Table 2.
Survival models through Cox proportional hazards utilizing clinical variables, ML-derived scores and comparator predictors (CCTA statistical score by De Graaf and Calcium Score)
| Predictors | HR | 95% CI | P value |
|---|---|---|---|
| Hybridized ML model | |||
| Age | 1.02 | .97–1.06 | .360 |
| Smoking | .95 | .41–2.18 | .897 |
| Diabetes | 1.06 | .69–1.62 | .772 |
| Hypertension | 1.03 | .48–2.18 | .934 |
| Dyslipidemia | .71 | .33–1.50 | .376 |
| ML-CCTA | 15.26 | 5.38–43.25 | < .0001 |
| ML-PET | 4.23 | 1.99–8.98 | .0002 |
| Expert-based model | |||
| Age | 1.03 | .99–1.08 | .129 |
| Smoking | 1.48 | .69–3.16 | .311 |
| Diabetes | 1.33 | .86–2.06 | .195 |
| HTN | 1.12 | .51–2.40 | .775 |
| DLP | .71 | .33–1.53 | .388 |
| Statistical CCTA score | 1.10 | 1.04–1.15 | .0006 |
| Clinical PET | 1.02 | .58–1.77 | .947 |
| Calcium score-based model | |||
| Calcium score | 1.16 | 1.03–1.30 | .0012 |
CI, confidence interval; HR, hazard ratio