I now have comprehensive data from multiple authoritative textbooks. Let me compile the complete answer:
Tumour Markers in Hepatobiliary and Pancreatic Cancer
Tumour markers in this region cover three main cancer types: hepatocellular carcinoma (HCC), cholangiocarcinoma (bile duct cancer), and pancreatic ductal adenocarcinoma (PDAC). Each has distinct markers with varying clinical utility.
1. Hepatocellular Carcinoma (HCC)
Alpha-Fetoprotein (AFP)
AFP is the principal and most widely used marker for HCC. It is a major fetal serum protein synthesised by the yolk sac, fetal hepatocytes, and the fetal GI tract. After birth, serum levels fall to near-zero; reappearance of high levels strongly suggests HCC (or hepatoblastoma).
- Reference range: < 20 ng/mL
- Diagnostic threshold: Values > 400 ng/mL in association with a liver mass are considered diagnostic in most cases. Markedly elevated levels (> 10,000 ng/mL) are essentially diagnostic in the right clinical context.
- Sensitivity: Elevated in > 90% of patients in high-prevalence regions (e.g., China/Southeast Asia). However, AFP may be normal in up to 40% of HCC patients in Western cohorts - sensitivity at the standard 20 ng/mL cut-off is approximately 60%.
- Specificity: AFP is also elevated in non-seminomatous germ cell tumours, pregnancy, regenerating livers in acute liver failure, and benign liver disease (cirrhosis, hepatitis) - so false positives occur.
- Surveillance: Used with ultrasound every 6 months for surveillance in high-risk patients (cirrhosis, chronic hepatitis B/C). A progressively rising AFP is highly suggestive of HCC.
- Prognosis: AFP > 1000 ng/mL predicts worse outcomes with liver transplantation. The HCC-MELD score combines AFP, MELD, and tumour size to predict post-transplant survival.
- Henry's Clinical Diagnosis and Management by Laboratory Methods (AFP section)
- Sleisenger and Fordtran's Gastrointestinal and Liver Disease (AFP/HCC section)
- Miller's Anesthesia, 10th ed., p. 1497-1498
AFP-L3 (Fucosylated AFP)
AFP is structurally heterogeneous due to differences in oligosaccharide side chains. Malignant hepatocytes produce AFP with unusual, highly fucosylated sugar chains not found in non-transformed hepatocytes.
- AFP-L3 (Lens culinaris agglutinin reactive fraction) is the fucosylated subfraction.
- Cut-off: > 10% of total AFP is considered positive.
- Improves specificity over total AFP, particularly when total AFP is in the borderline range (10-200 ng/mL).
- Generally less sensitive than total AFP but more specific for HCC.
- Studies have not confirmed that AFP-L3 has superior sensitivity or specificity for early HCC compared to AFP alone.
- FDA-approved for risk stratification in the US, but not for routine HCC surveillance.
- Sleisenger and Fordtran's, Fucosylated AFP section
Des-Gamma-Carboxyprothrombin (DCP) / PIVKA-II
DCP (also known as Protein Induced by Vitamin K Absence or Antagonist-II, PIVKA-II) is an abnormal prothrombin produced by a defect in post-translational carboxylation in malignant hepatocytes.
- Raised in most patients with HCC.
- High specificity for HCC.
- Shorter half-life than AFP - therefore more useful for monitoring treatment response and recurrence after ablation, resection, or TACE.
- Also associated with worse survival.
- Combined AFP + PIVKA-II gives higher sensitivity and specificity than either alone.
- FDA-approved for risk stratification, not screening.
- Miller's Anesthesia, 10th ed., p. 1497-1498
- Sleisenger and Fordtran's, Des-γ-Carboxy Prothrombin section
- Yamada's Textbook of Gastroenterology, 7th ed.
GALAD Model
A composite diagnostic model incorporating:
- Gender
- Age
- AFP-L3%
- AFP
- DCP (PIVKA-II)
GALAD shows promise in phase II (case-control) biomarker studies and outperforms any single marker, but requires phase III/IV validation before routine clinical use.
- Yamada's Textbook of Gastroenterology, 7th ed., Surveillance section
Emerging HCC Biomarkers
Active research is exploring:
- Circulating tumour DNA mutations
- Differentially methylated DNA regions
- microRNAs (miRNA) and long non-coding RNAs (lncRNA)
- Exosomal biomarkers (tumour cell-derived exosomes show differential biomolecule expression)
- VEGF (elevated serum VEGF correlates with metastasis and shorter survival in HCC)
2. Cholangiocarcinoma (Bile Duct Cancer)
No highly sensitive or specific serum tumour marker exists for cholangiocarcinoma. Diagnosis relies primarily on imaging and histology/cytology.
CA 19-9
- The most commonly elevated marker in cholangiocarcinoma.
- Major limitation: CA 19-9 is falsely elevated by biliary obstruction and cholangitis (the very conditions that cholangiocarcinoma causes), severely limiting its diagnostic accuracy.
- In patients with Primary Sclerosing Cholangitis (PSC), a cut-off of 100 IU/mL has sensitivity of 89% and specificity of 86% for cholangiocarcinoma.
- Elevated perioperative CA 19-9 levels are independent predictors of poor survival in patients with resectable cholangiocarcinoma.
CEA (Carcinoembryonic Antigen)
- Bile CEA levels are reportedly elevated in cholangiocarcinoma but not in benign diseases (other than intrahepatic stones).
Combined CA 19-9 + CEA Index (King's College Index)
- The King's College group developed a combined index incorporating both CEA and CA 19-9.
- This combined approach achieves similar or improved sensitivity for cholangiocarcinoma in PSC compared to CA 19-9 alone.
- Yamada's Textbook of Gastroenterology, 7th ed., Cholangiocarcinoma section
- Current Surgical Therapy, 14th ed.
3. Pancreatic Ductal Adenocarcinoma (PDAC)
CA 19-9 - The Primary Marker
CA 19-9 is the only FDA-approved biomarker for pancreatic ductal carcinoma. It is a carbohydrate antigenic determinant - the sialylated form of the Lewis A (LeA) blood group antigen - expressed on circulating mucins.
Key performance data:
| Parameter | Value |
|---|
| Sensitivity (symptomatic patients) | 75-85% (some sources 79-80%) |
| Specificity (symptomatic patients) | 82-90% |
| Cut-off (standard) | > 37 U/mL |
| High-specificity cut-off | > 100 U/mL (specificity ~98%) |
| Positive predictive value (general population screening) | 0.9% - not useful for screening |
Critical limitations:
- Lewis antigen negativity: ~5-10% of the population lacks the enzyme to produce CA 19-9. These patients will never have detectable levels regardless of tumour burden.
- Biliary obstruction raises CA 19-9 independently of cancer (false positives).
- Elevated in other cancers: biliary (95%), stomach (5%), colon (15%), liver/HCC (7%), lung (13%).
- Least sensitive for small, early-stage tumours - not effective for screening.
Clinical uses of CA 19-9 in PDAC:
- Staging: Preoperative levels < 100 IU make occult metastatic disease unlikely; staging laparoscopy may be omitted. Levels > 130 U/mL predict tumour unresectability (HR 2.7).
- Prognosis: Higher levels correlate with higher tumour stage and poorer outcome. Preoperative values > 50 U/mL are associated with higher recurrence risk.
- Treatment monitoring: Falling CA 19-9 during chemotherapy/radiotherapy is a surrogate for clinical response. A rising level in the absence of biliary obstruction suggests progressive disease.
- Recurrence detection: A post-resection rise in CA 19-9 precedes clinical or imaging evidence of recurrence.
- After curative resection, patients whose CA 19-9 returns to normal survive longer than those whose levels remain elevated.
- Sabiston Textbook of Surgery, CA 19-9 section, p. 1221-1222
- Henry's Clinical Diagnosis and Management by Laboratory Methods
- Mulholland and Greenfield's Surgery, 7th ed.
CEA (Carcinoembryonic Antigen)
- Elevated in only 40-45% of pancreatic carcinoma patients - inferior to CA 19-9.
- Used as a secondary/adjunct marker.
- Combined CA 19-9 + CEA + osteopontin (OPN) + hepatocyte growth factor (HGF) + KRAS DNA probes: a multi-marker approach has shown improved sensitivity for early-stage pancreatic cancer detection.
CA 19-5 and CA 50
- CA 50 lacks a fucose residue compared to CA 19-9 and can be produced in Lewis-negative individuals (complementing CA 19-9 in those patients).
- CA 19-5 reacts with both Le^a and sialyl-Le^a epitopes.
- CA 19-9 and CA 50 are complementary - their simultaneous use may improve sensitivity in detecting pancreatic and other carcinomas.
Emerging Pancreatic Cancer Markers
- GPC-1 (Glypican-1): Exosomal glypican-1 has shown diagnostic promise - a 2024 meta-analysis (PMID: 38470244) assessed circulating exosomal GPC-1 as a diagnostic and prognostic marker in pancreatic cancer.
- Exosomal markers: A 2025 systematic review (PMID: 39953585) evaluated the diagnostic accuracy of exosomes in pancreatic cancer.
- sialylated tumour-related antigen (sTRA): A newly discovered glycan produced by a different subset of pancreatic adenocarcinomas than those that produce CA 19-9 - potentially complementary.
- KRAS oncogene: Expressed in > 90% of pancreatic cancers; DNA probes for mutant KRAS in serum are being studied for early detection.
Summary Table
| Cancer | Marker | Primary Use | Key Limitation |
|---|
| HCC | AFP | Surveillance, diagnosis, prognosis | Low sensitivity for early HCC; elevated in benign liver disease |
| HCC | AFP-L3 | Risk stratification, improved specificity | Less sensitive; not validated for early HCC alone |
| HCC | DCP (PIVKA-II) | Diagnosis, treatment monitoring, recurrence | Also elevated in vitamin K deficiency |
| HCC | GALAD score | Investigational surveillance | Requires phase III/IV validation |
| Cholangiocarcinoma | CA 19-9 | Adjunct to imaging | Elevated by biliary obstruction/cholangitis (false positives) |
| Cholangiocarcinoma | CEA + CA 19-9 index | PSC surveillance | Limited data; no proven screening efficacy |
| PDAC | CA 19-9 | Staging, monitoring, recurrence detection | Not a screening tool; falsely elevated in biliary disease; 10% patients Lewis-negative |
| PDAC | CEA | Adjunct | Only elevated in 40-45% of cases |
| PDAC | CA 50 / CA 19-5 | Complementary to CA 19-9 | Limited clinical adoption |
Key principle: No single marker in this region is sufficiently sensitive or specific for population-level screening. Their greatest clinical value lies in monitoring treatment response, detecting recurrence, and prognostication rather than primary diagnosis.