Abstract 523P
Background
Early detection of cancer significantly improves patient outcomes. Current MCED tests often rely on single biomarkers, which can limit diagnostic accuracy. Multimodal analysis of ctDNA utilizing various molecular signatures presents a chance for enhanced diagnostic performance. This study explores the integration of methylation, fragmentomics, and hotspot mutations to improve MCED across diverse cancer types.
Methods
A single-blood draw workflow was developed, employing low-pass genome-wide bisulfite sequencing and targeted amplicon-based sequencing using a 700-hotspot panel to profile both epigenetic and genomic changes in plasma cell-free DNA. This workflow was retrospectively validated in a case-control cohort of 242 non-metastatic patients with five cancers (breast, colorectal, gastric, liver, and lung) and 304 healthy individuals.
Results
Hotspot mutations were detected in 123 out of 242 (50.83%) cancer patients, with the highest detection rate in liver cancer (96.55%, 28/29), followed by colorectal cancer (59.32%, 35/59) and lung cancer (50.00%, 14/28). Most mutations were found in TP53, PIK3CA, KRAS, APC, CTNNB1, EGFR, and ARID1A. Cancers with low tumor mutation burden (TMB), such as breast and gastric cancer, showed moderate detection rates of 31.25% (20/64) and 41.94% (26/62), respectively. In contrast, epigenetic signatures, including methylation, fragment length, and motif end profiles, demonstrated higher sensitivities for these low TMB cancers, with 51.56% (33/64) for breast and 62.90% (39/62) for gastric cancer. Combining methylation and mutation markers significantly improved early-stage cancer detection across all five types, with an overall sensitivity of 77.27% and a specificity of 97.70%. Enhanced sensitivities were observed for colorectal cancer (81.36%, 48/59) and lung cancer (75%, 21/28).
Conclusions
This study underscores the potential of a multimodal assay that combines genetic and epigenetic alterations for the improved early detection of various cancers. Despite the promising results, further validation in larger cohorts is necessary to support broader clinical applications.
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Gene Solutions JSC.
Disclosure
All authors have declared no conflicts of interest.