Alpha-Fetoprotein (AFP) Tumor Marker
Normal Range and Diagnostic Thresholds
AFP normal range is <8-12 ng/mL, with age-adjusted upper limits of 9 ng/mL for patients <40 years and 13 ng/mL for patients ≥40 years. 1, 2
Key Diagnostic Cut-offs:
- 20 ng/mL: Sensitivity 41-70%, specificity 80-94% for HCC—misses 30-59% of cases 1
- 200 ng/mL: Sensitivity drops to 22%, but specificity approaches 97-100%—highly reliable when elevated but misses 78% of cases 1, 3
- 400 ng/mL: Highly specific for HCC (approaching 100%), but only 18% of HCC patients reach this level 3, 2
Primary Clinical Applications
Hepatocellular Carcinoma (HCC)
AFP should NEVER be used alone for HCC screening or diagnosis due to poor sensitivity—always combine with imaging. 1, 3
Critical Limitations:
- 46% of HCC patients have completely normal AFP (<20 ng/mL), even with large tumors 1, 3
- Only 10-20% of early-stage HCC presents with abnormal AFP 1, 3
- Two-thirds of HCCs <4 cm have AFP <200 ng/mL 3
Surveillance Strategy (High-Risk Populations):
The AASLD recommends ultrasound combined with AFP measurement every 6 months for surveillance in cirrhotic patients and chronic hepatitis B carriers. 1, 3
Target populations requiring surveillance: 1
- All cirrhotic patients (any etiology: HBV, HCV, NASH, alcohol, PBC, hemochromatosis, alpha-1 antitrypsin deficiency)
- Chronic HBV carriers: Asian men >40 years, Asian women >50 years, African/North American blacks, family history of HCC
- All HBsAg-positive patients regardless of age, ALT, or disease severity
Combining ultrasound with AFP increases early-stage HCC detection from 45% to 63% compared to ultrasound alone. 1, 3
Diagnostic Algorithm for Detected Nodules:
For nodules ≥1 cm on surveillance ultrasound: Proceed to dynamic contrast-enhanced CT or MRI 1
Diagnosis without biopsy is possible when: 1, 3
- AFP ≥200 ng/mL AND
- Typical imaging features (arterial hyperenhancement with portal venous/delayed washout) AND
- Patient has cirrhosis or chronic liver disease
AFP trends are critical: Rising AFP in a step-like manner strongly suggests HCC, even if absolute values remain <200 ng/mL 1, 3
Progressive AFP elevation ≥7 ng/mL/month has 71.4% sensitivity and 100% specificity for HCC 1
Performance in HBsAg-Positive Patients:
AFP-based surveillance demonstrated 37% reduction in HCC-related mortality in randomized trials of HBsAg-positive patients. 1, 3
At 20 ng/mL cutoff in HBV carriers: 1, 3
- Sensitivity: 50-75%
- Specificity: >90%
- Negative predictive value: >99%
- Positive predictive value: Only 9-30%
Germ Cell Tumors (GCTs)
In young adults with elevated AFP, immediately evaluate for germ cell tumors before considering other diagnoses. 2
Diagnostic Workup:
- Measure serum β-hCG and LDH concurrently 2
- Obtain testicular ultrasound with 7.5 MHz transducer (even if examination normal) 2
- Perform CT chest/abdomen/pelvis for retroperitoneal or mediastinal masses 2
Key Clinical Points:
- Pure seminoma NEVER produces AFP—if elevated, assume undetected nonseminomatous elements 2
- AFP elevation occurs in 10-20% of stage I nonseminomatous GCT and 40-60% of advanced disease 2
- AFP isoforms in NSGCT are characterized by a-sialo (non-sialylated) species 4
Surveillance Post-Treatment:
- Measure AFP, β-hCG, LDH before each treatment cycle 2
- Post-treatment: every 2-4 months (year 1), every 3-4 months (year 2), annually for ≥10 years 2
Interpretation Pitfalls and False Positives
Benign Causes of AFP Elevation:
Benign liver disease (hepatitis, cirrhosis, hepatic toxicity) is the most common non-malignant cause of elevated AFP in adults. 2, 5
- Cirrhosis of any etiology: Fluctuating AFP reflects flares of liver disease, not necessarily cancer 3
- Chronic HBV/HCV infections: Can elevate AFP without malignancy 2, 5
- Pregnancy: Physiologic elevation, particularly 15-20 weeks gestation 3, 2
- Other malignancies: Intrahepatic cholangiocarcinoma, colon cancer metastases, lymphoma 3
Critical Diagnostic Algorithm for Elevated AFP:
For ALL adults with elevated AFP: 2
- Measure β-hCG and LDH (evaluate for GCT)
- Check liver function tests (ALT, AST, bilirubin, alkaline phosphatase)
- Obtain hepatitis panel (HBsAg, anti-HBs, HBcAb, HCV antibodies)
- Perform testicular ultrasound in young adults (especially males)
- Obtain multiphasic contrast-enhanced CT or MRI of abdomen
- Consider CT chest/abdomen/pelvis in young adults for extragonadal GCT
AFP Isoforms and Glycoforms (Advanced Diagnostics)
Disease-specific AFP isoforms can be identified by isoelectric focusing, with different glycosylation patterns distinguishing HCC from other conditions. 4, 6
- HCC-associated isoforms (Band +II): Mono-sialylated glycoproteins 4
- Benign liver disease (Band +I): Di-sialylated species 4
- NSGCT (Band +III): A-sialo (non-sialylated) species 4
- LCA-reactive AFP: Specific marker for HCC 6
- Con A non-reactive AFP: Marker for yolk-sac tumors and GI malignancies 6
Complementary Biomarkers
PIVKA-II (Des-gamma-carboxy Prothrombin/DCP):
EASL acknowledges PIVKA-II as complementary to AFP, but it lacks sufficient validation for routine use in most Western countries. 3
- Specificity: 94.8-100%—highly reliable when elevated 3
- Optimal cut-offs: 40 mAU/mL for screening, 250 mAU/mL for diagnosis 3
- Limitations: May preferentially detect portal vein invasion (later-stage disease), false positives with vitamin K deficiency or warfarin use 3
- AASLD and Western guidelines do not recommend routine clinical use 3
Combined Markers:
Combining AFP with novel biomarkers (AFP-L3, GP73, PIVKA-II) significantly improves diagnostic accuracy for HCC. 5
Common Pitfalls to Avoid
- Never rely on AFP alone—up to 46% of HCC patients have normal AFP 1, 3
- Don't dismiss rising AFP trends—progressive elevation is virtually diagnostic even below thresholds 1, 3
- Don't ignore normal AFP in high-risk patients—continue imaging-based surveillance 1, 3
- Always evaluate young adults for GCT first when AFP is elevated 2
- Imaging findings are more definitive than AFP levels alone for HCC diagnosis 2
- Western guidelines (AASLD, EASL) do not recommend AFP alone for surveillance—the 6-8% improvement in detection doesn't offset 80% cost increase and false positives 3