Table 2.

Summary of Logistic Regression fits for Training, Test Sets.

Independent Variables	# Variables	F Value	p-value	R2	AUC (Train)	AUC (Test) [95% LL, 95%UL]
3PC+Vol, Train-Test	2	5.63	0.0598	0.496	0.861	0.969 [0.882-1.00]
Ecc+ Reg+3PC, Train+Test	3	6.29	0.0984	0.541	0.889	0.906 [0.702-1.00]
Ecc+ Reg+3PC+Vol, Train+Test	4	6.98	0.137	0.586	0.944	0.938 [0.791-1.00]

Analysis of Training and Test Cases, Summary of Best Regression fits of combinations of eccentricity, SCR, and Volume to Gleason score. AUC, Area Under Curve; R2, coefficient of determination; LL, Lower Limit Confidence Interval; UL Upper Limit 95% Confidence Interval; ECC, eccentricity(0.45 ACE Threshold); Mod_Reg, Modified Regularization SCR; Reg, Regularized SCR; Vol, Volume (0.65 ACE threshold).