What Is Tumor Fraction and Why Is It Important in Cancer?

Tumor fraction is the proportion of cell-free DNA in your bloodstream that comes from a tumor rather than from normal, healthy cells. When cancer is present, tumor cells shed fragments of their DNA into the blood, mixing with the much larger pool of DNA that all cells release as part of normal turnover. Measuring how much of that total pool is tumor-derived gives clinicians a single number that reflects tumor burden, predicts outcomes, and tracks whether treatment is working. The concept is deceptively simple, but the biology behind it and the clinical implications that follow are reshaping how oncologists manage advanced cancers.

How Tumor DNA Ends Up in Your Blood

Every cell in your body releases small fragments of DNA into the bloodstream as it ages and dies. This background material is called cell-free DNA, or cfDNA. Most of it comes from blood-forming cells and is completely normal. When a tumor is present, it adds its own DNA to that pool, and these tumor-specific fragments are called circulating tumor DNA, or ctDNA. Tumor fraction is simply the ratio of ctDNA to total cfDNA.¹

For years, researchers assumed that ctDNA entered the blood mainly through cell death, either the programmed kind or the chaotic breakdown that happens when tumors outgrow their blood supply. That picture has become more complicated. Studies of cultured cancer cells have found that a substantial portion of the DNA they release is actively secreted, not a byproduct of cells dying.² The released DNA appears to be associated with protein complexes, suggesting the process is deliberate rather than accidental.³ These molecules, broadly termed circulating tumor nucleic acids, include both DNA and RNA fragments released through a mix of passive and active mechanisms.⁴

This matters because it means tumor fraction is not a straightforward proxy for how many tumor cells are dying. A small, highly active tumor might shed more DNA than a larger, slower-growing one. Blood supply to the tumor, the type of cancer, and even where the tumor sits in the body all influence how much ctDNA makes it into circulation. Understanding these drivers helps explain why two patients with similar-sized tumors can have very different tumor fractions.

How Tumor Fraction Is Measured

Measuring tumor fraction starts with a standard blood draw. The plasma is separated from the blood cells, and the free-floating DNA fragments are extracted. From there, several laboratory approaches can estimate what share of those fragments came from the tumor.

One widely used method relies on shallow whole-genome sequencing, a technique that reads the genome at low depth but across its full length. A computational tool called ichorCNA then looks for patterns of chromosomal gains and losses, the kinds of large-scale copy-number changes that are common in cancer but absent in normal cells. Because this approach does not require any prior knowledge of the patient’s specific tumor mutations, it can be run from a single blood sample in a cost-effective way.⁵ Studies in liver cancer,⁶ lung cancer,⁷ and other solid tumors have validated this approach.

Another strategy uses targeted sequencing or digital droplet PCR to track specific mutations known to be present in the tumor. These methods measure the variant allele frequency, essentially how often a mutant version of a gene appears relative to the normal version in the blood sample. The two measurement units, variant allele frequency and mutant molecules per milliliter, do not always agree perfectly. Discordance tends to crop up when sequencing depth is too shallow or when the total cfDNA concentration is unusually high.⁸

A newer frontier involves fragmentomics, the study of how cfDNA fragments break apart. Researchers have discovered that the fragmentation patterns around certain genomic sites correlate with whether those sites are methylated, a chemical modification that differs between tumor and normal DNA. One study demonstrated that analyzing these cleavage patterns could distinguish patients with liver cancer from those without, with very high accuracy, and could improve the predictive power of screening for nasopharyngeal cancer.⁹ Further work has shown that epigenomic modifications shape chromatin states, which in turn influence fragmentation patterns, and that these fragment-level signals correlate with tumor DNA fractions.¹⁰ The practical appeal is clear: fragmentomics could potentially infer tumor fraction and even tissue of origin from the physical characteristics of DNA fragments, without needing to know the patient’s mutations in advance.

Higher Tumor Fraction, Worse Prognosis

Across a range of cancers, higher tumor fraction at diagnosis is consistently linked to shorter survival. The signal is strong enough that some researchers argue tumor fraction should be treated as an independent prognostic marker, much like tumor stage or performance status.

In advanced non-small cell lung cancer, a prospective study of 878 patients found that those with a tumor fraction of at least 1% had significantly worse outcomes. Their median progression-free survival was about 13 months and overall survival about 18 months, both substantially shorter than in patients whose tumor fraction fell below that threshold.¹¹ In advanced prostate cancer, the relationship is even more dramatic: patients with baseline tumor fraction above 30% had roughly five times the risk of treatment failure and nearly six times the risk of death compared with patients whose tumor fraction was below 2%.¹²

Studies in urothelial cancer have pointed in the same direction. Patients with high tumor fraction had roughly double the risk of disease progression compared with those with lower levels.¹³ These findings span different cancer types, treatment settings, and measurement techniques, which strengthens the case that tumor fraction captures something biologically fundamental about how aggressive a cancer is and how much disease is present.

Tracking Treatment Response in Real Time

One of the most practical uses of tumor fraction is watching it change over time during treatment. A scan can tell you whether a tumor shrank, but scans are snapshots taken weeks or months apart. Tumor fraction offers a faster read. If treatment is working, the tumor sheds less DNA and the fraction drops; if the cancer is progressing, the fraction climbs.

In a phase 3 trial of sotorasib plus panitumumab for colorectal cancer with a specific KRAS mutation, reductions in ctDNA tracked closely with radiographic tumor shrinkage. Patients who cleared their ctDNA early were more likely to respond, while those who did not achieve clearance were unlikely to benefit. The decline in ctDNA correlated with improved progression-free and overall survival regardless of which treatment arm patients were assigned to.¹⁴

A similar pattern emerged in advanced urothelial cancer treated with a combination of enfortumab vedotin and pembrolizumab. Among patients whose ctDNA cleared, the objective response rate was about 87%. For those whose ctDNA decreased but did not fully clear, it dropped to roughly 56%, and for those whose ctDNA rose, responses were uncommon at around 38%.¹⁵ The speed of this information is the key advantage. Researchers have proposed that frequent ctDNA sampling within a compact window around the start of treatment could offer rapid predictions of longer-term response, potentially within days or weeks rather than months.¹⁶

Across multiple randomized trials and cancer types, the ctMoniTR project found that early decreases in ctDNA were consistently linked to improved outcomes at the individual patient level, though trial-level associations were more modest.¹⁷ This large-scale effort lends confidence that ctDNA dynamics are not just a quirk of one cancer type or one treatment but reflect a general biological signal.

Detecting What’s Left After Surgery

After a tumor is surgically removed, the question every patient asks is whether any cancer cells remain. Traditional imaging often cannot detect microscopic residual disease. This is where ctDNA-based minimal residual disease testing comes in. If tumor-derived DNA fragments are still detectable in the blood after surgery, the patient is at elevated risk of relapse.¹⁸

The clinical promise is substantial. If ctDNA clearance after surgery reliably identifies patients who are cured, those patients might be spared the side effects of unnecessary chemotherapy. Conversely, patients who remain ctDNA-positive could be steered toward more aggressive treatment. Growing evidence supports the prognostic value of ctDNA-based minimal residual disease detection across solid cancers, though the clinical utility for guiding treatment decisions is still being established in prospective trials.¹⁹ The stakes are high: getting this right could mean fewer patients undergoing chemotherapy they do not need, and more patients receiving treatment that prevents a recurrence before it becomes visible on a scan.

Where Tumor Fraction Measurement Gets Tricky

Tumor fraction is a powerful concept, but it comes with significant real-world complications. The first is simply that not all cancers shed much DNA into the blood. Brain tumors are a prime example. The blood-brain barrier limits how much tumor DNA reaches the general circulation, so patients with brain cancers often have very low or undetectable ctDNA yields, making tumor fraction estimates unreliable.²⁰ Early-stage cancers of any type also tend to shed less DNA, which is why most current clinical applications focus on advanced disease.

A second complication is clonal hematopoiesis, a phenomenon in which blood-forming stem cells acquire mutations as people age. These mutations can show up in cfDNA and mimic tumor-derived signals, leading to false-positive results. If a blood test flags a mutation as coming from a tumor when it actually arose from aging blood cells, it could lead to unnecessary treatment changes or misguided clinical decisions.²¹ Modern assays are increasingly designed to filter out these background mutations, but the problem is not fully solved.

The third challenge is the blood sample itself. How blood is collected, stored, and processed before cfDNA extraction all affect the results. Factors like the type of collection tube, the speed and timing of centrifugation, and the DNA extraction method can meaningfully change the cfDNA yield.²² One study found higher cfDNA yields with certain purification kits and specific centrifugation speeds,²³ while another found that varying the speed and timing of a second centrifugation step had little effect.²⁴ The lack of full standardization means that a tumor fraction result from one lab may not be directly comparable to a result from another, especially near the lower limits of detection.

Tumor Fraction and Immunotherapy Biomarkers

Tumor fraction does not just tell you how much tumor DNA is floating around. It also determines whether other blood-based biomarkers are trustworthy. One prominent example involves blood-based tumor mutational burden, a measure of how many mutations a tumor carries. Tumor mutational burden has emerged as a predictor of response to immune checkpoint inhibitors, the drugs that unleash the immune system against cancer. But measuring it from blood rather than tissue only works when there is enough tumor DNA in the sample.

A large Japanese study found that blood-based and tissue-based tumor mutational burden agreed well when the plasma tumor fraction was at least 1%. Below that threshold, the blood-based estimate became unreliable. Among patients with high blood-based tumor mutational burden and a tumor fraction of at least 10% who were treated with checkpoint inhibitors, there was a meaningful survival advantage compared with those who had low mutational burden at the same tumor fraction level.²⁵ The implication is clear: when reporting blood-based biomarkers, labs and clinicians need to check whether the tumor fraction is high enough for the result to mean anything. A “low mutational burden” finding in a sample with barely detectable ctDNA might just mean there was not enough tumor DNA to count mutations accurately.

Tracking How Tumors Evolve

Cancers are not static. Over time, especially under the selective pressure of treatment, tumor cell populations change. Some subpopulations are killed off while resistant clones expand. Tissue biopsies capture a snapshot of one spot in the tumor at one moment. Blood-based ctDNA, by contrast, collects fragments from potentially every tumor site in the body, making it better suited to capture this heterogeneity.

By tracking which mutations appear, disappear, or change in frequency over serial blood draws, researchers can reconstruct how a tumor is evolving. The quantitative levels of different mutations in plasma can reveal whether a resistant subclone is emerging before it causes visible disease progression on a scan. This information can guide decisions about whether to switch treatments or add a second drug to target the new clone.²⁶ Real-world use of this approach is still maturing, but it represents one of the most compelling long-term promises of tumor fraction monitoring: not just detecting cancer but watching it adapt in something close to real time.

Pediatric Cancer and the Evidence Gap

Most of what we know about tumor fraction comes from adult cancers. Childhood cancers are biologically different in ways that matter for ctDNA detection. Many pediatric tumors have fewer mutations than their adult counterparts, which can make mutation-based ctDNA detection harder. The tumor types are different, the treatment protocols are different, and the body’s normal cfDNA background differs in children. While ctDNA has shown value in adults for mutation profiling, risk stratification, and treatment response monitoring, equivalent tests have not been validated for routine use in pediatric oncology.²⁷ Research is ongoing, and mutation-agnostic approaches like fragmentomics may eventually bridge some of this gap, but for now, tumor fraction’s clinical utility is largely an adult-cancer story.

When Imaging and Blood Tests Tell Different Stories

A real-world pan-cancer study found that changes in tumor fraction during immune checkpoint inhibitor treatment could provide response information earlier than imaging alone, supporting the potential of serial monitoring to inform treatment decisions.²⁸ This is particularly relevant with immunotherapy, where tumors sometimes appear to grow on scans before shrinking, a phenomenon called pseudoprogression. A rising tumor on a scan combined with a falling tumor fraction might reassure clinicians that the treatment is actually working and the apparent growth is an immune response flooding the tumor site. Conversely, a shrinking tumor on imaging paired with a rising tumor fraction could signal that disease is progressing in locations the scan does not cover well.

These discordant scenarios are where tumor fraction adds the most value: not replacing scans, but providing a complementary biological readout that fills in the gaps imaging cannot. As assays become more standardized and prospective trials accumulate, the question is shifting from whether tumor fraction is informative to how best to integrate it into routine clinical decision-making alongside the tools oncologists already use.

The Cost Question

Liquid biopsy is sometimes presented as a cheaper, less invasive alternative to tissue biopsy. The “less invasive” part is undeniably true: a blood draw is simpler and safer than a needle biopsy of a lung tumor or liver metastasis. The cost picture is more nuanced. A cost-effectiveness analysis in advanced breast cancer found that adding liquid biopsy to conventional treatment was actually more expensive and yielded only a tiny incremental benefit in quality-adjusted survival.²⁹ That study examined one specific clinical scenario in one country, and the economics shift depending on the cancer type, the treatment landscape, and the assay used. Shallow whole-genome sequencing, for instance, is considerably cheaper than deep targeted panels. As the technology matures and competition increases among test providers, costs are expected to fall. But for now, the assumption that liquid biopsy automatically saves money does not hold universally, and cost remains a real barrier to adoption in many healthcare systems.