The idea that a simple blood draw could reveal the molecular profile of a patient's cancer — without needle, scalpel, or biopsy needle — once seemed futuristic. Today, liquid biopsy is an FDA-cleared clinical tool used in tens of thousands of oncology practices worldwide. In the United States alone, circulating tumor DNA (ctDNA) testing is guideline-recommended for multiple cancer types, with several assays approved by the FDA as companion diagnostics for targeted therapies. The field has grown from a research curiosity to a clinical standard in under a decade — and its applications continue to expand rapidly.
This guide explains the biology behind liquid biopsy, the different analytes that can be measured (ctDNA, circulating tumor cells, exosomes), the clinical scenarios where liquid biopsy is currently used, how its results compare to tissue biopsy, and the emerging role of minimal residual disease (MRD) monitoring in early-stage cancer.
What Is Circulating Tumor DNA?
When cancer cells die — through apoptosis, necrosis, or active secretion — they release fragments of their DNA into the bloodstream. This cell-free DNA (cfDNA) circulates briefly in plasma before being cleared by the kidneys and liver, with a half-life of approximately 15–60 minutes. In patients with cancer, a fraction of this cfDNA derives from tumor cells and carries the somatic mutations, copy number alterations, and epigenetic signatures of the tumor — this fraction is called circulating tumor DNA (ctDNA).
The challenge is detecting ctDNA against a background of normal cell-free DNA, which is present in vastly higher quantities. In patients with metastatic cancer, ctDNA may represent 1–80% of total cfDNA depending on tumor burden and vascularity. In patients with localized cancer or early-stage disease, ctDNA may represent as little as 0.01–0.1% of cfDNA — requiring extremely sensitive assays to detect. In patients after curative surgery with no detectable disease, ctDNA may be present at concentrations near the limits of current detection (parts per billion of plasma DNA).
The term "liquid biopsy" is broader than ctDNA and encompasses several other blood-based analytes. Circulating tumor cells (CTCs) are intact cancer cells shed into the bloodstream from primary or metastatic sites; they are rare (often 1–10 per mL of blood versus billions of normal blood cells) and require specialized capture technology (such as the FDA-cleared CellSearch system). Tumor-educated platelets (TEPs) acquire RNA from tumor cells and can provide information about tumor molecular profiles. Exosomes and extracellular vesicles released by tumor cells contain RNA, proteins, and DNA with potential diagnostic value. In current clinical practice, ctDNA is by far the most widely used and clinically validated liquid biopsy analyte.
How ctDNA Is Detected: Assay Types
Multiple technologies have been developed to detect ctDNA, differing in their sensitivity, the breadth of mutations they can detect, and their clinical validation status.
Targeted droplet digital PCR (ddPCR) detects specific known mutations with very high sensitivity (variant allele frequencies as low as 0.01–0.1%) but only interrogates single pre-specified variants. It is useful for monitoring known mutations identified from prior tissue biopsy — for example, tracking an EGFR T790M resistance mutation in a patient on an EGFR inhibitor — but cannot discover new variants.
Next-generation sequencing (NGS)-based plasma panels simultaneously interrogate hundreds to thousands of genomic loci, identifying somatic mutations, copy number alterations, fusions, and other alterations without prior knowledge of the tumor's molecular profile. FDA-approved examples include Guardant360 CDx (Guardant Health) and FoundationOne Liquid CDx (Foundation Medicine), both cleared as companion diagnostics for multiple targeted therapies. These assays use unique molecular identifiers (UMIs) to suppress sequencing errors and achieve variant allele frequency detection limits of approximately 0.1–0.5%.
Tumor-informed MRD assays represent the most sensitive class, designed specifically for minimal residual disease detection after curative-intent treatment. These assays use prior tissue sequencing data to design patient-specific panels targeting dozens to hundreds of tumor-specific mutations simultaneously, achieving sensitivities of 1 in 100,000 to 1 in 1,000,000 variant alleles — far exceeding standard NGS panels. Examples include Signatera (Natera) and Personalis NeXT Personal.
Current Clinical Applications
1. Companion Diagnostic for Targeted Therapy Selection
The most established clinical use of ctDNA is identifying actionable mutations when tissue biopsy is insufficient or unavailable. Guardant360 CDx and FoundationOne Liquid CDx are FDA-approved as companion diagnostics for multiple targeted therapies, allowing mutation detection from blood to guide treatment selection. In NSCLC, liquid biopsy detects EGFR mutations, ALK fusions, ROS1 fusions, KRAS G12C, and other driver alterations from plasma with concordance rates of 70–90% versus tissue biopsy in metastatic disease. The NILE trial demonstrated that liquid biopsy-guided treatment selection in metastatic NSCLC was non-inferior to tissue biopsy-guided selection for identifying patients eligible for targeted therapy, establishing liquid biopsy as an acceptable alternative when tissue is unavailable.
2. Resistance Mutation Monitoring
ctDNA has become the standard approach for identifying resistance mutations that arise during targeted therapy. The clearest example is EGFR T790M detection in NSCLC patients progressing on 1st/2nd-generation EGFR TKIs. In the AURA series of trials establishing osimertinib for T790M-positive disease, plasma T790M testing detected approximately 70% of T790M-positive patients identified by tissue biopsy — sufficient sensitivity to guide treatment decisions in patients where re-biopsy is technically challenging or clinically risky. Plasma testing now often precedes re-biopsy in clinical practice, with tissue biopsy reserved for plasma-negative patients given the false-negative rate.
3. Treatment Response Monitoring
Serial ctDNA measurements during treatment can track tumor dynamics — potentially earlier than imaging. Rapid ctDNA clearance after treatment initiation (within 2–4 weeks) has been associated with better outcomes in multiple tumor types including NSCLC, breast cancer, and lymphoma. Conversely, rising ctDNA during treatment — sometimes detectable weeks to months before radiographic progression — may serve as an early indicator of emerging resistance, potentially allowing treatment modification before clinical progression. This application is currently investigational and not yet widely integrated into clinical decision-making, but large prospective studies are testing ctDNA-guided treatment changes versus standard-of-care imaging.
4. Minimal Residual Disease Detection After Curative Treatment
Perhaps the most transformative potential application is postoperative MRD monitoring in early-stage cancers. Multiple studies have shown that ctDNA detection after curative-intent surgery predicts relapse with high specificity — patients who have detectable ctDNA post-surgery relapse at much higher rates than ctDNA-negative patients, often months to over a year before imaging detects recurrence. In colorectal cancer, the DYNAMIC trial demonstrated that ctDNA-guided decision-making for adjuvant chemotherapy (giving chemo only to ctDNA-positive patients, sparing ctDNA-negative patients) was non-inferior in terms of recurrence-free survival and reduced chemotherapy use by ~40% — a landmark result supporting ctDNA-guided de-escalation.
Liquid Biopsy vs. Tissue Biopsy: Complementary, Not Competitive
A common misconception is that liquid biopsy will replace tissue biopsy. In practice, the two approaches are complementary. Tissue biopsy provides information unavailable from blood: histologic characterization of the tumor (e.g., distinguishing adenocarcinoma from squamous cell carcinoma), PD-L1 protein expression by immunohistochemistry, tumor microenvironment assessment, and higher sensitivity for detecting low-frequency alterations in low-tumor-burden settings. Tissue biopsy remains the gold standard for initial diagnosis and tumor characterization.
Liquid biopsy offers advantages in situations where tissue biopsy is impractical: repeat sampling over time for resistance monitoring, profiling of metastatic disease that captures genomic heterogeneity across multiple sites (a single tissue biopsy captures only one site), assessment in patients with inaccessible lesions, and lower procedural risk. The two modalities increasingly function in an integrated workflow — initial tissue diagnosis, liquid biopsy for serial monitoring and resistance detection, repeat tissue biopsy when liquid biopsy is negative and clinical suspicion for progression is high.
Multi-Cancer Early Detection: The Frontier
Multi-cancer early detection (MCED) tests aim to screen for multiple cancer types simultaneously from a single blood draw, using methylation patterns, fragmentomics, and other epigenetic signatures in cfDNA to identify cancer signals and predict tumor origin. The most advanced clinical program is Galleri (GRAIL), which uses methylation-based machine learning to detect signals from 50+ cancer types in asymptomatic individuals. The NHS-Galleri trial in the United Kingdom — enrolling 140,000 participants — is the largest prospective MCED trial and will provide critical data on performance in real-world screening populations.
Early validation data show promising sensitivity for stage II–IV cancers (approximately 65–75% sensitivity with high specificity of >99%) but lower sensitivity for stage I disease (~20–40%), where ctDNA shedding is minimal. The clinical utility of MCED — whether earlier detection translates to improved outcomes, how to manage positive results without obvious tumor, and cost-effectiveness — remains to be established from ongoing prospective trials.
Key takeaway: Liquid biopsy has moved from research tool to clinical standard in less than a decade, and its applications continue to expand. For patients with advanced cancer, plasma ctDNA testing is now a recommended first step for identifying actionable mutations when tissue is unavailable, and a standard approach for detecting resistance mutations. For early-stage patients post-surgery, MRD monitoring is increasingly used to guide adjuvant therapy decisions. Multi-cancer early detection represents the most ambitious application — potentially screening for dozens of cancer types simultaneously — though its clinical validation is still in progress.