Skip to main content
. 2022 Feb 24;9:833283. doi: 10.3389/fmed.2022.833283

Figure 1.

Figure 1

The workflow of radiomics analysis. (A) The tumor region of interests (ROIs), which were manually segmented by an experienced radiologist from CT images with the lung window and contrast-enhanced mediastinal window, respectively; (B) the high throughput image features were extracted automatically from each ROI and the radiomic signatures were selected from them; (C) Random survival forests (RSFs) models were established for the progression-free survival (PFS) prediction.