Table and Figure legends
Table 1. Patient characteristics (n = 257).
Table 2. Clinical endpoints of patients (n = 257).
Table 3. The prediction accuracy of treatment outcomes and hematologic toxicities.
Figure 1 . Schematic of data collection and analysis in the study.
Figure 2. The receiver operating characteristic (ROC) curves for discriminating between the two groups of patients divided by tumor complete response (a), 5-year overall survival (b), hypohemoglobin (c), and leucopenia (d) in the training and validation sets. The performance of the training set was evaluated in a five-fold cross validation.
Figure 3. The Kaplan-Meier Curves of patients in the high-risk and low-risk groups stratified using the threshold value of the predicted scores in the clinical model (a), the radiomic model (b), and the clinical and radiomic model (c). The threshold values were defined by the ROC curve of the training set. Patients censored < 5 years were merged with the validation test when plotting the Kaplan-Meier Curves.
Figure 4. Boxplots and data distribution of the scores of all patients predicted by the clinical model (a), the radiomic model (b), and the clinical and radiomic model (c) for four clinical endpoints.