From the Pulmonary, Allergy and Critical Care Division (Drs Courtright and Manaker), Hospital of the University of Pennsylvania (Drs Courtright and Manaker), and Department of Medicine (Dr Manaker), University of Pennsylvania.
CORRESPONDENCE TO: Katherine Courtright, MD, Pulmonary, Allergy and Critical Care Division, Hospital of the University of Pennsylvania, Gates Bldg, 806 W, 3400 Spruce St, Philadelphia, PA 19104; e-mail: firstname.lastname@example.org
FINANCIAL/NONFINANCIAL DISCLOSURES: The authors have reported to CHEST the following conflicts of interest: Dr Manaker has received fees as a grand rounds speaker, lecturer, consultant, and expert witness on documentation, coding, billing, and reimbursement from hospitals, physicians, departments, practice groups, professional societies, insurers, and various attorneys. In March 2011, he received $5,400 from Aetna Inc for consultation on diagnosis coding. He serves on the Hospital Outpatient Panel, a federal advisory commission to the Centers for Medicare & Medicaid Services; serves on the Contractor Advisory Committee for Novitas Solutions, Inc, a Medicare contractor; and chairs the Practice Expense Subcommittee of the American Medical Association, Specialty Society Relative Value Unit Update Committee. Dr Manaker also serves on the board of directors of ACCP Enterprises, Inc, a wholly owned, for-profit subsidiary of the American College of Chest Physicians. Dr Courtright has reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.
Primarily detected in an advanced stage, lung cancer is the leading cause of cancer-related death in the United States and remains one of the most costly cancers, with dismal 5-year survival rates.1,2 The National Lung Screening Trial (NLST) demonstrated a 20% relative (1% absolute) reduction in lung cancer mortality by annual screening over 3 years with low-dose CT (LDCT) scanning compared with chest radiography (CXR).3 These results and the US Preventive Services Task Force (USPSTF) grade B recommendation4 fueled pressure for national insurance coverage of lung cancer screening with LDCT scanning. In contrast, the Medicare Evidence Development and Coverage Advisory Committee concluded that the apparent benefits of LDCT scan screening did not clearly outweigh the harms among Medicare beneficiaries. While awaiting the Centers for Medicare & Medicaid Services decision on coverage of LDCT scan screening (expected in 2015), we consider the extent and quality of the data currently available to make a reliable determination of the risk-benefit profile for this preventive service.
The NLST is the largest and the only one of three randomized clinical trials (RCTs) to show a mortality benefit of LDCT scan screening for lung cancer.5 However, the NLST results may not generalize to the community setting. The NLST comprised almost all large academic medical centers and specialty cancer centers with highly trained thoracic radiologists.3 The NSLT surgical mortality rate, fourfold lower than previous reports from national databases, is only partially explained by these centers also having board-certified thoracic surgeons performing diagnostic procedures.6,7
In this ideal study setting, the screening adherence in the NLST exceeded 90% across the 3 study years in both groups,3 an unrealistic expectation outside a rigorous RCT. Fewer participants in the CXR group returned for later screening rounds than in the LDCT scan group, raising the possibility that the CXR group accumulated more time for a nodule to metastasize before detection. Finally, participants with positive LDCT scan findings were less likely to undergo a diagnostic evaluation than those with positive CXR findings, potentially leading to fewer early stage lung cancer diagnoses detected on LDCT scan.8
Limited availability of such specialized resources and wider practice variations in the community setting will likely reduce the mortality benefit as seen in the NLST.3 This reduction may be further exacerbated by the healthy volunteer effect often seen in clinical trials that biases away from the null hypothesis.9
The NLST eligibility criteria did not identify who would most likely benefit from LDCT scan screening.3 A modified lung cancer risk prediction model from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial had significantly greater sensitivity and positive predictive value, with preserved specificity, than the NLST criteria for lung cancer detection.10 Particularly germane to the pending Centers for Medicare & Medicaid Services decision on Medicare coverage of LDCT scan screening are the demographics of NLST participants.3 Less than one-third of NLST participants were ≥ 65 years or older, and none were aged > 74 years. In a post hoc stratified data analysis, the significant reduction in both lung cancer and all-cause mortality disappeared for the group aged > 65 years.11 Limitations of post hoc analyses notwithstanding, these findings should particularly pique Medicare policymakers.
Using NLST eligibility criteria and national surveys of smoking prevalence, coverage of LDCT scan screening is estimated to create > $5 billion of US health-care expenditures annually12; however, self-reported smoking status is often underestimated,13 thereby erroneously reducing these projections. Conversely and also troubling is the perverse incentive for stakeholders to overstate smoking status to obtain coverage for LDCT scan screening, thereby increasing costs for a group without demonstrated benefit.
The Mayo Lung Project showed that overdiagnosis does occur in lung cancer screening,14 estimated at 10% based on criteria resembling that of the NLST.15 In addition, the broncheoloalveolar carcinoma incidence in NLST3 was twice that previously reported in a national sample.16 Discovering a preponderance of typically more-indolent tumor types by LDCT scan will increase the overdiagnosis rate without a concomitant benefit.
False-positive findings confer potential harm from unnecessary interventions and undue anxiety for patients. False-positive results occurred with astonishing frequency in both arms of the trial, but significantly more so in the LDCT scan group.3 Similarly, there were four times as many confirmed lung cancers in the LDCT scan group. Although the initial NLST results do not specify the time gap between a true-positive nodule detection and the diagnostic or therapeutic follow-up, the question remains whether lead time bias extends beyond the NLST exclusion of individuals with a suspicious lesion found on an initial screening CT scan. Premature adoption of LDCT scan screening for lung cancer risks ignores the hard lessons learned from the rapid widespread adoption of prostate-specific antigen screening.17
Finally, radiation-induced cancer is an important hazard for anyone undergoing serial LDCT scan screening because the risk is exacerbated in ever smokers and those aged ≥ 55 years.18 Regardless of whether the low radiation exposure for LDCT scanning is an underestimate, the most common follow-up diagnostic procedures for a suspicious nodule detected in the NSLT were diagnostic chest CT scan and PET scan, which, respectively, deliver five and nine times more radiation than LDCT scanning.5 Without long-term follow-up or consistent reporting of the frequency of additional imaging in the three screening RCTs, the radiation-induced cancer risk from LDCT scan screening cannot be reliably estimated or measured.
As for prostate cancer, breast cancer screening guidelines have come under recent scrutiny. Treatment advances and higher-than-predicted overdiagnosis rates from long-term follow-up have dramatically reduced the previously estimated breast cancer mortality risk reduction for average-risk women undergoing annual mammography.19 With only a single positive LDCT scan screening trial compared with seven breast cancer screening RCTs, it would be remiss not to consider the areas where NLST data are lacking. Regardless of the pending national coverage decision on lung cancer LDCT scan screening, from the natural histories of breast and prostate cancer screening programs, it is clear that we must periodically reassess available evidence.
Under the Patient Protection and Affordable Care Act, private health plans must provide coverage with a waiver of copayments and deductibles for any new preventive services recommended by the USPSTF with a grade A or B. Medicare has no such coverage mandate. In this era of unacceptably high annual health-care expenditures and in the spirit of value-driven care, delaying a coverage decision seems appropriate until a cost-effectiveness analysis based on the NLST data and USPSTF recommendations is completed. Using the number needed to screen of 320 reported in the NLST,3 LDCT scan screening has been predicted to cost approximately an additional $240,000 per additional lung cancer death avoided, assuming 75% screening adherence and accounting for overdiagnosis.20 More recently, simulations based on the USPSTF recommendations projected a comprehensive implementation program to have a 5-year Medicare budget impact of $27.4 billion, or nearly 1% of total annual Medicare spending.12
Many proponents of national coverage for LDCT scan screening argue, probably correctly, that another large-scale RCT is unlikely. However, policymakers increasingly rely on high-quality evidence obtained from pragmatic clinical trials that avoid many of the systematic biases associated with RCTs.21 A pragmatic clinical trial could address the generalizability concern of the NLST and allow for collection of other important patient-centered health outcomes related to cancer screening. A stepped-wedge design with entire centers as the unit of randomization allows all participating centers adequate implementation time and addresses post-NLST clinical equipoise concerns because ultimately, all centers receive the intervention.22
Finally, if national coverage is to proceed now, some reasonable constraints could be based on the current available evidence. Initial private health plan coverage of LDCT scan screening for those meeting NLST smoking criteria and aged < 65 years should have the greatest potential benefits. Alternatively, akin to solid organ transplantation, regionalized screening programs at academic centers would comport with the NSLT design. Either one or a combination of these approaches would allow for collection of long-term follow-up data to inform future decisions on implementing LDCT scan screening programs for lung cancer in the broader community setting.
Become a CHEST member and receive a FREE subscription as a benefit of membership.
Individuals can purchase this article on ScienceDirect.
Individuals can purchase a subscription to the journal.
Individuals can purchase a subscription to the journal or buy individual articles.
Learn more about membership or Purchase a Full Subscription.
Institutional access is now available through ScienceDirect and can be purchased at myelsevier.com.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 4
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a reminder to the email address on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.