Introduction
Methods
Patient population
Image acquisition and reconstruction
Visual analysis
Automated analysis
Invasive coronary angiography and fractional flow reserve
Statistical analysis
Results
Parameter | Total group (n = 206) | Derivation group (n = 103) | Validation group (n = 103) | P for subgroups |
---|---|---|---|---|
Demographics
| ||||
Age, years | 58.2 ± 8.7 | 57.8 ± 9.3 | 58.6 ± 8.1 | 0.501 |
Male | 131 (64%) | 64 (62%) | 67 (65%) | 0.666 |
BMI | 27.0 ± 3.7 | 27.2 ± 3.9 | 26.9 ± 3.6 | 0.547 |
LBBB | 6 (3%) | 3 (3%) | 3 (3%) | 0.990 |
CAD risk factors
| ||||
Hypertension | 95 (46%) | 49 (48%) | 46 (45%) | 0.677 |
Hyperlipidemia | 83 (40%) | 41 (40%) | 42 (41%) | 0.888 |
Diabetes | 33 (16%) | 16 (16%) | 17 (17%) | 0.850 |
Current smoker | 40 (19%) | 19 (18%) | 21 (20%) | 0.726 |
Smoking history | 99 (48%) | 49 (48%) | 50 (49%) | 0.890 |
Family history of CAD | 105 (51%) | 49 (48%) | 56 (54%) | 0.332 |
Significant CAD | 92 (45%) | 44 (43%) | 48 (47%) | 0.577 |
Diagnostic performance of visual and standard automated assessment
Sensitivity | Difference with expert | Specificity | Difference with expert | Diagnostic accuracy | Difference with expert | |
---|---|---|---|---|---|---|
Expert | 56.5 (45.8–66.8) | 93.9 (87.8–97.5) | 77.2 (70.8–82.7) | |||
SSS NC | 66.3 (55.7–75.8) | P = 0.108 | 68.4 (59.1–76.8) | P < 0.001* | 67.5(60.6–73.8) | P = 0.012* |
SSS AC | 67.0 (56.4–76.5) | P = 0.064 | 71.9 (62.7–79.9) | P < 0.001* | 69.8 (63.0–76.0) | P = 0.063 |
SDS NC | 80.0 (70.3–87.7) | P < 0.001* | 50.0 (40.5–59.5) | P < 0.001* | 63.2 (56.2–69.9) | P = 0.003* |
SDS AC | 86.5 (77.6–92.8) | P < 0.001* | 49.1 (39.6–58.7) | P < 0.001* | 65.5 (58.5–72.0) | P = 0.023* |
S-TPD NC | 62.0 (51.2–71.9) | P = 0.405 | 73.7 (64.6–81.5) | P < 0.001* | 68.4 (61.6–74.7) | P = 0.020* |
S-TPD AC | 64.8 (54.1–74.6) | P = 0.152 | 76.3 (67.4–83.8) | P < 0.001* | 71.2 (64.5–77.3) | P = 0.134 |
I-TPD NC | 65.6 (54.8–75.3) | P = 0.108 | 69.9 (60.6–78.2) | P < 0.001* | 68.0 (61.1–74.3) | P = 0.027* |
I-TPD AC | 65.9 (55.3–75.6) | P = 0.108 | 65.8 (56.3–74.4) | P < 0.001* | 65.8 (58.9–72.3) | P = 0.005* |
Optimizing automated assessment
Diagnostic performance of optimized automated assessment
Sensitivity | Difference with expert | Specificity | Difference with expert | Diagnostic accuracy | Difference with expert | |
---|---|---|---|---|---|---|
Expert | 52.1 (37.2–66.7) | 92.7 (82.4–98.0) | 73.8 (64.2–82.0) | |||
SSS NC | 52.1 (37.2–66.7) | P = 1.000 | 83.6 (71.2–92.2) | P = 0.227 | 68.9 (59.1–77.7) | P = 0.383 |
SSS AC | 68.1 (52.9–80.9) | P = 0.077 | 76.4 (63.0–86.8) | P = 0.035* | 72.5 (62.8–80.9) | P = 1.000 |
SDS NC | 50.0 (34.9–65.1) | P = 1.000 | 80.0 (67.0–89.6) | P = 0.092 | 66.3 (56.3–75.4) | P = 0.210 |
SDS AC | 66.7 (51.1–80.0) | P = 0.077 | 76.4 (63.0–86.8) | P = 0.035* | 72.0 (62.1–80.5) | P = 1.000 |
S-TPD NC | 45.8 (31.4–60.8) | P = 0.508 | 92.7 (82.4–98.0) | P = 1.000 | 70.9 (61.1–79.4) | P = 0.607 |
S-TPD AC | 83.0 (69.2–92.4) | P = 0.001* | 65.5 (51.4–77.8) | P < 0.001* | 73.5 (63.9–81.8) | P = 1.000 |
I-TPD NC | 46.8 (32.1–61.9) | P = 0.774 | 78.2 (65.0–88.2) | P = 0.057 | 63.7 (53.6–73.0) | P = 0.076 |
I-TPD AC | 76.6 (62.0–87.7) | P = 0.008* | 60.0 (45.9–73.0) | P < 0.001* | 67.6 (57.7–76.6) | P = 0.418 |