Table 5.

Linear regression and Pattern recognition analysis with significant features from the VIBE_FA30.

Linear Regression of Significant Features	AUC	Sensitivity	Specificity	PPV	NPV	Accuracy	Cut-Off
Linear regression of the textural features extracted from the VIBE_FA30 with respect to the front of tumor growth	0.55	0.88	0.56	0.77	0.74	0.76	8.81
Linear regression of the textural features extracted from the VIBE_FA30 with respect to the tumor budding	0.65	0.96	0.64	0.82	0.91	0.84	0.56
Linear regression of the textural features extracted from the VIBE_FA30 with respect to the mucinous type	0.26	1.00	0.04	0.64	1.00	0.64	−0.17
Linear regression of the textural features extracted from the VIBE_FA30 with respect to the recurrence presence	0.79	0.90	0.66	0.47	0.95	0.72	0.27
Pattern Recognition Analysis with Significant Features	Dataset	AUC	Accuracy	Sensitivity	Specificity	Training time [sec]	Model Type and Parameters
KNN	Training set	0.96	0.90	0.91	0.89	13.4	Weighted KNN; number of neighbors:10; distance metric: Euclidean; distance weight: squared inverse
	Validation set	0.95	0.80	0.67	1
	Training set	0.94	0.93	0.84	0.96	8.3
	Validation set	0.94	0.89	0.89	0.89
	Training set	0.93	0.91	0.96	0.73	7.51
	Validation set	0.89	0.88	0.89	0.8
	Training set	0.9	0.94	0.98	0.84	8.4
	Validation set	0.85	0.91	0.94	0.8