Skip to main content

Table 4 Comparison of performance of the HRnV models (based on 5-fold cross-validation), HEART, TIMI, and GRACE scores in predicting 30-day major adverse cardiac events (MACE). The cut-off values were defined as the points nearest to the upper-left corner on the ROC curves

From: Utilizing machine learning dimensionality reduction for risk stratification of chest pain patients in the emergency department

  AUC (95% CI) Cut-off Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) NPV % (95% CI)
Stepwise 0.887 (0.859–0.916) 0.3140 79.4 (74.3–84.4) 78.8 (75.4–82.3) 62.8 (57.5–68.2) 89.4 (86.7–92.2)
PCA 0.899 (0.872–0.926) 0.2881 85.4 (81.0–89.8) 78.5 (75.0–81.9) 64.1 (59.0–69.3) 92.3 (89.9–94.7)
KPCA 0.896 (0.869–0.923) 0.3489 81.8 (77.0–86.6) 82.1 (78.9–85.3) 67.3 (62.0–72.6) 90.9 (88.4–93.4)
LSA 0.899 (0.872–0.926) 0.2884 85.4 (81.0–89.8) 78.6 (75.2–82.1) 64.3 (59.1–69.5) 92.3 (89.9–94.7)
GRP 0.896 (0.868–0.923) 0.2965 85.0 (80.6–89.5) 78.5 (75.0–81.9) 64.0 (58.8–69.2) 92.1 (89.6–94.5)
SRP 0.898 (0.871–0.925) 0.2940 84.6 (80.1–89.1) 79.6 (76.2–82.9) 65.1 (59.9–70.3) 92.0 (89.5–94.4)
MDS 0.901 (0.874–0.928) 0.3095 83.4 (78.8–88.0) 81.6 (78.3–84.8) 67.1 (61.8–72.4) 91.6 (89.1–94.1)
Isomap 0.888 (0.860–0.917) 0.3468 78.5 (73.4–83.7) 82.7 (79.5–85.8) 67.1 (61.7–72.5) 89.5 (86.9–92.2)
LLE 0.898 (0.870–0.925) 0.3140 85.0 (80.6–89.5) 79.4 (76.0–82.8) 65.0 (59.8–70.2) 92.2 (89.7–94.6)
HEART 0.841 (0.808–0.874) 5 78.9 (73.9–84.0) 72.8 (69.1–76.5) 56.7 (51.4–61.9) 88.5 (85.5–91.4)
TIMI 0.681 (0.639–0.723) 2 63.6 (57.6–69.6) 58.4 (54.3–62.5) 40.8 (35.9–45.7) 78.0 (74.0–82.1)
GRACE 0.665 (0.623–0.707) 107 64.0 (58.0–70.0) 60.8 (56.7–64.9) 42.4 (37.3–47.4) 78.9 (75.0–82.8)
  1. AUC area under the curve, CI confidence interval, PPV positive predictive value, NPV negative predictive value, HEART History, ECG, Age, Risk factors and Troponin, TIMI Thrombolysis in Myocardial Infarction, GRACE Global Registry of Acute Coronary Events