Publication
Article
Current Evidence for the Use of Emerging Radiologic Technologies for Disease Screening
Author(s):
Recent technologic advances in the field of radiology haveresulted in the availability of several new tests with potentialapplications for disease screening. Presently, these tests are beingmarketed directly to patients as noninvasive means to providepeace of mind that they are disease free. Such assurance is appealingto many individuals, and some are willing to spend up to$1500 to choose from a menu of available diagnostic options.Given that a physician's referral is unnecessary, many healthcareproviders are unaware that such testing has taken place until theirpatients present to them with abnormal test results. In this review,we examine the evidence supporting the use of electron beamcomputed tomography for coronary artery disease screening, spiralcomputed tomography of the chest for lung cancer screening,computed tomographic colonography for colon cancer screening,and total-body computed tomography for general screening.Although some of these modalities show promise for the future,there is insufficient evidence to support the use of any of thesetesting methods for secondary prevention. The potential for harmassociated with false-positive test results, false-negative testresults, undue anxiety, and radiation exposure exists but requiresfurther study to quantify actual risk.
(Am J Manag Care. 2005;11:385-392)
Several laboratory and radiologic tests have beendeveloped to assist clinicians in screening fordisease. Some of these tests, such as mammography,serum cholesterol, and Papanicolaou smears,have shown enough benefit to be recommended forentire populations.1 Other tests, such as chest radiographsand serum tumor markers for ovarian cancer(cancer antigen 125), have failed to demonstrate significantimprovements in outcome to warrant generalacceptance.1-3
Recent advances in radiology have led to the developmentof several screening tests designed for earlydetection of disease. Unlike modalities developed earlier,these tests have gained popularity among the public,possibly because of intense, direct-to-consumer advertisingefforts. Some institutions have responded to thisconsumer demand, recognizing a potential niche in thehealthcare market. Although the use of these testsseems to have plateaued since 2002, the public continuesto display enthusiasm toward these screening tests,4which frequently can be obtained without a physician'sreferral. Some patients may choose to include theirphysicians in the matter before spending significantamounts of money (up to $1500) for such services. Insome cases, however, clinicians can be faced withpatients wishing to discuss abnormal results of tests thatthey never ordered.
This review will enable the practitioner to engage inmeaningful dialogue regarding 4 popular radiologic diagnosticscreening tests, including electron beam computedtomography (EBCT) of the heart, spiral computedtomography (CT) of the lung, virtual computed tomographiccolonography (CTC), and total-body CT.Informed consent before proceeding with these testsentails weighing the risks and benefits, so emphasis isplaced on available outcomes data and the potential forharm to the individual. Before describing specific testmodalities, we present a general overview of the evaluationprocess of screening tests.
EVALUATING A SCREENING TEST
The cutting-edge technology of modern radiologicimaging may capture the attention of patient and physicianalike, but it is incumbent on the physician to seebeyond the enthusiasm and perform a rational analysisof the screening test in context. The Charter on MedicalProfessionalism espouses the view that physicians havea duty to "create new knowledge and ensure its appropriate use," while advocating the appropriate allocationof resources and the "scrupulous avoidance of superfluoustests and procedures."5 This view is consistentwith guidelines provided in the American MedicalAssociation Code of Ethics.6 There are several componentsto consider when analyzing the merits of atest, including the characteristics of the disease forwhich one is screening, the statistical performance ofthe screening test, the effectiveness of the interventionto be performed if the test result is positive, thenature of the patient population to be screened, andthe overall effect of screening on morbidity and mortality(Table 1).
To begin, the disease must be capable of causing sufficientmorbidity and mortality to justify performing thetest. Many diseases, such as diverticulosis or gallstones,can be easily detected with screening tests but areunlikely to cause significant harm in most patients.Next, the prevalence of the disease must be high in thepopulation under consideration, or the disease must beof such significance that identification of even a smallnumber of cases can justify the costs. Finally, the diseasemust be amenable to treatment in its asymptomaticstate.
The screening test should be accurate, safe, andinexpensive. Accuracy describes the ability of the testto discriminate between healthy and diseased subjectsand is usually expressed in terms of sensitivity andspecificity. Generally, the initial test should be highlysensitive to minimize false-negative results. The testshould also be highly specific, or a confirmatory testshould be available to exclude false-positive test results.Sensitivity and specificity are dependent on the cutoffvalue used to define a positive test result. For example,with respect to prostate-specific antigen testing, a cutoffvalue of 10 ng/mL would result in low sensitivity andhigh specificity, while a cutoff value of 1 ng/mL wouldresult in high sensitivity and low specificity. The relationshipof sensitivity and specificity to the cutoff valueis described by the receiver operating characteristiccurve, which defines the accuracy of the test.7
The safety of a screening test is of great importance,as most patients being evaluated by a screening test arenot ill and therefore may not benefit from the test. Animportant risk in radiologic screening tests is the potentialfor harm from radiation exposure.8,9 In fact, someexperts have suggested that written and verbal consentregarding the risks of radiation be obtained before theprovision of CT screening procedures.10 The comfort ofthe test is also a consideration, as patients who are wellmay be unwilling to subject themselves to significantdiscomfort. The cost of the test includes not only the feefor the initial test but also any charges generated by follow-up and treatment. For example, follow-up of anindeterminate nodule identified by CT scan mayrequire periodic scans or biopsies, which may potentiallycause additional complications. Indirect costsinclude the time the patient must take to perform thetest and the discomfort and inconvenience experiencedby the patient. These costs must be multiplied by thelarge numbers of individuals who may be eligible forscreening. Therefore, the total cost to society maybe enormous compared with the benefits that mightbe obtained.
The physician should also consider the overall utilityof the screening test. In the event that the test resultis positive, an intervention will generally be performed.The intervention should be more effective or cheaperthan treatment offered after the disease becomes symptomatic.It must also be acceptable to the patient.Genetic testing for breast cancer susceptibility, forexample, must take account of the fact that the onlycertain preventive option for a woman with a positivetest result is bilateral mastectomy. The characteristicsof the patient population also affect the utility of thetest. The prevalence of the disease determines the predictivevalue of a positive test result. Given a diseasewith 10% prevalence, a test with a specificity of 90%will yield a positive predictive value of51%, while the same test performed on adisease with 0.1% prevalence will yield apositive predictive value of only 1%. Theprevalence of many diseases increaseswith age, resulting in poorer test performanceamong younger patients. Incontrast, older age and comorbiditydecrease the potential longevity gainsfrom screening.
The ultimate utility of a test is bestmeasured by its ability to decrease morbidityand mortality. Although this maybe determined using observational studies (eg, cohort or case control), the randomized controlledtrial remains the preferred approach. Biasesinherent in screening tests can affect survival measures,casting doubt on the validity of screening testeffectiveness. Observational studies are prone toselection bias (ie, participants in a screening programmay lead healthier lives or otherwise differ from thosewho choose not to participate in the program), leadingto the misattribution of the participants' health tothe screening program. In experimental and observationalstudies, improper measurement of the outcomemay result in lead-time and length-time bias. Leadtimebias refers to an apparent increase in survivalamong persons with the disease that is actually solelydue to earlier diagnosis. For example, a patientwho is screened for lung cancer may live 2 yearsfrom the time of diagnosis, while one who is diagnosedonly when the disease becomes clinicallyapparent may survive for only 6 months after diagnosis,even if the course of the disease is identical inboth. Length-time bias occurs when periodic testingreveals a disproportionate number of slowly progressivecases of the disease, while rapidly fatal casesprogress to death between screening intervals andare not detected. Similarly, overdiagnosis bias refersto the detection of subclinical disease (pseudodisease)that would not become evident until the patientdied of another cause.
ELECTRON BEAM COMPUTED TOMOGRAPHYFOR CORONARY HEART DISEASESCREENING
Calcification of the coronary arteries is a commonfinding in advancing atherosclerosis and has been correlatedwith the extent of coronary artery disease inautopsy studies.11,12 Quantification of coronary arterycalcification by EBCT has been suggested as a screeninginstrument for coronary artery disease in the asymptomaticpatient. Electron beam computed tomographicscanners allow for imaging of the heart and coronaryarteries without motion artifact.13 A coronary artery calciumscore (CACS) is calculated by measuring the densityof the coronary tree in several standardized areas.
Because older individuals have higher mean amountsof coronary artery calcification than younger individuals,and men have higher amounts than women, CACSresults should be compared with those of individuals ofthe same sex and similar age to best estimate coronaryartery disease risk. A CACS of zero is associated with avery low risk of an acute cardiac event.14 Young smokers,however, are an important exception; they mayhave underlying coronary disease without coronaryartery calcification, so their actual risk of myocardialinfarction may be missed by EBCT screening.15Although the rate of cardiac events is greater forpatients with very high absolute calcium scores, thepercentile rank of an individual's calcium score may bea more useful predictor of actual risk than the overallCACS.16 For example, a young individual with a moderateCACS but a high percentile for age and sex would beat higher risk than an older individual with the sameCACS and a lower matched percentile.
Because a large number of myocardial infarctionsoccur in patients who were previously asymptomatic,direct-to-consumer advertising efforts for EBCT havefocused on its ability to predict cardiac risk. Severallarge studies16-21 have attempted to define the role ofEBCT as a screening test for coronary artery disease inpatients with risk factors. All of these studies suggestthat the presence of coronary artery calcifications correlateswith the risk of experiencing a cardiac event(Table 2). The risk is greatest for patients with calciumscores in the highest age-and sex-matched quartiles.
Data regarding the superiority of EBCT over traditionalcardiac risk factor assessment are conflicting. Astudy21 of 1196 asymptomatic individuals at high riskfor myocardial infarction found that EBCT did notenhance risk prediction beyond that provided byFramingham risk tables (Table 2). However, 3 prospectivecohort studies16-18 of asymptomatic individualsscreened with EBCT demonstrated that it was more predictiveof risk than traditional cardiac risk factors.
Some physicians may choose to recommend EBCTfor patients reluctant to accept the need for lifestylechanges or medications designed to minimize their cardiacrisk, but data in support of such an approach areinconsistent. Wong et al22 administered follow-up healthbehavior surveys to a group of 703 asymptomaticpatients who underwent initial EBCT. After adjustmentfor other major cardiac risk factors, the presence ofcoronary calcium was significantly associated with newaspirin use, new cholesterol medication, physician consultation,weight loss, and decreased dietary fat intake.It was also strongly associated with increased worry.Conversely, a study by O'Malley et al23 suggested thatcoronary calcification screening was not associatedwith improvement in cardiovascular risk factors or withincreased anxiety. This study was done among asymptomaticyoung military personnel, who may be less fearfulof their mortality and less susceptible to anxietyfrom disease labeling than older individuals with higherdisease prevalence.
It has not been demonstrated that knowledge ofCACSs would ultimately prevent cardiac morbidity andmortality. Information obtained by EBCT may, in fact,represent a form of lead-time bias. Optimal managementbeyond modification of traditional risk factors inpatients with detectable coronary calcifications has yetto be determined. Population-based screening for coronarycalcification has not been found to be cost-effective.24 In summary, although high EBCT calcium valuesappear to correlate with an increased risk of cardiacevents, and although values of zero make the presenceof significant coronary disease unlikely, the lack of outcomesdata regarding the effect of interventionsprompted by an abnormal scan result precludes recommendingthe use of EBCT as a mass screening tool.
SPIRAL COMPUTED TOMOGRAPHYFOR LUNG CANCER SCREENING
As the leading cause of cancer death in the UnitedStates, lung cancer fulfills many of the prerequisitesfor a condition that should benefit from screening(Table 1). Lung cancer is common, with a prevalenceof 170 000 new cases per year. It causes significant morbidityand mortality, resulting in 160 000 deaths peryear.25 In early stages, it is potentially curable, suggestingthe possibility of significant benefit from an adequatescreening tool. Although neither screening withchest radiographs nor screening with sputum cytologicanalysis has been found to reduce mortality,2 spiral CTscanning has been advanced as a potential new tool forlung cancer screening.
Spiral CT scanning of the chest is rapid and requiresless radiation than conventional CT scans. Images areobtained in a single breath hold (about 15-20 seconds)at end inspiration following hyperventilation and do notrequire administration of intravenous contrast material.While chest radiographs have a nodule size detectionthreshold of approximately 1 cm, spiral CT can accuratelydetect nodules as small as 5 mm. The necessaryequipment is readily available and potentially portablein the form of mobile screening units.26
To date, 5 cohort studies26-30 have examined the useof low-dose thoracic CT scans in screening high-riskpopulations (Table 3). These studies demonstrate clearsuperiority of CT scanning for detection of noncalcifiedlung nodules. The Early Lung Cancer Action Project27and the Anti-Lung Cancer Association Project28 foundCT scanning to be 3 to 4 times more sensitive thanchest radiographs. Nodules detected by CT scanningwere, as predicted, smaller in mean diameter than thosedetected by chest x-ray. Finally, a higher percentage ofCT-detected cancers were stage I at the time of surgicalstaging. To date, 5-year survival data are only availablefrom the Anti-Lung Cancer Association Project trial,28which has reported a survival rate of 76%.
The effect that earlier detection and improved 5-yearsurvival rates will have on overall mortality remainsunclear. Randomized controlled investigations of chestx-ray screening for lung cancer performed in the 1970sdemonstrated enhanced detection of earlier-stage cancersand increased 5-year survival rates but failed toshow improvement in mortality.2 The Mayo LungProject enrolled more than 9000 patients and performedquarterly chest x-ray and sputum cytologicanalysis on almost 5000 patients.31 This trial demonstratedsignificantly increased lung cancer detection,tumor resectability, and survival in the interventiongroup compared with the control group (5-year survival,33% vs 15%). However, the overall mortality rate wasnot significantly different (3.2 vs 3.0 per 1000 patientyears).Lead-time bias, overdiagnosis bias, and contaminationhave been cited as explanations for thediscrepancy. It remains to be seen if the superior sensitivityof CT scanning, with even higher percentages ofstage I cancers detected, will translate into reducedmortality.
One major concern regarding the use of CT screeningfor lung cancer is that the high false-positive ratewill result in the performance of many unnecessarylung biopsies. The 5 trials to date26-30 have demonstratedvarying false-positive rates, from 5%26 togreater than 50%29,30 of participants. In the Anti-LungCancer Association Project trial, 35 of 71 biopsiesperformed showed benign findings.28 In contrast, inthe Early Lung Cancer Action Project study,27 the fearof unwarranted biopsy was not realized; of 29 biopsies,only 2 demonstrated false-positive results.However, to achieve this rate in the Early LungCancer Action Project protocol, 20% of all participantswere subjected to extended follow-up withrepeat CT scans (mean, 5 scans per patient during 2years). Therefore, the toll that false-positive findingswill have on patients, physicians, and the healthcaresystem is expected to be substantial. A computer-simulatedcost-benefit analysis concluded that, given thecurrent level of data and probable presence of length-timeand overdiagnosis bias, lung cancer screeningwith CT scanning is not cost-effective.32 This model,however, has been challenged by others who believethat screening with low-dose CT compares favorablywith cost-effectiveness ratios of other acceptedscreening programs.33,34
In an effort to reach a definitive conclusion onwhether CT scanning will reduce mortality and do so atan acceptable cost, the US National Cancer Institute isconducting the National Lung Screening Trial in which50 000 high-risk patients will be randomized to receivechest x-ray or spiral CT scanning every 4 months. Thetrial has been designed to detect a 20% reduction inmortality and should be completed by 2009. Similar trialsare also under way in several European countries.Until clear evidence of a decrease in mortality is demonstrated,spiral CT scanning cannot be recommended forsecondary prevention of lung cancer.
COMPUTED TOMOGRAPHICCOLONOGRAPHY FOR COLON CANCERSCREENING ("VIRTUAL COLONOSCOPY")
Colon cancer is the second leading cause of cancerdeath in the United States. Early detection has beenshown to improve survival rates.35,36 Current methodsto screen for colon cancer are imperfect. Fecal occultblood testing has been shown to improve mortality35 butlacks specificity.37 Flexible sigmoidoscopy providesdirect visualization of lesions but misses proximal neoplasms.37,38 Even colonoscopy, the gold standard, missesup to 6% of polyps greater than 10 mm in size and upto 13% of polyps between 6 and 9 mm.39,40 Given thelimitations and invasive nature of many of the currentscreening methods, proponents of virtual colonoscopyhave portrayed it as an effective alternative for coloncancer screening at the population level.
The bowel preparation for CTC is similar to that ofconventional colonoscopy, although patients may alsobe asked to ingest barium, gastrograffin, or both to tagresidual stool and luminal fluid. During the procedure,patients lie in the lateral decubitus position, and anenema tip is inserted in the rectum. The colon is insufflatedand distended to mild patient discomfort. Whileholding their breath, patients are scanned in supine andprone positions. Data are analyzed with advanced imagingsoftware to achieve 3-dimensional endoluminalreconstruction. The entire procedure lasts approximately10 to 20 minutes. No sedation is required, andpatients may resume normal activities immediately.
Two recently published studies with conflictingresults demonstrate the current limitations and thepotential future utility of CTC. Cotton et al41 comparedvirtual colonoscopy with conventional colonoscopy in615 patients. This study found that radiographic imagingwas significantly inferior to standard colonoscopyin detection of potentially significant lesions. Thesensitivity of CTC in detecting patients with lesionsat least 6 mm and 10 mm in size was 39% and 55%,respectively. In contrast, Pickhardt et al42 used a multisliceCT scanner, advanced 3-dimensional reconstructionsoftware, advanced training for radiologists,and unique stool and liquid tagging methods toachieve sensitivities more comparable to standardcolonoscopy. The study reported sensitivities of 89%and 94% for lesions at least 6 mm and 10 mm, respectively,in size. Most other studies,43-46 almost all performedin higher-risk groups, failed to demonstratesuch impressive results.
The fact that most of the published virtualcolonoscopy results have been of mediocre accuracy forsmall lesions is not surprising. Collapsed loops of colonthat hinder detection are especially prevalent in therectum and sigmoid. Diverticulosis, thickened or complexhaustral folds, undissolved medications, andretained stool lead to errors in interpretation. Flatpolyps and lesions at "hidden" locations (in flexures orat the convergence of 2 folds) can result in false-negativetest results. Software imperfections and radiologistinexperience may present obstacles to accuracy. In onestudy,47 sensitivity increased from 32% after 25 caseshad been reviewed to 92% after 75 cases had beenreviewed by the radiologist.
Compared with the gold standard of conventionalcolonoscopy, virtual colonoscopy offers several advantages.It is quick and minimally invasive, it allows theradiologist to view otherwise inaccessible areas of colon,and it does not require patient sedation. However, it stillrequires bowel cleansing and exposes patients to ionizingradiation. It is a screening procedure only; if alesion is found, patients should proceed to conventionalcolonoscopy for biopsy and resection, if necessary.Most important, the procedure has proven inconsistentin its ability to detect clinically significant lesions lessthan 10 mm in size.
With the adoption of any new technology, there arecost-benefit issues to consider. Sonnenberg et al48 estimatethat virtual colonoscopy costs $3656 more perlife-year saved than conventional colonoscopy. To becompetitive with conventional colonoscopy, virtualcolonoscopy must be offered at one half of the cost orsecure 15% to 20% higher patient participation thanconventional colonoscopy. These targets have not yetbeen attained.
Virtual colonoscopy has excellent potential for coloncancer screening in average-risk populations and maybe equivalent to conventional colonoscopy when performedwith the use of the latest techniques andadvanced training. Nevertheless, cost-effectiveness andreproducibility outside the study setting have yet to bedemonstrated. Until these issues have been addressed,the limitations of virtual colonoscopy preclude advocacyof its use for population-based colon cancer screening.
TOTAL-BODY COMPUTED TOMOGRAPHICSCAN ("VIRTUAL PHYSICAL")
Given the advances in CT technology in detection ofcoronary artery calcification, lung cancer, and coloncancer, the era of total-body CT scanning has emerged.This technology is being advertised directly to patientsas a way to screen for early-stage disease and to givepatients the peace of mind of being disease free. Patientsmay choose from a menu of scans that include imagingof the brain, chest, abdomen, and vasculature. Rapid,3-dimensional scans can be completed within 15 minutes;if EBCT of the heart and virtual colonoscopy arealso desired, additional time, charges, and preparationmay be required. Most tests are done without intravenouscontrast, although techniques vary betweencenters. Radiologists typically discuss the results withpatients shortly after testing is complete. Frequently,patients with abnormal results are referred to their primaryphysicians for further evaluation. The total costof these scans can range between $300 and $1500,depending on the tests chosen. Because most third-partypayers do not routinely cover such expenses,physician and facility reimbursement is directlyreceived from the patient at the time the service isprovided.
To date, no clinical trials have examined outcomesin patients undergoing total-body CT scanning forscreening purposes. Data on the prevalence of false-positiveand false-negative results have not been published,to our knowledge. A retrospective analysis ofabdominal CT scans performed for suspectednephrolithiasis demonstrated a high rate of incidentalfindings of greater than 40%.49 Thoracic CT scans forlung cancer screening in a Mayo Clinic study30 yieldeda 14% rate of incidental findings.
Although some centers market their ability to detecta large number of diseases at an early stage, most ofthese conditions are likely to be benign. Undue anxiety,significant radiation exposure, unnecessary evaluationand treatment, wasted healthcare expenditures, and arisk for iatrogenic complications may result from indiscriminatetotal-body scanning. The lack of sufficientscientific evidence precludes the uniform recommendationof this modality for routine screening.
CONCLUSIONS
With the move to the market model in healthcareduring the past decade, greater attention is now beingapplied to consumer satisfaction. Physicians andhealthcare systems are finding an increasing need to beresponsive to consumer demand, especially with regardto emerging technologies. The ability of new technologiesto diagnose problems in their early stages is appealing,yet current data do not support use of EBCT, spiralCT, CTC, or total-body CT scanning for populationscreening. There is also concern that anxiety and harmmay result if an ill-defined abnormality is detected.Invasive follow-up testing increases the physical risksand financial costs to the patient and ultimately to thehealthcare system. Although the initial examinationsare usually out-of-pocket expenses, follow-up testingprocedures for potential abnormalities are generallycovered services that can potentially add significantfinancial insult to an already stressed insurance industry.A cost-benefit advantage has yet to be demonstratedfor any of the modalities discussed. Patients mustalso be cognizant of the risk of being labeled with a preexistingcondition by insurers, although the discoveredabnormality may be benign.
The recent advances in radiologic imaging are excitingand offer some promise as methods for diseasescreening. Current evidence, however, does not supporttheir widespread use. Further data regarding testaccuracy, safety, cost, and effect on morbidity andmortality are necessary before adopting these CT-basedmodalities as tools for early detection of diseasein asymptomatic individuals.
From the Division of General Internal Medicine, Johns Hopkins School of Medicine,Baltimore, Md.
Address correspondence to: Bimal H. Ashar, MD, Division of General InternalMedicine, Johns Hopkins School of Medicine, 10753 Falls Road, Suite 325, Lutherville, MD21093. E-mail: bashar@jhmi.edu.
Guide to Clinical Preventive Services.
1. US Preventive Services Task Force. 2nded. Baltimore, Md: Williams & Wilkins; 1995.
N Engl J Med.
2. Patz EF Jr, Goodman PC, Bepler G. Screening for lung cancer. 2000;343:1627-1633.
Curr Opin
Oncol.
3. Paley PJ. Ovarian cancer screening: are we making any progress? 2001;13:399-402.
JAMA.
4. Schwartz LM, Woloshin S, Fowler FJ Jr, Welch HG. Enthusiasm for cancerscreening in the United States. 2004;291:71-78.
Ann Intern Med.
5. ABIM Foundation. American Board of Internal Medicine; ACP-ASIMFoundation. American College of Physicians-American Society of InternalMedicine; European Federation of Internal Medicine. Medical professionalism inthe new millennium: a physician charter. 2002;136:243-246.
6. American Medical Association. Principles of medical ethics. Available at:http://www.ama-assn.org/ama/pub/category/2512.html. Accessed January 3, 2005.
Statistical Methods in Diagnostic
Medicine.
7. Zhou XH, Obuchowski N, McClish D. New York, NY: John Wiley & Sons Inc; 2002.
Lancet.
8. Berrington de Gonzalez A, Darby S. Risk of cancer from diagnostic x-rays: estimatesfor the UK and 14 other countries. 2004;363:345-351.
Radiology.
9. Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure during cardiacCT: effective doses at multi-detector row CT and electron-beam CT. 2003;226:145-152.
AJR Am J Roentgenol.
10. Nickoloff EL, Alderson PO. Radiation exposures to patients from CT: reality,public perception, and policy. 2001;177:285-287.
Z Kardiol.
11. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary calcification:insights from sudden coronary death victims. 2000;89(suppl 2):49-53.
J Am Coll Cardiol.
12. Sangiorgi G, Rumberger JA, Severson A, et al. Arterial calcification and notlumen stenosis is highly correlated with atherosclerotic plaque burden in humans:a histologic study of 723 coronary artery segments using nondecalcifying methodology.1998;31:126-133.
J Am Coll Cardiol.
13. Agatston AS, Janowitz WR, Hidner FJ, Zusmer NR, Viamonte M Jr, DetranoR. Quantification of coronary artery calcium using ultrafast computed tomography.1990;15:827-832.
J Am Coll Cardiol.
14. Georgiou D, Budoff MJ, Kaufer E, Kennedy JM, Lu B, Brundage BH.Screening patients with chest pain in the emergency department using electronbeam tomography: a follow-up study. 2001;38:105-110.
Circulation.
15. Schmermund A, Baumgart D, Gorge G, et al. Coronary artery calcium inacute coronary syndromes: a comparative study of electron-beam computedtomography, coronary angiography, and intracoronary ultrasound in survivors ofacute myocardial infarction and unstable angina. 1997;96:1461-1469.
Circulation.
16. Raggi P, Callister TQ, Cooil B, et al. Identification of patients at increased riskof first unheralded acute myocardial infarction by electron-beam computedtomography. 2000;101:850-855.
JAMA.
17. Greenland P, LaBree L, Azen SP, Doherty TM, Detrano RC. Coronary arterycalcium score combined with Framingham score for risk prediction in asymptomaticindividuals. 2004;291:210-215.
Circulation.
18. Kondos GT, Hoff JA, Sevrukov A, et al. Electron-beam tomography coronaryartery calcium and cardiac events: a 37-month follow-up of 5635 initially asymptomaticlow- to intermediate-risk adults. 2003;107:2571-2576.
J Am Coll Cardiol.
19. Arad Y, Spadaro LA, Goodman K, Newstein D, Guerci A. Prediction of coronaryevents with electron beam computed tomography. 2000;36:1253-1260.
Am J Cardiol.
20. Wong ND, Hsu JC, Detrano RC, Diamond G, Eisenberg H, Gardin JM.Coronary artery calcium evaluation by electron beam computed tomography andits relation to new cardiovascular events. 2000;86:495-498.
Circulation.
Circulation.
21. Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does not accuratelypredict near-term future coronary events in high-risk adults [published correctionsappear in 2000;101:697, 1355]. 1999;99:2633-2638.
Am J Cardiol.
22. Wong ND, Detrano RC, Diamond G, et al. Does coronary artery screening byelectron beam computed tomography motivate potentially beneficial lifestylebehaviors? 1996;78:1220-1223.
JAMA.
23. O'Malley PG, Feuerstein IM, Taylor AJ. Impact of electron beam tomography,with or without case management, on motivation, behavioral change, and cardiovascularrisk profile: a randomized controlled trial. 2003;289:2215-2223.
Am Heart J.
24. O'Malley PG, Greenberg BA, Taylor AJ. Cost-effectiveness of using electronbeam computed tomography to identify patients at risk for clinical coronary arterydisease. 2004;148:106-113.
CA Cancer J Clin.
25. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer statistics,2003. 2003;53:5-26.
Lancet.
26. Sone S, Takashima S, Li F, et al. Mass screening for lung cancer with mobilespiral computed tomography scanner. 1998;351:1242-1245.
Lancet.
27. Henschke CI, McCauley DI, Yankelevitz DF, et al. Early Lung Cancer ActionProject: overall design and findings from baseline screening. 1999;354:99-105.
J Clin
Oncol.
28. Sobue T, Moriyama N, Kaneko M, et al. Screening for lung cancer with lowdosehelical computed tomography: Anti-Lung Cancer Association Project. 2002;20:911-920.
Radiology.
29. Diederich S, Wormans D, Semik M, et al. Screening for early lung cancer withlow-dose spiral CT: prevalence in 817 asymptomatic smokers. 2002;222:773-781.
Am J Respir Crit Care Med.
30. Swensen SJ, Jett JR, Sloan JA, et al. Screening for lung cancer with low-dosespiral computed tomography. 2002;165:508-513.
Cancer.
31. Fontana RS, Sanderson DR, Woolner LB, et al. Screening for lung cancer: acritique of the Mayo Lung Project. 1991;67(suppl):1155-1164.
JAMA.
32. Mahadevia PJ, Fleisher LA, Frick KD, Eng J, Goodman SN, Powe NR. Lungcancer screening with helical computed tomography in older adult smokers: adecision and cost-effectiveness analysis. 2003;289:313-322.
Chest.
33. Chirikos TN, Hazelton T, Tockman M, Clark R. Screening for lung cancer withCT: a preliminary cost-effectiveness analysis. 2002;121:1507-1514.
Chest.
34. Wisnivesky JP, Mushlin AI, Sicherman N, Henschke C. The cost-effectivenessof low-dose CT screening for lung cancer: preliminary results of baseline screening.2003;124:614-621.
N Engl J Med.
N Engl J Med.
35. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancerby screening for fecal occult blood: Minnesota Colon Cancer Control Study[published correction appears in 1993;329:672]. 1993;328:1365-1371.
N Engl J Med.
36. Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A case-control studyof screening sigmoidoscopy and mortality from colorectal cancer. 1992;326:653-657.
N Engl J Med.
37. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G;Veterans Affairs Cooperative Study Group 380. Use of colonoscopy to screenasymptomatic adults for colorectal cancer. 2000;343:162-168.
N Engl J Med.
38. Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Riskof advanced proximal neoplasms in asymptomatic adults according to the distalcolorectal findings. 2000;343:169-174.
Gastroenterology.
39. Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomasdetermined by back-to-back colonoscopies. 1997;112:24-28.
J
Natl Cancer Inst.
40. Hixson LJ, Fennerty MB, Sampliner RE, McGee D, Garewal H. Prospectivestudy of the frequency and size distribution of polyps missed by colonoscopy. 1990;82:1769-1772.
JAMA.
41. Cotton PB, Durkalski VL, Pineau BC, et al. Computed tomographic colonography(virtual colonoscopy): a multicenter comparison with standard colonoscopy fordetection of colorectal neoplasia. 2004;291:1713-1719.
N Engl J
Med.
42. Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic virtualcolonoscopy to screen for colorectal neoplasia in asymptomatic adults. 2003;349:2191-2200.
N Engl J Med.
43. Fenlon HM, Nunes DP, Schroy PC III, Barish MA, Clarke PD, Ferrucci JT. Acomparison of virtual and conventional colonoscopy for the detection of colorectalpolyps. 1999;341:1496-1503.
Eur Radiol.
44. Luboldt W, Fletcher JG, Vogl TJ. Colonography: current status, research directionsand challenges: update 2002. 2002;12:502-524.
Gastroenterology.
45. Pineau BC, Paskett ED, Chen GJ, et al. Virtual colonoscopy using oral contrastcompared with colonoscopy for the detection of patients with colorectal polyps.2003;125:304-310.
Gastroenterology.
46. Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evaluation ofcomputed tomographic colonography for screen detection of colorectal polyps.2003;125:311-319.
Am J Gastroenterol.
47. Spinzi G, Belloni G, Martegani A, Xangiovanni A, Favero CD, Minoli G.Computed tomographic colonography and conventional colonoscopy for colon diseases:a prospective, blinded study. 2001;96:394-400.
Am J Gastroenterol.
48. Sonnenberg A, Delco F, Bauerfeind P. Is virtual colonoscopy a cost-effectiveoption to screen for colorectal cancer? 1999;94:2268-2274.
Am J Emerg Med.
49. Messersmith WA, Brown DFM, Barry MJ. The prevalence and implicationsof incidental findings on ED abdominal CT scans. 2001;19:479-481.
