Integrative cfDNA profiling from low-pass whole-genome sequencing enables tissue-of-origin prediction in cancer.
A blood test combining multiple DNA signals accurately identifies where a cancer started in 4 out of 5 patients, helping doctors choose the right treatment.
A stacked ensemble classifier integrating 11 cell-free DNA feature types—spanning genomic, fragmentomic, methylation, and microbial signals—from low-pass whole-genome sequencing achieves 80% top-1 accuracy across 17 cancer types in an independent validation cohort of 1,221 patients. Performance is retained at low tumor fraction (71%) and in cancers of unknown primary (73% accuracy), directly addressing the major clinical unmet need for non-invasive tissue-of-origin identification.
What the study was
- Study design
- original_research
- Population
- Cancer patients across 17 tumor types including cancers of unknown primary; multi-center China
- Sample size
- 3035
- Category
- Liquid Biopsy / Early Detection
- Maturity
- Validated
- Journal
- Mol Biomed
Why it surfaced
11-feature cfDNA ensemble achieves 80%/90% top-1/top-2 accuracy across 17 cancers in independent validation (n=1,221); low tumor fraction performance (71%) enables early/residual disease application; 73% CUP accuracy addresses major unmet clinical need; Geneseeq industrial translational pathway.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.