Table 4.
Comparison of SNRs and CNRs from three methods with mDIXON coronary MRA
| Conventional | CS | DL-CS | |
|---|---|---|---|
| Mean SNR | 58.88 ± 16.98 | 59.24 ± 14.71 | 76.06 ± 15.93a,b |
| RCA | 62.42 ± 20.38 | 61.73 ± 16.50 | 79.65 ± 18.77a,b |
| LAD | 56.19 ± 14.89 | 57.48 ± 17.08 | 73.40 ± 17.07a,b |
| LCX | 58.02 ± 20.92 | 58.51 ± 15.77 | 75.12 ± 17.88a,b |
| Mean CNR | 17.61 ± 7.71 | 19.00 ± 7.87 | 29.70 ± 8.67a,b |
| RCA | 21.10 ± 11.31 | 22.36 ± 10.34 | 34.61 ± 12.95a,b |
| LAD | 15.85 ± 10.13 | 16.91 ± 10.90 | 26.02 ± 11.38a,b |
| LCX | 15.88 ± 11.51 | 17.73 ± 9.68 | 28.48 ± 11.37a,b |
CNR contrast-to-noise ratio, CS compressed sensing, DL deep learning, LAD left anterior descending artery, LCX left circumflex, RCA right coronary artery, SNR signal-to-noise ratio
a DL-CS vs conventional, p < 0.001
b DL-CS vs CS, p < 0.001