The National Lung Screening Trial: Where Do We Go From Here?

Denise R. Aberle, MD

Skepticism about the use of low-dose helical computed tomography (LDCT) for early screening of lung cancer has abated in light of recent findings from the National Lung Screening Trial (NLST), which revealed that participants who received these scans had a 20% lower risk of dying from lung cancer than participants who received standard chest x-rays. Now that the data are in, healthcare institutions are considering early computed tomography (CT) screening programs for at-risk patients. This will necessitate the development of recommended standards for the use of these programs.

Denise R. Aberle, MD, one of the principal investigators in the trial, summarized key points from the NLST in August at the 12th International Lung Cancer Congress. In this exclusive interview, Aberle, vice chair of research and professor of both Radiological Science and Bioengineering at UCLA, addressed the challenging issues that need to be considered now that these pivotal data have been released.

Discuss some of the skepticism that you encountered—both prior to and during the NLST trial— regarding the relationship between CT screening and improved mortality.

Randomized trials were conducted in the 1970s using chest x-ray (CXR) and sputum to determine whether screening would reduce lung cancer mortality. A primary observation was that CXR/sputum screening resulted in a significant increase in the diagnosis of early-stage lung cancers, but no decrease in the detection of late-stage cancers and no reduction in lung cancer mortality. Many have attributed these results to overdiagnosis, meaning that screening identified very indolent lung cancers that would not result in death, while failing to detect cancers at an early stage that would otherwise have progressed to advanced stage and death.

Similar concerns have been raised with CT screening and the NLST. The skepticism regarding the NLST was based on the concern that screening would result in overdiagnosis and would subject individuals with positive screens to unnecessary and potentially harmful additional procedures (eg, radiation, biopsies, surgery, emotional distress), while not achieving a reduction in lung cancer mortality.

I believe that scientific skepticism is what keeps medical research honest. In fact, it was because of skepticism that this randomized trial was so necessary.

Can you describe your specific involvement in the NLST?

The NLST was sponsored by the National Cancer Institute (NCI) and represents the union of two research teams: a grant to the American College of Radiology Imaging Network (ACRIN) sponsored by the Cancer Imaging Program, Division of Cancer Treatment and Diagnosis; and a contract called the Lung Screening Study administered through the Division of Cancer Prevention. I was the PI for the ACRIN group and was responsible for 23 of the 33 sites that participated in the NLST. Christine Berg, MD, chief of the Early Detection Research Group, was the project officer for the NLST Lung Screening Study.

The primary objective of the NLST—to determine whether LDCT reduced lung cancer mortality relative to CXR in high-risk individuals—was harmonized across all 33 NLST sites. However, beyond this fundamental epidemiologic question, ACRIN addressed a host of companion questions that examine the following issues: (1) the effects of screening—and of positive screening results—on patients’ quality of life and anxiety levels; (2) a formal cost-effectiveness analysis of the NLST; (3) the creation of a biospecimen archive for purposes of validating biomarkers of risk and of early lung cancer; and (4) the effects of screening on smoking behaviors and beliefs.

As the ACRIN PI, I was centrally involved in the coordination of all substudies. I helped to determine the hundreds of data elements to be captured, developed the case report forms, and established all ACRIN study procedures. I worked with ACRIN’s regulatory team to develop site auditing and monitoring procedures and participated in some of the audits. I ran several monthly ACRIN NLST calls with site investigators, study coordinators, our NCI sponsor, and the various executive committees; assembled specific committees to ensure optimal operations and procedures (ie, the medical physics committee and the protocol and operations committee); ran the biannual NLST Working Group meetings where we discussed progress, problems, and procedures; represented ACRIN in our biannual meetings of the Data and Safety Monitoring Board and our NLST Oversight Committee; cajoled individual investigators and collaborators and functioned as cheerleader for the team; and helped to conduct the NLST at my own site at UCLA.

I was fortunate to work with an excellent team of individuals and institutions in this trial. My colleagues at ACRIN headquarters; Brown Biostatistics; the NCI Cancer Imaging Program and our NCI project officer, Barbara Galen, MSN; the DSMB and Oversight Committee; and the NLST Lung Screening Study investigators were equal partners in this effort.

How can healthcare institutions best initiate CT programs for the early detection of lung cancer?

The flip side of a randomized trial is implementation and dissemination—the goal is to preserve the benefits of the screening intervention, while minimizing the risks. Implementation has many challenges. Healthcare institutions need to approach LDCT screening as a multidisciplinary program that attempts to minimize deaths from lung cancer and other smoking-related diseases. Radiologists, pathologists, surgeons, pulmonologists, primary care physicians, oncologists, and medical physicists are all required to provide an integrated program that can responsibly manage high-risk individuals.

By implementing these multidisciplinary programs, we hope to achieve several important outcomes: (1) high-quality screening interventions that incorporate aggressive smoking cessation and sustained abstinence programs; (2) standardized approaches for image acquisition, interpretation, and communication of results; (3) consistent mechanisms to follow up with individuals who receive positive screens; (4) judicious diagnostic evaluations based on the specific screening result and the evidencebased resources available at the site; and (5) proper documentation of outcomes for purposes of continuous quality improvement (ie, numbers of positive screens; numbers and types of follow-up; relationships of lung cancer diagnosis to screen interpretation; and outcomes of the patients who were screened). This takes a lot of coordination and oversight.

What are some of the challenges that the institutions will face when implementing these programs?

Screening with LDCT has the potential to overwhelm our institutions. The infusion of LDCT into busy radiology practices and the management of the patients who have been screened will be resource-intensive. How we implement this without exhausting our workforce could be challenging.

While the NLST provides some clarification on many questions about LDCT screening for lung cancer, there are several outstanding questions that remain. First, we do not know the “ideal risk profile” of individuals who would benefit from screening. For example, should individuals younger than the NLST cohort (aged 55-74 years) be screened if they have a family history of lung cancer or chronic obstructive pulmonary disease? Second, are there better ways to define a positive screen? Specifically, can we reduce the positivity rates without significantly compromising our rates of screen-detected lung cancer? Third, we’re still trying to determine the best diagnostic strategies for a given indeterminate nodule. For instance, are there biological tests (biomarkers) in blood or other specimens that can better predict which patients with CT-detected lung nodules should be evaluated more aggressively with biopsy? Finally, we still need to determine how far we can reduce radiation exposure with LDCT without compromising image quality.

The answers to these questions will likely not be determined with randomized trials. I do think that many of the outstanding concerns can be addressed through research registries in which individuals agree to have their medical data collected longitudinally. By using standardized data elements and collecting standardized information, we can use the observational data collected through routine practice to provide insights. Although much less expensive than randomized trials, research registries will still require funding (ie, from federal, philanthropic, and industry sources). However, given the relatively limited attention consistently given to lung cancer research, I am concerned that this will not be considered a priority for the health of the American people. That would be an unwise and potentially costly mistake.

You said that implementation of public policy regarding CT screening for lung cancer will be challenging. Can you elaborate on this statement?

A major challenge of LDCT screening for lung cancer will be implementation across the entire at-risk population, regardless of socioeconomic standing. Screening costs are not trivial—not just the screening exam itself, but the medical, economic, and emotional costs of evaluating positive screens. Screening exams are not reimbursed by Medicare or most insurance plans. Therefore, implementation will initially be greater among the better educated, higher socioeconomic portions of the population. The socioeconomically disadvantaged will bear a disproportionate burden with respect to lung cancer and smoking-related disease. A major ethical obligation is the implementation of screening uniformly across the population at risk.

In addition, how do we ensure that screening centers responsibly manage patients? Early on, we will be reliant on selfregulation. The balance of risks to benefits will be unfavorable should we see the growth of entrepreneurial screening centers that take the money, but do not employ rigorous quality-control procedures.

How do you envision overcoming these hurdles?

First, the radiology community will need to modify workflows to accommodate the infusion of these additional screens. This may require additional radiologists or the transfer of some of the image analysis to technically trained staff, using specialized software for nodule detection and characterization. As I previously alluded, standardization of all aspects of the screening process is important to ensure maximum quality at minimal risk.

Secondly, the primary providers will have to adjust to the implementation of LDCT screening. Many will need to be convinced that LDCT screening is effective and cost-effective. Providers will also need to modify workflows as they incorporate education on lung cancer screening into patient encounters, and will have to balance the merits of LDCT screening relative to other evidence-based prevention and early detection interventions.

Finally, the community at large will need to understand the risks and benefits of LDCT screening—the potential risks of cumulative radiation, the practical implications of positive screens, the likelihood of false positive screens, the potential for unnecessary diagnostic evaluations that carry risk and expense, and even the possibility that screening will miss their cancer or will detect a lung cancer that would never have caused symptoms or death (ie, overdiagnosis). This educational process needs to diffuse across all social strata. Groups that have historically been victims of unethical medical practices may need time to believe that lung cancer screening in high-risk individuals is an important health concern, and they can trust the medical community to implement screening wisely. In sum, the challenges for the community are awareness, access, and informed decision making.

You stated that we need robust programs on smoking cessation and risk modification. What channels and strategies might help to achieve these goals?

Smoking cessation and risk modification is a science. The experts in this arena must be at the table when screening programs are in development. How smoking cessation is introduced, how to maintain sustained abstinence, and how the screening process can be used to reinforce healthy behaviors need to be better studied and implemented.

What are the best follow-up measures for positive screenings? And how can complications from these follow-ups be avoided?

The NLST did not specifically set out to assess best diagnostic strategies. The trial investigators achieved consensus on guidelines for evaluation based on nodule size, nodule consistency, and nodule changes over time. There is still work to be done on this. Ultimately, the best diagnostic strategies will take into consideration what the evidence shows from observational study as well as what specific resources are available at a particular institution.

Complications can be minimized if experts in a given specialty are the providers of care. The NLST experienced a relatively low rate of surgical complications because most of the centers had dedicated thoracic surgeons managing the patients requiring surgery. This same principle will apply to all specialties that deal with patients screened for lung cancer.

When will we see additional data regarding how this screening influences patient habits and beliefs?

The data on the relationship of screening and positive screens to smoking habits is currently being analyzed and will hopefully be published in the coming year.

How soon will US society see a reduction in lung cancer mortality as a direct result of these new CT lung cancer screening programs?

A reduction in lung cancer mortality should be realized over the coming decade, but will depend on the degree of penetration of LDCT screening in medical practice as well as the combined impact of screening and modifications in health habits.