Table 3.
Results of the clinical, image, and combination experiments for the radiomics approach for predicting the good reperfusion [post-modified Thrombolysis in Cerebral Infarction (eTICI) ≥ 2b].
| Methods | AUC 95% CI | F1-Score | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|
| Clinical experiment | ||||||
| RFC | 0.53 (0.51–0.55) | 0.71 (0.68–0.74) | 0.79 (0.74–0.84) | 0.26 (0.19–0.32) | 0.64 (0.60–0.69) | 0.42 (0.36–0.48) |
| SVM | 0.54 (0.53–0.56) | 0.39 (0.08–0.70) | 0.32 (0.01–0.64) | 0.73 (0.44–1.02) | 0.68 (0.65–0.72) | 0.39 (0.35–0.43) |
| LR | 0.54 (0.51–0.56) | 0.61 (0.57–0.66) | 0.59 (0.54–0.64) | 0.44 (0.39–0.50) | 0.64 (0.61–0.68) | 0.39 (0.34–0.43) |
| XGB | 0.51 (0.50–0.54) | 0.63 (0.57–0.69) | 0.63 (0.55–0.71) | 0.37 (0.30–0.45) | 0.63 (0.58–0.68) | 0.37 (0.33–0.41) |
| NN | 0.51 (0.50–0.53) | 0.70 (0.62–0.79) | 0.81 (0.60–1.03) | 0.19 (0.03–0.41) | 0.63 (0.59–0.67) | 0.37 (0.32–0.43) |
| Image experiment | ||||||
| RFC | 0.54 (0.52–0.56) | 0.75 (0.74–0.75) | 0.91 (0.81–1.01) | 0.11 (0.01–0.22) | 0.64 (0.59–0.68) | 0.42 (0.35–0.50) |
| SVM | 0.55 (0.53–0.57) | 0.70 (0.61–0.79) | 0.79 (0.55–1.03) | 0.25 (0.03–0.53) | 0.64 (0.60–0.69) | 0.41 (0.37–0.46) |
| LR | 0.53 (0.50–0.57) | 0.61 (0.57–0.64) | 0.57 (0.53–0.61) | 0.47 (0.40–0.54) | 0.65 (0.61–0.69) | 0.39 (0.32–0.46) |
| XGB | 0.53 (0.50–0.56) | 0.64 (0.60–0.68) | 0.65 (0.54–0.75) | 0.39 (0.28–0.49) | 0.64 (0.61–0.68) | 0.40 (0.33–0.46) |
| NN | 0.53 (0.50–0.56) | 0.67 (0.65–0.69) | 0.69 (0.66–0.73) | 0.36 (0.32–0.40) | 0.65 (0.61–0.69) | 0.41 (0.35–0.46) |
| Combination experiment | ||||||
| RFC | 0.57 (0.55–0.59) | 0.75 (0.71–0.78) | 0.89 (0.85–0.93) | 0.15 (0.10–0.19) | 0.64 (0.60–0.68) | 0.45 (0.38–0.52) |
| SVM | 0.57 (0.54–0.61) | 0.63 (0.59–0.66) | 0.58 (0.55–0.61) | 0.52 (0.46–0.58) | 0.68 (0.63–0.72) | 0.42 (0.38–0.47) |
| LR | 0.57 (0.54–0.60) | 0.63 (0.60–0.66) | 0.59 (0.57–0.62) | 0.50 (0.46–0.55) | 0.67 (0.63–0.72) | 0.42 (0.38–0.46) |
| XGB | 0.57 (0.55–0.58) | 0.59 (0.55–0.64) | 0.54 (0.46–0.61) | 0.55 (0.46–0.63) | 0.67 (0.63–0.71) | 0.41 (0.36–0.46) |
| NN | 0.53 (0.51–0.55) | 0.66 (0.62–0.70) | 0.68 (0.63–0.72) | 0.37 (0.32–0.43) | 0.65 (0.60–0.70) | 0.40 (0.37–0.43) |
Average over 5-fold cross-validation. RFC, random forest classifier; SVM, support vector machine; LR, logistic regression; XGB, gradient boosting; NN, neural networks. AUC, area under the curve; NPV, negative predictive value; PPV, positive predictive value.