Home Liquid Biopsy and ctDNA CancerSEEK Test: Multi-Cancer Blood Test, Tumor Markers, and Early Detection Research

CancerSEEK Test: Multi-Cancer Blood Test, Tumor Markers, and Early Detection Research

By
Vita Library
-
1
CancerSEEK combines tumor DNA mutations and protein markers for multi-cancer detection. Learn what research shows, how results are interpreted, and why follow-up matters.

CancerSEEK is a research multi-cancer blood test developed to detect cancer-related DNA mutations together with circulating protein tumor markers. The original assay was designed to identify signals from eight common cancers and then estimate the most likely organ of origin. Its early results drew attention because it combined several biomarkers in one blood sample rather than screening for a single cancer. However, CancerSEEK should be understood as a research platform and an important step in the development of newer multi-cancer early detection tests, not as a stand-alone diagnosis. A positive result requires imaging and, when a suspicious lesion is found, usually tissue confirmation. A negative result cannot rule out cancer. The assay has also changed over time as the technology moved from academic development into larger prospective studies and commercial development, so performance figures from one CancerSEEK version should not automatically be applied to newer multi-cancer tests.

  • CancerSEEK combines tumor-related DNA mutations with protein biomarkers rather than measuring one tumor marker alone.
  • The original research version targeted eight cancers and reported an overall median sensitivity of about 70% in patients already known to have nonmetastatic cancer.
  • Specificity exceeded 99% in the original case-control study, but real-world screening performance is lower and depends on the tested population.
  • A positive CancerSEEK result is not a cancer diagnosis; follow-up imaging and diagnostic testing are required.
  • CancerSEEK research does not replace mammography, cervical screening, colorectal screening, low-dose lung CT, or other recommended screening.

Table of Contents

What CancerSEEK Is and How It Works

CancerSEEK is a multi-analyte blood test. Instead of asking whether one marker is high, it combines information from several types of cancer-associated signals. The original Johns Hopkins research assay evaluated mutations in circulating cell-free DNA and levels of selected serum proteins, then used an algorithm to estimate both the probability of cancer and the likely tissue of origin.

The DNA portion looks for somatic mutations—acquired changes that can occur in tumor cells. Tumor DNA released into plasma is only a fraction of the total cell-free DNA in blood, so the assay needs highly sensitive methods to detect rare mutant fragments among much more abundant normal DNA.

The protein portion adds a second type of evidence. Cancer can alter the concentrations of circulating proteins, including traditional tumor markers and other cancer-associated proteins. A single protein marker usually lacks enough sensitivity and specificity for broad population screening, but combining several proteins with mutation data can improve classification.

The original CancerSEEK model used mutations across 16 genes and eight protein biomarkers for cancer detection. A separate machine-learning step incorporated additional protein information and sex to help localize the cancer signal. The goal was not to identify every molecular feature of a tumor. It was to answer two screening questions:

  1. Is a cancer-associated signal present?
  2. If so, where is the cancer most likely to be located?

That makes CancerSEEK different from a broad ctDNA mutation panel, which is usually ordered after a cancer diagnosis to find treatment-related genomic alterations.

Cancers and Biomarkers Studied With CancerSEEK

The landmark 2018 CancerSEEK study evaluated people with clinically detected, nonmetastatic cancers of the ovary, liver, stomach, pancreas, esophagus, colorectum, lung, and breast. These eight cancers were chosen partly because several lack effective population screening programs.

The assay did not perform equally across all cancer types. In the original study, median sensitivity across the eight cancers was about 70%, but sensitivity varied substantially by organ and stage. Five cancers—ovary, liver, stomach, pancreas, and esophagus—had reported sensitivities from 69% to 98% in that study. Breast cancer sensitivity was notably lower.

Those figures need context. The study compared patients already known to have cancer with healthy controls. That is useful for establishing proof of concept, but it is easier than screening thousands of asymptomatic people in whom cancer prevalence is low and many tumors, if present, are very small.

CancerSEEK also used familiar protein markers in a new way. The original panel included proteins such as cancer antigen 125, carcinoembryonic antigen, cancer antigen 19-9, prolactin, hepatocyte growth factor, osteopontin, myeloperoxidase, and tissue inhibitor of metalloproteinases 1. These proteins are not specific enough to diagnose cancer individually. Their value in CancerSEEK came from combining them with mutation results and an algorithm.

That distinction is important for interpreting the name “tumor marker.” A tumor marker can contribute to a multi-marker prediction model without becoming a reliable screening test by itself. The algorithm considers the pattern across multiple inputs rather than treating one elevated protein as proof of malignancy.

What CancerSEEK Research Has Shown

CancerSEEK moved through two especially important stages of research: an initial case-control study and a later prospective screening study known as DETECT-A.

The 2018 proof-of-concept study

The first major CancerSEEK report included 1,005 patients with one of eight nonmetastatic cancers and 812 healthy controls. The test was positive in a median of about 70% of the cancer cases, and only seven healthy controls tested positive, corresponding to specificity above 99% in that study.

The assay also attempted to narrow the cancer to a small number of possible organs. Localization is clinically important because a screening blood test that merely says “some cancer may be present” can trigger a broad and potentially burdensome diagnostic search.

The strongest caution is that participants with cancer had already been diagnosed. This design does not reproduce the intended screening setting, where most people do not have cancer and early tumors may shed extremely little material into blood.

The DETECT-A prospective study

DETECT-A tested a CancerSEEK-based strategy in approximately 10,000 women aged 65 to 75 who did not have a known cancer. Positive blood results were confirmed and followed by PET-CT imaging when indicated. The study showed that multi-cancer blood testing could be incorporated into a real diagnostic workflow and could identify cancers that were not found through standard screening.

Blood testing first detected 26 cancers in the prospective study. Some were localized or regional, and a number of participants underwent surgery with curative intent. The study also documented false positives and the downstream testing needed to resolve them.

Longer follow-up has added useful context. A 2024 analysis reported multiyear outcomes for 26 participants whose cancers were first detected by CancerSEEK. After a median follow-up of about 4.4 years, 13 were alive and cancer-free. All eight participants with treated stage I or II disease were alive and cancer-free at the reported follow-up. These are encouraging observations, but they do not prove that CancerSEEK reduces cancer mortality. The cohort was small, and there was no randomized comparison showing fewer cancer deaths because of testing.

Another 2024 analysis followed 98 DETECT-A participants whose positive CancerSEEK result was ultimately classified as false positive after PET-CT and clinical workup. Ninety-five remained cancer-free over a median follow-up of 3.6 years. This provides reassurance about that particular diagnostic pathway, while also showing why a positive screening signal cannot be treated as a diagnosis.

What Positive and Negative Results Mean

CancerSEEK does not have a simple normal range such as “0 to 5 units.” It produces an algorithmic classification based on multiple DNA and protein measurements.

A positive result means the combined signal passed the test’s cancer-detection threshold. It does not tell you that a tumor definitely exists, how large it is, or what stage it is. The result must be interpreted in the context of the version of the assay, the screening population, and the follow-up process.

A negative result means no cancer signal was detected above the threshold. It does not exclude cancer. Early-stage tumors may shed too little DNA, some cancers are less detectable in blood, and the assay only searches for the biological features included in its design.

This is a common principle across blood-based multi-cancer early detection tests: high specificity is valuable, but sensitivity is not uniform across cancer types and stages.

The probability that a positive result represents real cancer is also affected by cancer prevalence. Even when specificity is very high, false positives matter in a screening population because most tested people are healthy. This is why positive predictive value in prospective screening is more useful to patients than a specificity figure alone.

CancerSEEK results should never override symptoms. A person with persistent bleeding, a new lump, unexplained weight loss, jaundice, progressive difficulty swallowing, or another concerning change needs diagnostic evaluation even if a prior screening blood test was negative.

Follow-Up After a Positive CancerSEEK Result

The purpose of follow-up is to convert a molecular screening signal into a confirmed diagnosis—or to establish that no cancer can be found. The DETECT-A study used a structured approach that included confirmatory blood testing and PET-CT, but real-world follow-up can vary by test and clinical situation.

A reasonable diagnostic sequence may include:

  1. Review the molecular result and predicted organ of origin. This helps focus the first round of testing.
  2. Assess symptoms, examination findings, and cancer risk. Clinical information can raise or lower concern for specific organs.
  3. Use targeted imaging or procedures. Depending on the suspected site, this may include CT, MRI, ultrasound, mammography, endoscopy, or another organ-specific test.
  4. Biopsy a suspicious lesion. Pathology remains the standard way to confirm most solid tumors before treatment.
  5. Plan follow-up when the workup is negative. Some cases can be closed after a convincing negative evaluation; others may need further testing because the molecular signal remains concerning.

A positive blood result can also reveal findings that are not invasive cancer. In DETECT-A follow-up, some participants were found to have clinically significant premalignant lesions. Other diagnostic workups may discover benign cysts, nodules, inflammatory conditions, or unrelated abnormalities.

The key practical point is that the blood draw is only the beginning of the pathway. Anyone considering multi-cancer screening should know in advance who will coordinate follow-up, what diagnostic tests are available locally, and how costs will be handled.

CancerSEEK does not require the type of preparation used for a colonoscopy or fasting metabolic test. In research protocols, however, blood collection, sample handling, and confirmatory testing follow strict procedures because cfDNA is present in very small amounts. A screening result should therefore be interpreted from the laboratory report rather than from an isolated protein-marker value obtained elsewhere. Repeating a conventional tumor marker such as CA-125 or CEA on its own does not reproduce the CancerSEEK algorithm.

Timing also changes what the result can mean. A tumor can be below the assay’s detection threshold at one blood draw and become detectable later. Conversely, an abnormal molecular signal may not lead to an immediately visible lesion on imaging. That mismatch is one reason prospective studies need defined follow-up periods rather than classifying every initial negative scan as a permanent false positive.

If a person has a strong hereditary cancer risk, CancerSEEK-type screening should not replace syndrome-specific surveillance. Someone with a pathogenic BRCA1 variant, Lynch syndrome, Li-Fraumeni syndrome, or another inherited predisposition may need earlier or more frequent imaging, endoscopy, or organ-specific screening. Multi-cancer blood testing can be discussed as an additional tool, but its value in each high-risk syndrome must be established separately. The same caution applies to cancer survivors. A new-cancer screening assay and a test designed to detect recurrence answer different questions, so follow-up after prior treatment should use the surveillance strategy recommended for that specific cancer rather than assuming one blood test can replace it.

Limitations and Current Clinical Status

CancerSEEK is best viewed as a major research milestone rather than a fixed, universally available clinical assay. The test described in the 2018 paper was an early version of technology that continued to evolve through Thrive Earlier Detection and later Exact Sciences. Exact Sciences has since developed newer multi-cancer screening products based on an expanded biomarker strategy; performance and indications for those newer products should be evaluated from their own data, not from the original CancerSEEK study alone.

Several limitations remain important:

  • The original sensitivity estimates came from people already known to have cancer. Screening populations are more challenging.
  • Sensitivity differs by cancer type and stage. A single overall number can hide weak performance for a particular cancer.
  • The test cannot diagnose or stage cancer. Imaging and pathology remain necessary.
  • False positives create additional procedures and anxiety. Even a highly specific assay can generate many false alarms at population scale.
  • A negative result can miss cancer. Low tumor shedding is a major biological limitation of blood-based early detection.
  • Mortality benefit is unproven. Finding more cancers or shifting stage at diagnosis is not automatically the same as preventing cancer deaths.

As of September 2026, no multi-cancer early detection test has FDA authorization as a population cancer-screening test in the United States. Some newer assays are offered as laboratory-developed tests, and the regulatory landscape is changing quickly. CancerSEEK itself is therefore more useful as the name of a foundational research platform than as a universal label for current commercial testing.

CancerSEEK Versus Other Liquid Biopsy Tests

CancerSEEK belongs to the broad liquid-biopsy field, but not every liquid biopsy has the same purpose. Understanding the intended use prevents major interpretation errors.

A standard ctDNA test used in a person with known cancer may look for actionable mutations, track treatment response, or detect molecular residual disease. Because the clinician already knows the tumor type, the assay can focus on a smaller and more specific clinical question.

A multi-cancer screening test has to search for an unknown cancer in an otherwise healthy person. It must maintain extremely high specificity while detecting tiny signals from many different tumor types. That is why screening assays increasingly use methylation, fragmentomics, proteins, or other broad biological patterns rather than relying only on mutation panels.

CancerSEEK also differs from CTC testing. A circulating tumor cell test tries to capture intact tumor cells in blood. CancerSEEK analyzes molecular signals—primarily DNA mutations and proteins—not intact cancer cells.

For patients, the most useful question is not “Which liquid biopsy is newest?” It is “What clinical decision is this test designed to support?” Screening, diagnosis, treatment selection, recurrence monitoring, and prognosis are different tasks. A test validated for one should not automatically be used for another.

The CancerSEEK research program helped establish that combining DNA with proteins could detect multiple cancers through a blood sample and could be integrated with imaging in a prospective screening workflow. Its legacy is therefore less about one permanent test configuration and more about demonstrating a practical path toward multi-cancer detection. The remaining challenge is to show, in large and appropriately controlled studies, that newer versions improve meaningful health outcomes while keeping false positives, unnecessary procedures, overdiagnosis, and cost at acceptable levels.

References

Disclaimer

CancerSEEK is an evolving research platform in multi-cancer early detection, and published performance depends on the assay version and study population. Its results cannot confirm or exclude cancer on their own and should not replace established screening or diagnostic evaluation. Medical decisions should be based on current clinical guidance and an individualized discussion with a qualified health professional.

RELATED ARTICLESMORE FROM AUTHOR