals at risk of lung cancer should have an opportunity to reach informed decisions about screening, and that doctors and patients must be given guidance about what is known and not known about the benefits, limitations, and potential harms associated with early lung cancer detection tests. Perhaps most important, the statement acknowledged that recent progress in imaging and identifying molecular markers associated with early disease have significantly increased the possibility that screening for lung cancer might be effective, and the research required to answer these questions is an urgent priority.

The Varese Conference represents the beginning of broad renewed interest in lung cancer screening, for which momentum was dramatically accelerated by publication in July 1999 of early results of lung cancer detection with low-dose helical CT in the Early Lung Cancer Action Project.7 Since that time numerous international workshops have been convened to address a broad range of issues related to lung cancer screening, with the single common denominator of discussing design issues related to prospective evaluation of spiral CT as a screening test. The National Cancer Institute recently announced the initiation of the Lung Screening Study, a year-long, $3 million pilot study of spiral computed tomography (CT) scans to determine the feasibility of a larger, longer randomized trial.8 At this time, whether or when a trial to evaluate lung cancer screening with spiral CT will take place is unknown, as is the number of years that would be required to assemble credible evidence to guide screening policy.

What is clear is that results from spiral CT studies are sufficiently encouraging to lead a growing number of institutions and facilities to promote CT screening to asymptomatic individuals at risk for lung cancer. Availability of CT screening likely will increase over time based on the current evidence, and largely independent of larger, organized initiatives designed to systematically gather data that could be the basis for health policy consensus. Given the high rate of positive results that occur in CT screening for lung cancer and the complexity of the algorithm established by the ELCAP group for following and evaluating small nodules detected by spiral CT, there is ample reason to be concerned about broad dissemination of lung screening outside of experienced, multi-specialty settings. For this reason, it is critically important that appropriate and influential professional organizations begin to create the guidance and foundation for best practices as we await more definitive data. This is not only necessary to ensure the most favorable early detection outcomes to the extent that they are possible, but also to avoid the significant harms that could be associated with false positive and false negative tests.

The magnitude of the disease burden, the rapid pace of technological change, and the anecdotal reports of growing promotion of spiral CT for lung screening means there is a critical need to arrive at a determination about the efficacy of early lung cancer detection as quickly as is feasible. In our judgement, this means the aggressive implementation of a sufficiently large randomized trial that is designed to produce results in a short period of time. It also means that methodological alternatives to randomized trials with death endpoints should be pursued as well. Each of these approaches should be pursued, pursued quickly, and pursued with international cooperation. Lung cancer is a global problem, and it warrants a global solution.

Ultimately, the preferred disease control strategy is the prevention of lung cancer through the elimination of tobacco use altogether, but for the foreseeable future a legacy of decades of tobacco use in a significant proportion of the world's population means we are still challenged to develop public health strategies to reduce deaths and suffering from those destined to develop lung cancer.

Ancillary