CT scan of a 72-year-old woman shows dominant pulmonary nodule (arrow) in the right lower lobe that proved at pathology to be adenocarcinoma. Source: American Journal of Roentgenology, Ginsberg et al, 2004. Image source: American Journal of Roentgenology, Ginsberg et al, 2004

The success of the National Lung Screening Trial (NLST) has not closed the book on lung cancer screening. Instead, it has spurred a host of questions related to overdiagnosis, costs and the potential for other screening methods, according to an article and editorial published Aug. 4 in New England Journal of Medicine. NLST researchers and the editorialist reached the same conclusion: screening CT is not ready for prime time.

In November 2010, NLST concluded early after demonstrating that low-dose CT screening of high-risk participants was associated with a 20 percent mortality reduction.

Lung cancer, the authors noted, remains a challenge with a low long-term survival rate. In the U.S., 94 million current or former smokers face increased risk for lung cancer and seven million persons meet NLST eligibility criteria. Previous studies of chest x-ray screening did not provide a mortality benefit, and although molecular markers are under investigation, they are not yet ready for clinical use.

The trial enrolled 53,454 high-risk participants between the ages of 55 and 74 beginning in August 2002 and randomized them to screening CT or x-ray annually for three years and then followed them through December 2009.

Radiologists interpreted the images in isolation and then in comparison with NLST priors, characterizing any non-calcified module at least 4mm as positive. Beyond that, there was variability in follow-up. “[Trial] radiologists developed guidelines for diagnostic follow-up, but no specific evaluation approach was mandated,” according to the writing members of the NLST team, led by Denise R. Aberle, MD, of University of California, Los Angeles.

Follow-up typically entailed further imaging and invasive procedures were infrequent. The number of positive screening results leading to diagnostic follow-up decreased from the first round of screening to subsequent rounds.

Few invasive work-ups led to a major complication; 0.06 percent of the positive lung CT screens that did not lead to a diagnosis of cancer resulted in a major complication, while 11.2 percent of those diagnosed with cancer were associated with a major complication. Corresponding percentages in the x-ray cohort were 0.02 percent and 8.2 percent.

NLST participants differed in important ways from the general population that meets NLST eligibility criteria, continued Aberle and colleagues. They were younger, had a higher level of education and were more likely to be former smokers.

The authors acknowledged several limitations to the trial, admitting that it could have been affected by the “healthy volunteer” effect, which is associated with results more favorable than would be found in the general community.

Importantly, current generation CT scanners are more sophisticated than those used during the trial. These systems could result in a larger reduction in mortality or a higher rate of false positive results. In addition, NLST participating organizations are recognized for their expertise in radiology and oncology. “It is possible that community facilities will be less prepared to undertake screening programs and the medical care associated with them,” wrote Aberle and colleagues.

Finally, NLST did not estimate reduced mortality associated with ongoing CT screening, which may be larger than the 20 percent observed after three rounds of CT screening.

Aberle and colleagues referred to three potentially harmful effects of CT screening: the high rate of false positive results, overdiagnosis of cancers that would otherwise remain indolent and the development of radiation-induced cancers among screened patients.

The researchers referred to several smaller ongoing trials in Europe that will likely be meta-analyzed; these trials are collecting types of data not collected by NLST and could answer additional questions such as best strategies for the follow-up and management of the nodules.

The researchers listed several ongoing questions, including:

• Will populations with risk profiles that are different from those of the NLST participants benefit?
• Are less frequent screening regimens equally effective?
• For how long should screening continue?
• Would the use of different criteria for a positive screening result, such as a larger nodule diameter, still result in a benefit?

The likely path to answering such questions includes modeling and microsimulation, rather than additional large-scale trials, explained the authors.

Aberle et al concluded, “Although some agencies and organizations are contemplating the establishment of lung cancer screening recommendations on the basis of the findings of the NLST, the current NLST data alone are, in our opinion, insufficient to fully inform such important decisions.” They pointed to screening and follow-up costs and hinted at the need for standardized screening schedules and follow-up protocols.

Currently, NLST investigators are examining quality-of-life effects, costs and cost-effectiveness of screening and plan to develop models to assess screening in a wide range of scenarios. In addition, there is potential for researchers to study molecular markers in blood, sputum and urine collected during NLST, which may help better target screening.

Editorialist Harold C. Sox, MD, of the department of medicine and the Dartmouth Institute at Dartmouth Medical School in West Lebanon, N.H., pointed to two differences between the CT and x-ray cohorts. Adherence to the protocol was slightly higher in the CT group and CT participants were less likely than those in the x-ray group to have a diagnostic work-up after the second and third rounds of screening.

However, Sox admitted that the potential effects of these differences probably would not have nullified the large mortality benefit of screening CT in the trial.

Nevertheless, he wrote, “The applicability of the results to the typical practice (external validity) is mixed.” He referred to the experience of the organizations in the trial as well as the demographic make-up of the cohort.

Sox also explored the case for overdiagnosis, and explained that 10 to 15 years of follow-up will be required to test the hypothesis that screening CT resulted in overdiagnosis.

He expressed his agreement with Aberle et al related to the development of future strategies, advising policymakers to wait for more information before endorsing screening programs. Key data include cost-effectiveness and the potential role of biologic markers.

Finally, he pointed out that patients who request CT screening should receive additional assistance in their efforts to quit smoking.