Skip to main content
. 2019 Jun 12;365:l1945. doi: 10.1136/bmj.l1945

Table 1.

Baseline patient characteristics and empirical diagnostic performance of computed tomography angiography to diagnose coronary artery disease, stratified by pretest probability category and scanner detector rows

Overall (n=5332) Pretest probability categories
0 to <10% (n=86) 10 to <20% (n=530) 20 to <30% (n=601) 30 to <40% (n=727) 40 to <50% (n=745) 50 to <60% (n=752) 60 to <70% (n=590) 70 to <80% (n=535) 80 to <90% (n=698) 90 to 100% (n=68)
Demographic characteristics (median (range) or No (%))
Median age (years) 61 (18-96) 47 (18-50) 56 (23-70) 59 (24-82) 57 (37-89) 55 (27-87) 63 (30-91) 70 (36-88) 55 (47-92) 66 (59-77) 80 (78-89)
Men 3473 (65) 0 29 (5) 211 (35) 509 (70) 576 (77) 507 (67) 391 (66) 484 (90) 698 (100) 68 (100)
Women 1859 (35) 86 (100) 501 (95) 390 (65) 218 (30) 169 (23) 245 (33) 199 (34) 51 (10) 0 0
Median body mass index 26.3 (14.3-57.1) 25.6 (17.9-39.3) 26.2 (14.3-47.3) 25.9 (16.1-44.8) 26.2 (16.9-41.8) 26.4 (17.5-45.2) 26.5 (17.2-57.1) 26.4 (15.5-56.2) 27.0 (17.5-42.5) 26.9 (16.9-56.7) 25.9 (16.8-35.2)
Clinical presentation (No)
Typical angina 1967 0 0 4 43 137 247 306 464 698 68
Atypical angina 1592 1 138 269 235 280 339 260 70 0 0
Non-anginal chest pain 796 38 162 157 188 158 80 12 1 0 0
Other chest discomfort 977 47 230 171 261 170 86 12 0 0 0
Diagnostic performance (No or %)
CAD prevalence (%)* 48.3 17.4 24.0 32.1 40.9 46.8 46.8 53.7 68.6 71.6 82.4
TP 2251 14 120 176 272 321 310 256 317 420 45
TN 2031 52 313 312 334 287 294 194 103 134 8
FP 728 19 90 96 96 109 106 79 65 64 4
FN 322 1 7 17 25 28 42 61 50 80 11
NDX† 554 13 58 50 54 67 76 79 60 85 12
NDX rate (%)† 10.4 15.1 10.9 8.3 7.4 9.0 10.1 13.4 11.2 12.2 17.6
PPV (%) 75.6 42.4 57.1 64.7 73.9 74.7 74.5 76.4 83.0 86.8 91.8
NPV (%) 86.3 98.1 97.8 94.8 93.0 91.1 87.5 76.1 67.3 62.6 42.1
Sensitivity (%) 87.5 93.3 94.5 91.2 91.6 92.0 88.1 80.8 86.4 84.0 80.4
Specificity (%) 73.6 73.2 77.7 76.5 77.7 72.5 73.5 71.1 61.3 67.7 66.7
Diagnostic accuracy (%) 80.3 76.7 81.7 81.2 83.4 81.6 80.3 76.3 78.5 79.4 77.9
LR+ 3.32 3.49 4.23 3.88 4.10 3.34 3.32 2.79 2.23 2.60 2.41
LR− 0.17 0.09 0.07 0.12 0.11 0.11 0.16 0.27 0.22 0.24 0.29
CT scanners with ≤64 detector rows (No or %)
No of patients 4666 80 452 529 651 634 637 530 472 619 62
CAD prevalence (%) 48.2 17.5 24.1 31.2 41.8 45.5 46.0 54.0 67.4 72.5 85.5
TP 1943 13 102 150 248 264 256 226 270 372 42
TN 1757 47 265 273 295 247 246 170 95 114 5
FP 662 19 78 91 84 99 98 74 59 56 4
FN 304 1 7 15 24 24 37 60 48 77 11
NDX† 538 13 54 50 50 65 75 77 58 84 12
NDX rate (%)† 11.5 16.3 11.9 9.5 7.7 10.3 11.8 14.5 12.3 13.6 19.4
PPV (%) 74.6 40.6 56.7 62.2 74.7 72.7 72.3 75.3 82.1 86.9 91.3
NPV (%) 85.2 97.9 97.4 94.8 92.5 91.1 86.9 79.9 66.4 59.7 31.3
Sensitivity (%) 86.5 92.9 93.6 90.9 91.2 91.7 87.4 79.0 84.9 82.9 79.2
Specificity (%) 72.6 71.2 77.3 75.0 77.8 71.4 71.5 69.7 61.7 67.1 55.6
Diagnostic accuracy (%) 79.3 75.0 81.2 80.0 83.4 80.6 78.8 74.7 77.3 78.5 75.8
LR+ 3.16 3.23 4.12 3.64 4.11 3.20 3.07 2.61 2.22 2.52 1.78
LR− 0.19 0.10 0.08 0.12 0.11 0.12 0.18 0.30 0.24 0.26 0.37
CT scanners with >64 detector rows (No or %)
No of patients 558 6 73 62 66 87 103 55 41 59 6
CAD prevalence (%)* 46.1 16.7 21.9 37.1 30.3 54.0 49.5 50.9 75.6 62.7 50.0
TP 240 1 16 21 19 44 46 27 29 34 3
TN 254 5 46 36 39 33 47 22 6 17 3
FP 47 0 11 3 7 7 5 5 4 5 0
FN 17 0 0 2 1 3 5 1 2 3 0
NDX† 16 0 4 0 4 2 1 2 2 1 0
NDX rate (%)† 2.9 0.0 5.5 0.0 6.1 2.3 1.0 3.6 4.9 1.7 0.0
PPV (%) 83.6 100 59.3 87.5 73.1 86.3 90.2 84.4 87.9 87.2 100
NPV (%) 93.7 100 100 94.7 97.5 91.7 90.4 95.7 75.0 85.0 100
Sensitivity (%) 93.4 100 100 91.3 95.0 93.6 90.2 96.4 93.5 91.9 100
Specificity (%) 84.4 100 80.7 92.3 84.8 82.5 90.4 81.5 60.0 77.3 100
Diagnostic accuracy (%) 88.5 100 84.9 91.9 87.9 88.5 90.3 89.1 85.4 86.4 100
LR+ 5.98 5.18 11.87 6.24 5.35 9.38 5.21 2.34 4.04
LR− 0.08 0 0 0.09 0.06 0.08 0.11 0.04 0.11 0.10 0

TP=true positives; TN=true negatives; FP=false positives; FN=false negatives; PPV=positive predictive value; NPV=negative predictive value; LR+=positive likelihood ratio; LR−=negative likelihood ratio; NDX=non-diagnostic results.

The empirical results were derived from raw data and thus differ from the results of the statistical model.

*

CAD prevalence was defined by coronary angiography.

Non-diagnostic results were included in the estimation of diagnostic accuracy as false positives if the reference standard was negative, and as false negative if the reference standard was positive.