Table 3 Case-level performance of the base U-Nets and metamodel for all scans in validation dataset. SAH1 base U-Net is model trained during this study. SAH2 is model which development and validation was previously described by thanellas et al.14. The best achieved metrics are bolded. TP = true positive, tn = true negative, fp = false positive, fn = false negative, fpr = false positive rate, npv = negative predictive value, ppv = positive predictive value, ci = confidence interval.
TP | TN | FP | FN | Sensitivity [95% CI] | Specificity [95% CI] | FPR [95% CI] | NPV [95% CI] | PPV [95% CI] | Accuracy [95% CI] | |
|---|---|---|---|---|---|---|---|---|---|---|
Metamodel - No Post-processing | 109 | 4131 | 3548 | 9 | 92.4% [87.6–97.2%] | 53.8% [52.7–54.9%] | 46.2% [45.1–47.3%] | 99.8% [99.6–99.9%] | 3.0% [2.4–3.5%] | 54.4% [53.3–55.5%] |
ICH Base U-Net | 115 | 2974 | 4705 | 3 | 97.5% [94.6–100.0%] | 38.7% [37.6–39.8%] | 61.3% [60.2–62.4%] | 99.9% [99.8–100.0%] | 2.4% [2.0–2.8%] | 39.6% [38.5–40.7%] |
IVH Base U-Net | 110 | 2221 | 5458 | 8 | 93.2% [88.7–97.8%] | 28.9% [27.9–29.9%] | 71.1% [70.1–72.1%] | 99.6% [99.4–99.9%] | 2.0% [1.6–2.3%] | 29.9% [28.9–30.9%] |
SAH1 Base U-Net | 118 | 1263 | 6416 | 0 | 100.0% [100.0–100.0%] | 16.4% [15.6–17.3%] | 83.6% [82.7–84.4%] | 100.0% [100.0–100.0%] | 1.8% [1.5–2.1%] | 17.7% [16.9–18.6%] |
SAH2 Base U-Net | 118 | 922 | 6757 | 0 | 100.0% [100.0–100.0%] | 12.0% [11.3–12.7%] | 88.0% [87.3–88.7%] | 100.0% [100.0–100.0%] | 1.7% [1.4–2.0%] | 13.3% [12.6–14.1%] |
