In patients with cirrhosis, hepatocellular carcinoma is best diagnosed non-invasively with multiphasic contrast-enhanced CT or MRI using standardized criteria (LI-RADS). Liver biopsy is reserved for indeterminate or atypical lesions and for lesions arising in a non-cirrhotic liver.^[1][2]

Multiphasic contrast-enhanced CT or MRI

• In a patient with cirrhosis (or other high-risk status, such as chronic HBV infection with advanced fibrosis), the diagnostic study of choice for a focal liver lesion detected on ultrasound surveillance or by an elevated alpha-fetoprotein is a multiphasic contrast-enhanced CT or MRI.^[1][2]
• The study must include a late arterial phase, a portal venous phase, and a delayed phase. MRI may use extracellular or hepatobiliary contrast agents.
• The imaging diagnosis of hepatocellular carcinoma is based on the characteristic vascular behavior of the tumor: non-rim arterial phase hyperenhancement (APHE) followed by washout on the portal venous and/or delayed phases.^[1][2]
• Imaging findings are reported using the American College of Radiology Liver Imaging Reporting and Data System (LI-RADS), which assigns each observation to a category (LR-1 to LR-5, LR-M, LR-TIV) with a standardized management recommendation.^[3]
• LI-RADS applies to adults with cirrhosis, chronic HBV infection, or current or prior HCC. It does not apply to children, to congenital hepatic fibrosis, or to cirrhosis due to vascular disorders (for example Budd-Chiari syndrome or cardiac cirrhosis).^[3]
• Non-invasive diagnosis without biopsy is accepted only in the at-risk population described above. In a non-cirrhotic liver, imaging criteria are not sufficiently validated, and histologic confirmation is generally required.^[1][2]

Contrast-enhanced ultrasound (CEUS)

• Contrast-enhanced ultrasound is an accepted complementary technique for characterizing selected nodules, for example when CT or MRI is contraindicated or inconclusive.
• CEUS cannot stage the whole liver or detect extrahepatic disease, and it may misclassify some intrahepatic cholangiocarcinomas as HCC.^[1]

Liver biopsy

• Liver biopsy is not the first-line test for HCC in cirrhosis. It is indicated when:^[1][2]
  • The lesion is atypical or indeterminate on imaging (for example LR-4 or LR-M observations after multidisciplinary review).
  • The lesion arises in a non-cirrhotic liver.
  • A tissue diagnosis would change management, for example differentiating HCC from intrahepatic cholangiocarcinoma, combined hepatocellular-cholangiocarcinoma, or metastasis to the liver.
  • Molecular or histologic information is needed in a clinical trial or systemic therapy context.
• The decision to biopsy is made by a multidisciplinary team. In patients who are potential candidates for resection or liver transplantation, the possibility of needle-tract seeding and the usefulness of the result should be weighed.^[1]
• A negative or non-diagnostic biopsy does not exclude HCC because of sampling error. Management is by repeat imaging, repeat biopsy, or multidisciplinary review.^[2]
• Techniques for obtaining tissue include:
  • Percutaneous ultrasound-guided or CT-guided core needle biopsy of the target lesion (the usual approach)
  • Transjugular approach, used when there is significant coagulopathy or ascites; it is generally better suited to sampling the background parenchyma than a focal lesion
  • Plugged percutaneous biopsy (embolization of the tract), used in moderate coagulopathy
  • Laparoscopic biopsy, used in selected cases
  • Endoscopic ultrasound (EUS)-guided biopsy, an uncommon option for selected lesions
• Risks of biopsy include:
  • Hemorrhage and hematoma
  • Pain
  • Biliary leak and peritonitis
  • Injury to adjacent organs
  • Tumor seeding along the needle tract (low but not negligible)
• Cirrhosis and portal hypertension increase bleeding risk. Platelet count, INR, and ascites are assessed before the procedure according to institutional protocol.

Immunohistochemistry and histopathology

• Features favoring HCC over a dysplastic nodule include loss of reticulin, CD34 capillarization (diffuse sinusoidal staining), stromal invasion, and a trabecular or pseudoglandular growth pattern.^[4]
• In well-differentiated hepatocellular neoplasms, positivity for at least two of three markers supports malignancy: Glypican 3, heat shock protein 70, and glutamine synthetase. This panel is intended for distinguishing early HCC from high-grade dysplastic nodules and is not a general diagnostic criterion for all HCC.^[4]
• Hepatocellular lineage markers (for example arginase-1 and HepPar-1) help distinguish HCC from metastasis. Cytokeratin 7 and cytokeratin 19 help identify cholangiocarcinoma and combined hepatocellular-cholangiocarcinoma.
• Histologic subtypes recognized by the WHO classification include steatohepatitic, macrotrabecular-massive, scirrhous, chromophobe, fibrolamellar, neutrophil-rich, and lymphocyte-rich HCC.^[5]

Alpha-fetoprotein and other serum markers

• Alpha-fetoprotein (AFP) is not sufficiently sensitive or specific to diagnose HCC by itself. It is used as a surveillance adjunct (a threshold of 20 ng/mL is commonly used) and as a prognostic and treatment-selection marker (for example, AFP of 400 ng/mL or more for ramucirumab eligibility).^[2][1]
• AFP-L3, des-gamma-carboxy prothrombin (DCP), and combined scores such as GALAD are under evaluation and are not used as stand-alone diagnostic tests in the guidelines cited here.

^[3][2] Recommendations vary slightly between guidelines and with the individual clinical context.

1. Surveillance: ultrasound every 6 months, with or without alpha-fetoprotein, in patients with cirrhosis and in selected patients with chronic HBV infection or advanced fibrosis. Abbreviated MRI or multiphasic CT is an option when ultrasound visualization is poor or the patient is at particularly high risk.^[2][1]
2. Abnormal surveillance result (nodule 1 cm or larger, new mass, or elevated AFP): obtain multiphasic contrast-enhanced CT or MRI.
   1. A nodule smaller than 1 cm on ultrasound is usually followed with repeat ultrasound at short intervals (for example 3-6 months), per local protocol.^[2]
3. Assign a LI-RADS category and follow the action in the table above.
4. Biopsy only for indeterminate or atypical observations, for non-cirrhotic liver, or when histology would change management.
5. Staging: complete tumor, liver-function, and performance status evaluation (see below), including chest and abdominal imaging for extrahepatic disease.^[1]
6. Multidisciplinary tumor board review before treatment decisions.^[1][2]

Non-invasive (imaging) criteria

• In an at-risk patient (see above), hepatocellular carcinoma may be diagnosed without biopsy if a multiphasic CT or MRI shows a nodule of 1 cm or larger with non-rim arterial phase hyperenhancement and washout appearance (in the portal venous phase, with or without the delayed phase). This corresponds to LI-RADS LR-5.^[3][2][1]
• Under the LI-RADS size-based rules:^[3]
  • For an observation of 10-19 mm, non-rim APHE with non-peripheral washout (or threshold growth) qualifies as LR-5; APHE with an enhancing capsule alone does not.
  • For an observation of 20 mm or larger, non-rim APHE plus one additional major feature (non-peripheral washout, enhancing capsule, or threshold growth) qualifies as LR-5.
• With hepatobiliary-agent MRI, washout must be assessed in the portal venous phase. Hypointensity in the hepatobiliary phase alone does not count as washout.^[1]

Histologic criteria

• A biopsy or resection specimen showing hepatocellular malignancy, with architectural and cytologic features of HCC and supportive immunohistochemistry, establishes the diagnosis (see above).^[4][5]

Staging in HCC integrates tumor burden, liver function, and patient fitness. The Barcelona Clinic Liver Cancer (BCLC) system is the one used in both AASLD and EASL guidance to link stage with treatment, whereas the AJCC TNM system describes anatomic extent and is mainly used for resected and transplanted patients.^[6][1][2]

Barcelona Clinic Liver Cancer (BCLC) 2022

• The 2022 update also formalizes treatment stage migration (moving to the next appropriate treatment option when the first-line option is unsuitable or fails) and incorporates the patient‘s individual risk-benefit profile.^[6]

Assessment of liver function

• Child-Pugh score uses bilirubin, albumin, INR, ascites, and encephalopathy.
• ALBI score is calculated from serum bilirubin and albumin only: (log₁₀ bilirubin in µmol/L × 0.66) + (albumin in g/L × −0.085). Grade 1 is ≤ −2.60, grade 2 is > −2.60 to ≤ −1.39, and grade 3 is > −1.39.^[7]
• MELD is used for transplant allocation. Portal hypertension is assessed by platelet count, varices, splenomegaly, and (where available) hepatic venous pressure gradient.^[1]

AJCC TNM staging (8th edition)

^[8]

^[8]

Historical staging systems (CLIP and Okuda)

• The Cancer of the Liver Italian Program (CLIP) and Okuda systems predate current treatment algorithms. They are of historical and prognostic interest and are not used by the guidelines cited here to allocate treatment.^[1][6]
• CLIP score (0-6 points): tumor morphology (uninodular and 50% or less of liver = 0; multinodular and 50% or less = 1; massive or more than 50% = 2); Child-Pugh class (A = 0, B = 1, C = 2); AFP (below 400 ng/mL = 0; 400 ng/mL or more = 1); portal vein thrombosis (absent = 0; present = 1).^[9]
• Okuda stage: four variables are scored (tumor involving more than 50% of the liver, ascites, albumin below 3 g/dL, bilirubin above 3 mg/dL). Stage I = none present, stage II = 1 or 2 present, stage III = 3 or 4 present.^[10]