A liquid biopsy cancer test analyzes tumor-related material in blood or another body fluid, most often fragments of circulating tumor DNA, or ctDNA, in plasma. It can identify actionable mutations when tissue is unavailable, reveal resistance changes at progression, and sometimes detect molecular residual disease after treatment. It is not a universal replacement for tissue biopsy. A blood sample may contain too little tumor DNA, especially in early-stage, low-volume, brain-only, lung-only, or peritoneal disease, so a negative result cannot reliably rule out a mutation or cancer. Positive results also need context because age-related blood-cell clones can release mutations into plasma that do not come from the tumor. The most useful report states the genes and alteration types covered, plasma tumor fraction, variant allele fraction, detection limit, and whether findings are likely tumor-derived. Interpretation depends on the clinical question: selecting therapy in advanced cancer, investigating resistance, monitoring a known mutation, or assessing recurrence risk after surgery are different uses with different evidence.
- A positive actionable ctDNA result can guide targeted treatment when the alteration fits the cancer type and the assay is clinically validated.
- A negative plasma result is not a true negative when tumor shedding is low; tissue testing should follow when feasible and treatment depends on finding a driver.
- Liquid biopsy usually tests plasma cell-free DNA, only a fraction of which may come from cancer.
- DNMT3A, TET2, ASXL1, and other blood-cell mutations may reflect clonal hematopoiesis rather than the solid tumor.
- ctDNA for minimal residual disease can estimate recurrence risk, but using it to start or stop treatment is not standard for every cancer.
Table of Contents
- What a liquid biopsy measures
- When ctDNA testing is used
- Collection and laboratory methods
- How to interpret a positive result
- Negative, uncertain, and conflicting results
- MRD, monitoring, and early detection
- Choosing a test and next steps
What a liquid biopsy measures
Cells release short DNA fragments into the bloodstream as they die and turn over. The total pool is called cell-free DNA, or cfDNA. Most cfDNA in a person with cancer still comes from normal blood cells. The tumor-derived portion is called circulating tumor DNA, or ctDNA. In some samples, ctDNA makes up more than 10% of cfDNA; in others, it is below 0.1% or undetectable.
A typical liquid biopsy separates plasma from blood cells, extracts cfDNA, and analyzes it for cancer-associated changes. Depending on the assay, these can include:
- single-nucleotide variants, such as EGFR L858R or KRAS G12C;
- small insertions and deletions, such as EGFR exon 19 deletions;
- gene amplifications, such as ERBB2 or MET amplification;
- rearrangements or fusions involving ALK, ROS1, RET, or other genes;
- splice alterations, including MET exon 14 skipping when the design supports it;
- microsatellite instability or a blood-based tumor mutational burden estimate; and
- methylation patterns or fragment features used by some early-detection and residual-disease tests.
The term “liquid biopsy” can also include circulating tumor cells, extracellular vesicles, tumor RNA, proteins, urine DNA, cerebrospinal-fluid DNA, or other body fluids. These tests are not interchangeable. A report should name the analyte and specimen rather than relying on the broad label.
Plasma can sample DNA released from several tumor sites at once. That may reveal heterogeneity missed by a small tissue biopsy. It also allows repeat testing without another invasive procedure. The tradeoff is loss of cellular and architectural information. Blood cannot show tumor histology, grade, immune-cell pattern, or whether a lung lesion is adenocarcinoma rather than small-cell cancer.
The amount of ctDNA depends on tumor burden, location, blood supply, biology, and recent treatment. Liver and bone metastases often shed more than small lung nodules, brain-only disease, or peritoneal tumors. Surgery, radiation, inflammation, strenuous exercise, and normal cell turnover can change total cfDNA without proportionally changing tumor DNA.
Variant allele fraction, or VAF, is the percentage of DNA reads carrying an alteration. A plasma VAF of 2% does not mean 2% of the body is cancer. It reflects the mixture of tumor-derived and normal cfDNA, plus copy-number changes and clonal structure. The highest credible tumor-derived VAF can provide a rough estimate of plasma tumor fraction, but it is not a direct measurement of tumor size.
A ctDNA test may be tumor-naive, meaning it uses a fixed panel without first sequencing the tumor, or tumor-informed, meaning it designs a personalized assay around mutations found in that patient’s tumor. The first approach is convenient for broad profiling; the second can improve specificity for molecular residual disease.
When ctDNA testing is used
The strongest routine use is genomic profiling of advanced cancer when tissue is unavailable, insufficient, risky to obtain, or too slow. Plasma testing is widely used in metastatic non-small cell lung cancer because treatment depends on finding drivers such as EGFR, ALK, ROS1, BRAF, MET exon 14, RET, NTRK, HER2, and KRAS G12C. A positive result for a validated target can shorten the time to treatment.
Liquid biopsy is also used at disease progression. Targeted therapy creates selection pressure, and resistant clones may emerge in different metastases. Plasma can reveal EGFR C797S, MET amplification, secondary ALK mutations, ESR1 mutations, RAS pathway changes, or other mechanisms. Tissue is still needed when the question includes histologic transformation, such as lung adenocarcinoma changing to small-cell carcinoma.
In metastatic breast cancer, plasma testing can detect ESR1 mutations that influence endocrine therapy and PIK3CA mutations that may guide targeted treatment. In colorectal cancer, it can identify RAS or BRAF mutations, track acquired resistance to anti-EGFR therapy, and find KRAS G12C. In prostate cancer, plasma can detect homologous-recombination repair alterations, although low tumor fraction and clonal hematopoiesis can complicate interpretation.
A third use is serial monitoring. When a tumor has a known mutation, falling ctDNA during treatment may accompany response, while a rise may precede radiographic progression. This is biologically compelling, but routine imaging and clinical assessment remain essential. A single rise can result from assay variability, tumor flare, or a new clone and should not trigger a major treatment change without confirmation.
A fourth use is molecular residual disease, often abbreviated MRD, after surgery or curative-intent treatment. Detectable ctDNA can indicate that microscopic cancer remains and is associated with a high risk of recurrence in colorectal, breast, lung, bladder, and other cancers. However, prognostic power does not automatically prove that changing treatment based on the test improves survival.
Liquid biopsy is not recommended as a general cancer screen in a person without symptoms outside established programs or trials. Multi-cancer early-detection tests are commercially available in some regions, but false positives, false negatives, tissue-of-origin uncertainty, overdiagnosis, and unproven mortality benefit remain important. Standard screening such as mammography, colon screening, cervical screening, and low-dose CT for eligible people should not be replaced.
It also should not be used to diagnose recurrence solely because a result is weakly positive without confirmation. The appropriate response may include repeat testing, imaging, review for clonal hematopoiesis, and examination of the original tumor profile.
Collection and laboratory methods
Most tests require one or two tubes of blood. Fasting is usually unnecessary. The laboratory’s collection kit matters because ordinary EDTA tubes require prompt plasma separation, often within several hours. Specialized preservative tubes stabilize white blood cells and allow longer shipping. Delayed processing can cause white cells to break apart and flood the sample with normal genomic DNA, diluting ctDNA.
The blood draw itself is low risk, but timing affects yield. A sample taken immediately after surgery may contain abundant cfDNA from tissue injury. MRD protocols therefore specify collection windows, often several weeks after surgery. A sample drawn during severe infection or inflammation may also have high background cfDNA.
Common laboratory strategies include:
| Method | Best suited for | Main limitation |
|---|---|---|
| Digital PCR | One or a few known mutations | Cannot discover unexpected alterations |
| Amplicon-based NGS | Focused hotspot panels with low DNA input | Uneven performance for fusions and copy number |
| Hybrid-capture NGS | Broad profiling of variants, indels, fusions, and copy number | More complex and may require more DNA |
| Tumor-informed MRD assay | Highly specific tracking of personal tumor variants | Requires tumor tissue and assay design time |
| Tumor-naive MRD or screening assay | No tumor tissue required | More dependence on statistical classifiers and background filtering |
Error suppression is essential because the clinically relevant signal may be one mutated fragment among thousands of normal fragments. Unique molecular identifiers label original DNA molecules so sequencing errors can be distinguished from true variants. Deep sequencing alone is not enough if errors are not controlled.
Analytic sensitivity is often stated as a limit of detection, such as 0.1% VAF. That number may apply only under ideal DNA input and to selected variant types. Fusions, amplifications, long insertions, and complex deletions can have different sensitivity from single-base substitutions. The report should not imply uniform performance across all genes.
Tumor fraction is a useful quality measure. If the assay estimates that no tumor DNA was present, a negative result is uninformative. Some reports label this “low shedding,” “ctDNA not detected,” or “indeterminate.” Others do not provide a direct estimate, so the clinician must infer it from detected variants and clinical burden.
Paired white-blood-cell sequencing improves specificity by identifying mutations that originate in blood cells. Without it, bioinformatic filters and variant patterns are used. Neither approach is perfect. Mutations in TP53, KRAS, PIK3CA, and other genes can occasionally arise in clonal hematopoiesis as well as tumors.
A broad tumor molecular profile may combine tissue and plasma. Tissue establishes histology and provides high sensitivity when tumor content is adequate; plasma adds speed and a wider view of metastatic heterogeneity. The two methods are complementary rather than competitors.
How to interpret a positive result
A positive result means the assay detected an alteration above its validated threshold. The first question is whether that alteration is likely to come from the tumor. The second is whether it is clinically actionable in that cancer, stage, and treatment line.
Reports commonly classify findings into tiers:
- alterations with an approved therapy for that tumor type;
- alterations with an approved therapy in another tumor type;
- findings supported by clinical trials or emerging evidence;
- resistance mutations; and
- variants of uncertain significance.
The same variant can have different meanings. EGFR L858R is a strong treatment marker in lung adenocarcinoma but not a general target in every cancer. BRAF V600E has tissue-specific therapy combinations. KRAS G12C treatment differs between lung and colorectal cancer. Actionability should never be inferred from the gene name alone.
A credible positive finding is more persuasive when it matches the tissue result, known tumor biology, or other alterations at similar VAFs. For example, a lung-cancer plasma sample with EGFR exon 19 deletion, TP53 mutation, and several copy-number changes has a coherent tumor pattern. An isolated DNMT3A mutation at 1% in an older adult is more likely to reflect clonal hematopoiesis.
Clonal hematopoiesis is the growth of a blood-cell clone carrying acquired mutations. It becomes more common with age and after chemotherapy or radiation. DNMT3A, TET2, ASXL1, PPM1D, TP53, JAK2, and spliceosome genes are common. These findings can be medically relevant to blood health and cardiovascular risk, but they should not be misreported as a solid-tumor mutation.
High VAF does not always mean a more important target. A truncal driver may have a high VAF because it is present in most cancer cells. A resistance mutation can be clinically decisive at low VAF because it represents a growing subclone. Copy-number changes can also distort percentages.
An unexpected variant that appears near 50% VAF may be germline. Plasma contains DNA from normal cells, so inherited variants can be visible. Tumor profiling cannot confirm hereditary cancer. A finding in BRCA1, BRCA2, PALB2, mismatch-repair genes, TP53, or another predisposition gene may warrant genetic counseling and confirmation in a dedicated germline specimen.
A positive MRD result after curative-intent treatment is not the same as a targetable mutation report. It means tumor-derived signal was detected and recurrence risk is increased. The size of the risk and the recommended response depend on cancer type, stage, assay, timing, and trial evidence.
Negative, uncertain, and conflicting results
A negative liquid biopsy means no reportable tumor alteration was detected in the plasma sample. It may reflect true absence, but it often means there was too little ctDNA to analyze. The distinction is critical.
False-negative risk is higher with:
- early-stage or low-volume disease;
- brain-only or central nervous system disease;
- small lung-only metastases;
- peritoneal or mucinous tumors;
- tumors with low cell turnover;
- treatment that recently reduced tumor burden;
- poor sample handling; and
- alterations that the assay detects less well, such as certain fusions or copy-number changes.
When advanced lung cancer needs a complete driver profile, a negative plasma result should be followed by tissue NGS if possible. The same principle applies whenever finding a mutation would change treatment. “No variants detected” is not equivalent to “all targetable alterations excluded.”
A variant of uncertain significance, or VUS, should not guide treatment. It is a real sequence change with insufficient evidence about function. Laboratories may reclassify variants as knowledge grows. Patients should not pursue an unproven drug solely because a VUS appears in a cancer-associated gene.
Tissue and plasma may conflict. Common patterns include:
| Pattern | Likely explanation | Usual response |
|---|---|---|
| Tissue positive, plasma negative | Low shedding | Trust validated tissue result; do not call mutation absent |
| Plasma positive, tissue negative | Heterogeneity, newer clone, old tissue, or clonal hematopoiesis | Review VAF, biology, white-cell data, and consider confirmation |
| Different resistance mutations | Multiple metastatic clones | Integrate both; consider broader treatment or trial |
| Plasma mutation at very low VAF only | Early subclone or artifact | Confirm before high-stakes action |
A failed test is different from a negative test. Failure may result from insufficient cfDNA, sequencing quality, contamination, or sample damage. The report should say whether a redraw is likely to help. Drawing more blood can increase DNA input, but it cannot force a non-shedding tumor to release ctDNA.
Cerebrospinal fluid may be more informative than plasma for brain tumors or leptomeningeal disease because the blood-brain barrier limits DNA release into blood. Urine can be useful for selected urinary-tract cancers. The best “liquid” depends on where the disease is located.
A changing result needs cautious interpretation. Clearance of a mutation during therapy is favorable in many studies but does not prove cure. A new mutation may signal resistance before scans change, but treatment should usually not be stopped on one molecular result without clinical correlation.
MRD, monitoring, and early detection
Molecular residual disease testing searches for tiny amounts of tumor DNA after surgery, radiation, or other curative-intent treatment. Imaging may show no cancer while ctDNA reveals microscopic disease. Across several solid tumors, a positive postoperative result is strongly associated with recurrence.
There are two main designs. Tumor-informed tests first sequence the resected tumor and track a personalized set of variants. This reduces the chance of mistaking clonal hematopoiesis for cancer. Tumor-naive tests use fixed panels, methylation, fragmentation, or combined signals and do not need tissue. They are faster to start but may be less personalized.
Timing matters. A draw too soon after surgery can be diluted by injury-related cfDNA. A draw too late can miss the chance to guide adjuvant treatment. Protocols often use a first sample about 2 to 8 weeks after surgery and repeat samples during surveillance, but exact schedules differ.
A positive MRD test is generally more informative than a negative one. Positive predictive value can be high, especially when the same signal persists or rises. Negative predictive value is lower because some tumors shed little DNA and recurrence can arise after an initially negative sample.
Clinical utility remains the central issue. A test can predict recurrence without proving that acting on it improves survival. Trials are evaluating whether ctDNA can identify people who need more chemotherapy, allow safe de-escalation, trigger earlier imaging, or select targeted therapy. Results are most mature in colorectal cancer, but routine recommendations still vary by stage and setting.
For surveillance, serial testing may detect recurrence months before imaging. Earlier detection is not automatically beneficial if no effective intervention exists at that stage. It can also create prolonged anxiety and repeated scans. Patients should know in advance what the oncology team will do with a positive or indeterminate result.
Multi-cancer early-detection assays use methylation or other signals to detect cancer and predict tissue of origin. Their performance is better for some advanced cancers than for small early tumors. A low prevalence of cancer in asymptomatic populations means even a highly specific test can generate false positives. Positive results often require extensive imaging and invasive procedures.
These assays should not replace proven screening. A person with a negative multi-cancer test still needs age- and risk-appropriate mammography, colon screening, cervical screening, and lung screening when eligible. Symptoms such as bleeding, a new mass, unexplained weight loss, or persistent pain also need standard evaluation regardless of a negative blood test.
Choosing a test and next steps
Choose the assay by the clinical question. Broad plasma NGS is appropriate when advanced cancer needs a treatment target. A focused digital-PCR test may be best for rapidly checking one known resistance mutation. A tumor-informed assay may be preferred for postoperative MRD. A screening assay should be used only with a clear pathway for confirmatory care.
Before ordering, ask:
- What decision will the result change?
- Is tissue available, and should tissue and plasma be tested together?
- Which genes, exons, fusions, copy-number changes, and biomarkers are covered?
- Does the assay report tumor fraction and detection limits by alteration type?
- How does it filter clonal hematopoiesis?
- What happens after a negative result?
- Is the MRD or screening use supported in this cancer and stage?
- Will insurance cover both the test and follow-up care?
After a positive advanced-cancer result, the oncology team should confirm that the alteration is an accepted biomarker for the tumor type and treatment line. A molecular tumor board can help with rare variants, conflicting evidence, or trial matching. A drug listed in the report is not automatically appropriate; health status, prior therapy, organ function, and other tumor features still matter.
After a negative result, review tumor burden and assay quality. If plasma tumor fraction was low or absent, tissue testing is usually the next step. If tissue is unavailable, a repeat plasma draw during clear progression may yield more DNA, but delaying necessary treatment solely to obtain a positive result is rarely appropriate.
After an unexpected hereditary finding, request genetic counseling. Confirmatory germline testing uses informed consent and addresses family implications. Do not test relatives based only on a plasma tumor report.
After an MRD result, ask the team to state the plan before repeating the test: observation, standard adjuvant treatment, trial enrollment, additional imaging, or confirmation. Serial testing without a response plan can add cost and anxiety without improving care.
Keep reports with specimen dates and treatment timelines. Plasma results can change quickly, so a report should be labeled as baseline, response, progression, or postoperative surveillance. The value of liquid biopsy comes from using a minimally invasive sample to answer a defined question while recognizing when tissue, imaging, or standard pathology remains necessary.
References
- ESMO recommendations on the use of circulating tumour DNA assays for patients with cancer: a report from the ESMO Precision Medicine Working Group 2022 (Position Statement)
- Liquid Biopsy for Advanced NSCLC: A Consensus Statement From the International Association for the Study of Lung Cancer 2021 (Consensus)
- ctDNA in the reading room: A guide for radiologists 2024 (Review)
- Cancer in a drop: Advances in liquid biopsy in 2024 2025 (Review)
- Circulating Tumor DNA: A Pan-Cancer Biomarker in Solid Tumors 2025 (Review)
- Japan society of clinical oncology position paper on appropriate clinical use of molecular residual disease (MRD) testing 2025 (Position Statement)
Disclaimer
This article is educational and does not replace diagnosis, screening, or treatment planning by an oncology team. A negative liquid biopsy cannot rule out cancer or a targetable mutation, and positive findings may require tissue, blood-cell, or germline confirmation. New or worsening cancer symptoms require standard medical evaluation even when a plasma test is negative.
