Table 2.
Characteristics of studies included in systematic review.
| Study, Year | Sample, Age, Sex | AIS Type | Tx | Onset-to-MRI Time | Outcome Criteria | Clinical Factors | MRI Markers | MRI Seq | RA Features | Predictive Models | AUC, 95% CI |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Quan et al. [33], 2021 | 110, 62, 70.9% male | first AIS in MCA territory, onset ≤72 h |
ivT, MT: 12 p | 26.5 ± 15.7 | 90 days unfavorable outcome mRS > 2 |
Age, gender, admission NIHSS | DWI-ASPECT score, ODs | FLAIR ADC | 6, TA, wavelet | Clinical | 0.79, 0.68–0.89 |
| Clinical + MRI | 0.78, 0.68–0.88 | ||||||||||
| ADC radiomics | 0.77, 0.62–0.83 | ||||||||||
| FLAIR radiomics | 0.73. 0.62–0.83 | ||||||||||
| ADC + FLAIR radiomics | 0.81, 0.73–0.89 | ||||||||||
| RA + Clinical + MRI | 0.92, 0.87–0.97 | ||||||||||
| Wang et al. [34], 2021 | 399, 67, 63.9% male | NR | NR | within 24 h after AIS onset | 90 days outcome mRS > 2 |
Age, 24-h NIHSS | Hemorrhage | DWI | 11, TA | Clinical model | 0.77, 0.71–0.84 |
| Radiomics model | 0.70, 0.64–0.77 | ||||||||||
| Clinical + radiomics | 0.80, 0.75–0.86 | ||||||||||
| Zhou et al. [35], 2022 | 311, 58, 72.7% male | Pen artery: 43.1%, cMCA: 28.6%, cACA: 5.5%, cPCA = 8.4%, ≥2 territories: 14.5% |
NR | <24 h: 6.1%24–72 h: 93.9% | 6-month good outcome (mRS ≤ 2), poor outcome (mRS > 2) | Age, gender, stroke history, DM, b-mRS, b-NIHSS | - | DWI, ADC | 7, first-order statistics, TA | Clinical model | 0.82, 0.77–0.87 |
| Radiomics model | 0.76, 0.70–0.82 | ||||||||||
| Clinical + radiomics | 0.86, 0.82–0.91 | ||||||||||
| Zhang et al. [36], 2022 | 103, 65, 64% male | Unilateral anterior circulation | NR | NR | 90 days outcome mRS > 2 |
Atrial fibrillation | - | ADC | 7, TA, wavelet, LGT | ADC | 0.60, 049–0.71 |
| tADC | 0.83, 075–0.91 | ||||||||||
| tADC + clinical | 0.86, 079–0.93 | ||||||||||
| Wang et al. [37], 2022 | 1003, 67, 67.9% m | Ant-circ: 68.5%, Post-circ: 28.5%, Both: 3% | NR | 72 h of AIS onset | 90 d outcome 1y AIS recurrence |
NR | - | DWI | 100, TA, wavelet | Radiomics model | 0.77, 0.75–0.80 |
| Clinical + radiomics | 0.84, 0.82–0.87 | ||||||||||
| Wang et al. [38], 2020 | 116, 64, 72% male | NR | NR | NR | 90 days outcome mRS > 2, stroke severity |
- | - | FLAIR, ADC | 15, first-order statistics, TA | RA features were not predictive of mRS. ADC-entropy and T2-FLAIR 0.75 quantile predicted AIS severity (AUC = 0.7, p = 0.01). | |
Tx = treatment, MRI Seq = MRI sequences for feature selection, MCA = middle cerebral artery, ivT = intravenous thrombolysis, MT = mechanical thrombectomy, OD = orthogonal diameters, TA = texture analysis, FLAIR = fluid-attenuated-inversion recovery, ADC = apparent diffusion coefficient, TA = texture analysis, tADC = texture analysis from ADC, Pen artery = penetrating artery, cor-MCA = cortical branches of middle cerebral artery, cor-ACA = cortical branches of anterior cerebral artery, cor-PCA = cortical branches of posterior cerebral artery, DM = diabetes mellitus, b-mRS = baseline mRS, b-NIHSS = baseline NIHSS, LGT = Laplacian of Gaussian transformation, NR = not reported.