Table 2.

Prognostic models (Lasso logistic regression) of the 15 imaging features (Table S4) without and with clinical features (Table S2), predicting locoregional recurrence-free survival, distant metastasis-free survival and overall survival, with the amount of patients (Table S7). The amount of features and the area under the curve (AUC) with the standard deviation (SD) is tabulated.

Logistic Regression Models
Outcome	Features	Patients	Imaging Features			Clinical Parameters + Imaging Features
			Features	AUC	SD	Features	AUC	SD
Locoregional recurrence-free survival	PRE	47	15	0.79	0.16	22	0.79	0.16
	INTRA	47	15	0.49	0.09	22	0.47	0.09
	Delta	47	15	0.76	0.14	22	0.77	0.14
	ALL	47	45	0.81	0.14	52	0.80	0.15
Distant metastasis	PRE	57	15	0.79	0.12	22	0.84	0.18
	INTRA	57	15	0.63	0.15	22	0.81	0.17
	Delta	57	15	0.52	0.12	22	0.82	0.15
	ALL	57	45	0.86	0.15	52	0.88	0.13
Overall survival	PRE	57	15	0.62	0.18	22	0.82	0.12
	INTRA	57	15	0.46	0.13	22	0.64	0.15
	Delta	57	15	0.48	0.11	22	0.66	0.15
	ALL	57	45	0.53	0.17	52	0.69	0.16