| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
1 From the Department of Radiology, University of Pittsburgh, Imaging Research, Suite 4200, 300 Halket St, Pittsburgh, PA 15213. Received April 26, 2005; final version accepted May 6. Address correspondence to the author (e-mail: gurd{at}upmc.edu).
More than 30 years ago when the Breast Cancer Detection Demonstration Project was performed in a large number of institutions as a "field trial," it was considered virtually a revolutionary concept (1). Many associated scientific and operational issues that needed to be addressed were unknown at the time. Despite valid concerns regarding the uncertainties of potential benefits, radiation risk, the ability to maintain participants' compliance, and the potential cost associated with periodic screening for early detection of breast cancer, the program was, in retrospect, one of the primary successful steps in a long series of events that led to the recommendation for annual mammographic screening. Although several similar issues are still being raised 30 years later, the fact is that screening for early detection of breast cancer has become but one factor in the great success we are beginning to observe; namely, a noticeable reduction in mortality associated with this disease (2–4). It took more than 20 years and numerous changes in our practices to come to this point. It is becoming clearer by the year that whatever we are doing collectively through better detection, intervention, and management it seems to be working. The controversy associated with mass screening for early detection of breast cancer, and the manner in which this controversy was addressed, should serve as an important lesson.
When considering screening for early detection of lung cancer, we should be able to learn from the processes undertaken in breast cancer screening and not have to repeat at least some of these steps before exhibiting strong support toward the initiation of such a program—at a minimum as a large field trial in a high-risk population. Lung cancer is a major public health concern that contributes substantially to morbidity and mortality, especially in the smoking and past-smoker populations. Computed tomography (CT) is a very effective tool in depicting nodules (masses) associated with most types of lung cancer, and the majority of studies that have attempted to assess the possible benefits of CT-based cancer screening found the expected increase in detection during the initial examination (the "prevalence" or "baseline" screening examination) (5). The average size of a cancer detected through this type of intervention, in particular during repeat screening (second examination and beyond), is smaller than that detected (frequently as an incidental finding) in the nonscreened population (6,7).
There are several aspects of screening for early detection of lung cancer that we can, for the most part, agree on—perhaps not in terms of absolute values but rather on a relative scale. First, while imaging-based lung cancer screening may not be the optimal solution for all types of lung cancers, we can agree that in most instances "smaller" at detection generally means "earlier" and, on average, also means detection at a lower stage. Because, on average, detection at a lower stage leads to a better outcome, this should be considered a strong point in favor of some type of screening if we can demonstrate that specific practices will actually result in earlier detection. Many physicians believe that the imaging modality of choice for earlier detection should be thin-section low-dose CT (8–10). I am of the belief that there are ample data from case-control and population-based studies that support a statement that well- and meticulously performed screening with posteroanterior chest radiography will also result in earlier detection, albeit probably not as early as with CT examinations (11,12).
Second, unfortunately, modeling the potential benefits (and risks) of lung cancer screening programs or performing well-controlled laboratory studies beyond the ones that have been or are already being performed is not likely to substantially improve what we already know. Each of these study designs is open to criticism in regard to their methods and the potential impact of biases on the ultimate validity of the results and conclusions. For example, the use during modeling of registry data (eg, data from the Surveillance, Epidemiology, and End Results, or SEER, registry [13]) to estimate the possible benefits of screening is likely to result in an underestimation because in general, lung cancer cases that are included in registries (even when adjusted for lesion size at detection) are likely to be more invasive than those detected through a screening program. This is particularly true for cancers of comparable size detected during repeat screening.
Third, data are needed in this area to determine the absolute risk from both radiation and practice-related complications (eg, biopsy, unnecessary treatment of benign disease) (14). Radiation risk for the specific population being screened and radiation risk at the exposure levels and dose rate used can be estimated with relatively large uncertainties but are not going to be substantially better (or more precisely) known in the near future. At the same time, practice-related risks due to intervention-related complications or treatment of benign disease will only be fully appreciated when screening (and all associated follow-up procedures) is widely performed in the field and the number of events increases substantially in every possible category of interest (14). Optimizing interventions and recommendations for acceptable management protocols will only be possible when this important step is taken.
Fourth, the importance of finding and closely monitoring (following up) part-solid and nonsolid nodules will only be defined when enough of these in different sizes and categories are discovered through screening programs. Similarly, the development of optimal management protocols for slow-growing cancers will require a long period of time, and there is no reason to delay screening until we know all (or even most of) the answers for these types of nodules.
Fifth, there is no doubt that the number of possible nodules that are being identified, and the relatively difficult task of determining whether these are changing (growing) over time, will make computer-aided detection (CAD) and characterization an extremely valuable tool in this area (15–19). Although CAD system performance may not be optimal at this time and ease of use may have resulted in but a sporadic use of these systems for detection and classification of lung nodules (masses), improvements in performance and widespread use of these systems are inevitable as they evolve. Large databases of CT examinations that reveal lung cancers must be made available to the research community if we are to achieve this objective in a timely manner.
Sixth, overdiagnosis is always an important issue when screening programs are initiated and will remain an issue for a while. Although overdiagnosis should not be ignored, it is but one of many confounding factors that need to be studied and addressed. However, overdiagnosis need not be the reason for delaying our support for a large field trial of screening.
Last, despite the difficulty in accepting this concept, I personally believe that initially we have to focus on defining as precisely as we can the potential benefits, if any, from chest CT screening. Hence, we should actually use diagnostic-quality images for this purpose by perhaps raising radiation levels (eg, to 
100 mAs), and, by using such images, allow for the optimal assessment of other findings (eg, coronary heart disease, chronic obstructive pulmonary disease) as well. After all, the primary objective of a screening program is the detection of treatable abnormalities that would not otherwise be detected for a while. We may ultimately find (or not find) that the detection of "incidental" findings during screening is as important as the primary goal, namely, the early detection of lung cancer.
The arguments for reducing radiation exposure levels are perhaps premature in this case. If we truly believe that low-dose thin-section CT scanning is as good as conventional-dose CT scanning for the diagnosis of all possible abnormalities that may be depicted in the field of view, we should consider changing our standard of practice to adhere to the "as low as reasonably achievable," or ALARA, concept. The most appropriate screening protocol could always be rereviewed and reassessed after we clearly demonstrate the magnitude of the expected benefit (if any) of a screening program that is performed with high-quality imaging. Mammographic screening was initially implemented with the best available technology at a time when radiation dose was much higher than that currently being applied. The recently computed and eloquently discussed possible risk of radiation-induced cancers in the screened population (20) should constitute but one factor in the consideration of wide-scale implementation of lung cancer screening. We should also consider that if indeed the benefit from earlier detection is substantial, detection of many of the radiation-induced cancers should have similar expected benefits as well. Namely, some (if not all) induced cancers will be detected earlier and hence will be more treatable. Reaffirmation of the "linear no-threshold risk model" (a model that remains heavily dependent on one population exposed in completely different circumstances) every so often, with accompanying statements that relatively large uncertainties (by as much as a factor of three) in estimated risk at low levels "cannot be excluded" (21), should not deter us from thoroughly exploring the potential of practices that could yield great benefits. If we do that, it is very likely that substantial advancements in technology and improvement in practices will be forthcoming, as we have seen in the area of breast cancer screening. The linear no-threshold risk model is clearly an adequate one for making public policy in radiation protection, but many of us remain doubtful that it represents the actual risk at these exposure levels and believe that, if anything, it should be seen as representing an upper limit of the risk.
We need to be mindful that after almost 30 years of extensive studies, uncertainties concerning the absolute risks and benefits of breast cancer screening remain despite the fact that it has become a standard practice and many millions of women undergo such screening each year. These uncertainties are an integral part of the field we are in and are likely to remain for many years to come. Critical reviews of our practices should be welcomed and hopefully will result in appropriate adjustments in our recommendations as we proceed. However, there are no compelling reasons why we should delay a large field trial of lung cancer screening in higher-risk populations.
Screening for early detection of breast cancer has put the radiologist at the center of an important clinical program. Despite the related economic issues, we are proud to play an important role in, and be major contributors to, the continuing successes being observed in this area. If we are meticulous, persistent, and patient, there is a high likelihood that this will also be the case in screening for early detection of lung cancer. It is our choice to proactively lead an exciting endeavor to begin lung cancer screening or to continue to study it for a long period to come. If we choose the latter course, we are likely to find out that the decision to make lung cancer screening happen had been ours all along and that substantially more direct and indirect information was available to us than had been available when screening for breast cancer started. Additionally, the estimated potential benefit in the targeted population is much larger than that of screening mammography (albeit at perhaps a greater risk and cost, as well). The probability that we will learn much more by continuing the "laboratory-type" studies for many years to come is relatively small, and continuing such studies will be done at a great cost in terms of both actual cost and time lost.
In summary, we should continue to study, assess, and learn from experience by initiating a large nonlaboratory screening program that is a true field trial. We can do that in a predefined population. In my own opinion, this population should consist of healthy individuals between 55 and 70 years of age with an extensive smoking history (eg, at least 30 pack-years), and we should establish initially recommended screening and follow-up (when needed) protocols (again in my own opinion, such protocols should include a minimum of two annual screening examinations followed by repeat screening every 2 or 3 years thereafter). If there is a concern about the radiation risk, as perhaps there should be, we can initiate the scanning at a later age (eg, 60 years), and, despite our sensitivity to the sex issue, initially study only men (20). Ideally, a large randomized clinical trial with imaged and nonimaged arms should yield the needed information, but it is extremely difficult to execute such a long-term study in today's environment without biasing the results because of "crossover" between arms and other external factors.
This editorial is not intended to generate agreement. It is obviously the opinion of one individual who experienced for many years similar deliberations associated with what turned out to be an important contribution to improved women's health. The hope is that this editorial will generate discussions that will accelerate a decision to initiate a large field trial for early detection of lung cancer with CT scanning. Having recently and carefully listened to comprehensive reviews of most of the available data regarding lung cancer screening from all over the world, particularly in Japan and the United States, I was very much aware of the repetition in the arguments (both pro and con) that were made approximately 30 years ago in a different yet so similar context of a public health screening initiative that was driven by an imaging procedure.
FOOTNOTES
Author stated no financial relationship to disclose.
See also the editorial by Reich in this issue.
References
Related Article
This article has been cited by other articles:
|
|
 |
|
  E. J. Wilck Computed Tomography Screening for Lung Cancer Ann. Thorac. Surg., February 1, 2008; 85(2): S699 - S700. [Full Text] [PDF]
|
|
|
|
 |
|
 W. C. Black, D. R. Aberle, C. D. Berg, and For the Executive Committee of the National Lung S Large Field Trial for Lung Cancer Screening: Putting the Wrong Cart before the Horse? Radiology, May 1, 2007; 243(2): 314 - 316. [Full Text] [PDF]
|
|
|
|
|
|
D. Gur Large Field Trial for Lung Cancer Screening: Putting the Wrong Cart before the Horse?--Commentary Radiology, May 1, 2007; 243(2): 317 - 318. [Full Text] [PDF]
|
|
|
|
|
|
J. D. Keen, J. E. Keen, M. A. Ganott, J. H. Sumkin, and D. Gur Does the Marginal Benefit Exceed the Marginal Cost? Radiology, April 1, 2007; 243(1): 299 - 300. [Full Text] [PDF]
|
|
|
|
 |
|
 H. Schoder and M. Gonen Screening for Cancer with PET and PET/CT: Potential and Limitations J. Nucl. Med., January 1, 2007; 48(1_suppl): 4S - 18S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
