Introduction
Materials and methods
Patients
CTA acquisition
Automatic tree labeling method
Coronary artery tree model
Main branch | Labels of main branch | Labels of side-branches | |
---|---|---|---|
LM | / | ||
Sub-trees | RCA | pRCA, mRCA, dRCA | Right posterior lateral (RPLB) branch RD type: right posterior descending (RPDA) artery |
LAD | pLAD, mLAD, dLAD | Two diagonal arteries (D1, D2) | |
LCX | pLCx, LCx | Two obtuse marginal (OM1, OM2) arteries, left posterior lateral (LPLB) branch Anterolateral (AL) artery or ramus intermedius (RI) arteries LD type: left posterior descending (LPDA) artery |
Main branch identification
Segment labeling
Labeling results refinement
Evaluation measures
Presence
Overlap
Accuracy for clinical risk scores
Statistical analysis
Results
Total (83) | |
---|---|
Age (years) | 59.9 ± 11.4 |
Gender (% male) | 61 (73%) |
Diabetes | 20 (24%) |
Hypertensiona
| 34 (41%) |
Hypercholesterolemiab
| 37 (45%) |
Family history of CAD* | 26 (31%) |
Smoking | 15 (18%) |
Obesity | 22 (27%) |
Presence
Overlap
Accuracy for clinical risk scores
RD | LD | |
---|---|---|
Patients (n = 56) auto/expert (%) | Patients (n = 27) auto/expert (%) | |
Identify three main branches | ||
N = 1 | 100.0% | 100.0% |
N = 2 | 98.2% | 98.1% |
N = 3 | 70.5% | 96.3% |
Proximal segments and LM | 100.0% | 100.0% |
SIS | ||
Mean(±SD) | 92.6% (±7.3%) | 93.3% (±8.1%) |
=100.0% | 37.5% | 46.3% |
≥90.0% | 68.8% | 70.4% |
≥80.0% | 98.2% | 90.7% |
<80.0% | 1.8% | 9.3% |
Leaman score | ||
Mean(±SD) | 96.0% (±3.8%) | 96.7% (±4.1%) |
=100.0% | 37.5% | 50% |
≥95.0% | 67.9% | 70.4% |
≥90.0% | 95.5% | 96.3% |
<90.0% | 4.5% | 3.7% |