Table 4.

Linear regression and Pattern recognition analysis with significant features from the VIBE_FA10.

Linear Regression of Significant Features	AUC	Sensitivity	Specificity	PPV	NPV	Accuracy	Cut-Off
Linear regression of the textural features extracted from the VIBE_FA10 with respect to the front of tumor growth	0.72	0.93	0.82	0.90	0.88	0.89	1.49
Linear regression of the textural features extracted from the VIBE_FA10 with respect to the tumor budding	0.78	0.84	0.84	0.94	0.65	0.84	1.54
Linear regression of the textural features extracted from the VIBE_FA10 with respect to the mucinous type	0.80	0.85	0.82	0.56	0.95	0.83	0.28
Linear regression of the textural features extracted from the VIBE_FA10 with respect to the recurrence presence	0.63	0.52	0.88	0.59	0.84	0.79	3.81
Pattern Recognition Analysis with Significant Features	Dataset	AUC	Accuracy	Sensitivity	Specificity	Training Time [sec]	Model Type and Parameters
KNN	Training set	0.96	0.91	0.84	0.95	8.7	Weighted KNN; number of neighbors:10; distance metric: Euclidean; distance weight: squared inverse
	Validation set	0.97	0.92	1	0.86
	Training set	0.89	0.93	0.81	0.97	3.9
	Validation set	0.9	0.93	0.73	1
	Training set	0.93	0.89	0.94	0.73	3.2
	Validation set	0.95	0.88	0.91	0.8
	Training set	0.91	0.93	0.99	0.77	9.21
	Validation set	0.97	0.94	0.9	0.91