Table 2.

Results for screening imaging models.

Study, Country, Outcome	No. of CPP*	AI methods	Predictors	Val. methods	Performance (AUC, Accuracy (Acc%), Sensitivity (SEN%), Specificity (SPE%), PPV/NPV (%), (95% CI))	Risk of Bias**: Participants/Predictors/Outcome/Analysis/Overall
Abdani et al. [30], Fast screening	219	DL, CNN	Imaging features	5-FCV	Acc 94	H	U	H	H	H
Ahammed et al. [5], Early detection	285	ML, DL, CNN, SVM, RF, k-NN, LR, GNB, BNB, DT, XGB, MLP, NC, perceptron.	Imaging features	10-FCV	AUC 95.52, Acc 94.03, SEN 94.03, SPE 97.01	H	H	H	H	H
Barstugan et al. [31], Early detection	53	ML, SVM	Imaging features	10-FCV	Acc 99.68, SEN 93, SPE 100	U	U	U	H	H
Wu et al. [55], China, Fast and accurately identification	368	DL	Imaging features	TTS	AUC 0.905, Acc 83.3, SEN 82.3	L	U	U	H	H
Wang et al. [56], China, Triage	1647	DL	Imaging features	Ext. val.	AUC 0.953 (95% CI 0.949–0.959), SEN 92.3 (95% CI 91.4–93.2), SPE 85.1 (84.2–86.0), PPV 79 (77.7–80.3), NPV 94.8 (94.1–95.4)	L	U	U	H	H

*CPP = COVID-19 Positive Patients, Abbreviations of medical terms included in this Table are provided in the Appendix.

**L: Low, H: High, U: Unclear