Table 2.
Formula vs operator | Formula vs deep learning | Operator vs deep learning | |||||||
---|---|---|---|---|---|---|---|---|---|
Mean diff | P-value | 95% CI of diff | Mean diff | P-value | 95% CI of diff | Mean diff | P-value | 95% CI of diff | |
Left main | |||||||||
Vessel length (cm) | 0.0 ± 0.1 | 0.68 | − 0.0, 0.1 | 0.0 ± 0.2 | 0.80 | − 0.1, 0.1 | 0.0 ± 0.1 | 0.90 | − 0.1, 0.1 |
Vessel Sharpness (all) (%) | − 1.9 ± 6.8 | 0.13 | − 4.5, 0.6 | − 2.1 ± 8.2 | 0.17 | − 5.2, 1.0 | − 0.1 ± 6.1 | 0.90 | − 2.4, 2.1 |
Left anterior descending | |||||||||
Vessel length (cm) | 0.3 ± 1.6 | 0.33 | − 0.3, 0.9 | 0.5 ± 2.0 | 0.23 | − 0.3, 1.2 | 0.2 ± 1.5 | 0.54 | − 0.4, 0.7 |
Vessel sharpness (all) (%) | − 0.9 ± 3.8 | 0.19 | − 2.4, 0.5 | 0.3 ± 4.5 | 0.69 | − 1.3, 2.0 | 1.3 ± 4.6 | 0.15 | − 0.5, 3.0 |
Vessel sharpness (first 4 cm) (%) | − 1.0 ± 3.8 | 0.14 | − 2.5, 0.4 | − 0.3 ± 4.4 | 0.68 | − 2.0, 1.3 | 0.7 ± 4.3 | 0.37 | − 0.9, 2.3 |
Left circumflex | |||||||||
Vessel length (cm) | 0.4 ± 1.5 | 0.19 | − 0.2, 0.9 | 0.6 ± 1.4 | 0.02 | 0.1, 1.1 | 0.2 ± 1.3 | 0.32 | − 0.3, 0.7 |
Vessel sharpness (all) (%) | 0.2 ± 4.0 | 0.84 | − 1.3, 1.6 | 0.6 ± 6.9 | 0.66 | − 2.0, 3.1 | 0.4 ± 6.1 | 0.72 | − 1.9, 2.7 |
Vessel sharpness (first 4 cm) (%) | 0.1 ± 4.5 | 0.92 | − 1.6, 1.8 | 0.9 ± 7.3 | 0.52 | − 1.9, 3.6 | 0.8 ± 6.3 | 0.51 | − 1.6, 3.1 |
Right coronary artery | |||||||||
Vessel length (cm) | 0.5 ± 1.7 | 0.10 | − 0.1, 1.2 | 0.4 ± 2.3 | 0.39 | − 0.5, 1.2 | − 0.2 ± 2.3 | 0.67 | − 1.0, 0.7 |
Vessel sharpness (all) (%) | − 0.6 ± 4.8 | 0.50 | − 2.4, 1.2 | 0.7 ± 7.4 | 0.61 | − 2.1, 3.5 | 1.3 ± 6.0 | 0.25 | − 0.9, 3.5 |
Vessel sharpness (first 4 cm) (%) | − 0.5 ± 5.1 | 0.57 | − 2.5, 1.4 | 0.2 ± 8.4 | 0.88 | − 2.9, 3.4 | 0.8 ± 7.5 | 0.57 | − 2.0, 3.6 |
Due to the vessel length of the left mainstem being substantially less than 4 cm, vessel sharpness for the whole vessel is displayed only. A paired t-test or a Wilcoxon Signed Rank test was used to compare differences between each method. A p value and 95% confidence interval are provided for each comparison. Data is displayed as mean ± SD. Statistically significant values (p<0.05) are indicated in bold