Table 7 Applications of radiomics-based treatment response prediction
Image modality | Number of patients | Cancer | Target | Number of radiomics features | Commercial or open-source software | Method | References |
|---|---|---|---|---|---|---|---|
US | 36 | HNC | Responses to radical radiotherapy | 1–3 | SPSS, Matlab | ML: Naïve Bayes, KNN, Leave-one-out CV SM: Shapiro–Wilk test, t-test, Mann–Whitney U-test, Kaplan–Meier analysis, log-rank test | [44] |
CT | 290 | HNSCC | Incomplete response to definitive radiotherapy or chemo-radiation | Unspecified | PyRadiomics, Python | ML: L1-penalized (LASSO) LR, MI, Grid search with CV, fivefold CV SM: F-test, ANOVA, Pearson correlation analysis | [45] |
CT | 27 | HNSCC | Lymph node response to IC | 3 | 3D Slicer, R | ML: LASSO, LR, fivefold CV SM: Pearson correlation analysis | [46] |
MRI | 272 | NPC | Tumor retraction to IC combined with concurrent chemo-radiotherapy | 7/12 | ITK-SNAP, Artificial Intelligence Kit, R | ML: mRMR, LASSO, LR, tenfold CV SM: ICC, Z-score | [47] |
MRI | 137 | Rectal cancer | Treatment response to NAC | 19 | ITK-SNAP, Python, R | ML: LASSO, LR SM: ICC, Pearson correlation analysis, Univariate analysis, Backward elimination, Chi-square test or Fisher’s exact test, the Kruskal–Wallis test | [48] |
MRI | 140 | Breast cancer | Pathologic complete response to NAC | 5 | ITK-SNAP, GE Healthcare, Python, R | ML: LASSO, LR, fivefold CV SM: ICC, ANOVA, t-test, Spearman correlation analysis. Mann–Whitney U test, Chi-square test or Fisher’s exact test, Hosmer–Lemeshow test, DeLong test | [49] |
MRI | 634 | Rectal cancer | T downstaging (ypT0-2) after NAC | Unspecified | SPSS, Precision Medicine Open Platform, R, SIMCA | ML: PCA, SVM, LR, LASSO, Partial least-squares discriminant analysis, RF SM: Mann–Whitney U test, Fisher’s exact test, Univariate analyses, Multivariate analyses, Pearson correlation analysis, ANOVA | [50] |