Table 3.
Radiomics predicts recurrence and metastasis of head and neck cancer
| Study | Number of patients | Tumor type | Imaging modality | Outcome, feature selection method, model | Machine learning algorithm |
|---|---|---|---|---|---|
| Zhang et al. 2019 87 | Train: 80pts Validation: 60pts |
NPC | T2, CE- T1 | LR-free survival. Radiomics score, Cox regression | Logistic Regression |
| Bogowicz et al. 2017 100 | Train: 93pts Validation: 56pts |
HNSCC | CE- CT | LC and HPV status. PCA in combination with univariable logistic regression. Multivariable logistic regression | Logistic regression, PCA |
| Martens et al. 2020 101 | Train: 103pts Validation: 71pts |
HNSCC | PET, low-dose-CT | LR, DM, OS. RadCat tool Cox regression analysis. Multivariable logistic regression | Logistic regression |
| Li et al. 2018 97 | Total: 306pts, 20 of whom developed with recurrence | NPC | CT, MR, PET | LR. PCA. Machine learning classifiers | PCA, ANN, KNN, SVM |
| Liu et al. 2019 90 | Total: 120pts | PTC | Preoperative ultrasound images | Metastasis. Support vector machine classifier | SVM |
| Wu et al. 2020 98 | Train: 141pts Validation: 96pts |
HNC | PET, CT | LR. PCA. Multivariate Cox proportional hazards regression | PCA |
| Zhou et al. 2020 95 | Total: 188pts | HNSCC | PET, CT | DM. Machine learning classifiers | SVM, DT and KNN |
| Bogowicz et al. 2017 102 | Train: 128pts Validation: 50pts |
HNSCC | PET, CT | LR. PCA and LASSO. Multivariable Cox regression | PCA, LASSO |
| Tan et al. 2018 89 | Train: 154pts Validation: 76pts |
ESCC | Arterial-phase CT | LMR. Rad-score, logistic regression | LASSO, logistic regression |
| Vallieres et al. 2017 103 | Total: 300pts | HNC | Pre-treatment FDG-PET and CT | LR and DM. Machine learning classifier | Random forests |
| Kwan et al. 2018 104 | Total: 300pts 36 DM pts | HPV-related Oropharyngeal Carcinoma | CT | DM. PyRadiomic. Radiomics score | Logistic regression |
| Park et al. 2019 92 | Train: 400pts Validation: 368pts |
PTC | Neck ultrasound | LNM. LASSO. Rad-score, LASSO regression | LASSO |
| Zhang et al. 2019 105 | Train: 360pts Validation: 120pts |
Non-metastatic T4 NPC | T1, T2, CE- T1 | LR, Rad-score, cox proportional hazards regression | Logistic regression |
| Zhang et al. 2019 99 | Total: 176pts | NPC | PET, CT | DM. LASSO. Multivariate logistic regression | Logistic regression |
| M.D. Anderson Cancer 88 | Train: 255pts Tune: 165pts Validation: 45pts |
HNC | CT, MRI, PET | 5-year LCR. Multivariable Cox regression | DT |
| Bahig et al. 2019 106 | Total: 176pts, 20 supraglottic , 5 pyriform sinus tumors | LHSCC | DECT | LRR. Univariate Cox regression | DT |
| Lu et al. 2019 93 | Train: 154pts Validation: 67pts |
PTC | non-contrast and venous CE- CT | LNM. SVM. Multivariable logistic regression | SVM, logistic regression |
| Qu et al. 2018 94 | Train: 90pts Validation: 91pts |
EC | MRI | LNM. LASSO. Multivariable logistic regression | LASSO, elastic net regression, logistic regression |
| Martens et al. 2020 107 | Train: 103pts Validation: 71pts |
HNSCC | 18F-FDG-PET, CT | LR, DM, OS. Rad- score. Multivariable survival regression | Logistic regression |
ANN: Artificial neural network; LCR: Local control rate; DECT: Dual-energy computed tomography; LHSCC: Larynx and hypopharynx squamous cell carcinoma.