Circulating tumor DNA (ctDNA) is fragments of DNA shed by cancer cells into the bloodstream. It’s a subset of a broader category called cell-free DNA (cfDNA), which is DNA released by all types of cells, both healthy and cancerous. The tumor-derived portion can make up anywhere from less than 0.1% to over 90% of the total cell-free DNA in a person’s blood, depending on the type, stage, and location of the cancer. Detecting and analyzing these tiny fragments is the basis of what’s commonly called a “liquid biopsy,” a blood draw that can reveal information about a tumor without surgery or a needle biopsy.
How Tumor DNA Ends Up in Your Blood
Cancer cells release DNA into the bloodstream through three main routes. The first is apoptosis, the body’s normal process of clearing out damaged or unwanted cells. When a cell undergoes apoptosis, its DNA is neatly chopped into small, predictable fragments. The second is necrosis, a messier form of cell death that happens when tumor cells outgrow their blood supply or are damaged by their harsh microenvironment. Necrosis produces larger, more irregular DNA fragments, sometimes thousands of base pairs long, because the breakdown is chaotic rather than orderly.
The third route is active secretion. Living tumor cells can package DNA into tiny membrane-bound bubbles called extracellular vesicles and release them into surrounding fluid. This process is energy-dependent and regulated, not a byproduct of dying cells. Researchers have observed tumor DNA appearing in cell cultures even when no cell death was detected, confirming that active secretion is a real and distinct mechanism. These vesicles may also play a role in helping tumors spread and resist treatment, though the full picture is still coming into focus.
ctDNA vs. Cell-Free DNA
Every person, healthy or not, has cell-free DNA floating in their blood. Your body constantly replaces old cells, and DNA fragments from that turnover end up in circulation. In someone with cancer, a fraction of that cell-free DNA comes from tumor cells and carries the genetic mutations, methylation patterns, and other alterations specific to the tumor. That fraction is ctDNA.
One physical difference helps researchers distinguish the two: tumor-derived fragments tend to be shorter, roughly 90 to 150 base pairs, while fragments from healthy cells cluster around 166 base pairs, the length of DNA wrapped around a single protein spool called a nucleosome. This size difference has become a useful signal for detection, because filtering for shorter fragments enriches the sample for tumor DNA.
How ctDNA Is Detected
Two main technologies drive ctDNA testing. Digital PCR works by splitting a blood sample into roughly 20,000 individual droplets and running a targeted reaction in each one, looking for a specific known mutation. It can detect mutations present at frequencies as low as 0.01% of the total DNA in a sample, making it extremely sensitive for tracking a single alteration. The tradeoff is that you need to know what you’re looking for ahead of time.
Next-generation sequencing (NGS) takes a broader approach, scanning panels of dozens or hundreds of genes simultaneously. It’s less sensitive than digital PCR for any individual mutation, but it can identify multiple alterations in a single test. This makes NGS better suited for profiling a tumor’s full genetic landscape, especially when the goal is matching a patient to a targeted therapy.
A newer approach called fragmentomics sidesteps mutation detection entirely. Instead of searching for specific genetic changes, it analyzes the patterns in how DNA is fragmented: the size distribution, the sequences at the fragment ends, and the spacing that reflects how DNA was packed inside the cell. Machine learning models trained on these patterns have achieved diagnostic accuracy above 0.9 (on a 0-to-1 scale) in distinguishing cancer patients from healthy individuals across multiple cancer types. This is particularly valuable for cancers that shed very little DNA into the blood, where mutation-based methods struggle.
What ctDNA Testing Is Used For
The most established use right now is guiding treatment decisions. The FDA-approved FoundationOne Liquid CDx test, for example, scans cell-free DNA for specific gene mutations that determine whether a patient is eligible for certain targeted therapies. It’s approved to identify BRCA mutations in ovarian and prostate cancer patients, ALK rearrangements in lung cancer, and PIK3CA mutations in breast cancer, among other indications. For patients whose tumors are hard to biopsy or whose tissue samples are insufficient, a blood draw can provide the same actionable information.
Monitoring treatment response is another major application. Because ctDNA has a short half-life in the bloodstream, its levels respond quickly to changes in tumor activity. If a treatment is working, ctDNA levels drop. If they start rising again, that can signal resistance or progression, sometimes before anything shows up on a scan. In one study, ctDNA detected disease progression nearly a month earlier than imaging (4.6 months versus 5.5 months on average).
Detecting Cancer That Stays Behind After Surgery
One of the most promising applications is detecting minimal residual disease (MRD), the small number of cancer cells that may remain after surgery and eventually cause a relapse. Standard imaging can’t see microscopic clusters of cancer cells, but ctDNA can sometimes reveal their presence through persistent or reappearing mutations in the blood.
The data here is striking. In the landmark TRACERx study of lung cancer patients, more than 99% of patients whose blood showed no detectable ctDNA after treatment remained cancer-free. When ctDNA was detectable, it preceded clinical detection of relapse by a median of 212 days, roughly seven months. In colorectal cancer research, patients with detectable ctDNA at one year after surgery had a dramatically higher risk of distant recurrence, and the median lead time from first detectable ctDNA to clinical recurrence was 18.9 months. That window could allow earlier intervention, potentially when the disease is more treatable.
Where ctDNA Testing Falls Short
The biggest limitation is that not all cancers shed DNA into the blood at detectable levels. Brain tumors are a well-known example. The blood-brain barrier restricts the movement of cell-free DNA from brain tissue into general circulation, making ctDNA levels in glioma patients often too low to detect reliably. Other factors matter too: tumor cell replication rate appears to influence ctDNA concentration more than tumor size alone, which means a slow-growing cancer can be harder to detect than a smaller but more aggressive one.
Concordance with tissue biopsies is also imperfect. When researchers compare the mutations found in a blood sample to those found in a piece of tumor tissue from the same patient, the overlap typically falls in the 40% to 50% range across solid tumors. In a recent ovarian cancer study, only 40.8% of mutations were found in both the tissue and blood samples. Some mutations appeared only in the tissue, and others appeared only in the blood. This doesn’t mean one method is wrong. Tissue biopsies sample a single spot in a tumor, while ctDNA reflects DNA shed from potentially multiple tumor sites throughout the body. Each captures a different, incomplete snapshot.
Early-stage cancers with low tumor burden present another challenge. When there’s very little tumor tissue, the amount of ctDNA in the blood can be vanishingly small, pushing even the most sensitive assays to their limits. This is one reason ctDNA-based cancer screening in otherwise healthy people hasn’t yet become standard practice, despite active research.
What a Liquid Biopsy Looks Like in Practice
From a patient’s perspective, a liquid biopsy is simply a blood draw, typically one or two tubes. The sample goes to a specialized lab where the cell-free DNA is isolated from the plasma (the liquid portion of blood after cells are removed), then analyzed using one of the technologies described above. Results usually come back within one to two weeks, depending on the complexity of the test.
The experience is dramatically different from a tissue biopsy, which can involve sedation, imaging guidance, recovery time, and the risk of complications like bleeding or infection. Liquid biopsies can also be repeated easily over time, making them well suited for serial monitoring. A patient finishing cancer treatment might have blood drawn at regular intervals for months or years, with each test checking whether ctDNA reappears as an early warning of recurrence.