A multicenter retrospective study of 2,065 colorectal cancer patients reported that CT deep radiomics can noninvasively predict homologous recombination deficiency (HRD) status. The model also stratified survival outcomes, linking imaging-derived signatures to genomic repair phenotypes. The work supports a growing pipeline concept: using routine imaging to approximate molecular risk factors that would otherwise require costly or limited tissue testing. For precision oncology programs, HRD prediction can influence therapeutic selection, particularly in contexts where DNA damage response status is clinically relevant. If prospectively validated, such signatures could help scale HRD-like stratification across broader patient populations.
Get the Daily Brief