Skip to main content
Insights into Imaging
Fig. 2 | Insights into Imaging

Fig. 2

From: A deep learning pipeline to simulate fluorodeoxyglucose (FDG) uptake in head and neck cancers using non-contrast CT images without the administration of radioactive tracer

Fig. 2

Area under receiver operation curves for four random forest models trained with a combination of radiomic features to classify CT regions based on FDG uptake. Experiment 1A compared regions of elevated versus negligible FDG uptake. Experiment 1B compared tumour regions of High versus Low FDG uptake. Each model was trained using a tenfold cross-validation method for 100 iterations on a selected group of 25 radiomic features. Following training, each of the 100 models is applied to the testing cohort to assess model performance. The statistical differences between each model are assessed using a one-way ANOVA. **p < 0.01; ***p < 0.001; ****p < 0.0001.

Back to article page