Lung cancer is the leading cause of cancer-related death in developed countries. To date, the only chance of cure is surgical resection in the early stage of disease, with better prognosis for small tumours compared with larger ones. Early detection could potentially decrease lung cancer mortality by diagnosing the disease at an earlier – and potentially more curable – stage. Therefore, lung cancer should be a good candidate for screening.

An important issue is the enrolment of the population into screening trials. About 85% of lung cancers are caused by smoking, and the risk of developing the disease is higher in current smokers than in former smokers or non-smokers. The primary method of prevention is to reduce smoking habits, while screening trials for early detection of lung cancer are considered a secondary method of prevention. Age is another risk factor for developing lung cancer.¹ The aim of screening is the detection of early-stage disease to be operated on, though old age could contraindicate surgery. Therefore, many screening studies have defined the age range for participants and enrol persons who are elderly but healthy enough for potential surgery.^2–10

A screening test must be capable of detecting the lung cancer at a point at which the course of its natural history can be altered through treatment. The screening test should be non-dangerous, nor should it produce too many false-positive results.¹¹ Moreover, its prevalence, specificity, sensitivity, accessibility, cost and associated morbidity must be reasonable.¹²

The most important outcome measure of a screening is the demonstration that the mortality rate from the disease is significantly lower in the total screened population compared with the cancer mortality rate in an equivalent or unscreened population. Individuals with cancer identified by screening will have longer survival times than those diagnosed via usual clinical detection. These apparent increased survival times do not always equate to a reduction in mortality from cancer. Survival from the time of diagnosis is not an appropriate measure of a diagnostic screening test and can be misleading because it is subject to lead-time bias, length-time bias and overdiagnosis bias.¹³ Therefore, all of these biases should be considered when designing a controlled screening trial and in evaluating the reported results.

The Past

The first screening test for lung cancer was the chest X-ray. In a trial performed in London in the late 1960s, 55,034 men were randomised to either chest X-ray every six months for three years or chest X-ray at the beginning and the end of the three-year period (control group). Lung cancer mortality was similar in both groups, with 62 deaths in the study arm and 59 in the control group.^14,15 In the early 1970s, four randomised, controlled trials integrated chest X-ray with sputum cytology, enrolling approximately 37,000 people. These studies reported an increased incidence of earlierstage lung cancers, more resectable cancers and improved five-year survival rates in the screened groups compared with the control groups (35 versus 15%). However, there was no statistically significant difference in mortality attributable to lung cancer between the two groups.^{2–5,16–20} Several critical analyses have been provided for these trials. Many may be invalid due to the absence of an unscreened study arm and thus no determination of true efficacy could have been made.²¹ Moreover, no women were included in any of these trials despite the fact that lung cancer incidence has risen among the female population. The people enrolled in these studies were not at high risk of lung cancer due to smoking history, allowing the randomisation of subjects with only one pack-year of smoking (one pack a day for one year).¹ The sample size of these trials was considered inadequate and, based on these biases, it is impossible to draw any conclusion from these studies.¹¹ To further determine the role of chest X-ray in lung cancer screening, a large, randomised, controlled trial – the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial – enrolled approximately 155,000 men and women aged between 55 and 74 years, but the results are not expected to be published before 2010.^9,22

The Present

Low-dose computed tomography (LDCT) used as a screening tool for lung cancer allows a low-resolution image of the entire thorax to be obtained in a single breath-holding and with low radiation exposure.²³ LDCT is able to detect approximately three to five times as many lung cancers compared with chest X-ray.^7,8 For these reasons, LDCT has emerged as a new and promising screening test for early detection of lung cancer.

Several studies are currently under way, but only preliminary prevalencescreening data are available.^{7,8,10,24–33} Considering the non-definitive results, final comments should not be made, but a few considerations are mandatory. LDCT detects more lung cancers than chest X-ray and mainly in stage I. The detection of early-stage disease does not relate to an improvement in life expectancy. In fact, LDCT may detect neoplastic nodules that could be indolent in their behaviour, so treating these patients cannot increase life expectancy. The detection of small lesions that do not grow has been referred to as overdiagnosis. Moreover, surgery can alter the natural history of the disease. LDCT detects many more non-cancerous than cancerous nodules, but 3D reconstructions with serial CT scans seem to reduce the number of invasive procedures performed on subjects with abnormality but who do not have lung cancer. In the aforementioned trials, non-calcified pulmonary nodules were found in around 5–50% of screened subjects, with a percentage of diagnosed lung cancer ranging from 0.4 to 2.7%. Based on these data, the ratio between the apparent false-positive and the true positive is high.^11,34 For these reasons, the issue of cost-effectiveness of screening with LDCT was addressed, concluding that a baseline LDCT scan for lung cancer screening is potentially highly cost-effective.^35–39

Currently, several large observational and randomised trials are ongoing or planned. Baseline findings of a randomised feasibility trial of lung cancer screening with LDCT versus chest X-ray were reported. A total of 3,318 heavy or long-term smokers were randomised. Thirty versus seven lung cancers were diagnosed, respectively.⁴⁰ More data should come out of the National Lung Screening Trial, which randomised nearly 50,000 current or former smokers to annual screening with LDCT versus chest X-ray for three years. The accrual ended in February 2004; however, the results are not expected to be available until 2009.⁴¹