Table 1. Average results from the leave-one-patient-out cross-validations for each model.

Model	Locality	mean ROC AUC	mean Dice	mean Sensitivity	mean Specificity	mean Prediction Time (s)
LR	Global	0.809±0.110**	0.322±0.218**	0.386±0.243**	0.959±0.047	0.055±0.031
LR	Local	0.861±0.109**	0.337±0.221**	0.448±0.254	0.955±0.041**	0.062±0.015*
LR	Hybrid	0.872±0.092	0.348±0.221	0.444±0.252	0.955±0.047**	0.126±0.038**
RF	Global	0.789±0.104**	0.319±0.215**	0.361±0.218**	0.965±0.037	29.762±6.857**
RF	Local	0.845±0.099**	0.311±0.208**	0.404±0.208**	0.956±0.030**	704.859±146.593**
RF	Hybrid	0.859±0.089**	0.353±0.220	0.415±0.231**	0.964±0.034	736.284±148.309**

Average ROC AUC values, Dice coefficients, sensitivity, specificity, and prediction time values for one dataset from the leave-one-patient-out cross-validation for each model. Best results according to each metric are highlighted in bold. Significant differences to this best-performing method computed with a one-sided paired student’s t-test are marked with a star (*) for a confidence interval of 95% (p < 0.05) and two stars (**) for a confidence interval of 99% (p < 0.01). Nominal p-values are reported without correction for multiplicity, similarly as in [23].