Table 2.

Summary of prognostic performance of US radiomics in head and neck oncology reported in the literature (studies have been summarized in alphabetical order).

Authors	Radiomics Platform	Number of Features	Statistical Analysis/Model Performance
Dasgupta et al., 2020 [38]	MATLAB	31	Recurrence vs. no recurrence kNN-based model a priori to treatment Training AUC: 0.74
Fatima et al., 2020 [39]	MATLAB	31	Recurrence vs. no recurrence Pre-treatment AUC: 0.71 1-week post-treatment AUC: 0.75 4 weeks post-treatment AUC: 0.81
Osapoetra et al., 2021 [40]	MATLAB	105	Prediction of clinical outcome (early responders vs. late responders vs. progressive disease) SVM AUC: 0.91
Park et al., 2019 [41]	MATLAB	40	Estimation of disease-free survival C-index: 0.78 (95% CI: 0.735, 0.829)
Tran et al., 2019 [42]	MATLAB	41	Complete vs. partial responders Univariate models kNN AUC: 0.81 (95% CI: 0.640, 0.980) naive-Bayes AUC: 0.87 (95% CI: 0.730, 1.010)
Tran et al., 2020 [43]	MATLAB	31	Complete vs. partial response Univariate kNN classifier 24 h post-RT AUC: 0.74 1-week post-RT AUC: 0.81 4 weeks post-RT AUC: 0.80