Key Takeaways
- “Liquid biopsy” is an umbrella term. It may analyse circulating tumour DNA (ctDNA), circulating tumour cells, RNA, proteins, vesicles or other material in blood, urine or another body fluid [1].
- Plasma genotyping can help identify treatment biomarkers in selected advanced cancers, especially when tissue is unavailable or time matters. It does not replace pathology, tumour architecture or every tissue-based test [3].
- A negative plasma result may mean “the tumour is not shedding enough DNA,” not “the alteration is absent.” Reflex tissue testing is often needed when feasible [2][3].
- Post-treatment ctDNA can predict recurrence risk in several cancers, but risk prediction and proof that changing treatment improves outcomes are different evidence levels.
- Some variants come from blood-cell clones or the germline rather than the tumour. The report must address variant origin before it drives cancer treatment or family testing [9].
Content
A blood draw is easier than a tissue biopsy, but the word “biopsy” can create the wrong expectation. A liquid test samples material shed into a body fluid; it usually cannot show tissue architecture, tumour grade, invasion, immune pattern or whether the needle reached the correct lesion. Its value depends on a tightly defined question.
The NCI definition includes tumour cells and fragments of DNA, RNA or other molecules in blood, urine or another fluid [1]. In current solid-tumour practice, many commercial “liquid biopsies” are plasma cell-free DNA tests. Only a fraction of cell-free DNA may come from the tumour; that fraction can be extremely small [5].
Four common uses need four different evidence discussions
Use · Question · Main limitation
Advanced-cancer genotyping · Is there a targetable or resistance alteration now? · low shedding, incomplete alteration coverage, uncertain variant origin
Molecular residual disease (MRD) · Is tumour-derived DNA detectable after curative-intent treatment? · negative does not prove zero disease; management utility is cancer- and assay-specific
Treatment monitoring · Is ctDNA changing during therapy? · timing and biological fluctuations; action thresholds may be unvalidated
Early detection/screening · Is there a cancer in a person without a diagnosis? · very low prevalence, false positives, false negatives, uncertain tissue of origin and downstream harm
A test validated for one row should not be marketed as proof for all four.
Advanced cancer: a complement to tissue
Validated, sufficiently sensitive plasma assays can identify actionable alterations in some advanced cancers. FDA approvals of liquid companion diagnostics illustrate how narrow the claim can be: a defined assay, specimen, tumour type, alteration and linked medicine [2]. The product’s current Chinese registration and intended use should likewise be checked before a report is treated as a companion diagnostic.
Tissue remains essential when diagnosis or histology is uncertain, when the biomarker requires protein expression or tissue context, or when the liquid assay performs poorly for the relevant fusion, copy-number change or tumour. ESMO recommends considering reflex tumour testing after a non-informative ctDNA result because false negatives occur [3].
Ask whether the test is tumour-agnostic profiling, a single-drug companion diagnostic or research screening for a trial. “A mutation was found” is not yet a treatment recommendation: the exact variant, tumour type, line of therapy, regulatory label and evidence level must match.
Why “not detected” has several meanings
The report should distinguish:
- no reportable variant detected;
- adequate tumour fraction with no target found;
- low or unmeasurable tumour fraction;
- sample failure or insufficient DNA;
- alteration type not covered or below the limit of detection.
Small-volume disease, treatment response and some anatomical sites can release little DNA into plasma. Technical sensitivity also depends on blood volume, tube, transport time, processing, extraction, panel design, sequencing depth and error suppression. FDA’s safety summary for an approved plasma companion test explicitly warns that a negative plasma result does not mean the tumour is negative and advises tissue confirmation when feasible [2].
Tissue and plasma may disagree without either being “wrong”
A tissue block samples one location at one time. Plasma may collect DNA from several lesions, but mainly from those shedding into circulation. Differences can reflect tumour evolution, treatment selection, spatial heterogeneity, low tumour fraction, assay coverage, an old tissue sample or a blood-cell variant.
Resolve a discordant result by checking dates, treatment between samples, tumour fraction, assay limits and variant allele fraction. When the result would trigger a high-risk or expensive therapy, molecular-pathology review and, if feasible, confirmatory tissue or orthogonal testing are more useful than choosing the more convenient report.
MRD: prognostic signal is not automatically a treatment instruction
After surgery or curative-intent therapy, ctDNA positivity can indicate a high relapse risk before imaging detects disease. Tumour-informed assays first sequence the tumour and then track patient-specific variants; tumour-naive assays use a preset or genome-wide signal without requiring the original tumour. The two designs have different turnaround, tissue requirements and false-positive controls [6].
Sampling time matters. Blood drawn too soon after surgery can contain abundant non-tumour cell-free DNA and dilute the signal. Serial testing may be more informative than one result, but only if the schedule and action rule were prospectively defined.
The DYNAMIC randomised trial in stage II colon cancer is an important, specific example. A ctDNA-guided strategy reduced adjuvant chemotherapy use compared with standard clinicopathological management without compromising recurrence-free survival; five-year follow-up remained similar between strategies [7][8]. This does not establish one universal MRD rule for stage III colon, lung, breast, pancreas or another assay. Use the exact disease, stage, time points and treatment algorithm from the evidence.
Monitoring and resistance: useful clues, incomplete control panel
Serial ctDNA may fall with response or reveal a resistance alteration. Yet imaging, symptoms, physical examination and laboratory results remain necessary. A transient rise can reflect timing or biology; a fall does not prove every lesion is responding. Brain-only, peritoneal or small-volume progression may be under-represented in plasma.
Before changing treatment on a molecular rise alone, ask whether the assay and decision threshold have prospective evidence, whether the action is routine or part of a trial, and what confirmatory imaging or tissue is planned. ESMO classifies several early-response and preclinical-resistance uses as research rather than routine care [3].
Screening healthy people has a different risk balance
In a low-prevalence population, even a highly specific test can produce false positives. A positive multi-cancer signal may not locate the tumour, and work-up can involve repeated imaging, invasive procedures and anxiety. A negative test cannot replace established screening such as cervical, colorectal, breast or lung screening in eligible groups.
NCI describes early detection as an active validation field and operates dedicated liquid-biopsy research programmes [4]. Do not interpret participation in an early-detection study as proof that a test reduces cancer mortality.
Not every mutation in plasma came from the cancer
Age-related clonal haematopoiesis can release mutated DNA from blood-cell clones. Variants in genes such as DNMT3A, TET2, ASXL1, TP53, JAK2 or others may therefore be misassigned to a solid tumour. Germline variants can also appear in cell-free DNA. Paired white-blood-cell analysis, bioinformatic filtering, tumour comparison and allele-pattern review help determine origin [9].
A possible inherited pathogenic variant needs confirmation in a suitable non-tumour sample through a genetics pathway before relatives are tested. A plasma tumour report should not be handed to family members as if it were a germline result.
The pre-analytical chain is part of test quality
Record the specimen type, collection tube, collection and processing times, transport temperature, plasma-separation method, laboratory, assay version and whether the sample met quality requirements. Haemolysis or delayed processing can release white-cell DNA and dilute ctDNA. Different laboratories may report allele fraction, tumour fraction and variants differently.
For cross-border care, confirm whether the Chinese laboratory accepts overseas tubes, whether the specimen may legally and technically cross the border, and whether the receiving trial requires testing in a central laboratory. A local result can be clinically useful yet still fail a trial’s protocol requirement.
Read the report as a decision document
A usable report should identify:
- specimen and collection date;
- assay, version, genes/regions and alteration classes;
- analytical sensitivity and important coverage limitations;
- tumour fraction or another measure of sample informativeness when available;
- each variant with nomenclature and allele fraction;
- likely tumour, clonal-haematopoiesis or germline origin where assessed;
- evidence tier tied to tumour type and therapy;
- negative-result caveat and tissue-reflex recommendation;
- whether the use is approved, guideline-supported or investigational.
Do not let a report’s list of “associated drugs” override pathology, tumour type, contraindications or the current drug label. For a clinical trial, verify the trial registry, protocol biomarker definition, test method, cut-off, screening window and whether repeat tissue is still required.
Medical disclaimer: This guide does not interpret an individual liquid-biopsy result or recommend treatment. Results require review with oncology and molecular-pathology professionals alongside tissue, imaging, stage, prior therapy and the exact assay validation.
FAQ
Can a liquid biopsy diagnose my cancer without tissue?
Sometimes it contributes molecular information, but it often cannot establish histology, grade or tissue architecture. Tissue is still required in many diagnostic situations.
Does “ctDNA not detected” mean I am cancer-free?
No. Disease may be present below the assay’s detection limit or may shed little DNA. Interpret the result with imaging, pathology, stage and sampling time.
Can liquid biopsy select a targeted drug?
In selected cancers, an appropriately validated companion or profiling assay can identify an actionable alteration. The exact variant, cancer, treatment line and drug label must match.
Is postoperative ctDNA testing standard for every cancer?
No. It is strongly prognostic in several settings, but evidence that ctDNA-directed treatment improves outcomes is still specific to particular diseases, assays and algorithms.
Why might the laboratory ask for white blood cells as well as plasma?
White-cell DNA helps distinguish tumour variants from clonal haematopoiesis or possible germline variants, reducing the risk of assigning a blood-cell mutation to the cancer.
Sources
- US National Cancer Institute — Definition of Liquid Biopsy
- US FDA — Safety and Effectiveness Summary for a Plasma Companion Diagnostic
- ESMO Precision Medicine Working Group — ctDNA Assay Recommendations
- US National Cancer Institute — Liquid Biopsy Consortium
- US National Cancer Institute — Why Low ctDNA Limits Liquid-Biopsy Sensitivity
- US FDA — Tumour-Informed and Tumour-Naive Liquid Biopsy Research for MRD
- New England Journal of Medicine — DYNAMIC Randomized Trial
- Nature Medicine — Five-Year Outcomes of the DYNAMIC Trial
- Clinical Cancer Research — False-Positive Plasma Genotyping From Clonal Haematopoiesis
Image Review
- Decision: Replaced with a topic-specific ImageGen hero and visually reviewed for medical relevance, obvious generation artifacts and bilingual reuse.
- Editorial note: The image is another generic clinician consultation with a test-tube and pathway icons. It does not show blood separation, plasma, cell-free DNA, sequencing, tissue–plasma comparison or residual-disease monitoring, so readers cannot distinguish liquid biopsy from an ordinary laboratory test. Replace it with a non-branded plasma tube, fragmented DNA and a clearly labelled decision pathway without fabricated patient results.