Introduction: The study was conducted to investigate how quantitative texture analysis can be used to non-invasively identify novel radiogenomic correlations with Clear Cell Renal Cell Carcinoma (ccRCC) biomarkers which are relevant in IO / VEGFi therapy as well as prognostic biomarker panels viz Clear code 34.
Methods: The Cancer Genome Atlas–Kidney Renal Clear Cell Carcinoma (TCGA-KIRC) open-source database was used to identify 190 sets of patient genomic data that had corresponding multiphase contrast-enhanced CT images in The Cancer Imaging Archive (TCIA-KIRC). Only CT images in which the tumor was more than 8 pixels were included for analysis. Twelve clinically relevant biomarkers were identified from the literature.2824 radiomic features spanning fifteen texture families were extracted from CT images using a custom-built software package in MATLAB. Robust radiomic features with strong inter-scanner reproducibility were selected. Random Forest (RF), AdaBoost and Elastic Net machine learning (ML) algorithms were used to evaluate the ability of the selected radiomic features to predict the presence of the previously identified biomarkers. ML analysis was repeated with cases stratified by stage (I/II vs. III/IV) and grade (1/2 vs. 3/4). 10-fold cross validation was used to evaluate model performance.
Results: Before stratification, radiomics predicted the presence of several biomarkers with weak discrimination (AUC 0.60-0.66). Among patients with high stage (III / IV), radiomics predicted the presence of TeffhighMyeloidlow gene expression subtype and high indel burden with acceptable discrimination (AUC 0.71 and 0.73, respectively). Among high-grade patients, radiomics predicted the presence of Teffhigh/Myeloidlow gene expression subtype with acceptable discrimination (AUC 0.71 and 0.72, respectively), and high indel burden with excellent accuracy (AUC 0.83). Additionally, radiomics predicted ClearCode34 risk class with acceptable accuracy in low-stage patients (AUC 0.73).
Conclusions: Radiomic texture analysis has the potential to identify a variety of clinically relevant biomarkers in patients with ccRCC and may be used in addition to biopsy results to predict prognosis and select treatment.
Source of Funding: Clinical and Translational Science Institute; American Cancer Society