Table 2 RQS rating of included studies
16 items according to 6 key domains | Range | Median (range) | Percentage of ideal score, n (%) | Adherence rate, n (%) |
|---|---|---|---|---|
Total 16 items | − 8 to 36 | 7 (− 3 to 18) | 7.3 (20.2) | 184 (38) |
Domain 1: protocol quality and stability in image and segmentation | 0–5 | 2 (0–2) | 1.6 (31.3) | 47 (15) |
Protocol quality | 0–2 | 1 (0–1) | 0.9 (46.7) | 28 (93) |
Multiple segmentations | 0–1 | 1 (0–1) | 0.6 (63.3) | 19 (63) |
Test–retest | 0–1 | 0 (0–0) | 0 (0) | 0 (0) |
Phantom study | 0–1 | 0 (0–0) | 0 (0) | 0 (0) |
Domain 2: feature selection and validation | − 8 to 8 | − 2 (− 8 to 6) | 0.9 (10.8) | 42 (70) |
Feature reduction or adjustment of multiple testing | − 3 to 3 | 3 (− 3 to 3) | 2.8 (93.3) | 29 (97) |
Validation | − 5 to 5 | − 5 (− 5 to 3) | − 1.9 (0) | 13 (43) |
Domain 3: biologic/clinical validation and utility | 0–6 | 1.5 (0–6) | 2.0 (33.9) | 47 (39) |
Non-radiomics features | 0–1 | 0.5 (0–1) | 0.5 (50.0) | 15 (60) |
Biologic correlations | 0–1 | 1 (0–1) | 0.6 (60.0) | 18 (60) |
Comparison with “gold standard” | 0–2 | 0 (0–2) | 0.8 (40.0) | 12 (40) |
Potential clinical utility | 0–2 | 0 (0–2) | 0.1 (6.7) | 2 (7) |
Domain 4: model performance index | 0–5 | 2 (1–4) | 2.1 (42.7) | 34 (38) |
Cutoff analysis | 0–1 | 0 (0–0) | 0 (0) | 0 (0) |
Discrimination statistics | 0–2 | 2 (1–2) | 1.9 (95.0) | 30 (100) |
Calibration statistics | 0–2 | 0 (0–2) | 0.2 (11.7) | 4 (13) |
Domain 5: high level of evidence | 0–8 | 0 (0–7) | 0.2 (2.9) | 1 (2) |
Prospective study | 0–7 | 0 (0–7) | 0.2 (3.3) | 1 (3) |
Cost-effectiveness analysis | 0–1 | 0 (0–0) | 0 (0) | 0 (0) |
Domain 6: open science and data | 0–4 | 0 (0–1) | 0.4 (10.8) | 13 (43) |