CA 19-9 and CEA
Support monitoring and diagnostic workup. They are not accurate or specific enough for population screening.
Evidence-led detection guide
Blood Test for Pancreatic Cancer
CA 19-9, Avantect and Galleri answer different questions. This guide compares their roles, evidence and limits without treating a blood test as a diagnosis.
Start with the direct answerClinical review status
Pending physician review. A named MD has not yet been supplied; no medical reviewer is represented in the page schema.
The direct answer
Can a Blood Test Detect Pancreatic Cancer?
No single blood test can diagnose pancreatic cancer on its own. CA 19-9 helps monitor known disease but is not recommended for screening. Newer cfDNA tests include pancreatic-specific Avantect and multi-cancer Galleri® (under FDA review). Neither is FDA-approved, and neither replaces imaging, biopsy or guideline surveillance.
Avantect
Looks for a pancreatic cancer signal using cfDNA and other biomarkers in higher-risk populations.
Galleri
Screens for a signal shared by more than 50 cancer types, including pancreas, from a blood draw.
Useful for monitoring, not screening
CA 19-9: What It Shows and Why It Falls Short for Screening
CA 19-9 is a Lewis blood-group antigen that can be shed by pancreatic tumor cells. It is used mainly to monitor treatment response and recurrence—not to screen for cancer. Its roughly 80% sensitivity and 68–91% specificity are too inconsistent for population screening.
The usual reference range is 0–37 U/mL, but interpretation belongs in clinical context. Cholestasis, pancreatitis and other gastrointestinal cancers can raise CA 19-9. NCCN and ASCO guidance does not recommend it as a pancreatic cancer screening test.
About 5–10% of people are Lewis-antigen-negative and cannot produce CA 19-9, even when cancer is present. CEA is sometimes measured alongside CA 19-9 but is less specific. CA 19-9 remains useful for following known disease and supporting an imaging-led workup.
A deep organ, often-silent disease
Why Pancreatic Cancer Is So Hard to Detect Early
The pancreas sits deep behind the stomach, and symptoms often appear after disease has spread. Only about 13% of cases are diagnosed while localized; American Cancer Society data place five-year survival at about 44% for localized disease, 17% for regional disease and 3% for distant disease.
Jaundice, unexplained weight loss, abdominal or back pain and new-onset diabetes can be warning signs, but none is specific to pancreatic cancer. Symptoms call for medical evaluation and diagnostic imaging—not a screening result alone.
Unlike colorectal, breast or eligible high-risk lung screening, there is no routine population screening pathway for average-risk adults. The ACS estimated about 67,530 new US cases and 52,740 deaths in 2026, which is why reliable early-detection research remains urgent.
High-risk surveillance is imaging-led
Who Gets Screened for Pancreatic Cancer Today?
People with certain inherited variants or a strong family history may qualify for surveillance with MRI/MRCP and endoscopic ultrasound (EUS) at an experienced center. Blood tests do not replace this pathway.
- Guidance from CAPS, ASGE and NCCN includes selected carriers of BRCA1, BRCA2, PALB2, ATM, CDKN2A and STK11 variants, hereditary pancreatitis, and familial pancreatic cancer.
- Surveillance often begins around age 45–50 or 10 years before the youngest diagnosis in the family, depending on the variant and guideline.
- Published surveillance cohorts report substantially more Stage I diagnoses and improved survival, but those observational results do not prove the same outcome for every participant.
- Most pancreatic cancers occur in people who do not meet high-risk criteria, leaving a major detection gap.
Inherited risk and current cancer are different questions
Read about BRCA1, BRCA2 and inherited pancreatic cancer risk. A genome does not detect existing pancreatic cancer, and the exact genes included in any service should be confirmed before ordering.
Pancreatic-specific cfDNA screening
Avantect: A Blood Test Designed for Pancreatic Cancer
Avantect from ClearNote Health analyzes cell-free DNA, including 5-hydroxymethylcytosine (5hmC) signatures, alongside genomic and glycan information to look for a pancreatic cancer signal. It is a CLIA-validated laboratory-developed test, not an FDA-approved or cleared diagnostic test.
A peer-reviewed 2023 blinded validation reported roughly 67–68% overall and early-stage sensitivity with 96.9% specificity. A February 2026 ClearNote press release reported 82.6% overall sensitivity, 76.8% for Stage I/II and 97.5% specificity for an enhanced test. Those newer figures are company-reported data on file and are not yet peer-reviewed.
The test is intended for selected higher-risk adults, including people with inherited risk or family history and some adults over 50 with new-onset type 2 diabetes. It is being studied in SAFE-D and PRECEDE. HLI sells Avantect; contact HLI about the Avantect pancreatic cancer test for current eligibility, availability and pricing.
Prospective evidence first
Galleri: Multi-Cancer Detection Including Pancreatic Cancer
Galleri analyzes cfDNA methylation patterns and screens for a cancer signal shared by more than 50 cancer types. Pancreas is in GRAIL's pre-specified group of 12, but no verified pancreas-only prospective sensitivity has been published.
69.8%PATHFINDER 2 sensitivity for the 12-cancer group—not pancreas alone
39.3%PATHFINDER 2 sensitivity across all cancers
99.6%PATHFINDER 2 specificity across all cancers
PATHFINDER 2 full results came from a prospective cohort of 35,878 adults age 50 or older. Positive predictive value was 60.3% and Cancer Signal Origin accuracy was 91.3% across the study. Group figures cannot be applied to pancreatic cancer alone.
In CCGA3, a case-control study of people already known to have cancer, secondary-reported pancreatic sensitivity was 83.7%. Case-control designs can overstate screening performance. The per-stage values below remain pending verification against the primary paper's Figure 3.
| Cancer stage | Sensitivity | Evidence note |
|---|---|---|
| Stage I | 61.9% | Case-control; people already known to have cancer; not a prospective screening rate |
| Stage II | 60.0% | Case-control; people already known to have cancer; not a prospective screening rate |
| Stage III | 85.7% | Case-control; people already known to have cancer; not a prospective screening rate |
| Stage IV | 95.9% | Case-control; people already known to have cancer; not a prospective screening rate |
| Overall | 83.7% | Case-control; people already known to have cancer; not a prospective screening rate |
NHS-Galleri produced mixed aggregate results: the primary endpoint was not met, while Stage IV diagnoses declined by more than 20% in screening rounds two and three. No pancreas-specific breakdown was published. A negative Galleri result does not rule out pancreatic cancer.
Different tools, different jobs
Avantect vs Galleri vs CA 19-9 — How They Compare
CA 19-9 monitors known disease. Avantect asks a pancreatic-specific screening question in higher-risk adults. Galleri asks a multi-cancer screening question. HLI sells Avantect and resells GRAIL's Galleri test; neither is FDA-approved.
| Feature | CA 19-9 | Avantect | Galleri |
|---|---|---|---|
| What it detects | CA 19-9 protein level | Pancreatic cancer cfDNA, including 5hmC signatures | A signal shared by more than 50 cancer types through cfDNA methylation |
| Primary purpose | Monitoring treatment and recurrence; supports diagnostic workup | Pancreatic-specific early-detection screening | Multi-cancer early-detection screening |
| Who it is for | People being evaluated or treated for pancreatic cancer | Higher-risk adults, including some people with inherited risk, family history or new-onset diabetes after 50 | Rx required; adults with elevated cancer risk, such as those 50 or older |
| Pancreatic sensitivity evidence | About 80%; not established for population screening | About 67–68% in a peer-reviewed 2023 validation; 82.6% in a 2026 company release using an enhanced version | 83.7% in CCGA3, a case-control study of people with known cancer; no verified prospective pancreas-only figure |
| Early-stage evidence | Not established for screening | About 68% for Stage I/II in the 2023 validation; 76.8% in the 2026 company release | Secondary-reported CCGA3 case-control figures: 61.9% Stage I and 60.0% Stage II; pending primary-figure verification |
| Specificity evidence | About 68–91%; inadequate for screening | 96.9% in the 2023 validation; 97.5% in the 2026 company release | 99.6% across all cancers in prospective PATHFINDER 2 |
| Lewis antigen dependent? | Yes; 5–10% of people cannot produce CA 19-9 | No | No |
| Regulatory status | Established tumor marker, but not a stand-alone diagnostic test | CLIA-validated laboratory-developed test; not FDA-approved or cleared | CLIA-validated laboratory-developed test under FDA review; not FDA-approved or cleared |
| Through HLI? | No | Yes; contact HLI for current availability and price | $798; not covered by Medicare |
Not a head-to-head comparison. The columns draw from different study designs and populations, including a peer-reviewed validation, a company press release, a case-control study and a prospective study. They cannot establish that one test performs better than another. If you qualify for MRI/EUS surveillance, neither blood test replaces it.
A blood signal is only a starting point
How Pancreatic Cancer Is Actually Diagnosed
A definitive diagnosis requires imaging—often pancreatic-protocol CT, MRI/MRCP or EUS—followed by tissue biopsy. CA 19-9, Avantect and Galleri may raise or lower suspicion, but none can confirm that cancer is present.
EUS can guide fine-needle aspiration (FNA) biopsy. ERCP is used selectively for duct evaluation and treatment, while laparoscopy may help stage disease. Pathology establishes whether a tumor is pancreatic ductal adenocarcinoma or another type.
Molecular testing on tumor tissue can include KRAS, MSI-H/dMMR, BRCA1/2 and NTRK fusions when relevant to treatment planning. These tests are performed after a diagnosis or strong clinical suspicion; they are not population screening tests.
Clear answers
Frequently Asked Questions
Can pancreatic cancer be detected by a blood test?
CA 19-9 can be elevated in pancreatic cancer but is not reliable enough for screening. Newer cfDNA blood tests such as Avantect and Galleri screen for cancer-associated signals, but their figures come from different studies. A positive result requires imaging and biopsy, and a negative result does not rule out pancreatic cancer.
What is the biggest indicator of pancreatic cancer?
Unexplained jaundice is a recognizable warning sign, especially for tumors in the head of the pancreas. Unexplained weight loss, new-onset diabetes after age 50, persistent back or abdominal pain, and changes in stool color also need medical evaluation. These symptoms are not specific to cancer and often appear only after a tumor has grown.
What blood levels are associated with pancreatic cancer?
CA 19-9 above the usual reference range of 37 U/mL can occur with pancreatic cancer, as can elevated bilirubin and liver enzymes when a tumor blocks the bile duct. Glucose may rise with new-onset diabetes. Each finding has non-cancer causes and cannot diagnose pancreatic cancer.
Is new-onset diabetes a sign of pancreatic cancer?
New-onset type 2 diabetes after age 50 can sometimes be an early sign, particularly without typical risk factors. Most new diabetes is not caused by cancer. ClearNote is studying Avantect in this population through SAFE-D; new or unexpected metabolic changes should be discussed with a clinician.
How does the Avantect test differ from the Galleri test?
Avantect is designed specifically for a pancreatic cancer signal and uses 5-hydroxymethylcytosine cfDNA signatures. Galleri screens for a signal shared by more than 50 cancer types using broader cfDNA methylation patterns. Their results come from different studies and populations, so the tests cannot be ranked against each other. HLI sells both.
Predisposition is not detection
Where Whole Genome Sequencing Fits
Whole genome sequencing does not detect existing pancreatic cancer. It may identify inherited risk that changes whether someone qualifies for MRI/EUS surveillance, but a clear report does not rule out future or current cancer.
A clinician or genetic counselor can interpret family history and decide which clinical genetic testing is appropriate. Do not assume a named gene is covered without checking the service's current report scope.
See what the $599 genome service includes. WGS asks about inherited predisposition; Avantect and Galleri ask about a current blood signal; MRI/EUS and biopsy address current disease. None substitutes for the others.
Sources and Evidence Notes
Key sources: Klein et al., Annals of Oncology (2021, CCGA3 case-control study); Song et al., Clinical Gastroenterology and Hepatology (2023, Avantect validation); ClearNote Health enhanced-performance release (February 2026, company-reported data on file); American Cancer Society, Cancer Facts & Figures 2026; NCCN Pancreatic Adenocarcinoma guideline; ASGE high-risk screening guideline; AGA Clinical Practice Update (2020); Schrag et al., Lancet (2023, PATHFINDER); GRAIL PATHFINDER 2 and NHS-Galleri releases (2026); FDA advisory committee materials (September 2026); EGTM tumor marker status report; and Luo et al. on Lewis-antigen-negative pancreatic cancer. Study percentages describe populations, not an individual outcome.