Table 3.
RQS results according to the six domains and overall, along with the classification accuracy achieved for the MGMT methylation status prediction with models based only on radiomic features
| Ref | First author | RQS | AUC | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Domain | Total score (%) | ||||||||
| 1 | 2 | 3 | 4 | 5 | 6 | ||||
| [32] | Calabrese E | 2 | −2 | 2 | 2 | 0 | 0 | 4 (11) | 0.70 |
| [33] | Chen S | 2 | −1 | 2 | 1 | 0 | 0 | 4 (11) | 0.90 |
| [34] | Crisi G | 1 | −2 | 3 | 2 | 0 | 0 | 4 (11) | 0.84 |
| [35] | Do DT | 3 | −2 | 2 | 2 | 0 | 2 | 7 (19) | 0.87 |
| [36] | Hajianfar G | 2 | −2 | 2 | 2 | 0 | 1 | 5 (14) | 0.78 |
| [37] | Haubold J | 1 | 5 | 3 | 1 | 0 | 0 | 10 (28) | 0.74 |
| [38] | Haubold J | 1 | −2 | 2 | 2 | 0 | 0 | 3 (8) | 0.76 |
| [39] | He J | 1 | −2 | 4 | 3 | 0 | 0 | 6 (17) | 0.83 |
| [40] | Huang WY | 2 | −2 | 6 | 4 | 0 | 0 | 10 (28) | 0.86 |
| [41] | Huang WY | 1 | −2 | 5 | 5 | 0 | 0 | 9 (25) | 0.83 |
| [42] | Jiang C | 2 | 6 | 4 | 2 | 0 | 0 | 14 (39) | 0.90 |
| [43] | Kihira S | 1 | 6 | 3 | 3 | 0 | 0 | 13 (36) | 0.70a |
| [44] | Kihira S | 1 | −1 | 5 | 2 | 0 | 0 | 7 (19) | 0.62a |
| [11] | Korfiatis P | 2 | −2 | 4 | 3 | 0 | 0 | 7 (19) | 0.85 |
| [14] | Le NQK | 3 | −2 | 2 | 2 | 0 | 2 | 7 (19) | 0.90 |
| [12] | Li ZC | 1 | 6 | 4 | 1 | 0 | 0 | 12 (33) | 0.88 |
| [45] | Lu Y | 1 | 5 | 4 | 2 | 0 | 0 | 12 (33) | 0.62a |
| [46] | Pasquini L | 1 | −2 | 4 | 3 | 0 | 0 | 6 (17) | 0.69 |
| [53] | Pease M | 2 | −2 | 2 | 3 | 0 | 0 | 5 (14) | 0.92 |
| [47] | Sasaki T | 2 | −2 | 3 | 2 | 0 | 0 | 5 (14) | 0.71 |
| [48] | Shboul ZA | 2 | 5 | 3 | 2 | 0 | 2 | 14 (39) | 0.70 |
| [15] | Sohn B | 1 | 5 | 3 | 3 | 0 | 0 | 12 (33) | 0.65 |
| [49] | Verduin M | 0 | 6 | 4 | 3 | 0 | 0 | 13 (36) | 0.63 |
| [50] | Vils A | 1 | 6 | 3 | 2 | 0 | 0 | 12 (33) | 0.67 |
| [51] | Wei J | 3 | 5 | 6 | 2 | 0 | 0 | 16 (44) | 0.90 |
| [52] | Xi YB | 1 | 6 | 3 | 2 | 0 | 0 | 12 (33) | 0.80a |
Domain 1: protocol quality and stability in image and segmentation (0–5 points); domain 2: feature selection and validation (−8 to 8 points); domain 3: biologic/clinical validation and utility (0–6 points); domain 4: model performance index (0–5 points); domain 5: High level of evidence (0–8 points); domain 6: open science and data (0–4 points)
cv cross-validation, AUC area under the curve, RQS Radiomics Quality Score, MGMT O6-methylguanine-DNA methyltransferase
aAUC was not available for this study and accuracy is reported instead