Skip to main content

Table 4 Sensitivity of the segmentation models

From: Automated segmentation of brain metastases in T1-weighted contrast-enhanced MR images pre and post stereotactic radiosurgery

Sensitivity

Baseline

FU1

FU2

FU3

FU4

FU5

FU6

Model 1 – nnU-Net trained with ETZ data only

0.93

0.71

0.63

0.53

0.53

0.55

0.51

Model 2 - nnU-Net trained with ETZ and public data

0.94

0.92

0.84

0.83

0.82

0.83

0.83

Model 3 - MedNeXt trained with ETZ data only

0.94

0.79

0.65

0.66

0.58

0.64

0.65

Model 4 - MedNeXt trained with ETZ and public data

0.94

0.91

0.84

0.84

0.80

0.81

0.82

  1. In bold is the value of the best sensitivity for the corresponding test set