Circulating tumor DNA, or ctDNA, is DNA shed by cancer cells into the bloodstream, where it mixes with the much larger pool of cell-free DNA released by healthy cells throughout the body. It carries the same genetic mutations and molecular signatures as the tumor itself, which means a standard blood draw can, in principle, reveal what is happening inside a cancer without a surgical biopsy. Over the past decade, ctDNA testing has moved from research curiosity to clinical tool, now used to match patients with targeted therapies, track whether treatment is working, and catch cancer recurrence months before imaging does.
How Tumor DNA Ends Up in Your Blood
Every cell in your body sheds DNA fragments when it dies, and those fragments float briefly in the plasma portion of your blood before being cleared by the liver and kidneys. Cancer cells are no exception. When tumor cells die through natural turnover, through immune attack, or through treatment-induced damage, they release their DNA into circulation. Tumors also actively secrete DNA-containing particles. The result is a tiny fraction of tumor-derived fragments mixed into a much larger background of normal cell-free DNA.
That fraction is often vanishingly small. In early-stage cancers, ctDNA can make up less than one percent of total cell-free DNA in the blood. Advanced cancers shed more, but even then the signal can be a needle in a haystack. What distinguishes ctDNA from the surrounding normal fragments are the mutations, methylation patterns, and other molecular marks unique to the tumor. Detection technologies have to be sensitive enough to pick those rare altered fragments out of a sea of normal ones.
ctDNA fragments also tend to be physically shorter than normal cell-free DNA. Research has shown that tumor-derived fragments peak at lengths between roughly 90 and 150 base pairs, while normal cell-free DNA clusters around 167 base pairs, the length of DNA wrapped around a single nucleosome.
Why Fragment Size and Half-Life Matter
The size difference between ctDNA and normal cell-free DNA is not just a curiosity. It has become a practical tool. By enriching for shorter fragments during laboratory processing, labs can boost the proportion of tumor DNA in a sample and improve the chances of detecting low-level mutations. Studies in both animal models and human cancers have confirmed that ctDNA fragments run shorter than their normal counterparts, with tumor-derived pieces typically falling in the 134 to 150 base-pair range versus about 167 for background DNA.1PubMed Central. Fragment Length of Circulating Tumor DNA Researchers have further found that ctDNA is enriched not only in fragments shorter than a single nucleosome but also in those shorter than a di-nucleosome, and that certain DNA-cutting patterns at fragment ends can help identify tumor-derived pieces.2PubMed Central. Enhanced detection of circulating tumor DNA by fragment size analysis
Once in the bloodstream, ctDNA does not linger. One study of lung cancer patients measured a median half-life of about 35 minutes, meaning half the ctDNA molecules are cleared from blood in roughly half an hour.3Clinical Cancer Research. Perioperative Dynamic Changes in Circulating Tumor DNA in Patients with Lung Cancer (DYNAMIC) That rapid turnover is actually a feature, not a bug. Because ctDNA refreshes so quickly, a blood sample provides a near-real-time snapshot of what the tumor is doing right now, rather than a historical record. If a tumor shrinks after treatment, ctDNA levels drop within hours. If it regrows, levels climb. That responsiveness is what makes ctDNA useful for monitoring.
How ctDNA Is Detected
Two broad categories of technology drive most clinical ctDNA testing today. The first is next-generation sequencing (NGS), which reads large panels of genes or even the entire genome from a blood sample. NGS can screen for hundreds of mutations at once, making it useful when clinicians want a broad survey of a tumor’s genetic landscape. The second is droplet digital PCR (ddPCR), which zeroes in on a small number of known mutations with extreme sensitivity. Each approach has trade-offs.
In a head-to-head comparison in non-metastatic rectal cancer, tumor-informed ddPCR detected ctDNA in about 59% of baseline plasma samples, compared to roughly 37% for an NGS-based panel, a statistically significant difference.4PubMed Central. Performance Comparison of Droplet Digital PCR and Next‐Generation Sequencing for Circulating Tumor DNA Detection in Non‐Metastatic Rectal Cancer That advantage in sensitivity makes ddPCR attractive for tracking known mutations over time, particularly when the ctDNA fraction is low. On the other hand, NGS casts a wider net and can discover new mutations that ddPCR would miss entirely. In lung cancer, ddPCR successfully identified about 71% of targetable driver mutations previously found by tissue sequencing, while also picking up a handful of mutations that tissue testing missed.5Translational Oncology. Up-front mutation detection in circulating tumor DNA by droplet digital PCR has added diagnostic value in lung cancer
In clinical practice, the choice between approaches often depends on the question being asked. A patient who needs broad molecular profiling to find a drug target usually gets NGS. A patient being monitored for recurrence after surgery, where the tumor’s mutations are already known, is often better served by a highly sensitive targeted assay like ddPCR.
Tumor-Informed Versus Tumor-Agnostic Assays
Within ctDNA monitoring, another important distinction is whether an assay is “tumor-informed” or “tumor-agnostic.” A tumor-informed test starts by sequencing the patient’s actual tumor tissue to identify mutations unique to that cancer, then designs a custom blood test to look for those specific mutations in plasma. A tumor-agnostic test skips the tumor sequencing step and instead scans the blood for a fixed panel of commonly mutated genes.
Tumor-informed approaches tend to be more sensitive because they know exactly what to look for and can filter out background noise more effectively. In one study of colorectal cancer patients, about 84% had at least one trackable mutation identified through a tumor-informed approach, compared to only 37% using a tumor-agnostic method after filtering out false signals from a phenomenon called clonal hematopoiesis.6PubMed Central. Tumor-informed or tumor-agnostic circulating tumor DNA as a biomarker for risk of recurrence in resected colorectal cancer patients A meta-analysis comparing the two approaches in colorectal cancer found that when blood was drawn at multiple time points (serial sampling), tumor-informed assays had a sensitivity of 88% for detecting recurrence versus 59% for tumor-agnostic assays, with no meaningful difference in false-positive rates.7Cancer Treatment Reviews. Comparative diagnostic accuracy of tumor-informed and tumor-agnostic circulating tumor DNA assays in colorectal cancer
The downside of tumor-informed testing is that it requires tumor tissue upfront, which is not always available, and the custom assay takes time to design. Tumor-agnostic tests can be ordered off the shelf and run immediately, which matters when a patient needs answers fast. Interestingly, when only a single blood draw was taken at a fixed time point rather than serial sampling, the gap between the two approaches narrowed and was no longer statistically significant.
Matching Patients to Targeted Therapies
One of the most established clinical uses of ctDNA is identifying the molecular targets a cancer carries so that the right drug can be chosen. In advanced non-small-cell lung cancer, for instance, guidelines now recommend testing for specific gene mutations like EGFR, ALK, and ROS1 before starting treatment. A liquid biopsy can return those results roughly 27 days faster than traditional tissue biopsy and with comparable accuracy. One large study found that for guideline-recommended biomarkers, liquid biopsy agreed with tissue results roughly 95% to 100% of the time, with no significant difference in survival outcomes between patients whose treatment was guided by blood versus tissue testing.8PubMed. Liquid Biopsy Versus Tissue Biopsy to Determine Front Line Therapy in Metastatic Non-Small Cell Lung Cancer (NSCLC)
That speed advantage is more than a convenience. In aggressive cancers, the weeks spent waiting for tissue biopsy results can mean delayed treatment. Liquid biopsy also avoids the risks and discomfort of an invasive tissue procedure, which is particularly relevant for tumors in hard-to-reach locations. A real-world analysis found that liquid biopsy returned positive findings in about 75% of patients across various tumor types, with an average turnaround of seven working days.9PubMed. Routine Clinical Liquid Biopsy Testing for Solid Tumors Delivers the Promise of Minimally Invasive Detection of Genomic Variants With a Faster Turnaround Time Concordance with tissue genotyping was around 62%, lower than the lung-cancer-specific figures, which reflects the broader mix of tumor types and stages in a general clinical population.
Detecting Leftover Cancer After Surgery
After a surgeon removes what appears to be all visible cancer, the critical question is whether microscopic disease remains. Traditional imaging often cannot detect tiny clusters of residual cancer cells. This is where ctDNA testing for molecular residual disease (MRD) has shown some of its most compelling results.
In a landmark study of stage II colon cancer patients who did not receive chemotherapy after surgery, ctDNA was detected in about 8% of them. Of those with detectable ctDNA, 79% went on to have their cancer recur within a median follow-up of 27 months. Among the patients whose ctDNA was undetectable after surgery, fewer than 10% relapsed. The risk of recurrence was roughly 18 times higher for the ctDNA-positive group.10PubMed Central. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer That enormous difference in risk has practical implications: it could help identify the small subset of early-stage patients who truly need chemotherapy while sparing those who are already cured by surgery alone.
Multiple studies have since reinforced these findings across different cancer types. In colorectal cancer specifically, ctDNA-based MRD detection has been shown to predict relapse, guide decisions about post-surgical chemotherapy, and track molecular changes over time.11PubMed Central. Circulating tumor DNA for MRD detection in colorectal cancer: recent advances and clinical implications Several ongoing clinical trials are now testing whether using ctDNA to guide chemotherapy decisions after surgery can improve survival compared to standard approaches.
Tracking Treatment Response in Real Time
Because ctDNA refreshes in the bloodstream so quickly, measuring its levels over the course of treatment provides a dynamic readout of whether a therapy is working. In patients with advanced solid tumors receiving chemotherapy, any decrease in ctDNA tumor fraction during treatment was associated with significantly longer time to needing a new therapy and better overall survival. Patients who achieved at least a 98% drop in ctDNA levels had the best outcomes. Conversely, patients whose ctDNA initially fell but then rose again had significantly worse results, and that rise preceded clinical progression by a median of about two months.12PubMed Central. Chemotherapy response monitoring with DNA methylation-based ctDNA tumor fraction
Similar patterns have appeared in studies of specific cancers. In patients with advanced esophagogastric cancer receiving immunotherapy, those who responded to treatment showed a dramatic drop in ctDNA levels within about 30 days, with a median ratio of post-treatment to baseline ctDNA of just 0.016. Patients whose cancer was stable or progressing showed no significant change.13Scientific Reports. Monitoring treatment response using an ultra-sensitive ctDNA assay in advanced esophagogastric cancer patients In colorectal cancer patients who had liver metastases removed and then received adjuvant chemotherapy, rising ctDNA levels during chemotherapy predicted recurrence in over 92% of cases, while falling levels predicted recurrence in about 64%.14PubMed Central. Dynamic monitoring of circulating tumor DNA to predict prognosis and efficacy of adjuvant chemotherapy after resection of colorectal liver metastases
Telling Pseudoprogression from Real Progression
Immunotherapy creates a distinctive clinical puzzle. When checkpoint inhibitors activate the immune system against a tumor, the resulting inflammation can make tumors appear larger on imaging scans, even though the cancer is actually responding. This phenomenon, called pseudoprogression, can look indistinguishable from genuine treatment failure on a CT scan. Stopping a working immunotherapy because of misleading imaging would be a serious error, but continuing a failing therapy is equally dangerous.
ctDNA offers a way through this dilemma. In melanoma patients treated with anti-PD-1 antibodies, researchers found that every patient experiencing pseudoprogression had a favorable ctDNA profile: their tumor DNA levels in the blood were either undetectable from the start, dropped to zero during treatment, or fell more than tenfold by week 12. In contrast, patients with true progression showed rising or persistently high ctDNA. The test was 90% sensitive and 100% specific in distinguishing the two scenarios.15JAMA Oncology. Association Between Circulating Tumor DNA and Pseudoprogression in Patients With Metastatic Melanoma Treated With Anti–Programmed Cell Death 1 Antibodies Rapid ctDNA decreases have also been observed in pseudoprogressive lung cancer patients treated with anti-PD-1 therapy, in contrast to dramatic ctDNA increases in truly progressive disease.16PubMed Central. Monitoring of KRAS-mutated ctDNA to discriminate pseudo-progression from true progression during anti-PD-1 treatment of lung adenocarcinoma
Catching Resistance Mutations Before They Spread
Cancers evolve. A targeted therapy that works brilliantly for months can fail when a small subpopulation of tumor cells carrying a resistance mutation expands under the selective pressure of the drug. Detecting that resistance early opens the door to switching therapies before the cancer fully escapes control. ctDNA is well suited to this task because it samples DNA from the entire tumor burden, including metastatic sites that a single tissue biopsy would miss.
In a study of patients on various targeted therapies, newly emerging resistance mutations were detected in the blood of up to 34% of patients, and those resistance mutations were polyclonal (arising from multiple independent subpopulations) in up to 14% of cases.17PubMed Central. Circulating tumor DNA landscape and prognostic impact of acquired resistance to targeted therapies in cancer patients This polyclonality is precisely why liquid biopsy adds value beyond tissue testing: a needle biopsy of one tumor deposit might miss resistance mutations sitting in a different metastasis entirely.
A vivid example comes from a gastrointestinal stromal tumor case. Tissue biopsy of the surgical specimen identified two KIT mutations. But plasma ctDNA analysis found those two plus three additional resistance mutations in different exons that were not detected in the bulk tumor tissue. When researchers went back and sequenced ten spatially separated regions of the surgical specimen, they found two of the three “plasma-only” mutations hiding in distinct parts of the tumor.18Molecular Cancer Therapeutics. Noninvasive Detection of ctDNA Reveals Intratumor Heterogeneity and Is Associated with Tumor Burden in Gastrointestinal Stromal Tumor In other words, the blood test painted a more complete picture of the tumor’s genetic diversity than any single biopsy could.
Tracking tumor evolution through the blood has also proven useful in metastatic breast cancer, where serial ctDNA analysis has allowed researchers to map how different metastatic sites respond differently to the same treatment and to pinpoint which specific mutations drive resistance at which site.19Nature Communications. Multifocal clonal evolution characterized using circulating tumour DNA in a case of metastatic breast cancer
Where False Positives Come From
ctDNA testing is not infallible, and one of the most important sources of error is a biological phenomenon called clonal hematopoiesis. As people age, stem cells in the bone marrow acquire mutations and expand into detectable clones. These mutations show up in cell-free DNA and can mimic tumor-derived mutations, potentially leading to a false-positive cancer signal. In one documented case, a patient under surveillance for esophageal cancer had mutations detected in ctDNA analysis that turned out to originate from clonal hematopoiesis rather than from cancer.20PubMed Central. Liquid biopsy in esophageal cancer: a case report of false-positive circulating tumor DNA detection due to clonal hematopoiesis
Some methylation-based ctDNA tests can also produce false positives when non-cancerous conditions produce overlapping methylation patterns. The good news is that false-positive rates are declining as databases improve and bioinformatics filtering gets better at distinguishing tumor-derived signals from background noise.21PubMed Central. False-Positive Liquid Biopsy Assays Secondary to Overlapping Aberrant Methylation from Non-Cancer Disease States Tumor-informed assays, as discussed earlier, largely sidestep this problem because they look for mutations already confirmed to be in the patient’s tumor rather than scanning for common cancer-associated mutations that might also arise from aging blood cells.
Multi-Cancer Early Detection
Perhaps the most ambitious frontier for ctDNA is screening apparently healthy people for multiple cancers at once. Multi-cancer early detection (MCED) tests analyze cell-free DNA from a blood draw and attempt to identify not only whether cancer is present but what organ it likely originated from. These tests leverage a combination of molecular signals: mutation patterns, methylation marks, and fragment-length profiles.22Trends in Cancer. What Is ctDNA and How Is It Used in Cancer Care?
A test called TriOx, developed by an Oxford University research team and reported in early 2025, uses deep whole-genome sequencing combined with machine learning to detect six cancer types (colorectal, esophageal, pancreatic, kidney, ovarian, and breast) in patients with symptoms. In testing against matched tumor biopsies, it achieved roughly 95% sensitivity and 89% specificity, and maintained strong performance even at ctDNA fractions as low as 0.7%.23PubMed Central. A Review of Circulating Tumor DNA (ctDNA) and the Liquid Biopsy in Cancer Diagnosis, Screening, and Monitoring Treatment Response The promise is tantalizing, particularly for cancers like pancreatic and ovarian that currently lack effective screening programs. But MCED tests remain early in their development, and large-scale trials are still needed to determine whether finding cancers this way actually reduces deaths.
The Cost and Coverage Question
ctDNA tests are not cheap. A comprehensive liquid biopsy panel can cost over a thousand dollars, and repeated testing over the course of treatment adds up. Modeling studies have tried to estimate whether ctDNA-guided treatment decisions justify their price tag. In advanced non-small-cell lung cancer, a simulation found that using ctDNA to guide therapy switching yielded about four extra weeks of survival at an additional cost of roughly $4,000, translating to an incremental cost-effectiveness ratio of about $88,000 per quality-adjusted life year. Under commonly used willingness-to-pay thresholds in the United States, there was about a 66% probability that this would be considered cost-effective.24The Journal of Liquid Biopsy. The potential survival gain and cost-effectiveness of circulating tumor DNA-guided treatment switching in advanced non-small cell lung cancer
For cancer screening, the economics look less favorable at present. A systematic review of health economic evaluations of ctDNA-based cancer screening found that only about 28% of studies concluded ctDNA screening was cost-effective, mostly for colorectal cancer and mostly when the blood test achieved higher screening uptake than conventional alternatives like colonoscopy.25PubMed Central. Health Economic Evaluations of Circulating Tumor DNA Testing for Cancer Screening: Systematic Review The remaining studies found ctDNA-based screening was not cost-effective relative to existing methods. This does not mean screening tests will never pencil out, but it suggests the technology needs to get cheaper, more sensitive, or both before it makes economic sense as a population-wide tool.
Insurance coverage adds another layer of complexity. Although ctDNA testing coverage has expanded in recent years, policies vary widely across private insurers and Medicare, creating inconsistent access that can influence which patients receive targeted therapies and which do not.26PubMed Central. Private Payer and Medicare Coverage Policies for Use of Circulating Tumor DNA Tests in Cancer Diagnostics and Treatment
Practical Considerations That Affect Results
The accuracy of any ctDNA test depends on what happens between the blood draw and the lab. Cell-free DNA is fragile, and white blood cells in a blood sample can burst during storage and flood the plasma with their own DNA, diluting the already-faint ctDNA signal. The type of blood collection tube matters: standard EDTA tubes maintain ctDNA stability for about six hours, while specialized cell-stabilizing tubes like Streck or CellSave tubes extend that window to about 48 hours by preventing white blood cell breakdown.27PubMed. Comparative analysis of circulating tumor DNA stability In K(3)EDTA, Streck, and CellSave blood collection tubes For hospitals that ship blood samples to off-site reference labs, that extra stability can make or break the test’s reliability.
Timing of the blood draw also matters in specific clinical contexts. After surgery, ctDNA can clear within hours, so drawing blood too early after an operation might still show residual signal from the tumor that was just removed. Many MRD protocols specify a waiting period, often a few weeks post-surgery, to allow surgical ctDNA to clear before checking whether any truly residual disease is present. The lung cancer study noted earlier found that patients who still harbored residual disease after surgery had a much slower ctDNA clearance rate than those who were disease-free, with half-lives of over 100 minutes versus about 30 minutes.3Clinical Cancer Research. Perioperative Dynamic Changes in Circulating Tumor DNA in Patients with Lung Cancer (DYNAMIC)
Methylation and the Next Wave of ctDNA Features
Most early ctDNA tests focused on detecting mutations: specific letter changes in the DNA code. But mutations are just one molecular signal. DNA methylation, a chemical modification that controls which genes are turned on or off, is increasingly recognized as a powerful complementary marker. Methylation patterns differ between cancer types and between cancer and normal tissue, and they can be detected in cell-free DNA circulating in the blood. Because aberrant methylation often appears early in cancer development, methylation-based ctDNA markers hold particular promise for early detection and screening, potentially flagging cancers before enough mutations have accumulated to be reliably picked up by mutation-focused panels.28PubMed Central. Detection and relevance of epigenetic markers on ctDNA: recent advances and future outlook
Fragment-length analysis, sometimes called fragmentomics, is another emerging layer. Rather than reading the DNA sequence itself, these approaches examine the sizes, end-motif patterns, and nucleosome-positioning signatures of cell-free DNA fragments. Combined with methylation data and mutation calling, fragmentomics is part of a move toward multimodal ctDNA analysis, where tests layer multiple signals on top of one another to squeeze more information out of the same blood draw. The TriOx test mentioned earlier is one example of this integrated approach, combining genomic and methylation data through machine learning to improve diagnostic accuracy.