Computed Tomography Radiomic Features Reflect ATP-Binding Cassette Subfamily G Member 2 Expression in Ovarian Serous Cystadenocarcinoma: A Radiogenomic Study in the Context of Chemoresistance
Retrospective radiogenomic study investigates CT-derived features and chemoresistance markers in ovarian cancer, suggesting imaging insights into molecular profiles.
Key Points
This study aims to explore the relationship between CT-derived radiomic features and ABCG2 expression in ovarian serous cystadenocarcinoma, focusing on chemoresistance mechanisms.
Retrospective analysis of 70 patients from the Cancer Genome Atlas with contrast-enhanced CT images and gene expression data.
Manual segmentation of tumors using Horos, with radiomic feature extraction from the largest tumor area on axial CT slices.
Statistical analysis conducted with univariate testing and exploratory modeling using LASSO and ridge regression.
Five radiomic features displayed nominally significant differences between ABCG2-positive and ABCG2-negative tumors, focusing on tumor shape and attenuation.
No features remained significant after correction for multiple comparisons.
Predictive performance of the models was limited, with area under the curve values of 0.508 (LASSO) and 0.598 (ridge regression).