National Lung Screening Trial: a breakthrough in lung cancer screening?

One of the most important yet contentious issues in the fight against the most lethal malignancy is the role of CT screening to detect early lung cancer. While several studies have suggested potential in this area, the results of others have indicated a lesser benefit. In a recent commentary, this author listed the lack of a breakthrough in CT screening for lung cancer detection as the greatest disappointment of the last 25 years in thoracic imaging. I added the qualifier ‘‘at least so far’’ because none of the studies published to that point was a randomized trial showing a mortality improvement, considered to be the highest level of evidence. Recent events, however, provide reason for greater optimism. The preliminary announcement demonstrating a significant advantage for the CT screened population of the National Lung Screening Trial (NLST), a randomized trial aimed at evaluating early lung cancer, is a notable milestone. This study of over 53,000 smokers and ex-smokers age 55 to 74 showed a 20.3% mortality reduction in lung cancer among those who underwent low dose helical CT screening as compared to the control group, whose baseline imaging was a chest radiograph. An all-cause mortality reduction of 7% was also found in the study group. While it is clear that the preliminary results of NLST break new ground, enthusiasm should be tempered by the fact that the full study has not yet been published. A recent paper outlined the study principles and methods and provides an excellent roadmap for how the study was performed. There are many gaps that a summary article will fill. A detailed tally of the composition of the patient population, based on age, gender, race, and ethnicity has been published, but it remains to be seen whether subgroup analysis will show a particular advantage for any of these groups using CT-based lung cancer screening. Moreover, implicit in the study design are factors that the scientific community will have to consider carefully before deciding what weight to assign to the results. The study population of 55 to 74 year old smokers and ex-smokers with more than 30 pack-years of smoking history is a somewhat restricted group. It does not provide evidence for or against screening younger patients, who on average have a lower prevalence of lung cancer but more productive years of life to be saved. Nor can the NLST results be applied to individuals with a less extensive smoking history. Lung cancer is increasingly recognized in patients who have no smoking history. In some instances, second-hand smoke, environmental exposure, family history or other known causes have been implicated. However, there are patients for whom no risk factor can be identified. The NLST results do not support CT-based lung cancer screening in this group of patients. The individuals who were enrolled in the study were mainly from large urban medical centers and were presumably highly motivated. It is less certain whether the results are equally applicable to a more general population. The participating centers themselves are mostly large medical centers with considerable expertise in running clinical trials often with subspecialty radiologists. It is difficult to be certain that more general screening programs could be run as effectively. The NLST may also have limitations that lead to underestimation of the benefit of screening. The control group in the study underwent annual chest radiography. However, the current standard of care recommendation for smokers and ex-smokers is no imaging intervention. Although previous randomized studies of lung cancer screening by chest radiography failed to show any advantage, these studies were not as optimally designed as the NLST and included smaller numbers of patients. If chest radiography, in fact, confers a small screening advantage over no screening, the benefit of CT based screening as compared to a non-screened population would be greater. Several other randomized low dose CT trials being performed in Europe, most notably the NELSON trial, use a no-screening strategy in the control arm. Moreover, Henschke have suggested that the full benefit for CT screening may not be achieved for as many as 10 years after baseline with continuous screening. NLST enrollees underwent a prevalence screening and 2 incidence screenings with a maximum of seven years of follow-up. Because a particular design was selected for the NLST based on practical and economic considerations, the study does not address the optimal length of screening.