OBJECTIVE. This study was performed to evaluate new scoring methods for quantitat-
ing coronary artery calcifications with helical CT and to compare the results with those of
quantitative coronary angiography in patients with suspected coronary artery disease.
SUBJECTS AND METHODS. Unenhanced dual-slice helical CT and coronary
angiography were performed within 24 hr of each other in 101 patients with symptoms of
coronary artery disease. Coronary artery calcifications with a density above 90 H were
identified on each slice and, with the same regions of interest, quantitative scoring was per-
formed at thresholds of 90 H (new) and 130 H (old). Two mathematical algorithms (one
new and one old) were evaluated for both thresholds (yielding four scoring systems). By
CT imaging, we defined disease as a score of greater than zero. By angiography, we
defined disease as a 50% or greater reduction in the luminal diameter of any major vessel.
Interobserver variations in calcification scoring were evaluated. Seventeen of our patients
also underwent a second, consecutive CT scan to determine reproducibility.
RESULTS. With the new threshold and the new algorithm, the sensitivity, specificity,
and accuracy of helical CT in predicting disease were 88%, 52%, and 76%, respectively.
We found a moderate positive association between the total CT calcification score and the
number of stenotic coronary arteries at angiography (Pearson’s correlation coefficient, .43;
p = .05 [analysis of variance]). The accuracy and the area under the receiver operating
characteristic curve were higher with the new threshold and the new algorithm. Interob-
server agreement in calcification scoring was high (intraclass correlation coefficient, .99
[Pt = 851), as was reproducibility (intraclass correlation coefficient, .94 [n = 17]). Reproduc-
ibility was higher when scoring was based on the new threshold and the new algorithm.
CONCLUSION. The quantity ofcoronary artery calcifications as measured by helical
CT correlated positively with obstructive coronary artery disease as measured by angiogra-
phy. Interobserver agreement and reproducibility were excellent. A new scoring method
AJR:167, August 1996 439
Measurement of Coronary Artery
Calcium with Dual-Slice Helical
CT Compared with Coronary
Angiography: Evaluation of CT
Scoring Methods, Interobserver
Variations, and Reproducibility
Lynn S. Broderick1
Joseph Shemesh2’3
Robert L. Wilensky4
George J. Eckert5
Xiao-hua Zhou5
William E. Torres6
Michael A. Balk7
WendyJ. Rogers1
Dewey J. Conces, Jr.1
Kenyon K. Kopecky1
Received September 25, 1995; accepted after revision
February 22, 1996.
1Department of Radiology, Indiana University School of
Medicine, Indiana University Hospital 0279, 550 N. Uni-
versity Blvd., Indianapolis, IN 46202-5253. Address corre-
spondence to L S. Broderick.
2The Cardiac Rehabilitation Institute, Sheba Medical
Center, Tel Aviv, Israel.
3Sackler School of Medicine, Tel Aviv University, Tel Aviv,
Israel.
4 Division of Cardiology, Department of Medicine, Indiana
University School of Medicine, Indianapolis, IN 46202.
tDivision of Biostatistics, Department of Medicine, Indi-
ana University School of Medicine, Indianapolis, IN
46202.
tDepartment of Radiology, Emory University School of
Medicine, Atlanta, GA 30322.
7Division of Cardiology, Department of Medicine, Emory
University School of Medicine, Atlanta, GA 30322.
AJR 1996;167:439-444
0361-803X/96/1672-439
© American Roentgen Ray Society
showed promise.
E lectron beam CT has been used
to evaluate the heart because it
can acquire axial images in 0.1
sec with electrocardiographic gating and
because the heart can be imaged with non-
overlapping 3-mm slices in 30-40 sec (a
breath-hold). The quantity of coronary
artery calcium as measured by unenhanced
electron beam CT correlates positively
with obstructive coronary artery disease as
measured by angiography. Because of this
correlation, CT has been used as a diagnos-
tic test to evaluate symptomatic patients
with suspected or known coronary artery
disease and also as a screening tool for
asymptomatic subjects [1-1 11.
Single-slice helical CT acquires a volume
of spatial information over several seconds
without electrocardiographic gating. These
data can be reconstructed into thin overlap-
ping axial images, with an acquisition time
for each image of approximately 0.6-1 .0 sec.
Dual-slice helical CT is the same as single-
slice helical CT, except that it acquires data
twice as fast because it has two parallel arcs
of detectors that are simultaneously irradiated
by a single anode. With helical CT (single or
dual slice), the heart can be imaged with 3-
mm slices within 30 sec (a breath-hold). If the
quantitation of coronary artery calcifications
can be performed with helical CT at an accu-
racy similar to that of electron beam CT, as