Liquid biopsy uses a blood sample (or other body fluid) to detect tumor-derived molecules — DNA, RNA, cells, or exosomes — circulating in the body. It has emerged as a complement to, and sometimes replacement for, tissue biopsy in cancer diagnosis and monitoring.
What’s actually being measured
Circulating tumor DNA (ctDNA)
Tumors continuously release DNA into the bloodstream as they grow and undergo apoptosis or necrosis. ctDNA carries the tumor’s genetic alterations — point mutations, copy number changes, methylation patterns, and structural rearrangements. Most clinical liquid biopsy applications focus on ctDNA.
Circulating tumor cells (CTCs)
Whole tumor cells that have entered the bloodstream. Rare (sometimes a single cell per mL of blood), but can be enriched and analyzed for proteins, transcriptomes, and metastatic potential.
Exosomes and extracellular vesicles (EVs)
Membrane-bound particles carrying proteins, lipids, and nucleic acids from the cell of origin. Active research area for biomarker discovery.
Cell-free RNA (cfRNA)
RNA fragments in plasma. Can identify tissue of origin and emerging clinical applications include early cancer detection.
How ctDNA is detected
The challenge: ctDNA is typically less than 1% of all cell-free DNA (cfDNA) in plasma — sometimes less than 0.1%. Detection requires high-sensitivity methods:
- Digital PCR (ddPCR, BEADS): Detects single molecules of known mutations. Excellent sensitivity for known targets.
- Targeted NGS panels with UMIs: Sequence cancer-relevant genes at very high depth. UMIs reduce error rates to detect mutations at <0.1% frequency.
- Whole-genome ctDNA sequencing: Lower depth but covers structural variants, copy number, and epigenetic features (methylation patterns).
- Methylation profiling: Tissue-specific methylation patterns enable tissue-of-origin inference for early cancer detection.
Major clinical applications
Treatment selection (companion diagnostics)
FDA-approved liquid biopsy assays guide targeted therapy decisions in lung, colorectal, breast, and other cancers — particularly when tissue is insufficient or unavailable.
Resistance monitoring
Detecting emergent resistance mutations (e.g., EGFR T790M in NSCLC, ESR1 in breast cancer) before clinical progression. Allows therapy switches earlier.
Minimal residual disease (MRD)
Detecting tumor-derived DNA after curative-intent treatment. ctDNA positivity predicts recurrence months before clinical or imaging detection. Increasingly used to stratify patients for adjuvant therapy.
Multi-cancer early detection (MCED)
Tests like Galleri analyze methylation patterns in cfDNA to detect cancer signals across many cancer types in asymptomatic patients. Performance varies by cancer stage and type — generally better for late-stage and certain biology-friendly cancers.
Limitations to be aware of
- Low sensitivity for early-stage disease: Small tumors shed less DNA. Detection rates for stage I cancers are often modest
- Tissue heterogeneity: ctDNA may not fully represent all tumor subclones
- Clonal hematopoiesis: Variants from blood cell clones can mimic tumor variants — requires paired buffy coat sequencing
- Negative results don’t rule out disease: Absence of detectable ctDNA doesn’t mean absence of cancer
- Standardization gaps: Pre-analytical variables (tube type, processing time) significantly affect results
Practical considerations
- Use Streck or PAXgene blood tubes — EDTA tubes degrade cfDNA quickly
- Process plasma within 4–6 hours when possible
- Pair tumor ctDNA testing with germline sequencing to filter inherited variants
- Match the assay to the question — broad MCED panels are different tools from MRD assays
The future
- Multi-analyte assays combining ctDNA, methylation, and protein markers
- Whole-genome ctDNA at population scale
- Liquid biopsy in non-cancer contexts (transplant rejection, autoimmune disease, prenatal diagnosis)
- Integration with imaging and clinical data via AI for risk stratification
Liquid biopsy isn’t replacing tissue biopsy across the board, but it has become a routine part of cancer care for selected applications. Match the test to the question, watch sensitivity carefully, and integrate results with clinical context.
