Segmentation of Tumor Images and Radiomics Feature Extraction

ITK-SNAP software (Version 3.6.0, www.itksnap.org) was used for segmentation of tumors. A defined polygonal region-of-interest was delineated on the center slice of the ccRCC on corticomedullary phase (CMP) and nephrographic phase (NP) images, avoiding covering the paratumoral renal parenchyma and perinephric fat ( Figure 2 ).

Manual segmentation of the tumor on the center axial slice of the clear cell renal cell carcinoma (ccRCC).

AK software (AnalysisKit 3.2.0; GE Healthcare, China) was used to extract a total of 396 radiomics features from the region-of-interest for one phase. The radiomics features are detailed in Supplementary Information 1.3 .

Inter- and intra-class correlation coefficients (ICCs) were calculated to estimate the inter-observer reliability and intra-observer reproducibility of features extraction. 20 cases of CT images containing six aggressive ccRCCs and 14 non-aggressive ccRCCs were randomly chosen; region-of-interest segmentation was drawn by one radiology resident (Reader 1, BK) and one radiologist (Reader 2, XW) independently; both were aware of the diagnosis of ccRCC but were blinded to the pathologic details. Reader 1 then repeated the contouring procedure 8 weeks after the initial analysis to assess the agreement of feature extraction. The remaining image segmentation was performed by Reader 1.