Comprehensive Analysis of Cell-Free DNA Fragmentation Across Cancer Stages

Abstract

Abstract Background Circulating cell-free DNA (cfDNA) in the bloodstream exhibits cancer-derived fragmentation patterns, providing a noninvasive diagnostic avenue for cancer patients. However, the alterations of cfDNA fragmentation patterns throughout cancer progression remain largely unexplored.Methods In this study, we investigated the relationship between cfDNA fragmentation profiles and cancer stages by analyzing a comprehensive dataset consisting of 214 whole-genome cfDNA samples across seven cancer types.Results Our analysis revealed that the association between cfDNA fragmentation patterns and cancer stages showcases variations in distinct cancer types. Notably, cfDNA fragmentation patterns in colorectal cancer (CRC) demonstrated a strong correlation with cancer stages. We validated this finding using an independent targeted cfDNA dataset with 29 CRC samples. Furthermore, we introduced "frag2stage", a machine learning model that leverages cfDNA fragmentation data to differentiate cancer stages of CRC. Through evaluation across two independent cfDNA datasets, our model consistently demonstrated its ability to distinguish CRC cancer stages with area under the curve (AUC) values ranging from 0.68 to 0.99.Conclusion Our analysis provides strong evidence that cfDNA fragmentation patterns might carry yet undiscovered genetic and epigenetic signals, highlighting their promising potential for broader diagnostic applications in oncology.