Best Measure of Lung Cancer Screening: Curability or Mortality?

OAK BROOK, Illinois—Although any steps taken to prevent cancer require many years before they influence outcomes, clinical investigators typically do not consider the time lag needed for prevention to elicit a measurable effect on life and health, opting instead to assess mortality in the short term, Claudia Henschke, PhD, MD, said at the Lung Cancer Workshop IV: Application of High Resolution CT Imaging Data to Lung Cancer Drug Development.

"We all agree that if people stopped smoking, they would be less likely to die of lung cancer. But if you were going to study this, you would randomize people into a smoking-cessation group and a continue-smoking group. Then you would have to follow them long enough to see the difference in the deaths of the two groups. When would that occur? It's not going to occur within the first 5 years. It will take many years to see a difference in death rate, 10 to 15 years and beyond. Likewise, if I find a lung cancer early and treat it, there will be a similar lag time to see an effect on mortality," said Dr. Henschke, professor of radiology, Weill Medical College of Cornell University and New York Presbyterian Hospital.

The International Early Lung Cancer Action Program (I-ELCAP), Dr. Henschke said, is just beginning to determine the length of time before low-dose CT screening and treatment for early-stage lung cancer show a positive difference in outcomes. She believes that studies such as the National Lung Screening Trial (NLST) are not examining the effect of lung cancer screening over a long enough follow-up period. The NLST follows individuals for 3 to 6 years after three rounds of screening. "One of the problems of the NLST is that when you stop screening, you go back to the usual death rate from lung cancer," she said.

Another problem is that clinical trials tend to focus on mortality, which is a population measure that is not relevant to individual patients, according to Dr. Henschke. A more telling measure, she said, is curability, which, in the case of lung cancer screening, is an estimate of the number of patients who may be cured by screening and prompt treatment, based on the Kaplan-Meier survival rate.

In a recent study, Bach et al (JAMA 297:953-961, 2007; see also ONI April 2007, page 1) concluded that CT screening may not reduce the risk of dying from lung cancer. Dr. Henschke pointed out that this study looked at deaths within the first 4 years after the start of screening, "when one cannot reasonably expect to see a reduction in death rate."

The I-ELCAP trial has been conducted in 44 centers in eight countries and enrolled 39,588 asymptomatic individuals for baseline and repeat low-dose CT screening. The study (N Engl J Med 355:1763-1771, 2006) showed that 13% of participants had a positive finding of cancer at baseline screening and 5% on repeat screening; 92% of subsequent biopsies revealed a diagnosis of cancer. A total of 85% of those diagnosed with cancer at baseline had clinical stage I lung cancer (86% on repeat screening), and 95% of these cancers were invasive to the adjacent parenchyma. "If we summarize the results from our study, we would say that we diagnosed 85% of cancers in stage I. If these patients had prompt treatment, then the curability rate was 92%," she said.

Further analysis, a year after the close of data submission to NEJM, examined survival 11 years after lung cancer screening. Using the Kaplan-Meier survival rate to estimate the curability of lung cancer after screening, Dr. Henschke found the overall curability of all cases regardless of the stage of lung cancer and treatment was 81%. Curability of patients with clinical stage I lung cancer who had prompt resection was 93%.

"An important part of understanding screening is that at baseline, you already have 15% of cancers in late stage. Screening will not help that 15%, and they will proceed to die in the initial years. But it's the later years that are important in a screening program, and therein lies controversy. When I find an early cancer, the death that is prevented by screening and resection comes many years after that, not in the initial years," she said.