Hepatocellular Carcinoma

• HCC is the 6th most common cancer worldwide and the 3rd leading cause of cancer-related death globally (after lung and colorectal cancer) ^[1][2].
• Globally ~900,000 new cases/year and ~830,000 deaths/year (2020 GLOBOCAN data), reflecting the very high case-fatality ratio.
• There is a striking geographic variation linked to the prevalence of underlying aetiologies:

• The rising incidence in Western countries is driven by the NAFLD/NASH epidemic (obesity, metabolic syndrome) and ageing HCV cohorts ^[2].

This is extremely high-yield for HKUMed exams.

• HCC is exceedingly common in Hong Kong ^[1][4].
• Second commonest cancer death in Hong Kong ^[4] (some older data cites 3rd — the ranking fluctuates between 2nd and 3rd with lung and colorectal cancer; as of recent data, it is the 3rd most common cause of cancer death in HK, but lecture slides from Prof Poon state second ^[4]).
• 5th most common cancer in HK overall by incidence ^[2].
• HCC is 6× more common than cholangiocarcinoma (HCC:CC = 6:1) ^[1].
• HCC constitutes 78.7% of primary malignant liver tumours in HK, CC = 9.7% ^[1].
• M:F = 4:1 ^[4] (some older data quotes 6:1 — the key point is a strong male predominance, partly explained by higher HBV carrier rates in males, androgen-related pathways, and higher rates of alcohol use).
• Most patients age > 50 years, but can occur in young patients ^[4].
• Peak age of mortality = 45–55 years in HK ^[1].
• 80% of HCC in Hong Kong are HBsAg-positive — i.e. Hepatitis B is the overwhelmingly dominant aetiology in HK ^[3][4].
• Frequent association with cirrhosis (80% in Hong Kong) ^[4].

• Insidious onset with rapidly progressing course which is fatal ^[1].
• Median survival of untreated HCC = 2–4 months (for symptomatic, advanced disease) ^[1].
• 5-year survival varies dramatically by stage:
  • Early stage (BCLC 0/A) with curative treatment: 50–70%
  • Intermediate stage (BCLC B): ~20%
  • Advanced/Terminal stage: < 5%

The liver is divided into 8 functionally independent segments (I–VIII), each with its own:

• Portal pedicle (portal vein branch, hepatic artery branch, bile duct)
• Hepatic venous drainage

This segmental anatomy is the basis for anatomical hepatic resection — you can remove segments independently because each is a self-contained unit.

The liver has a unique dual blood supply, which is essential for understanding HCC biology and treatment:

1. Portal vein (~75% of hepatic blood flow): carries nutrient-rich, relatively deoxygenated blood from the GI tract
2. Hepatic artery (~25% of hepatic blood flow): carries oxygenated blood from the coeliac trunk

Why does this matter for HCC?

• Normal hepatocytes derive most of their blood supply from the portal vein.
• HCC tumour nodules derive > 90% of their blood supply from the hepatic artery (neoangiogenesis is predominantly arterial).
• This differential supply is the basis for:
  • Arterial enhancement on CT/MRI (tumour "lights up" in arterial phase → washes out in portal venous/delayed phase — the hallmark imaging feature of HCC)
  • Transarterial chemoembolisation (TACE) — you deliver chemotherapy and embolic material via the hepatic artery, selectively starving the tumour while relatively sparing normal liver parenchyma

Three major hepatic veins (right, middle, left) drain into the inferior vena cava (IVC). This is relevant because:

• HCC has a high propensity for venous invasion (portal and hepatic veins) ^[4]
• Spread through portal vein branches → intrahepatic metastasis (the most common route of intrahepatic spread)
• Spread through hepatic vein branches → IVC → right heart → lungs ^[1]
• Portal vein tumour thrombus (PVTT) is one of the most significant adverse prognostic features

The liver is enveloped by Glisson's capsule (a thin layer of connective tissue derived from the peritoneum). The liver parenchyma itself has no pain fibres — pain only occurs when Glisson's capsule is stretched or invaded by tumour. This explains why:

• Small HCCs are painless (they don't stretch the capsule)
• Large HCCs cause RUQ pain ± referred right shoulder pain (diaphragmatic irritation via phrenic nerve, C3-C5) ^[2]
• Ruptured HCC causes sudden severe abdominal pain and peritonism

• Hilum (porta hepatis): where the portal vein, hepatic artery, and common hepatic duct enter/exit. HCC spread to hilar lymph nodes occurs here ^[1].
• Diaphragm: the liver dome is in close proximity; large HCCs can invade or irritate the diaphragm.

Hepatitis B virus is responsible for ~80% of HCC cases in Hong Kong ^[3][4].

Why does HBV cause HCC? Two mechanisms:

1. Indirect (via cirrhosis): Chronic HBV infection → chronic hepatic inflammation → hepatocyte necrosis and regeneration → progressive fibrosis → cirrhosis → HCC

   • Risk of cirrhosis = 20–30% of chronically infected adults will develop cirrhosis over 10–15 years ^[1]

2. Direct oncogenic effect (unique to HBV, not HCV):

   • HBV DNA integrates into the host hepatocyte genome — this can activate proto-oncogenes or disrupt tumour suppressor genes ^[5]
   • HBV-encoded X antigen (HBxAg), Pre-S1, Pre-S2 proteins are responsible for carcinogenesis ^[1]:
     • HBx protein → transactivates cellular oncogenes, inhibits p53 tumour suppressor function, promotes cell proliferation, inhibits DNA repair, activates NF-κB and MAPK pathways
     • Pre-S mutants → cause ER stress, oxidative DNA damage, and genomic instability
   • Because of this direct oncogenic effect, HCC complicating HBV can present on a non-cirrhotic liver (20% of cases) ^[1]

Risk factors for HCC in HBV carriers:

• Male sex
• Older age (> 40 years)
• Family history of HCC
• High HBV DNA viral load (> 2000 IU/mL)
• HBeAg positivity
• Genotype C (common in HK)
• Co-infection with HCV, HDV, or HIV
• Active inflammation (elevated ALT)
• Presence of cirrhosis
• Aflatoxin exposure (synergistic with HBV)

• Common in Japan and Western countries ^[4] but accounts for only ~4% of HCC in HK ^[1].
• HCC complicating HCV ALWAYS presents on a cirrhotic liver (100%) ^[1] — HCV is an RNA virus that does not integrate into host DNA (unlike HBV). Therefore HCV causes HCC purely through the indirect pathway: chronic inflammation → cirrhosis → HCC.
• HCV causes HCC at a rate of ~1–4% per year in cirrhotic patients.
• With the advent of direct-acting antivirals (DAAs), HCV can now be cured in > 95% of cases, but the risk of HCC persists even after viral eradication (especially if cirrhosis is already established).

• Cirrhosis is present in 70–90% of HCC cases overall; 80% in HK ^[1][4].
• Any cause of cirrhosis can lead to HCC, because the final common pathway is: chronic hepatocyte injury → regenerative hyperplasia → dysplastic nodules → HCC.
• The annual risk of HCC development in a cirrhotic liver is 1–6%/year depending on the aetiology.

• Accounts for ~4% of HCC cases in HK ^[1] but is a major cause in Western countries.
• HCC develops on top of alcoholic cirrhosis (100%) — alcohol does not have a direct oncogenic effect comparable to HBV ^[1].
• Mechanism: Chronic alcohol → acetaldehyde (toxic metabolite) → oxidative stress + lipid peroxidation → hepatocyte injury → inflammation → fibrosis → cirrhosis → HCC.
• High but NOT moderate level of alcohol consumption is the risk factor ^[1].
• Alcohol can also act synergistically with HBV or HCV to accelerate cirrhosis and HCC development.

• Now renamed MASLD (2023 nomenclature), encompassing:
  • Non-alcoholic fatty liver (NAFL) / metabolic dysfunction-associated steatotic liver (MASL)
  • Non-alcoholic steatohepatitis (NASH) / metabolic dysfunction-associated steatohepatitis (MASH)
• The fastest-growing cause of HCC worldwide, driven by the global obesity/metabolic syndrome epidemic.
• NASH-related HCC can occur without cirrhosis in up to 20–30% of cases — similar to HBV. This is thought to be due to the pro-inflammatory, pro-oxidative environment of steatohepatitis driving direct DNA damage.
• Key metabolic risk factors: obesity, type 2 diabetes mellitus, dyslipidaemia, metabolic syndrome ^[1][2].

• Mycotoxins produced by Aspergillus flavus which contaminates corn, soybeans, and peanuts ^[1].
• Leads to mutation of the p53 tumour suppressor gene (specifically the R249S hotspot mutation — an arginine-to-serine substitution at codon 249) ^[1].
• Risk factor in Africa and rural areas of China but NOT Hong Kong ^[1][4].
• Aflatoxin has a synergistic effect with HBV: the combination of HBV + aflatoxin exposure increases HCC risk multiplicatively (up to 60× compared to neither alone).

• Autoimmune hepatitis (AIH): chronic inflammation → cirrhosis → HCC
• Primary biliary cholangitis (PBC): progressive bile duct destruction → biliary cirrhosis → HCC
• Primary sclerosing cholangitis (PSC): biliary stricturing and inflammation → biliary cirrhosis → both cholangiocarcinoma (more common) and HCC

• Hereditary haemochromatosis: excess iron deposition → oxidative stress → hepatocyte injury → cirrhosis → HCC. Iron itself is also a direct mutagen via Fenton chemistry (Fe²⁺ + H₂O₂ → OH• free radicals).
• Wilson's disease: copper overload → oxidative liver damage → cirrhosis → HCC (though interestingly, the risk of HCC in Wilson's disease is lower than expected, possibly because excess copper has some anti-tumour properties).
• α1-antitrypsin deficiency: misfolded protein accumulates in hepatocyte ER → ER stress → chronic liver injury → cirrhosis → HCC.

HCC develops through a well-characterised multistep process, beautifully illustrated on the lecture slide showing the progression from chronic liver disease to HCC ^[5]:

Chronic Liver Injury → Hepatocyte Proliferation → Fibrosis/Cirrhosis →
Hyperplastic Nodule → Dysplastic Nodule (Low → High Grade) →
Well-Differentiated HCC → Moderately Differentiated → Poorly Differentiated HCC

Let's walk through each step:

1. Chronic liver injury (HBV, HCV, alcohol, NASH, etc.) → ongoing hepatocyte death
2. Stellate cell activation → extensive scarring (collagen deposition) → fibrosis → cirrhosis ^[5]
3. Hepatocyte regeneration under proliferative pressure → telomere shortening with each cell division ^[5]
4. Hyperplastic nodules → still polyclonal, no dysplasia
5. Moderate genomic instability → Dysplastic nodules → monoclonal, contain cytological atypia ^[5]
6. Telomerase reactivation → cells escape senescence and gain replicative immortality ^[5]
7. Marked genomic instability → Loss of p53 and other tumour suppressors → HCC ^[5]
8. Progressive dedifferentiation: well-differentiated → moderately differentiated → poorly differentiated

HCC is a vascular tumour with a high propensity for venous invasion (portal and hepatic veins) ^[4].

• Neoangiogenesis: HCC secretes high levels of VEGF, creating a rich network of abnormal arterial vessels → this is why HCC is a hypervascular tumour
• Arterial blood supply: as hepatocytes become dysplastic and then malignant, they progressively lose their portal venous supply and become increasingly dependent on hepatic artery branches via unpaired arteries (arteries not accompanied by portal tracts). This arterialisation is a hallmark of HCC.
• Sinusoidal capillarisation: normal liver sinusoids are fenestrated (with gaps); in HCC, sinusoids become capillarised (endothelium becomes continuous) → contributes to the imaging characteristics.

This is the most widely used staging system for HCC because, unlike TNM, it integrates tumour burden + liver function + performance status and links directly to treatment recommendations.

This is indispensable for HCC management. It uses 5 parameters ("ABCDE" mnemonic: Albumin, Bilirubin, Clotting/PT-INR, Distension/ascites, Encephalopathy):

Developed at Queen Mary Hospital, HKU — tailored for the Hong Kong/Asian HBV-dominant population. It differs from BCLC by being more aggressive with surgical/locoregional treatments in intermediate-advanced stages, reflecting the better outcomes observed in Asian centres with high surgical expertise. It classifies patients based on:

• ECOG performance status
• Child-Pugh classification
• Tumour extent (intra/extrahepatic, vascular invasion)

• The most common route of spread
• Via portal venous circulation — tumour invades a portal vein branch, and tumour cells seed downstream portal territories
• This is why you often see satellite nodules around the main tumour and in other segments

• High propensity for portal vein invasion ^[4]
• Can extend from segmental branches → main portal vein → even into the superior mesenteric vein
• Consequences:
  • Worsens portal hypertension (→ variceal bleeding, ascites)
  • Makes the non-tumorous liver completely dependent on hepatic arterial supply
  • Major caution for TACE, especially main portal-vein occlusion: loss of both inflows risks ischaemic liver failure. Branch involvement requires individual expert assessment

• Spread through hepatic vein branches → IVC → right atrium
• Can form a tumour "tongue" extending into the IVC (similar to renal cell carcinoma)
• Lung metastasis via hepatic vein dissemination ^[4]

• Lung (most common extrahepatic site) — via hepatic veins/systemic circulation
• Bone — often osteolytic, can present with pathological fractures or cord compression
• Lymph nodes (porta hepatis) ^[1] — regional spread
• Peritoneal metastasis ^[4] — especially after rupture
• Adrenal glands ^[1]
• Brain — rare but possible
• Usually death precedes extensive metastasis ^[1] — most patients die of liver failure or variceal bleeding before metastases become clinically dominant

The lecture slide from Prof Poon lists 4 major clinical presentations ^[4]:

1. Subclinical (detected on screening with AFP and USG in HBsAg carriers or cirrhotic patients)
2. RUQ pain, hepatomegaly
3. Decompensation of cirrhosis (ascites, variceal bleeding, hepatic encephalopathy)
4. Intraperitoneal haemorrhage (rupture of HCC)
5. Paraneoplastic features: fever, hypercalcaemia, hypoglycaemia, erythrocytosis

• Ultrasound (USG): sensitivity ~60–80% for early HCC. Limited by operator dependence, body habitus, and cirrhotic liver background.
• Alpha-fetoprotein (AFP): see below.
• CUHK-HCC score: a validated risk prediction model incorporating age, albumin, bilirubin, cirrhosis, HBV DNA level ^[2] — helps stratify which HBV carriers need more intensive screening.

• What it is: A congenital vascular malformation composed of dilated blood-filled vascular spaces lined by endothelium. It is the most common benign liver tumour ^[2][6].
• Why it enters the differential: Both haemangioma and HCC are hypervascular on imaging. Both can present as incidental liver masses. A large haemangioma can cause RUQ pain and hepatomegaly.
• Key distinguishing features:

• What it is: A hyperplastic regenerative nodule that forms in response to a pre-existing arterial malformation (an aberrant "feeding" artery). It is not a true neoplasm — it's a hyperplastic response to abnormal blood flow ^[6].
• Why it enters the differential: Arterially enhancing liver mass in a young/middle-aged patient. Can be mistaken for HCC especially if a background liver disease is present.
• Key distinguishing features:

• What it is: A benign proliferation of hepatocytes (a true neoplasm, unlike FNH). The name "hepat-oma" literally means liver tumour — confusingly, some older texts use "hepatoma" for HCC, but strictly it should refer to this benign lesion ^[6].
• Why it enters the differential: Arterially enhancing solid liver mass, can be large, and importantly can rupture (like HCC) causing haemoperitoneum. Also has malignant potential → can transform into HCC.
• Key distinguishing features:

• What it is: Adenocarcinoma of the bile duct epithelium (cholangio- = bile duct, carcinoma = malignant epithelial tumour). > 90% are adenocarcinoma ^[7][8].
• Intrahepatic cholangiocarcinoma (ICC) accounts for 5–20% of primary liver malignancy ^[7] and is the second most common primary liver cancer after HCC ^[2].
• Why it enters the differential: Intrahepatic cholangiocarcinoma presents as a liver mass, often in a patient with chronic liver disease (HCV, PSC), and can be difficult to distinguish from HCC on imaging.

• Already discussed in the pathology section but worth reiterating as a differential:
• Histologic variant of HCC; most common in younger women, good prognosis ^[4]
• Not associated with HBV or cirrhosis ^[2]
• AFP is typically normal (because the well-differentiated tumour cells don't secrete AFP efficiently)
• Imaging: large mass with central calcified scar (unlike FNH, the scar does NOT enhance on delayed phase)
• If you see a large liver mass in a young patient without cirrhosis and normal AFP → think fibrolamellar HCC

• Epithelioid haemangioendothelioma (EHE): a rare low-to-intermediate grade vascular malignancy
• Presents as multiple liver lesions, often at the periphery with capsular retraction (a fairly characteristic finding)
• Can mimic metastatic disease on imaging
• Distinguished from HCC by: no association with HBV/cirrhosis, normal AFP, IHC positive for vascular markers (CD31, CD34, Factor VIII)
• In children, infantile haemangioendothelioma is a distinct entity (benign vascular tumour of infancy)

• Extremely rare, highly aggressive malignant vascular tumour
• Associated with Thorotrast (thorium dioxide — an old radiological contrast agent), vinyl chloride, arsenic, and anabolic steroids
• Presents with rapidly enlarging liver mass, haemoperitoneum (rupture), consumptive coagulopathy
• Prognosis is dismal

• Rare; liver is more commonly involved secondarily in systemic lymphoma
• Can present as a solitary mass or diffuse infiltration
• Distinguished from HCC by: LDH elevation, normal AFP, imaging may show a "low density" mass, and diagnosis requires biopsy with IHC (CD20+ for B-cell lymphoma)

The mnemonic from the senior notes is helpful: "Some GU Cancers Produce Bumpy Lumps" ^[2]:

• Stomach
• GU: kidney, ovary, uterus
• Colon (most common — accounts for ~1/3 of liver metastases) ^[2]
• Pancreas
• Breast
• Lung

Commonest site is from GI tract (portal venous circulation): colorectal, stomach, pancreas ^[9].

The diagnosis of HCC can be established non-invasively when ALL of the following are met:

1. Appropriate at-risk population: adults with cirrhosis or at-risk chronic HBV (or prior HCC), applying the chosen imaging algorithm. Chronic HCV/MASLD without cirrhosis does not automatically qualify for imaging-only diagnosis
2. Nodule ≥ 1 cm on surveillance imaging
3. Characteristic enhancement pattern on one dynamic contrast-enhanced imaging study (triphasic CT or dynamic MRI):
   • Non-rim arterial phase hyperenhancement (APHE): the nodule "lights up" brighter than the surrounding liver in the arterial phase
   • Non-peripheral "washout": the nodule becomes hypo-attenuating/intense relative to the surrounding liver in the portal venous and/or delayed phase
   • Additional supporting features (for MRI): enhancing "capsule" appearance on delayed phase, restricted diffusion, T2 mild hyperintensity

If both APHE + washout are present → HCC is diagnosed. No biopsy required.

LI-RADS was developed by the ACR (American College of Radiology) to standardise the reporting of liver observations in at-risk patients. It assigns a category from LR-1 to LR-5:

CT/MRI LI-RADS v2018: non-rim APHE is required for LR-5. For a 10–19 mm observation, non-peripheral washout or threshold growth can establish LR-5; capsule alone cannot. At ≥20 mm, one additional major feature (non-peripheral washout, enhancing capsule or threshold growth) suffices. Apply the complete size-dependent table and population exclusions. With gadoxetate, assess washout in the portal venous phase; hepatobiliary hypointensity alone is not washout. ACR core algorithm.

• AFP >400 ng/mL is concerning but is not diagnostic of HCC ^[4] in the right clinical context (HBV carrier, cirrhosis)
• However, normal AFP level in 30% of patients (i.e. < 20 ng/mL) ^[4] — so a normal AFP does not exclude HCC
• AFP alone is insufficient for diagnosis — it must be combined with imaging

Surveillance should be performed at intervals of 6 months (doubling time of HCC ≈ 3–4 months) ^[1][2][10].

Surveillance targets: cirrhosis in patients who could benefit from HCC treatment, and selected noncirrhotic chronic HBV according to age, sex, origin, family history and validated risk. AASLD examples for people from endemic countries are men >40 and women >50; earlier surveillance may be indicated by family history or other risk. Child–Pugh C patients generally benefit only if transplant-eligible. HCV cure does not remove surveillance need in established cirrhosis. AASLD guidance.

Screening method: USG + AFP every 6 months ^[2][4][10]

Rationale for 6-month interval ^[1]:

• Based on expected HCC growth rates (doubling time ≈ 3–4 months ^[2])
• No difference in outcomes compared with 3-month interval surveillance ^[1]
• Better survival compared with 12-month interval surveillance ^[1]
• USG has higher sensitivity than AFP alone ^[1]

The data from the HKU screening study (Chan AC et al., Ann Surg 2008) ^[10] demonstrates the clear advantage of screening vs symptomatic presentation:

This is the power of screening — patients are caught with smaller tumours, better liver function, and less advanced disease, translating into more curative treatment options.

Subcentimetre observations use 3–6-month reassessment, not an automatic six-month wait. Stable findings on two short-interval examinations can return to routine surveillance. AFP ≥20 ng/mL or a concerning rise may prompt CT/MRI even with no ultrasound-visible mass. US LI-RADS 2024 update.

Blood tests: CBP, LFT, alpha fetoprotein (A markedly elevated level is concerning but not diagnostic of HCC, normal AFP level in 30% of patients i.e. < 20 ng/mL), hepatitis B and C serology ^[4].

Biopsy is unnecessary when an observation meets definitive imaging criteria in an eligible at-risk population. It is appropriate for LR-M, selected LR-4/indeterminate lesions, a mass outside the validated imaging-only population, or when tissue changes treatment. Resectability does not itself prohibit biopsy. Discuss route and diagnostic yield with radiology/hepatobiliary specialists; negative biopsy does not always exclude HCC.

Bleeding, seeding and non-diagnostic sampling are relevant risks. INR >1.2 is not a universal contraindication: INR poorly predicts procedural bleeding in cirrhosis. Assess platelet count, anticoagulants, renal failure, infection, lesion location and procedure/route rather than applying that isolated cutoff. AASLD procedural guidance.

Faculty pathology checkpoint (liver pathology deck pp. 13–29): malignant hepatocytes may form trabeculae/pseudoglands and produce bile; stromal invasion helps distinguish early HCC from dysplasia. HepPar-1, arginase-1 and glypican-3 support hepatocellular differentiation as a panel. Cholangiocarcinoma forms glands in desmoplastic stroma and often expresses CK7/CK19. CK7 or CK19 alone cannot identify a metastasis's primary site; use morphology and panels such as TTF-1/Napsin A, GATA3/ER, or CDX2/SATB2 in clinical context.

Once HCC is confirmed, staging determines treatment allocation.

The HKLC classification of liver tumour status ^[1]:

Published HKLC preferred-treatment allocation ^[1][3][11]:

Key difference from BCLC: HKLC is more aggressive in selected patients. It still offers resection for Stage 2B intermediate tumours (Child A, no EVM) and TACE for Stage 3B locally advanced tumours without EVM, whereas BCLC would usually channel these groups toward non-surgical therapy earlier. This reflects the fact that multiple tumours and limited intrahepatic vascular invasion are NOT always treated as absolute contraindications to resection ^[2] in Hong Kong practice.

BCLC algorithm — for reference, different from local management approach ^[2]:

TARE uses Yttrium-90 microspheres (TheraSphere) which induce extensive tumour necrosis with acceptable safety profile ^[1].

• Radioactive microspheres are directly injected into hepatic artery branches that supply the tumour and deliver doses of high-energy, low-penetration radiation selectively to the tumour ^[1]
• Y-90 is a pure β-emitter with a tissue penetration of only ~2.5 mm — so radiation is highly localised to the tumour bed

Key advantage over TACE:

• Preferred over TACE ONLY in the setting of HCC complicated by malignant branch or lobar portal vein thrombosis ^[1] — less arterial ischaemia induced by radioembolisation because of the smaller particle size, suggesting it should be safer in the setting of portal vein thrombosis ^[1]
• This is a critical distinction: Main portal-vein occlusion increases the risk of ischaemic liver failure with TACE. Selected branch involvement may be treated in expert MDT pathways; TARE can be considered with suitable liver reserve and dosimetry because its effect is predominantly radiation rather than embolic occlusion.

Histotripsy is a novel focused ultrasound ablation technique. Unlike RFA, microwave ablation, HIFU, or SBRT, it is:

• Non-invasive — no needle track
• Non-thermal — less heat-sink limitation near vessels in theory
• Non-ionising — no radiation dose to normal liver
• Mechanical — creates cavitation bubbles that liquefy the targeted tissue

It has FDA-authorised device availability for non-invasive destruction of liver tumours based on early primary and metastatic liver tumour data, including the #HOPE4LIVER trial, but this approval is for liver tumour destruction and does not by itself establish HCC-specific survival benefit or a standard place in HCC algorithms ^[19].

Current HCC role:

• May be considered only in highly selected patients through MDT discussion, especially when standard locoregional options are unsuitable.
• Potential future roles include bridge/downstaging before transplant, treatment near vessels/bile ducts, and combination with systemic immunotherapy, but evidence is still early.
• Not currently a routine HKLC / Hong Kong consensus treatment category for HCC management.

• Increased intra-tumoural pressure with occlusion of hepatic veins by tumour thrombi or invasion
• Rapid tumour growth and necrosis
• Vascular dysfunction → friable vessels

• Sudden onset of epigastric pain with abdominal distension
• Hypovolaemic shock
• Increased likelihood of peritoneal seeding
• Prognosis poor: mortality 25–100%

First-choice treatment for ruptured HCC is transarterial embolisation (TAE) ^[4][2]:

• TAE if portal vein is patent and reasonable liver function ^[2]
• Complications of TAE: liver failure, embolism, arterial dissection ^[2]

If uncontrolled bleeding — laparotomy ^[4]:

• Perihepatic packing + suture plication (small bleeding site only)
• Absolute alcohol injection
• Selective hepatic artery ligation (HAL)
• Staged liver resection ^[2]

Acute management principles ^[2]:

• Admit ICU
• NPO, IV access, fluid resuscitation ± blood products transfusion, monitor vitals, I/O, CVP
• CBC, LRFT, clotting, cross-match, HBsAg (may need entecavir cover)
• CXR, USG
• CT scan when stabilised to rule out portal vein thrombosis ^[2]

This is the single most dramatic and life-threatening complication of HCC. Rupture occurs in 5–10% of HCC patients ^[2] and carries a mortality of 25–100% ^[2].

Pathogenesis ^[2]:

• Tend to occur in large and peripherally located tumours — subcapsular tumours have less surrounding parenchyma to contain them
• Increased intra-tumoural pressure with occlusion of hepatic veins by tumour thrombi or invasion — venous outflow obstruction raises pressure within the tumour, like a balloon with a blocked outlet
• Rapid tumour growth and necrosis — the tumour outgrows its blood supply, leading to central necrosis with softening and structural weakness
• Vascular dysfunction → friable vessels — HCC neoangiogenesis produces structurally abnormal, thin-walled vessels that lack the smooth muscle and elastic layers of normal vessels; these are prone to spontaneous rupture

Clinical features ^[1][2]:

• Leads to intraperitoneal haemorrhage ^[1]
• Presents with severe abdominal pain with peritoneal signs and shock ^[1]
• Sudden onset of epigastric pain with abdominal distension ^[2]
• Hypovolaemic shock — tachycardia, hypotension, pallor, diaphoresis, altered consciousness
• Increased likelihood of peritoneal seeding ^[2] — rupture exposes tumour cells directly to the peritoneal cavity, dramatically increasing the risk of peritoneal carcinomatosis. Even patients who survive the acute event face a much worse long-term prognosis.

Management ^[1][2][4]:

• Treatment of choice is transarterial embolisation (TAE) ^[1][4] — if portal vein is patent and liver function is reasonable
• Uncontrolled bleeding should proceed directly to laparotomy ^[1][4]:
  • Perihepatic packing + suture plication
  • Absolute alcohol injection
  • Selective hepatic artery ligation (HAL)
  • Staged liver resection

HCC is a vascular tumour with a high propensity for venous invasion (portal and hepatic veins) ^[4]. Portal vein tumour thrombus is found in 30–60% of HCC patients at diagnosis or during disease course.

Why does HCC invade the portal vein?

• HCC grows into the path of least resistance. The portal vein branches run through the liver parenchyma immediately adjacent to hepatocytes. Malignant hepatocytes can directly invade through the thin sinusoidal walls into portal venule tributaries, then grow "upstream" along the portal vein like ivy growing along a pipe.

Consequences of PVTT (each explained from first principles):

Spread through hepatic vein branches and spread to IVC, right heart and lungs ^[1][4].

• Tumour can grow as a "tongue" of tissue extending from hepatic veins into the IVC and even into the right atrium
• Budd-Chiari-like syndrome: hepatic venous outflow obstruction → hepatic congestion, worsened ascites, hepatomegaly
• Right atrial tumour extension: can cause acute right heart obstruction → cardiogenic shock, sudden death
• Pulmonary tumour embolism: tumour fragments can embolise to the lungs → acute dyspnoea, right heart failure

Spread of HCC ^[3]:

• Local invasion: portal vein, hepatic vein, bile duct
• Lymphatic spread: one quarter of patients
• Transperitoneal spread: rare
• Haematogenous spread: lung and bone

Local invasion into bile duct ^[3]:

• HCC can invade the biliary tree causing obstructive jaundice (though this is NOT common ^[2])
• Haemobilia — tumour eroding into bile ducts causes bleeding into the biliary system → melaena, jaundice, RUQ pain (Quincke's triad)
• Biliary obstruction can precipitate ascending cholangitis (Charcot's triad: fever, jaundice, RUQ pain)

These were covered extensively in the Clinical Features section but represent a category of "complications" in their own right:

• Hypoglycaemia — can be severe and recurrent, causing seizures, LOC, and even death if not recognised. Occurs via IGF-2 secretion and high metabolic tumour demand ^[2][3]
• Hypercalcaemia — via PTHrP secretion; can cause cardiac arrhythmias, delirium, renal failure ^[2][3][4]
• Erythrocytosis — secondary polycythaemia from ectopic EPO; increases blood viscosity → risk of thrombosis ^[2][4]
• Diarrhoea — from VIP secretion; causes dehydration and electrolyte derangement ^[2][3]

Decompensation of cirrhosis: ascites, variceal bleeding, hepatic encephalopathy ^[4].

• The cirrhotic liver cannot adequately clear ammonia (from GI bacterial metabolism and deamination of amino acids)
• Ammonia crosses the blood-brain barrier → astrocytes convert it to glutamine (via glutamine synthetase) → ↑ intracellular osmolality → astrocyte swelling → cerebral oedema → altered consciousness
• HCC worsens this by: (a) further reducing functional hepatic mass, (b) portal vein thrombosis creating more portosystemic shunting of ammonia-rich blood
• Precipitants in HCC patients: GI bleeding (blood = protein load in gut → ↑ ammonia production), infection (SBP), constipation, sedatives

• The cirrhotic liver has ↓ synthesis of clotting factors (II, V, VII, IX, X, fibrinogen) AND ↓ synthesis of anticoagulant proteins (protein C, S, antithrombin)
• Thrombocytopenia from hypersplenism compounds this
• Result: "rebalanced" haemostasis that is fragile — patients can bleed OR clot unpredictably
• In the context of HCC: surgery, biopsy, or ruptured HCC can provoke life-threatening haemorrhage

• Cirrhotic ascites provides a culture medium for bacteria; impaired hepatic reticuloendothelial function (↓ Kupffer cell activity) fails to clear portal bacteraemia
• Presents with fever, abdominal pain, worsening ascites, encephalopathy
• Diagnosis: ascitic fluid PMN ≥ 250/mm³
• Common in HCC patients because they often have refractory ascites requiring frequent paracentesis (entry point for infection)

Overall post-hepatectomy outcomes ^[1][4][11]:

• Operative mortality: 1–5%
• Morbidity: ~30%
• 5-year recurrence rate: ~50% — due to "field cancerisation" effect ^[1][4]