Skip to main content
. 2021 Jul 13;49(2):527–538. doi: 10.1007/s00259-021-05430-z

Table. 4.

Performances of the baselines vs the best-performing model trained on 3 GS groups (GGG 1–3, GGG 4, GGG 5) and tested on 2 GS groups (GGG 1–3, GGG 4–5) (top, grey: baseline models; bottom, orange: double-image radiomics), expressed as their balanced accuracies (mean and standard deviation, in percentages), and sensitivities by class (t-test comparisons: patient baseline vs radiomics baseline: p = 0.203; patient baseline vs PET + ADC: p = 0.011; radiomics baseline vs PET + ADC: 0.029)

graphic file with name 259_2021_5430_Tab4_HTML.jpg