285

Fast Computed Tomography Detection ofCoronary Calcification in the Diagnosis

of Coronary Artery DiseaseComparison With Angiography in Patients <50 Years Old

James A. Fallavollita, MD; Alan S. Brody, MD; Ivan L. Bunnell, MD;Krishna Kumar, MD; John M. Canty, Jr, MD

Background The predominant cause of coronary arterycalcification is atherosclerosis. Although fast x-ray computedtomography (CT) has been demonstrated to be a sensitivetechnique to detect coronary calcification, the increasing prev-alence of calcification with age has been associated with a lowspecificity for identifying obstructive atherosclerosis. We hy-pothesized that the specificity of this test would be improved ina younger patient population, making it more useful in thediagnosis of coronary artery disease.Methods and Results We compared fast CT-detected cal-

cification with coronary angiography in 106 patients under theage of 50 years. Nonenhanced fast CT scans consisting of 20contiguous 3-mm tomograms of the proximal coronary arterieswere obtained during a single breath hold. A positive scan wasdefined as 4 contiguous voxels (21 mm2) of density >130Hounsfield units in the region of the epicardial coronaryarteries. Calcification detected by fast CT had an 85% sensi-tivity to predict patients with significant coronary artery

number of previous investigations have estab-A lished a strong association between proximal

coronary artery calcification and atherosclero-SiS.1-3 Initial studies using conventional fluoroscopy haveshown that the detection of calcium was highly specificbut only moderately sensitive in predicting the presenceof angiographically significant stenosis at cardiac cath-eterization.4-8 This has raised speculation that moresensitive methods of detection could potentially beuseful in the noninvasive diagnostic evaluation of pa-tients with suspected coronary artery disease (CAD).Results with improved fluoroscopic detection tech-niques have been encouraging in this regard.9-"1

Several recent investigations have capitalized on thepotentially improved densitometric sensitivity of x-raycomputed tomography (CT) to detect coronary calcifi-cation.12-15 Fast CT appears to be the most promisingCT approach, since it is able to obtain contiguous gateddiastolic tomograms of the proximal coronary arteriesduring a single breath hold and avoid misregistration of

Received May 3, 1993; revision accepted September 9, 1993.From the Departments of Medicine, Physiology, and Radiology

at the State University of New York at Buffalo School of Biomed-ical Sciences, the Childrens Hospital of Buffalo, Buffalo GeneralHospital, and the Erie County Medical Center, Buffalo, NY.

Reprint requests to John M. Canty, Jr, MD, Associate Professorof Medicine, University at Buffalo Clinical Center, Room CC-172,462 Grider St, Buffalo, NY 14215.

disease (250% diameter stenosis), with a specificity of 45%.Although the sensitivity to detect multivessel disease was 94%,the sensitivity to detect single-vessel disease was 75%. Chang-ing the threshold for defining a positive fast CT scan from 4 to2 contiguous voxels produced a small improvement in sensitiv-ity, to 88%, but reduced specificity to 36%.

Conclusions Although the specificity to detect angiograph-ically significant coronary disease with fast CT improves in ayounger patient population, it continues to be relatively low. Incontrast to older patient populations, a small but significantnumber of patients <50 years old with angiographically signif-icant coronary artery disease do not have coronary calcifica-tion demonstrated by fast CT. Thus, caution should be used inexcluding significant coronary artery disease on the basis of anegative fast CT study. (Circulation. 1994;89:285-290.)Key Words * tomography * atherosclerosis * coronary

calcification

tomograms because of motion artifacts. Initial studieswith fast CT have indicated that the presence of calci-fication increases dramatically as age exceeds 50 years.Thus, although the sensitivity of fast CT to detectdisease was very high (ranging between 96% and100%), its specificity was low (26% to 28%).14

Since the prevalence of CT-detected calcium appearsto increase markedly with age, we hypothesized that fastCT would be more suitably applied to evaluating youngerpatients, provided that the sensitivity to detect diseaseremained high. To test this, we compared the results ofcalcification detected by fast CT with those by coronaryangiography in patients <50 years of age.

MethodsThe protocol was approved by the investigational review

committee of the State University of New York at Buffalo.

Patient PopulationThe study group consisted of 108 patients <50 years of age

(range, 25 to 49 years) who underwent coronary angiographyfor the clinical evaluation of CAD. Patients were contacted bytelephone after angiography and, after informed consent wasobtained, underwent fast CT. The fast CT scans were ob-tained 59±29 days after angiography.

Coronary AngiographyCoronary angiograms were performed by each patient's

cardiologist for routine clinical indications. All coronary an-

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

286 Circulation Vol 89, No 1 January 1994

TABLE 1. Study Population Characteristics (n=106)

Number %

Male 78 74

White 96 91

Diabetes mellitus 13 12

Family history ofcoronary artery disease 67 63

Cholesterol .200 mg/dL 67 63

Hypertension 46 43

Tobacco abuse 57 54

giograms were reviewed by an observer blinded to the resultsof the CT study. One hundred of the angiograms were

available for review by a second blinded observer. Eachcoronary artery was visually categorized as being normal,insignificantly stenosed, or significantly diseased. Normal ar-

teries were defined as those having no luminal irregularities onangiography. Insignificant stenoses were defined as thoseinvolving <50% of the normal luminal diameter. SignificantCAD was defined as a .50% diameter stenosis. Calipermeasurements were used to confirm categorization of lesions.In cases of disagreement between observers, the lesions were

reviewed by a third blinded observer, and final categorizationwas determined by consensus.

Fast CT ScansNonenhanced fast CT scans were performed on an Imatron

C-100 CT scanner (Imatron Inc, South San Francisco, Calif) inthe high-resolution mode with a 100-millisecond scan time anda 3-mm slice thickness. The field of view was 26 cm, with a

resultant voxel size of 0.5x0.5x3 mm. A scout image was

obtained to identify the area of interest, and 20 consecutivetomograms were obtained in a single breath hold. Scans began1 cm caudad to the carina and encompassed the proximal leftand right coronary arteries. The slices were electrocardio-graphically triggered to obtain each image at the same point indiastole for each of 20 consecutive heart beats as previouslydescribed.14 Total radiation exposure was <0.005 Gy perpatient.Each CT study was interpreted by two independent observ-

ers. Significant coronary calcification was defined as 21 mm2(4 voxels) of contiguous area of CT density >130 Hounsfieldunits (Hu). A lesion score was calculated by multiplying thearea in square millimeters by a "density factor" as previouslydescribed.14 The density factor was determined from the peakdensity within the area of calcification and was equal to 1 forlesions of 131 to 199 Hu, 2 for lesions of 200 to 299 Hu, 3 forlesions of 300 to 399 Hu, and 4 for lesions that were >400 Hu.

Calcific lesions were assigned to the distribution of the leftmain, left anterior descending, left circumflex, or right coro-

nary artery by their anatomic orientation. A total calciumscore was calculated by summing the lesion scores in all 20tomograms.14 The results reflect the average scores of the twoobservers. When there was disagreement regarding the pres-ence or absence of significant calcification within an individualvessel, scans were subjected to review by a third blindedobserver and settled by consensus.

Data AnalysisStandard estimates of sensitivity, specificity, and positive

and negative predictive value were obtained by use of two-by-two contingency tables.16 All values are the mean±SD. Base-line characteristics between study groups were compared by X2and ANOVA,16 with significance defined at the 0.05 level.

ResultsTechnically adequate fast CT studies were obtained

in 106 of the 108 patients. One study had a respiratorymotion artifact, and another was reconstructed with aninappropriate field of view. Demographic variables aresummarized in Table 1. There were 78 men and 28women, with an average age of 43.6±5.4 (SD) years.Ninety-one percent of the population was white. Riskfactor data were obtained from official catheterizationreports and from review of each patient's medicalrecord. There was an average of 2.4±1.2 coronary riskfactors per person.The results of coronary angiography and the distri-

bution of fast CT-detected coronary calcification aresummarized in Table 2. Significant CAD was identifiedin 59 patients (56%). Of the remaining 49 patientswithout significant disease, 20 (19%) had luminal irreg-ularities (<50% diameter reduction) and 27 (25%) hadsmooth, luminally normal coronary arteries. There wasno significant difference in the age of patients with andwithout angiographically significant disease (44.1±5.0versus 42.9±5.8 years). Coronary calcification was de-tected in 76 (72%) of the patients. Those having coro-nary calcification were more likely to be older (P=.002),male (P=.001), and hypercholesterolemic (P=.015).The prevalence of smoking and hypertension was simi-lar in each group. The overall sensitivity of coronarycalcification to detect significant CAD was 85%, with aspecificity of 45% (Fig 1). The positive predictive valueof coronary calcification for detecting significant CADwas 66%. The negative predictive value was 70%.Subgrouping patients by sex or by age (<40 versus >40years) revealed no statistically significant differences.

TABLE 2. Detection of Coronary Calcification With Fast Computed Tomography vsAnglographic Disease

Coronary Calcification

Number % Present Absent

Slgnificant coronary artery disease 59 56 50 9

One-vessel disease 28 26 21 7

Two-vessel disease 14 13 13 1

Three-vessel disease 17 16 16 1

Insignificant coronary disease 47 44 26 21

Luminal irregularities 20 19 13 7

Normal coronary arteries 27 25 13 14

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

Fallavollita et al Fast CT in the Diagnosis of CAD in Young Patients 287

Percent100-

80-

60-

40-

20-

0-

M Sensitivity Specificity

85%

Sensitivity SpecificityFiG 1. Bar graph showing sensitivity and specificity of fast com-puted tomography (CT)-detected coronary calcification to predictangiographically significant coronary artery disease. Although thesensitvty of CT was fairly high, 9 of 59 patents with angiographi-cally significant disease did not have detectable calcification.

Nine of the 59 patients with significant CAD hadfalse-negative fast CT studies. Fig 2 shows the preva-lence of calcification in angiographic subsets. The sen-sitivity of CT was particularly low in patients withsingle-vessel disease and averaged 75%. This groupaccounted for seven of the nine false-negative CTstudies. The sensitivity to detect multivessel disease washigher and averaged 94% for patients with two- andthree-vessel disease combined. Fig 3 shows the effect ofreducing the number of contiguous voxels required todefine a positive CT study on calculations of sensitivityand specificity. When this was reduced from 4 to 2voxels, there was an increase in sensitivity from 85% to88%, but specificity fell from 45% to 36%. A 1-voxelthreshold increased sensitivity to 93%, but specificitywas only 17%.

Fig 4 shows the prevalence of fast CT-detectedcalcification in individual vessels subgrouped into nor-mal, insignificant disease, and significant disease. Twoleft main arteries were too short to be accuratelycategorized. Of the remaining 422 vessels, 106 (25%)were significantly diseased, 132 (31%) had insignificantdisease, and 184 (44%) were luminally normal. Asshown in Fig 4, CT-detected calcification was not invari-ably present in vessels with 250% stenosis. The absenceof calcification in 30% of significantly diseased vessels

100 Sensitivity 93% 94%

80 75%

60

40-

20-

0-Single Vessel Doubie Vessel Triple Veseel

FIG 2. Bar graph showing sensitivity of fast computed tomog-raphy-detected coronary calcification by number of angiograph-ically diseased vessels. The sensitivity to predict muftivesseldisease was very high (94% for two- and three-vessel diseasecombined) but was substantially lower for single-vessel disease.Seven of the 28 patients with single-vessel disease had nodetectable calcification.

Percent100-85% 86% 88% 93%

4 Voxeis 3 Voxeis 2 Voxeis 1 Voxei

FIG 3. Bar graph showing sensitivity and specificity of fastcomputed tomography-detected coronary calcification for theprediction of significant coronary artery disease as a function ofthe number of voxels used to define a positive scan. Changingthe criteria for a positive scan produced small increases insensitivity. As predicted, specificity fell. The largest drop ob-served was for a 1-voxel threshold.

was approximately similar to the rate of false-negativeCT studies (25%) in patients having single-vesseldisease.The distribution of calcium scores as a function of the

angiographic extent of disease is depicted in Fig 5.Individual scores have been plotted on a logarithmicscale to encompass the wide range of scores we ob-served within various angiographic subgroups. The av-erage calcium score increased as the angiographic ex-tent of disease increased (normal, 3+5; <50% stenosis,17±31; one-vessel disease, 78±143; two-vessel disease,126±161; three-vessel disease, 285±345). Although theaverage calcium score increased with the extent ofdisease, there was considerable overlap between angio-graphic subgroups.

Interobserver VariabilityThere was agreement in categorization of patients as

to the presence or absence of significant angiographicdisease in 97% of the studies (Table 3). On a vessel-by-vessel basis, there was agreement in categorization ofnormal and insignificant and significant disease in 80%of the vessels reviewed (317 of 398). Most of thevariability in categorization by vessel involved the dis-tinction between luminal irregularities and smooth,luminally normal arteries.

Prevalence of Calcification1001

80-

60-

40

709L

AACL

201 15%

Luminally Normal ( 50% Stenosis > 50% Stenosis

FIG 4. Bar graph showing prevalence of fast computed tomog-raphy-detected coronary calcification as a function of the angio-graphic severity of disease in individual vessels. There was aprogressive increase in the prevalence of calcification as theseverity of disease within individual arteries increased. Calcifica-tion was not invariably associated with angiographic disease andwas sometimes found in smooth, luminally normal arteries.

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

288 Circulation Vol 89, No 1 January 1994

1,000-

100-

CalciumScore 10-

n=106

a

*~~~~~$ 0

*

% 4

$ 949

aL aLu _, _ * ,

Normal <50% 1 VD 2VD 3VD

FIG 5. Plot showing relation among individual calcium scoresas a function of the degree of angiographic disease. Calciumscores are shown on a logarithmic scale for clarity. Although themean scores for each group increased with increasing extent ofdisease, there was considerable individual patient overlap (.).This precluded reliable subgroup separation by individualscores. Normal indicates smooth, luminally normal arteries;<50%, luminal irregularities; 1 VD, one-vessel disease; 2 VD,two-vessel disease; and 3 VD, three-vessel disease.

All 106 fast CT studies were reviewed by two inde-pendent observers. There was agreement regarding thepresence or absence of coronary calcification in 103 ofthe 106 studies (97%, Table 4). On an individual vesselbasis, there was agreement as to the presence or ab-sence of calcification in 391 of 424 vessels (92%). Totalcalcium scores showed more interobserver variability.Calcium scores agreed by .1 in 68 of the 106 scans(64%). The regression relation for scores between ob-servers was y= 1.05x+ 7.83, r2=.87. The average differ-ence in scores between observers was 16%. Most of thevariability in interpretation arose from uncertainty as towhether significant calcifications were within or outsidethe expected position of the vessel. This most frequentlyoccurred in the distal segments of the right and circum-flex arteries, which were variably delineated by low-density epicardial fat. Intraobserver variability was de-termined in 29 scans (3 of which had no calcification).The mean difference in calcium scores was 8%.

DiscussionThere are several new findings in our study. First,

although fast CT showed a very high sensitivity todetect the presence of angiographically significant mul-tivessel disease, a significant number of patients withsingle-vessel CAD failed to have detectable calcification(25%). This was similar to the frequency with which CTcalcification was absent in individual arteries with an-

TABLE 3. Categorization by Coronary Angiography:Interobserver Variability (n=100)

Observer 1

<50% Stenosis .50% StenoslsObserver 2

<50% Stenosis 46 1.50% Stenosis 2 51

TABLE 4. Categorization by Fast ComputedTomography: Interobserver Variability (n=106)

Observer 1

No Calcification CalcificationObserver 2

No calcification 30 0

Calcification 3 73

giographically significant stenoses (30%). Although thespecificity was better than that demonstrated in studiesof older populations, it continued to be relatively low.

Methodological LimitationsOur investigation retrospectively enrolled patients

who had angiography for routine clinical indications.Although our study group is clearly not representativeof the general population in this age group, it should berepresentative of younger patients referred for thediagnostic evaluation of suspected CAD who wouldmost likely benefit from this noninvasive test. A secondpotential limitation is that we restricted our fast CTscans to the basal 6 cm of the coronary arteries.Although this reduced radiation exposure and simpli-fied the conduct of the scans by allowing them to beperformed in a single breath hold, we would not havedetected calcification in more distal segments. Never-theless, pathological studies' 3 have shown that calcifi-cation isolated to the distal coronary arteries is rela-tively infrequent. When the additional yield of use of 40contiguous 3-mm tomograms was assessed in a previousstudy,14 only 1 of 58 patients had calcification restrictedto the lower 20 levels of the scan. Extrapolating fromthese results suggests that this limitation could haveaccounted for no more than 2 of the 9 patients withdisease and a negative scan in our study.

Findings in the Present StudyWe found that the sensitivity of fast CT to detect

angiographically significant disease was lower than thatpreviously reported in populations that included pa-tients >50 years old.13-'5 This was largely because of itsrelative inability to detect patients with single-vesseldisease, in whom the sensitivity was only 75%. A similarlimitation in the correlation between calcification andangiographic disease was found when we examined theprevalence of calcification in individual arteries. Only70% of significantly diseased vessels had detectablecalcification by fast CT. The prevalence of calcificationfell to 44% in vessels with <50% stenosis and to 15% inluminally normal arteries. These findings are consistentwith the notion that the prevalence of calcificationvaries as a function of the anatomic severity of disease.This is similar to the conclusion reached in studiesexamining this issue with cardiac fluoroscopy.4,8 Similarfindings have also been noted in several preliminaryreports1718 that show increases in CT calcium scoreswith increasing angiographic extent of disease. Unfor-tunately, in young patients, the wide scatter in quanti-tative estimates observed within each angiographic sub-group does not support the notion that a cutoff calciumscore can separate these groups with any degree ofcertainty (Fig 5).

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

Fallavollita et al Fast CT in the Diagnosis of CAD in Young Patients 289

The specificity of fast CT-detected calcification inpatients without significant disease was only moderatelyimproved in this study by restriction of its use to youngerpatients. The high prevalence of calcification in patientswith luminal irregularities but stenoses <50% couldhave accounted for some of the false-positives in our

study. Nevertheless, even when any degree of luminaldisease was considered an "abnormal" angiographicstudy, the specificity improved only to 52%. The lowspecificity could be related to image noise or couldreflect the ability of fast CT to detect atheroscleroticdisease at a stage when there is no luminal narrowing.Previous studies have demonstrated an excellent prog-nosis for patients with smooth, luminally normal angio-grams as opposed to those with insignificant CAD.'9-21Margolis and colleagues6 showed that fluoroscopic cal-cification was an independent predictor of increasedmortality in patients with significant disease, but theydid not separate patients without disease into those withand without luminal irregularities. Thus, the prognosticsignificance of coronary calcification in patients withinsignificantly diseased arteries remains undefined.

Relation of the Present Study to Previous StudiesOur results differ from studies using fast CT in

somewhat older patient populations, in which the pres-ence of calcification was correlated with clinicallyand/or angiographically defined CAD. In a large seriesof patients evaluated with fast CT, Agatston et al14found that the detection of calcification by fast CT hada sensitivity of 96% in patients with either a previousmyocardial infarction or significant angiographic dis-ease. In the subset of their 311 patients <50 years old,only 20 had documented CAD, and the prevalence ofsignificant angiographic disease in their "normal" sub-jects was unknown. Sensitivity in patients <50 years oldranged from 88% to 100% (average, 90%), and speci-ficity ranged from 61% to 75% (average, 66%). Al-though they found that the sensitivity of fast CT was

higher in patients >50 years old (96% to 100%), thespecificity fell to between 26% and 28%, making the testof questionable value in older populations because ofthe high prevalence of coronary calcification.

In a preliminary report of a large cohort of olderpatients (average age, 61+± 12 years), all of whom hadangiography, the same group found that the sensitivityof fast CT was 95%, with a surprisingly high specificityof 76%.'7 This study included an undefined number ofpatients who were enrolled from centers in Japan. Thepotential influences of ethnicity on calcification andatherosclerosis may have increased their specificity,since the age of onset of atherosclerosis in Japan may belater than that of the US population.Breen et al15 correlated the presence of fast CT-

detected calcification with angiography in 100 patients<60 years old (mean, 47±8 years). In contrast to thepreviously noted studies, as well as our own, theirprotocol used 40 contiguous 3-mm scan slices, as op-posed to 20. In addition, they defined the minimum area

required for a significant calcification as 2 adjacentvoxels having a CT density .130 Hu. Using thesecriteria, they found sensitivity to be 100%, with a

specificity of 47%. When we used 2 voxels to define a

positive CT calcification in our study, the sensitivityimproved only modestly to 88%, with a reduction in

specificity to 36%. Differences in our population, in-cluding the lower mean age and the fact that we hadmore than twice as many women in our study, may haveaccounted for the reduction in sensitivity. The indepen-dent influence of including 20 additional caudal tomo-grams cannot be addressed, since Breen et al did notreport the frequency of CT studies demonstrating onlydistal calcification. Based on the prevalence of isolateddistal calcification in previous studies, this is unlikely toexplain the difference between our studies. Finally,differences in the ages of patients with and withoutanatomic disease as well as the extent of CAD, whichwere not reported in their study, could have led todifferent results.A recent postmortem study by Mautner et a122 dem-

onstrated that histopathological calcium was present in7% to 9% of coronary segments from patients withsignificant CAD who had all undergone coronary arterybypass graft surgery. Importantly, 21 of the 22 patientshad histopathological evidence of one or more calciumdeposits in the native coronary circulation. Althoughthe correlation between histopathological calcificationand fast CT-detected calcium remains unknown, arecent pathological study by Simons et a123 correlatedcross-sectional area reduction in histological sectionswith fast CT-detected calcifications from 13 excisedpostmortem hearts. They used 1 voxel of CT density>130 Hu to define significant calcification and noted a59% sensitivity and a 90% specificity for detecting anyhistopathological coronary artery (disease being de-fined as .75% area stenosis). This reinforces the notionthat atherosclerotic disease is not invariably associatedwith calcification. They were also able to correlate thedegree of "calcium burden" (voxels of CT density > 130Hu) with plaque area and luminal area stenosis. Al-though the spread of their data was large, the absence ofcalcium had a negative predictive value of 97.5% for theabsence of significant pathological atherosclerotic dis-ease. Although encouraging, the clinical applicability ofa single high-density voxel in patients appears to belimited by the additional image noise caused by beamhardening and scatter encountered in vivo. In our study,the use of a 1-voxel threshold increased the sensitivity to93%, but specificity fell to 17%.Only one previous study has examined the ability of

coronary calcification detected with cardiac fluoroscopyto predict disease in a patient population of similar age.Loecker et al"l compared the diagnostic use of cardiacfluoroscopy, exercise ECG, and `1Tl scintigraphy in alargely asymptomatic male military population under-going diagnostic evaluation for suspected CAD. Al-though their average age was similar to that in our ownstudy (40.2+5.0 years), 13% of the population under-going angiography was >50 years old. Since a positivenoninvasive test was required before cardiac catheter-ization was performed, only approximate sensitivity andspecificity for angiographically defined disease wereavailable. Fluoroscopically detected calcification had asensitivity for predicting significant disease (.50% lu-minal diameter stenosis) of 66%, with a specificity of77%, and was as accurate as `'Tl scintigraphy in pre-dicting the presence of significant CAD.

Clinical ImplicationsThe clinical utility of detecting coronary calcification

with fast CT in the noninvasive evaluation of CAD has

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

290 Circulation Vol 89, No 1 January 1994

not been established. Unfortunately, a negative studydoes not reliably exclude anatomically significant CADin patients <50 years old. Furthermore, the specificity islow and is only modestly improved by restricting its useto this age group. In screening of populations with aneven lower pretest likelihood of disease than thosereferred for angiography, the diagnostic accuracy of fastCT for predicting significant CAD would be expected tobe even lower. Even though fast CT has the potentialadvantage of being quantified, further studies will berequired to determine whether calcium scores can beused to provide any diagnostic or prognostic informa-tion over that provided by the presence or absence ofcalcification alone.

Scanning 20 levels and using 4 contiguous voxels>130 Hu in a single tomogram to define a positive CTstudy results in an overall sensitivity in detecting signif-icant CAD that, although imperfect, appears to becomparable to noninvasive tests that assess physiologi-cal end points. Prospective studies directly comparingthe efficacy and cost-effectiveness of CT in relation toother noninvasive tests and/or testing algorithms shouldbe performed before its clinical use is recommended.

AcknowledgmentsWe would like to thank Robert Kufchak and Debra Nurmi

for their technical assistance and Nina Goss for her secretarialassistance.

References1. Blankenhorn DH, Stern D. Calcification of the coronary arteries.Am J RoentgenoL 1959;81:772-777.

2. Frink JR, Achor RW, Brown AL, Kincaid OW, Brandenburg RO.Significance of calcification of the coronary arteries. Am J Cardiol.1970;26:241-247.

3. McCarthy JH, Palmer FJ. Incidence and significance of coronaryartery calcification. Br Heart J. 1974;36:499-506.

4. Hamby RI, Tabrah F, Wisoff BG, Hartstein ML. Coronary arterycalcification: clinical implications and angiographic correlates. AmHeart J. 1974;87:565-570.

5. Aldrich RF, Brensike JF, Battaglini JW, Richardson JM, Loh IK,Stone NJ, Passamani ER, Ackerstein H, Seningen R, Borer JS,Levy RI, Epstein SE. Coronary calcifications in the detection ofcoronary artery disease and comparison with electrocardiographicexercise testing: results from the National Heart, Lung, and BloodInstitute's Type II Coronary Intervention Study. Circulation. 1979;59:1113-1124.

6. Margolis JR, Chen JTT, Kong Y, Peter RH, Behar VS, Kisslo JA.The diagnostic and prognostic significance of coronary artery cal-cification: a report of 800 cases. Radiology. 1980;137:609-616.

7. Hung J, Chaitman BR, Lam J, Lesperance J, Dupras G, Fines P,Bourassa MG. Noninvasive diagnostic test choices for the eval-uation of coronary artery disease in women: a multivariate com-parison of cardiac fluoroscopy, exercise electrocardiography andexercise thallium myocardial perfusion scintigraphy. J Am CollCardioL 1984;4:8-16.

8. Uretsky BF, Rifkin RD, Sharma SC, Reddy PS. Value of fluo-roscopy in the detection of coronary stenosis: influence of age, sex,and number of vessels calcified on diagnostic efficacy.Am Heart J.1988;115:323-333.

9. Detrano R, Markovic D, Simpfendorfer C, Franco I, Hollman J,Grigera F, Stewart W, Ratcliff N, Salcedo EE, Leatherman J.Digital subtraction fluoroscopy: a new method of detectingcoronary calcifications with improved sensitivity for the predictionof coronary disease. Circulation. 1985;4:725-732.

10. Detrano R, Simpfendorfer C, Day K, Salcedo EE, Rincon G,Kramer JR, Hobbs RE, Shirey EK, Rollins M, Sheldon WC.Comparison of stress digital ventriculography, stress thallium scin-tigraphy, and digital fluoroscopy in the diagnosis of coronary arterydisease in subjects without prior myocardial infarction. Am JCardiol. 1985;56:434-440.

11. Loecker TH, Schwartz RS, Cotta CW, Hickman JR. Fluoroscopiccoronary artery calcification and associated coronary disease inasymptomatic young men. JAm Coll Cardiol. 1992;19:1167-1172.

12. Reinmuller R, Lipton MJ. Detection of coronary artery calcifi-cation by computed tomography. Dynamic Cardiovasc Imaging.1987;1:139-145.

13. Tanenbaum SR, Kondos GT, Veselik KE, Prendergast MR,Brundage BH, Chomka EV. Detection of calcific deposits incoronary arteries by ultrafast computed tomography and corre-lation with angiography. Am J Cardiol. 1989;63:870-872.

14. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M,Detrano R. Quantification of coronary artery calcium usingultrafast computed tomography. J Am Coll Cardiol. 1990;15:827-832.

15. Breen JF, Sheedy PF, Schwartz RS, Stanson AW, Kaufmann RB,Moll PP, Rumberger JA. Coronary artery calcification detectedwith ultrafast CT as an indication of coronary artery disease.Radiology. 1992;185:435-439.

16. Snedecor GW, Cochran WG. Statistical Methods. Ames, Iowa: IowaState University Press; 1967.

17. Agatson AS, Janowitz WR, Aizawa N, Gasso J, Hildner F,Viamonte M, Prinease R. Quantification of coronary calciumreflects the angiographic extent of coronary artery disease. Circu-lation. 1991;84(suppl II):II-159. Abstract.

18. Leimgruber PP, Judge TP, Fuho BE, Viren FK, Shields JP. Cor-relation between coronary calcification assessed by ultrafast CTand angiographically documented coronary artery disease. JAmColl Cardiol. 1993;21:54A. Abstract.

19. Proudfit WL, Bruschke VG, Sones FM Jr. Clinical course ofpatients with normal or slightly or moderately abnormal coronaryarteriograms: 10 year followup of 521 patients. Circulation. 1980;62:712-717.

20. Kemp HG, Kronmal RA, Vlietstra RE, Frye RI, and Participantsin the Coronary Artery Surgery Study. Seven year survival ofpatients with normal or near normal coronary arteriograms: aCASS registry study. JAm Coll Cardiol. 1986;7:479-483.

21. Epstein SE, Quyyumi AA, Bonow RO. Sudden cardiac deathwithout warning. N Engl J Med. 1989;321:320-324.

22. Mautner SL, Lin F, Mautner GC, Roberts WC. Comparison inwomen versus men of composition of atherosclerotic plaques innative coronary arteries and in saphenous veins used as aorto-coronary conduits. JAm Coll Cardiol. 1993;21:1312-1318.

23. Simons DB, Schwartz RS, Edwards WD, Sheedy PF, Breen JF,Rumberger JA. Noninvasive definition of anatomic coronaryartery disease by ultrafast computed tomographic scanning: aquantitative pathologic comparison study. JAm Coll Cardiol. 1992;20:1118-1126.

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from

J A Fallavollita, A S Brody, I L Bunnell, K Kumar and J M Canty, Jrartery disease. Comparison with angiography in patients < 50 years old.

Fast computed tomography detection of coronary calcification in the diagnosis of coronary

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1994 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.89.1.285

1994;89:285-290Circulation. 

http://circ.ahajournals.org/content/89/1/285World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer this process is available in the

click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on Novem

ber 13, 2017http://circ.ahajournals.org/

Dow

nloaded from