Ten years ago, it was nearly impossible to go through the day without seeing an advertisement for whole-body CT screening. Today it’s a different story. The radiation dose exposure from CT scanning has come under intense scrutiny and the value of CT screening must be proved in rigorous trials before many payors, especially Medicare, will consider reimbursing for an exam.
Lung cancer screening
There are primarily two research camps when it comes to lung cancer screening with CT: Those who rely on data from randomized controlled trials (RCTs) to determine whether CT reduces mortality, and those who rely on observational single-arm studies to determine whether CT improves survival.
While this debate rages on, those in the observational camp stand by their data. The International Early Lung Cancer Action Program (I-ELCAP), for example, screened more than 30,000 people, found nearly 500 cancers, of which 85 percent were stage I (N Engl J Med 2006;355:1763-1771). Henschke et al estimated a 10-year survival rate of 88 percent for these patients with stage I cancer caught early. For approximately 300 patients whose stage I cancer was resected within one month of diagnosis, researchers estimated a 10-year survival rate of 92 percent. While patients were followed for 40 months, researchers estimated a mortality reduction of 80 percent at 10 years.
G. Scott Gazelle, MD, PhD, director of the Institute for Technology Assessment at Massachusetts General Hospital in Boston, and colleagues conducted a comprehensive microsimulation model using actual data from a single-arm Mayo Clinic (Rochester, Minn.) lung cancer CT screening study (Radiology 2008;248:278-287). They found a 10-year survival rate of 87 percent, consistent with the 88 percent found in I-ELCAP. However, whereas I-ELCAP estimated a mortality reduction of 80 percent, Gazelle’s model found an 8 percent relative reduction in lifetime lung cancer mortality with five screening exams, and a 14 percent reduction with 10 screening exams. “With single-arm study designs, changes in mortality cannot be quantified,” they concluded.
Proponents of randomized control trials await the outcome from two trials. The National Lung Screening Trial (NLST) is a multi-million, multi-year U.S. study that randomized 53,000 participants to either low-dose CT or chest x-ray. The NLST is designed to detect as little as a 20 percent decrease in mortality. Results should be available this year. The NELSON randomized controlled trial (Netherlands, Belgium and Denmark) randomized approximately 20,000 participants to low-dose CT screening or no screening. The study will have an 80 percent power to detect a mortality reduction of 25 percent. Results should become available within two years.
“It’s not possible to answer all the relevant questions about lung cancer screening with one clinical trial, because of the various patient populations and characteristics,” Gazelle says. Those questions include: Will those screened derive any mortality benefit? If so, how great is it? What is the cost relative to the gains, relative to the costs of other therapies?
Gazelle says the two RCTs mentioned above were powered for a fairly substantial mortality reduction and “even though they are large, expensive and ambitious, they may end up being inconclusive.” Modeling, on the other hand, can tackle large and diverse populations and should be used in conjunction with RCTs, he says.
Currently, no physician-based medical organizations recommend CT lung cancer screening. They cite the conflicting data from previous RCTs and the risks associated with lung biopsy and surgery, as well as a considerable number of false positive lesions that culminate in downstream costs.
Colon cancer screening
“It’s clear that screening for colon cancer is worthwhile,”