Table 1.
Summary of general study characteristics.
| Training validation/ Testing | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study ID |
Country of corresponding author | Study type | Index test | Date source | Eligibility criteria | Reference standard | Common type of pneumonia | Number of COVID-19 vs. other pneumonias | AUC | Type of validation | Number of COVID-19 vs. other pneumonias | SEN | SPC |
| Ardakani 2020 | Iran | R | CT | Single hospital | Yes | RT-PCR | Atypical, viral pneumonia | 86 vs. 69 | 0.999 | Random split | 22 vs.17 | 1.00 | 0.99 |
| Ardakani 2021 | Iran | R | CT | Single hospital | Yes | RT-PCR | Atypical and viral pneumonia | 244 vs.244 | 0.988 | Random split | 62 vs.62 | 0.935 | 0.903 |
| Ali 2021 | turkey | R | CXR | Single database | No | NA | Viral pneumonia |
146 vs.901 | NR | 3fold CV | 73 vs.444 | 0.973 | NR |
| Han2021 | Korea | R | CT | 2datasets | No | NA | Viral pneumonia, bacterial pneumonia, fungal pneumonia |
164 vs.320 | NR | External validation | 21 vs.40 | 0.997 | 0.959 |
| Di2020 | China | R | CT | 5hospitals | No | RT-PCR | CAP | 1933 vs. 1064 | NR | 10 fold CV | 215 vs.118 | 0.932 | 0.840 |
| Bai 2020 | China | R | CT | 10hospitals | Yes | RT-PCR | Pneumonia of other origin |
377 vs.453 | NR | Random split | 42 vs.77 | 0.950 | 0.960 |
| Panwar 2020 | Mexico | R | CXR | 3datasets | No | NA | Pneumonia | 133 vs.231 | NR | Random split | 29 vs.85 | 0.966 | 0.953 |
| Kang 2020 | China | R | CT | 3hospitals | No | RT-PCR | CAP | 1046 vs. 719 | NR | Random split | 449 vs.308 | 0.966 | 0.932 |
| Liu 2021 | China | R | CT | 2hospitals | Yes | RT-PCR | Viral infections, mycoplasma infections, chlamydia infections, fungus infections, co-infections |
66 vs.313 | 1.000 | External validation | 20 vs.20 | 0.850 | 0.900 |
| Chen 2021 | China | R | CT | Single hospital | Yes | RT-PCR | Other types of pneumonia | 54 vs.60 | 0.984 | Random split | 9 vs.11 | 0.816 | 0.923 |
| Song 2020 | China | R | CT | 2hospitals | Yes | RT-PCR | CAP | 66 vs.66 | 0.979 | External validation | 15 vs.20 | 0.800 | 0.750 |
| Sun 2020 | China | R | CT | 6hospitals | No | RT-PCR | CAP | 1196 vs. 822 | NR | 5fold CV | 299 vs.205 | 0.931 | 0.899 |
| Wang 2021 | China | R | CT | 3hospitals | Yes | RT-PCR | Other types of viral pneumonia |
74 vs.73 | 0.970 | External validation | 17 vs.17 | 0.722 | 0.751 |
| Zhou 2021 | China | R | CT | 12hospitals | Yes | RT-PCR | Influenza pneumonia | 118 vs.157 | NR | External validation | 57 vs.50 | 0.860 | 0.772 |
| Azouji2021 | Switzerland | R | CXR | 7datasets | No | NA | MERS, SARS | 338 vs.222 | NR | 5fold CV | 85 vs.56 | 0.989 | NR |
| Cardobi 2021 | Italy | R | CT | Single hospital | No | swab test | Interstitial pneumonias | 54 vs.30 | 0.830 | Random split | 14 vs.17 | 0.570 | 0.930 |
| Yang 2021 | China | R | CT | Single hospital | No | RT-PCR | Other pneumonias | 70 vs.70 | NR | 10fold CV | 20 vs.20 | 0.942 | 0.854 |
| Chikontwe 2021 | Korea | R | CT | Single hospital | No | RT-PCR | Bacterial pneumonia | 38 vs.49 | NR | Random split | 30 vs.39 | 1.000 | 0.975 |
| Zhu 2021 | China | R | CT | 6hospitals | No | RT-PCR | CAP | 1345 vs. 924 | NR | 10fold CV | 150 vs.103 | 0.913 | 0.910 |
| Xie 2020 | China | R | CT | 5hospitals | Yes | RT-PCR | Bacterial infection, Viral infection |
227 vs.153 | NR | prospective RWD | 243 vs.73 | 0.810 | 0.820 |
| Qi 2021 | China | R | CT | 3hospitals+dataset | Yes | RT-PCR | CAP | 127 vs.90 | NR | 10fold CV | 14 vs.10 | 0.972 | 0.940 |
| Wang 2020 | China | R | CT | 7hospitals | Yes | RT-PCR | Bacterial pneumonia, Mycoplasma pneumonia, Viral pneumonia, Fungal pneumonia | 560 vs.149 | 0.900 | External validation | 102 vs.124 | 0.804 | 0.766 |
| Yang 2020 | China | R | CT | 8hospitals | No | RT-PCR | CAP | 960 vs.628 | 0.976 | External validation | 1605 vs. 452 | 0.869 | 0.901 |
| Wu 2020 | China | R | CT | 3hospitals | No | RT-PCR | Other pneumonia | 294 vs.101 | 0.767 | Random split | 37 vs.13 | 0.811 | 0.615 |
| Zhang 2021 | China | R | CT | 3hospitals | No | RT-PCR | CAP, influenza, mycoplasma pneumonia | 72 vs.127 | 0.987 | 5fold CV | 31 vs.62 | 0.879 | 0.887 |
| Xin 2021 | China | R | CT | 2hospitals | Yes | swab tests | CAP | 34 vs.48 | NR | 5fold CV | 9 vs.12 | 0.957 | 0.984 |
| Guo 2020 | China | R | CT | 2hospitals | No | RT-PCR | Seasonal flu，CAP | 8 vs.42 | 0.970 | External validation | 11 vs.44 | 0.889 | 0.935 |
| Fang2020 | China | R | CT | 2hospitals | Yes | nucleic acid detection | Viral pneumonia | 136 vs.103 | 0.959 | Random split | 56 vs.34 | 0.929 | 0.971 |
| Xia 2021 | China | R | CXR | 2hospitals | Yes | nucleic acid | Influenza A/B pneumonia |
246 vs.44 | NR | Random split | 266 vs.62 | 0.869 | 0.742 |
| Huang2020 | China | R | CT | 15hospitals | Yes | RT-PCR | Viral pneumonia | 62 vs.64 | 0.849 | 5fold CV | 27 vs.28 | 0.778 | 0.786 |
| Wu 2021 | China | R | CT | Single hospital | Yes | nucleic acid | Other infectious pneumonia |
76 vs.77 | NR | 5fold CV | 19 vs.19 | 0.809 | 0.842 |
| Chen2021 | China | R | CT | 2hospitals | No | RT-PCR | Viral pneumonia | 81 vs.81 | 0.807 | Random split | 27 vs.19 | 0.733 | 0.822 |
Abbreviations: AUC: area under the curve; CAP: community acquired pneumonia;CT: Computed tomography; CV: cross-validation;CXR: Chest X-Ray; R: retrospective; RT-PCR: Reverse transcriptase polymerase chain reaction; RWD: Real-world dataset; SEN: sensitivity;SPC: specificity