Standardization, coordination needed to maximize lung cancer screening benefits
The US Preventive Services Task Force lung cancer screening guidelines have the potential to substantially reduce lung cancer mortality.
Ninety percent of patients with lung cancer die of the disease, making it the leading cause of cancer-related death in the United States. About half of patients have advanced disease at the time of diagnosis.
“Historically, we haven’t had anything that could help us detect lung cancer early,” Denise R. Aberle, MD, professor of radiology and bioengineering at the David Geffen School of Medicine at UCLA, told HemOnc Today. “We now have the potential to make a significant dent in lung cancer-related mortality. These guidelines provide a positive step in the right direction.”
Denise R. Aberle
The guidelines — issued in December — recommend adults aged 55 to 80 years with a 30 pack-year smoking history who either still smoke or quit within the prior 15 years undergo annual screening with low-dose CT. More than 8 million American adults meet eligibility criteria.
Yet, the benefits of annual screening must be balanced with risks, which include the potential for overdiagnosis, a high frequency of false-positive results and increased radiation exposure. Some researchers also have questioned whether the guidelines — based on results of clinical trials — are ready for real-world implementation.
“There are a lot of questions that remain to be answered that the US Preventive Services Task Force (USPSTF) has touched on, but on which there is still too little evidence to provide guidance,” Aberle said. “It remains to be seen how this will be fully implemented in the field. Screening requires a level of expertise, coordination across multiple and different types of providers, and the standardization of interpretation and management algorithms, all of which will influence both the cost and performance of CT screening.”
HemOnc Today spoke with several clinicians about the advantages and potential harms of annual lung cancer screening, the challenges medical centers might face when they implement screening programs, and the need for further refinement and consensus in guidelines.
A range of guidelines
The USPSTF is the latest entity — following the American Cancer Society, National Comprehensive Cancer Network, American Lung Association (ALA) and others — to recommend some form of lung cancer screening with low-dose CT.
The USPSTF guidelines are based in part on the results of the National Lung Screening Trial (NLST), which included 53,454 adults aged 55 to 74 years who were at high risk for lung cancer. Researchers randomly assigned participants to annual low-dose CT or single-view posteroanterior chest radiography for 3 years.
The results, published in 2011 in The New England Journal of Medicine, showed 247 lung cancer deaths per 100,000 person-years in the CT group and 309 lung cancer deaths per 100,000 person-years in the radiography group. Researchers calculated a 20% (95% CI, 6.8-26.7) reduction in lung cancer mortality with low-dose CT.
In a paper published in Cancer in 2013, Ma and colleagues determined that if a similar screening approach was implemented for the estimated 8.6 million (95% CI, 8 million-9.2 million) eligible US adults, it would have the potential to save 12,250 (95% CI, 10,170-15,671) lung cancer deaths each year.
Source: Photo courtesy of University of Michigan Health System
“Considering lung cancer has the highest mortality rate of all cancers, there is a large public health need for lung cancer screening and a test that works,” Ella A. Kazerooni, MD, MS, professor of radiology and director of cardiothoracic radiology at the University of Michigan Health System and their site primary investigator for the NLST, told HemOnc Today. “The USPSTF recommendation will allow third-party payer coverage under the Affordable Care Act as a required covered benefit and further a coverage decision by CMS through a Medicare Evidence Development & Coverage Advisory Committee review. Together, this will make a huge impact on the lives of people who are at risk for lung cancer.”
The NLST results also showed low-dose CT was associated with a 6.7% (95% CI, 1.2-13.6) reduction in the risk for all-cause mortality.
“The scans not only detect lung cancer but coronary calcification, which is a marker of heart disease, and emphysema,” Kazerooni said. “Cardiovascular disease is the leading cause of death in the United States, and respiratory disease is the third leading cause of death, and it’s reasonable to conclude that it is through the detection of these kinds of abnormalities there is a benefit in all-cause mortality.”
Although NLST eligibility was capped at 74 years, the USPSTF extended the age limit for its recommendation to 80 years based on results of modeling studies that suggested a correlation between older age and increased benefit.
Peter Bach, MD, MAPP, a physician, epidemiologist and director of the Center for Health Policy and Outcomes at Memorial Sloan-Kettering Cancer Center, questioned why the recommendation would be based on model speculation rather than empirical data.
“The recommendation creates a one-size-fits-all across patients groups within risk categories, even though we have a decade of research with model studies showing the risk will vary greatly within this population, and that the risk will be directly related to the probability of benefit,” Bach said. “In addition, the extrapolation up to age 80 is very concerning, especially considering the NLST was grossly under-represented for older people, in general.”
Aberle, the national principal investigator for the American College of Radiology Imaging Network component of the NLST, also expressed concerns.
“For the most part, the task force’s guidelines used the criteria of the NLST, except that they extended the upper age limit,” she said. “That covers a wide swath of individuals, many of whom will never get lung cancer. We need better criteria to identify those at highest risk of lung cancer, including biological tests that are only now emerging and being validated.”
The NCCN guidelines recommend screening for individuals aged as young as 50 years who have additional risk factors besides smoking — largely drawn from statistical modeling studies — such as radon exposure, occupational exposure and a family history of lung cancer.
Otis W. Brawley
“As smoking history and age increase, the relative risk reduction and benefits from screening also increase,” Otis W. Brawley, MD, chief medical officer for the ACS, told HemOnc Today. “One can utilize that principle to say younger adults would have no benefit, if any. I don’t know any responsible experts in the field who would want to screen people under the age of 50. Those who do may not respect that screening is a science of its own.”
Benefit vs. risk
The science behind screening is based on a comprehensive analysis of benefits vs. risk. Due to the potential for harm, the American Academy of Family Physicians has determined there is “insufficient evidence to recommend for or against lung cancer screening.”
Despite this recommendation, a study survey of 962 primary care physicians and general internists conducted from 2006 to 2007 by Klabunde and colleagues suggested PCPs “frequently order lung cancer screening tests for asymptomatic patients.” Results — published in 2012 in Annals of Family Medicine — showed 55% of physicians ordered a chest radiograph, 22% ordered a low-dose spiral CT and fewer than 5% ordered sputum cytology.
A study by Harris and colleagues, published in JAMA Internal Medicine, suggested there are four domains of potential harms from lung cancer screening: physical harms, psychological harms, financial strain and opportunity costs.
“As screening and earlier treatment delays death for only 20% of patients destined to die from lung cancer, 80% of screened people with fatal cancer will die at the same time they would have without screening,” Harris and colleagues wrote. “Early detection from screening has caused these patients to live longer with the diagnosis. ... They experience harm rather than benefit from screening.”
Psychological harms — such as anxiety or distress — can occur due to meeting eligibility criteria for screening, the “labeling” of cancer, the workup associated with incidental findings and overtreatment, they wrote.
“I would never underestimate anxiety associated with screens because it’s real and it definitely negatively impacts quality of life,” Aberle said. “Part of helping to minimize or alleviate that anxiety is establishing strong communication with patients and their health care providers or patient navigators who have expertise in this area.”
Other risks of screening include false-positive results and overdiagnosis.
Of all screens conducted during the 3-year NLST study period, 24.2% were deemed positive. Of these, 96.4% eventually were identified as false positives.
Fungal infections, such as San Joaquin Valley fever, have been mistaken for lung cancer. Tuberculosis or scars from viral pneumonia also can be misread on CT scans.
After additional imaging, 2.5% of all patients with positive results required bronchoscopy, needle biopsy, thoracoscopy or another invasive diagnostic procedure, NLST researchers found. Overall, the evaluation of positive results resulted in 61 complications and six deaths.
William C. Black
William C. Black, MD, chest radiologist at Dartmouth-Hitchcock Norris Cotton Cancer Center and professor of radiology at Geisel School of Medicine at Dartmouth College, said it is more difficult to quantify the effect of overdiagnosis from CT screens.
A modeling study conducted as part of the USPSTF review estimated a 10% to 12% rate of overdiagnosis, meaning the nodules identified would never have posed a threat. Yet, an analysis of NLST data by Patz and colleagues indicated a higher incidence of overdiagnosis from estimates based on approximately 7 years of follow-up. Their results, published in JAMA Internal Medicine, showed an 18.5% (95% CI, 5.4-30.6) risk that any lung cancer detected from low-dose CT was overdiagnosis, a 22.5% (95% CI, 9.7-34.3) risk that a detected non–small cell lung cancer was overdiagnosis, and a 78.9% (95% CI, 62.2-93.5) risk that a detected bronchioaveolar lung cancer was overdiagnosis. However, estimates that any cancer detected from low-dose CT was overdiagnosis with lifetime follow-up decreased to 11% (95% CI, 7-15).
“One problem with overdiagnosis is that we can never identify who is actually overdiagnosed,” Black said. “Those who are overdiagnosed are going to suffer the burden of the disease label and treatment. That’s a very serious harm, although it affects a much smaller percentage of patients than false positives.”
Also, the risk for overdiagnosis is not exclusive to lung cancer screening.
“Overdiagnosis is an inherent part of any screening program,” Kazerooni said. “It’s really a matter of how much overdiagnosis you will tolerate to be able to have a benefit.”
Efforts must be made to improve the accuracy of screening to better distinguish lung cancers from other abnormalities.
“You wouldn’t want to reduce the frequency with which you detect things in the lung, but instead enhance the probability that when you detect something, it’s actually a lung cancer,” Bach said. “In patients who have abnormalities detected that aren’t lung cancer, they cannot benefit because you have not found lung cancer. They are just harmed.”
Thus, additional guidelines are needed to better interpret and manage results from CT scans.
“Everyone is essentially in agreement with how screening could work, but there’s a lot ambiguity on the methods,” Black said. “We need to make it clear how findings should be interpreted and how patients should be managed. Right now, there’s not much consensus in the literature.”
The most widely used guidelines in radiology are the Fleischner guidelines, which set the threshold size for repeating CT in high-risk patients with detected pulmonary nodules larger than 4 mm. Researchers in the NLST also used this threshold for determining nodules “suspicious for” lung cancer, which are initially classified as positive.
However, the American College of Radiology is developing radiology guidelines — expected to be published in the next few months — that would increase the threshold to 6 mm. The guidelines, titled LungRADS, include a structured reporting and management tool for radiologists that is similar to mammography’s BI-RADS, according to Kazerooni, who chairs the American College of Radiology’s Committee on Lung Cancer Screening.
The threshold adjustment may reduce the risk for false-positive results by half, Black said. Radiologists also need defined coding systems to recommend various degrees of workup.
“We are trying to codify the findings the radiologist should classify between benign and highly suspicious,” Kazerooni said. “We need a standardized reporting scheme to have a consistent set of language and terminology to maximize the benefits of lung cancer screening and minimize the downstream testing when not necessary.”
Definitions of how to audit the practice should be another component of the reporting process.
“We need to define a positive screening rate, lung cancer diagnosis rate and the percentage of patients we should be calling back for extra tests,” Kazerooni said. “As we have more sites performing lung cancer screening, we’ll get a better sense of what these benchmarks should be, which will probably vary geographically.”
Aberle said programs should establish registries to collect data and track the implementation and effectiveness of screening. These data are critical to inform screening implementation as experience increases nationally, she said.
Yet, because risks remain, it is vital that patients are aware of them so they can decide whether they should undergo screening, Brawley said. The USPSTF guidelines suggest “the decision to begin screening should be the result of a thorough discussion of the possible benefits, limitations, and known and uncertain harms.” Lung cancer screening guidelines from the ACS place greater emphasis on the need for patient–physician conversations.
“Because we have strong evidence of harm, ethically, there must be some type of informed consent,” Brawley said. “People need to understand the benefit-to-risk ratio, and that the ratio is likely to be different in other environments, such as in community hospitals that didn’t originally participate in the NLST.”
Therefore, ACS guidelines state that “clinicians with access to high-volume, high-quality lung cancer screening and treatment centers should initiate a discussion about screening” with patients who are eligible, rather than automatically recommending screening for all patients who fit the high-risk criteria.
Guidelines from the ALA also emphasize informed decision making, suggesting “the choice to undergo lung cancer screening must be an individual one, and the ALA should ensure that every patient has the information they need to make an informed decision.”
Many clinicians indicated the USPSTF’s mention of informed decision making may not be enough.
“The USPSTF guidelines didn’t say we should ‘offer’ screening, they just recommended screening,” Black said. “It’s a subtle, but important distinction. Historically, we haven’t been offering screening to eligible people. Instead, we’ve been telling eligible people they should get screened.”
Another concern about the USPSTF recommendation centers on whether the results of the NLST will be replicated in the real-world setting.
“There was extraordinarily high compliance and follow-up in the NLST. The population also skewed wealthy, educated and young,” Bach said. “There remain a lot of questions about how well the results of the NLST could be generalized to the population at large.”
The results of the NLST often are compared with the results of the Early Lung Cancer Action Project (ELCAP), conducted by Henschke and colleagues, which found CT was more effective at detecting noncalcified pulmonary nodules than chest radiography (23% vs. 7%).
However, recent analyses of ELCAP data published in Annals of Internal Medicine indicated positive detection would have decreased to 10.2% (95% CI, 9.8-10.6) had 6 mm thresholds been used, and further decreased to 4% (95% CI, 3.7-4.2) with 9 mm thresholds. Those approaches would have translated to 36% to 75% reductions in workup.
Researchers also await results from the population-based, randomized NELSON trial, which used software to evaluate nodules based on their volume or volume-doubling time. Interim results from the trial, published in 2009 in The New England Journal of Medicine, indicated 2.6% of participants had a positive scan in the first round of screening, and 1.8% had a positive scan in the second round. The overall lung cancer detection rate was 0.9%.
Lung cancer screening programs require interdisciplinary coordination to be effective, Aberle said. This requires integrated practice among radiologists, pulmonologists, medical oncologists and thoracic surgeons.
The fact that the NLST was conducted in 33 highly rated, NCI-designated cancer centers presents another concern about the benefit of screening in other environments.
In a paper published in 2013 in the Journal of the American College of Radiology, McKee and colleagues described initial results of the screening program implemented at Lahey Hospital and Medical Center in Burlington, Mass. The hospital followed NCCN screening criteria and analyzed data after the first 500 patients had been screened.
The rate of positive nodule detection after one CT scan was 25% at Lahey Hospital vs. 27% during the first year of the NLST. However, the rate of clinically significant incidental findings at Lahey was about half that observed during the first year of the NLST (5.6% vs. 10.2%).
“There’s about 20 years of health services research which have shown that the care and outcomes of patients who are treated in major centers with high volumes and NCI designations tend to do better than the population at large, and that’s where the NLST was conducted,” Bach said. “Institutions with different results may not be screening correctly, or they may be screening the wrong populations, and that’s a reason to be cautious.”
However, nearly 25% of NLST sites were not academic sites, and many of ELCAP’s 25 screening sites are in the community setting, Kazerooni said.
Setting raises the issue of cost, according to Brawley.
“Some hospitals are offering reduced-cost or even no-cost screening as a loss leader,” Brawley said. “You can make up the cost of reduced screening with the recall rate. Instead of having the 25% recall rate of the NLST, community hospitals are having a 50% recall rate. It’s ironic that the more expert the radiologist, the lower the amount of income billed you get on this test. This presents a tremendous potential for abuse.”
An actuarial analysis by Pyenson and colleagues, published in 2012 in Health Affairs, found the cost of lung cancer screening per life-year saved was $18,862. After converting calculations from prior studies into 2012 dollars, that figure compared favorably with the $31,309 to $51,274 estimated cost per life-year saved for mammography, and the $18,705 to $28,958 estimate for colonoscopy.
Black, who is in the process of publishing cost data he has presented to the NCI, found lung cancer screening as conducted in the NLST can be cost-effective.
“It’s certainly plausible that screening can be done in a cost-effective way from a societal perspective,” Black said. “However, a societal perspective is not actually what drives practice. What drives practice is how the providers are paid. All of those payments and incentives have to be properly aligned for us to really perform lung cancer screening in a cost-effective way.”
Due to the USPSTF recommendation and the requirements of the Affordable Care Act, screening should be covered by third-party payers by January 2015.
“This recommendation is going to lead to insurance mandates, and insurance mandates will lead to measurements of performance, and those measurements might lead to more attempts to screen adults rather than educate them,” Bach said. “These are things I consider to be net harmful, or not as beneficial as a more nuanced set of recommendations would have been.”
Encouraging patients to change detrimental behaviors should be a component of lung cancer screening programs, and provides another means to reduce cost.
“If we are successful in engaging patients who come for screening in smoking cessation programs, we can achieve much higher levels of cost-effectiveness,” Aberle said. “The screens themselves are not the only means of reducing lung cancer mortality. The screening event should be a teachable moment — a means of reducing their risk for continued smoking, and ultimately for lung cancer.” — by Alexandra Todak
Aberle DR. N Engl J Med. 2011;354:395-409.
American Academy of Family Physicians. Summary of Recommendations for Clinical Preventive Services. 2014. Available at: www.aafp.org/dam/AAFP/documents/patient_care/clinical_recommendations/cps-recommendations.pdf. Accessed on Feb. 18, 2014.
American Lung Association. Providing Guidance on Lung Cancer Screening to Patients and Physicians. 2012. Available at: www.lung.org/lung-disease/lung-cancer/lung-cancer-screening-guidelines. Accessed on Feb. 18, 2014.
Harris RP. JAMA Intern Med. 2014;174:281-285.
Henschke CI. Ann Intern Med. 2013;158:246-252.
Henschke CI. Lancet. 1999;354:99-105.
Klabunde CN. Ann Fam Med. 2012;10:102-110.
Ma J. Cancer. 2013;119:1381-1385.
MacMahon H. Radiology. 2005;237:395-400.
McKee BJ. J Am Coll Radiol. 2013;10:586-592.
Moyer VA. Ann Intern Med. 2013;doi:10.7326/P14-9009.
Patz EF. JAMA Intern Med. 2014;174:269-274.
Pyenson BS. Health Aff. 2012;31;770-779.
Van Klaveren RJ. N Engl J Med. 2009;361:2221-2229.
Wender R. CA Cancer J Clin. 2013;63:106-117.
Wood DE. J Natl Compr Canc Netw. 2012;10:240-265.
For more information:
Denise R. Aberle, MD, can be reached at UCLA Department of Radiology, 10833 Le Conte Ave., BL-428 CHS, Los Angeles, CA 90024; email: email@example.com.
Peter Bach, MD, MAPP, can be reached at Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10065; email: firstname.lastname@example.org.
William C. Black, MD, can be reached at Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03756; email: email@example.com.
Otis W. Brawley, MD, can be reached at Winship Cancer Institute, The Emory Clinic ‘C’, 1365 Clifton Road NE, Atlanta, GA 30322; email: firstname.lastname@example.org.
Ella A. Kazerooni, MD, MS, can be reached at University of Michigan Radiology, Cardiovascular Center, Floor 5, Room 5482, 1500 E. Medical Center Drive, SPC 5868, Ann Arbor, MI 48109; email: email@example.com.
Disclosure: Aberle, Bach, Black, Brawley and Kazerooni report no relevant financial disclosures.
Do the risks associated with radiation exposure outweigh the benefits of annual lung cancer screening?
Radiation is a risk associated with annual CT, although it is hard to quantify that risk.
Low-dose CT scans pose risks, particularly for current smokers who potentially will have annual CT scans for up to 25 years per the US Preventive Services Task Force (USPSTF) recommendation. That’s a considerable number of CT scans, and the more CT scans you get, the higher the concern for radiation exposure is. However, we have not been able to clearly ascertain the cancer risk from annual CT beyond estimates and extrapolations. Once you start doing low-dose CT scans in people at high risk for cancer, it will be hard to determine whether a cancer that develops is from smoking, the CT scan or some play between the two. It is going to be hard to sort out and determine the risks from radiation exposure.
As the technology improves, it is likely the radiation dose will decline, and therefore the risk from each individual test also will decline. However, as radiation decreases from each individual scan, there is a tendency to conduct more scans. The overall accumulation is what needs to be evaluated, not the radiation dose on a single test.
In addition to the risk posed by the radiation and the CT scan itself, an even greater risk is the potential harms from overaggressive follow-up of false positives. The study on which the recommendation was based used a conservative protocol for follow-up of incidental or abnormal findings. The follow-up included a repeat scan, which, of course, is another source of radiation exposure. But it did not immediately move to biopsies and surgeries. It is highly questionable whether that is going to be followed in the community. My biggest concern, shared by my colleagues at the American Academy of Family Physicians, is the unknown harms that are likely to occur because of overaggressive follow-up of incidental findings on these tests. Informed decision making is a vital component to annual screening; however, given the scares associated with cancer diagnosis, the risks associated with screening are difficult concepts for patients to understand. Society as a whole tends to reflexively accept screening as something positive, without always considering the potential harms that can result.
The other concern myself and others in the academy hold is that the USPSTF recommendation states current smokers should undergo annual CT while they are smoking, unless they have quit smoking for at least 15 years. Because of the Affordable Care Act’s coverage for preventive services, the scans will be free, although there will be cost for follow-up biopsies and repeat testing. We’re concerned about the message that gives to smokers, which is: Just keep smoking, and we’ll keep screening you and provide you some protection. That really is not true. Screening is associated with only a 20% reduction, at best, in deaths from lung cancer. The message should be: Stop smoking.
Doug Campos-Outcalt, MD, MPA, is chair of the department of family, community and preventive medicine at the University of Arizona College of Medicine, Phoenix. He also is a member of the American Academy of Family Physicians. He can be reached at University of Arizona, 550 E. Van Buren St., Campus P.O. Box: 245105, Building 1, Room 1272, Phoenix, AZ 85004; email: firstname.lastname@example.org. Disclosure: Campos-Outcalt reports no relevant financial disclosures.
The benefits of annual low-dose CT lung cancer screening, done in a structured program similar to the National Lung Screening Trial, clearly outweigh the risks of radiation exposure.
Frank C. Detterbeck
A multi-society, comprehensive, systematic review concluded that screening 10,000 individuals at elevated risk for lung cancer with annual low-dose CT will prevent approximately 30 lung cancer deaths but cause approximately three radiation-induced cancers, typically 10 to 20 years later (Bach PB. JAMA. 2012;307:2418-2429). However, a structured program that minimizes further imaging and biopsies is crucial. The radiation dose of the screening low-dose CT is low (less than half the dose of annual environmental background radiation), but the radiation from subsequent diagnostic imaging studies may be substantial. This is particularly concerning because low-dose CT frequently detects small nodules.
Nevertheless, lung cancer screening studies have shown that we can manage this well without excessive imaging, but structure is crucial. A management algorithm can effectively soothe the anxiety stirred up by the many small findings and counter the push from anxious patients for more imaging or invasive procedures than what is really needed. The keys to effective implementation of screening are the thoughtful, judicious evaluation of the low-dose CT and communication of this interpretation with the patient — not the low-dose CT scan itself or jumping to aggressive intervention. Furthermore, selection of individuals who are clearly at elevated risk also is crucial; extending screening to lower-risk individuals exponentially reduces the benefits, and screening younger individuals increases the risks for radiation exposure over time.
The evidence shows that lung cancer screening, done in an organized program and involving appropriately selected patients, has benefits that outweigh the risks of radiation exposure. The outcomes of unorganized screening (ie, loosely selected individuals, scans interpreted by general radiologists, findings managed by primary care physicians) have not been defined, but they are likely to be highly variable and much less positive. Therefore, all the thoracic societies and the US Preventive Services Task Force recommend that screening be done in a structured program. In such a setting, the risks of radiation exposure are dwarfed by the reduction in lung cancer deaths.
Frank C. Detterbeck, MD, FACS, FCCP, is a professor of surgery and chief of thoracic surgery at Yale University, and associate director of the Yale Cancer Center. He can be reached at Thoracic Surgery, 333 Cedar St., P.O. Box 208062, New Haven, CT 06520-8062; email: email@example.com. Disclosure: Detterbeck reports no relevant financial disclosures.