Table 2.
Radiomics quality score (RQS) assessment of included studies.
| Study criteria | Korfiatis et al., 2023 (26) | Liu, 2020 (21) | Chen et al., 2023 (53) | Cao et al., 2023 (37) | Qu et al. 2024 (27) | Ma, 2020 | Si et al., 2021 (25) | Park et al., 2023 (28) | Ramaekers et al., 2024 (29) | Viriyasaranon et al., 2023 (35) | Mandal et al., 2024 (33) | Liao et al., 2025 (36) | Thanya et al., 2025 (24) | Degand et al., 2025 (38) | Nadeem et al., 2025 (23) | Average score |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Image protocol quality (+1 or +2) | 1 | 1 | 1 | 2 | 1 | 2 | 1 | 2 | 2 | 2 | 2 | 2 | 1 | 2 | 1 | 1.53 |
| Multiple segmentation (+1) | 0 | 1 | 0 | 0 | 0 | 1 | 1 | 0 | 1 | 0 | 0 | 1 | 2 | 2 | 0 | 0.6 |
| Phantom study (+1) | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 2 | 0.27 |
| Imaging at multiple time points (+1) | 0 | 0 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 2 | 2 | 0.47 |
| Feature reduction or adjustment for multiple testing (−3 or +3) | 3 | 0 | 3 | 3 | -3 | 0 | 3 | 0 | 0 | 1 | 3 | 3 | 1 | 2 | 2 | 1.4 |
| Multivariable analysis (+1) | 1 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 1 | 0.4 |
| Biological correlates (+1) | 1 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.27 |
| Cut-off analysis (+1) | 0 | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 2 | 0.67 |
| Discrimination statistics (+1 or +2) | 1 | 1 | 1 | 2 | 1 | 2 | 1 | 2 | 2 | 1 | 2 | 2 | 2 | 2 | 2 | 1.6 |
| Calibration statistics (+1 or +2) | 0 | 0 | 0 | 1 | 0 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | 0 | 0 | 0 | 0.6 |
| Prospective study (+7) | 0 | 0 | 0 | 7 | 0 | 0 | 0 | 7 | 0 | 0 | 7 | 1 | 0 | 0 | 0 | 1.47 |
| Validation (−5 to +5) | 3 | 2 | 3 | 3 | 2 | 2 | 3 | 3 | 2 | 2 | -5 | 3 | 4 | 5 | 2 | 2.27 |
| Comparison to ‘gold standard’ (+2) | 2 | 2 | 2 | 2 | 0 | 2 | 2 | 2 | 2 | 0 | 2 | 2 | 2 | 2 | 0 | 1.6 |
| Potential clinical application (+2) | 2 | 2 | 2 | 2 | 1 | 2 | 3 | 2 | 2 | 1 | 2 | 2 | 1 | 2 | 1 | 1.8 |
| Cost-effectiveness analysis (+1) | 1 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 |
| Open science and data (+1 to +4) | 1 | 0 | 1 | 3 | 0 | 4 | 1 | 3 | 4 | 1 | 3 | 1 | 0 | 1 | 0 | 1.53 |
| Total score (−8 to +36) | 15 | 9 | 13 | 23 | 6 | 23 | 14 | 20 | 24 | 9 | 18 | 20 | 14 | 22 | 16 | 16.4 |
RQS scores ranged from 6 to 24 (maximum 36), with a converted mean of 45.6%. Strengths included image protocol quality and discrimination statistics, while common limitations involved lack of phantom studies, biological correlates, cost-effectiveness analyses, and open science practices. Interpretation: These limitations may reduce reproducibility and hinder translation of deep learning models into clinical practice.