Hepatocellular Carcinoma

[1]

Meet the patient

A 58-year-old man with known HBV cirrhosis, Child-Pugh A, has been on 6-monthly surveillance for years and feels entirely well. His routine liver ultrasound this visit flags a new 2.5 cm right-lobe nodule, and his AFP has climbed off baseline to 220 ng/mL. He is pain-free, ascites-free, and bewildered that anyone is worried. [1]

Hold two questions and the whole topic falls into place: can the tumour be cut out (is it resectable)? and can the remaining liver survive the cut (is the liver function good enough)? Every scan, score, and guideline below exists to answer those two questions in the right order. [1]

Overview — HCC is a disease of the cirrhotic liver

Hepatocellular carcinoma (HCC) is a primary malignant epithelial neoplasm of the hepatocyte — the commonest primary liver cancer (75 to 85 percent), dwarfing intrahepatic cholangiocarcinoma (10 to 15 percent). [1]

The one concept that earns more marks than any other: HCC is overwhelmingly a disease of the cirrhotic liver — about 80 to 90 percent arise in a liver already scarred by chronic injury. Etymology nugget: cirrhos is Greek for tawny, orange-yellow — the colour Laënnec saw in the gnarled, nodular organ, and the colour that names the very field in which HCC grows. [1]

That single fact has three consequences you meet on every ward round: the patient is usually known before the cancer declares itself (the basis for surveillance); the functional reserve of the surrounding liver decides treatment as much as the tumour (the basis for Child-Pugh, ALBI, and the future liver remnant); and preventing cirrhosis prevents HCC (the basis for vaccination, antivirals, and lifestyle). [1]

[1]

Liver anatomy — the dual supply is the foundation everything rests on

Every surveillance image you read, every resection you plan, and every TACE you book exploits one anatomical fact: the liver has a dual blood supply, and HCC flips it. [1]

The dual blood supply

• Portal vein — roughly 75 percent of inflow, nutrient-rich venous blood from gut, spleen and pancreas.
• Hepatic artery (from the coeliac axis) — roughly 25 percent of inflow, oxygenated systemic blood, yet it supplies about half the liver's oxygen. [1]

Venous drainage runs through the three hepatic veins (right, middle, left) into the IVC; the groove between the right and middle hepatic veins marks the main portal scissure (Cantlie line) splitting right from left functional lobes. [1]

The arterialisation concept — one physiology, three uses

As a regenerative nodule turns malignant, it loses its portal supply and grows abnormal arterial (neoangiogenic) feeders. That single switch is exploited in three completely different ways: [1]

1. Diagnosis — the tumour brightens in the arterial phase and washes out in portal/delayed phases (it has no portal supply to hold the contrast). Arterial hyperenhancement plus washout is the non-invasive signature.
2. Locoregional treatment — TACE pumps chemotherapy down the tumour's arterial feeder then embolises it, sparing the portal-fed parenchyma.
3. Systemic therapy — the angiogenic switch (driven by VEGF and FGF) is the target of bevacizumab and the multi-kinase inhibitors. [1]

Couinaud segmentation — the surgical map

Resections follow the Couinaud classification — eight independent segments, each carrying its own portal pedicle (portal vein, hepatic artery, bile duct) and drained by a hepatic vein, numbered I to VIII clockwise from the visceral surface. [1]

[1]

Standard resections follow Couinaud planes: right hepatectomy removes V to VIII; left hepatectomy II to IV; extended right (trisegmentectomy) IV to VIII; extended left II to IV plus V and VIII; left lateral sectionectomy II and III. This is why intra-operative ultrasound and segmental anatomy are non-negotiable on the table. [1]

Epidemiology — the geography of HBV, and the Western rise of NASH

HCC is the 6th most common cancer and the 3rd leading cause of cancer death worldwide, killing over 800,000 people a year, and its map tracks its viruses. [1]

[1]

The male predominance (2 to 4:1) reflects higher HBV carriage, alcohol, smoking, and androgen-driven tumour promotion. Peak age is 50 to 70 in low-HBV regions but 20 to 40 in high-HBV, aflatoxin-exposed Africa — where HCC in young adults without cirrhosis is a recognised pattern. [1]

Risk factors — HEPATOMA, and the cirrhosis rule

Most HCC risk factors act through one final common pathway — chronic injury and cirrhosis. A smaller group (HBV, aflatoxin, haemochromatosis) is directly oncogenic regardless. Sort every risk factor into those two buckets and the list memorises itself. [1]

Hepatitis B — the global number one

HBV causes over half of all HCC worldwide and raises risk 15 to 20-fold — and it breaks the cirrhosis rule in ways juniors forget: [1]

• HCC develops in HBV carriers without cirrhosis — unlike HCV — which is why HBV carriers themselves qualify for surveillance.
• Direct oncogenesis by HBV DNA integration into the host genome (chromosomal instability, insertional mutagenesis) decades before cancer appears — so antivirals reduce but never abolish risk.
• HBx protein transactivates proliferative genes and blunts p53-mediated apoptosis.
• Aflatoxin synergy with HBV multiplies (not adds to) risk. [1]

The highest-risk HBV profile: high viral load (HBV DNA over 2000 IU/mL), HBeAg positivity, genotype C, male sex, older age, family history, African descent. [1]

Hepatitis C — cirrhosis-dependent

HCV causes HCC almost exclusively through cirrhosis; once cirrhosis is established the risk runs 1 to 4 percent per year. Direct-acting antivirals cure HCV (SVR over 95 percent) and cut risk sharply — but surveillance continues after SVR, because residual risk persists in the established cirrhosis. [1]

NASH / MASLD — the Western rising tide

NASH is the fastest-growing cause of HCC in the West, riding the obesity and diabetes epidemics. Diabetes and obesity are independent risks even beyond cirrhosis (hyperinsulinaemia, IGF signalling, adipose IL-6/TNF-alpha), and NASH-HCC patients are often diagnosed later because they sit outside traditional surveillance. [1]

Alcohol

Alcohol accelerates fibrosis toward cirrhosis and synergises with viral and metabolic risk; alcohol-related cirrhosis carries roughly 1 percent HCC risk per year, and cessation lowers but does not erase risk once cirrhosis is established. [1]

Aflatoxin B1 — the fungal carcinogen

A fungal toxin from Aspergillus flavus/parasiticus contaminating stored grains (maize, groundnuts) in hot, humid sub-Saharan Africa and Southeast Asia, aflatoxin forms DNA adducts producing the signature TP53 codon 249 (R249S) mutation. Aflatoxin plus HBV is among the most powerful carcinogenic synergies known. [1]

Hereditary haemochromatosis

Untreated HFE (C282Y homozygous) iron-overload cirrhosis carries one of the highest per-cirrhosis HCC risks — up to 20 to 30 percent lifetime; iron-driven reactive oxygen species damage DNA, and venesection before cirrhosis largely abolishes the risk. [1]

Other hereditary and metabolic causes

• Alpha-1 antitrypsin deficiency (ZZ), hereditary tyrosinaemia type I, glycogen storage diseases (Ia, III), porphyria cutanea tarda.
• Wilson disease — cirrhosis occurs but HCC is relatively uncommon (possible protective copper effect). [1]

Other chronic liver diseases

PBC (with advanced fibrosis), autoimmune hepatitis with cirrhosis, haemophilia with transfusion-acquired HCV, and Budd-Chiari with established cirrhosis all raise risk. [1]

Behavioural and environmental factors

Tobacco modestly raises risk and synergises with alcohol and viral hepatitis; anabolic steroids/androgens and betel quid contribute in some populations. [1]

Pathophysiology and molecular biology — the march from regenerative nodule to HCC

The multistep carcinogenesis model — the cirrhosis pathway

The dominant cirrhosis sequence marches regenerative nodule to low-grade dysplastic to high-grade dysplastic (nodule-in-nodule) to overt HCC, carrying two structural changes with it: [1]

1. Arterialisation — loss of portal tracts, gain of arterial neoangiogenesis (exactly what imaging detects).
2. Microvascular invasion — the propensity to invade portal/hepatic venous branches, seeding intrahepatic satellites and driving post-resection recurrence. [1]

Molecular pathways — the high-yield mutations

HCC is genetically heterogeneous, but a handful of recurrent driver pathways dominate and are increasingly exam-relevant: [1]

• TERT promoter (~60 percent) — the commonest driver; reactivates telomerase, often an early event in dysplastic nodules.
• Wnt/beta-catenin (CTNNB1, ~30 percent) — stabilises beta-catenin, driving proliferative transcription; these tumours are well-differentiated, chemoresistant, and immune-cold (poor checkpoint-inhibitor response).
• TP53 (~25 to 30 percent) — especially HBV- and aflatoxin-driven, with the signature aflatoxin R249S codon-249 mutant; aggressive, poorly differentiated.
• Chromatin remodelling (ARID1A, ARID2), PI3K/AKT/mTOR (PIK3CA, PTEN, FGF19), and NFE2L2/NRF2 (constitutive antioxidant stress response, chemoresistance). [1]

Host-tumour and immune factors

The HCC microenvironment is immunosuppressive — tumour-associated macrophages, regulatory T cells, and PD-L1 expression — which is the biological rationale for combining checkpoint blockade (atezolizumab, anti-PD-L1) with anti-VEGF (bevacizumab), the latter also reversing VEGF-driven immunosuppression. [1]

[1]

Clinical presentation — a cirrhotic who suddenly decompensates

In a cirrhotic under surveillance, HCC is found asymptomatically; symptoms mean advanced disease. When it does declare itself, expect: [1]

• Right upper quadrant pain, weight loss, early satiety — the mass effect of a large tumour.
• Hepatomegaly with a hard nodular edge; a bruit over the liver is occasionally audible (tumour vascularity).
• Decompensation of known cirrhosis — new ascites, jaundice, encephalopathy; haemorrhagic ascites suggests tumour bleeding.
• Paraneoplastic syndromes — non-islet-cell hypoglycaemia (IGF-II), erythrocytosis (erythropoietin), hypercalcaemia (PTHrP), diarrhoea, dermatomyositis.
• Acute collapse — haemoperitoneum from tumour rupture (a surgical emergency, commoner in large superficial and HBV-related tumours) or acute-on-chronic liver failure. [1]

Examination blends signs of chronic liver disease (palmar erythema, spider naevi, gynaecomastia, clubbing, Dupuytren, ascites, caput, asterixis) with the tumour itself; a friction rub or bruit over the liver is the classic exam finding. [1]

Surveillance — the 6-monthly ultrasound, for the right patient only

Surveillance exists to catch HCC at a curable stage (BCLC 0/A) — but only in the at-risk, never the general public. [1]

Who gets surveilled

AASLD and EASL recommend 6-monthly liver ultrasound +/- AFP for: [1]

1. All cirrhosis of any cause (Child-Pugh A, B, and selected C listed for transplant) — annual HCC incidence 1 to 8 percent.
2. HBV carriers without cirrhosis if any of: Asian men over 40, Asian women over 50, all African descent over 20, first-degree family history, or high HBV DNA (over 2000 IU/mL) with active hepatitis over 40 (men)/50 (women).
3. Non-cirrhotic NAFLD/NASH with advanced fibrosis (F3/F4).
4. Stage 4 PBC (advanced fibrosis). [1]

Who does NOT benefit

Child-Pugh C patients who are not transplant candidates (surveillance cannot change their outcome), HBV carriers below age/sex thresholds, and HCV patients without cirrhosis after SVR (low residual risk). [1]

How it is done

• Liver ultrasound every 6 months — the interval reflects HCC's volume doubling time (a 1 cm tumour takes 4 to 6 months to reach 2 cm); sensitivity 60 to 80 percent, lower in obesity and nodular cirrhosis.
• AFP adds sensitivity; over 400 ng/mL is highly suggestive, but AFP is neither sensitive nor specific alone (false positives in pregnancy, hepatitis, germ-cell tumours; false negatives in small well-differentiated tumours) — a rising trend beats any single value.
• Any nodule over 1 cm on ultrasound is characterised by multiphase CT or MRI (see LI-RADS).
• Emerging markers DCP/PIVKA-II and AFP-L3 add sensitivity but are not yet universal. [1]

Diagnosis — non-invasive by design: do not biopsy an LR-5

The mantra: a nodule in a cirrhotic liver is HCC until washout rules it out. [1]

In cirrhosis, a nodule over 1 cm showing arterial phase hyperenhancement then portal venous or delayed washout on multiphase CT or MRI is diagnostic of HCC — EASL and AASLD endorse this non-invasive diagnosis, and biopsy is usually not required. [1]

The classic trap: biopsying an LR-5 lesion. A cirrhotic nodule that already meets LR-5 is diagnostic without a needle — and percutaneous biopsy carries a 1 to 3 percent risk of needle-track tumour seeding. Biopsy is reserved for atypical imaging, non-cirrhotic livers, or equivocal LR-3/LR-4 observations; never biopsy a lesion you already know is HCC. [1]

Staging workup

Once HCC is confirmed, staging sets treatment: [1]

• Multiphase CT chest/abdomen/pelvis (or CT chest + MRI liver) for tumour burden, portal vein tumour thrombus, and extrahepatic spread (lung, bone, adrenal).
• MRI liver if CT is indeterminate.
• Bone scan if bone pain or raised ALP (metastases are osteolytic, painful).
• AFP baseline (and to guide later ramucirumab eligibility), Child-Pugh/ALBI, ECOG performance status, and full bloods, coagulation, liver and renal biochemistry, viral markers. [1]

LI-RADS — the imaging gatekeeper (LR-5 = treat, no biopsy)

The Liver Imaging Reporting and Data System (LI-RADS) stratifies observations in at-risk patients from LR-1 (definitely benign) to LR-5 (definitely HCC) — and only LR-5 is diagnostic without biopsy. [1]

The five major features

LI-RADS scores five major features: [1]

1. Nodule size (under 10 mm / 10 to 19 mm / 20 mm and over).
2. Arterial phase hyperenhancement (APHE) — non-rim enhancement unequivocally greater than liver in the arterial phase.
3. Washout — falling enhancement on later phases, becoming hypoenhanced.
4. Enhancing capsule — smooth capsular rim on portal/delayed phases.
5. Threshold growth — at least 50 percent enlargement within 6 months, or a new 10 mm-plus observation within 24 months. [1]

The LI-RADS categories

[1]

Staging — three axes: tumour, liver, performance

No single score dominates HCC, because prognosis rides three independent axes — tumour burden, liver function, and performance status — and treatment must respect all three. BCLC is the one that ties them most directly to a treatment. [1]

Child-Pugh — functional liver reserve

The Child-Pugh (Child-Turcotte-Pugh) score estimates hepatic reserve on five parameters and underpins almost every HCC decision: [1]

Interpretation:

• Child-Pugh A (5 to 6) — well-compensated; eligible for resection and full systemic therapy.
• Child-Pugh B (7 to 9) — significant impairment; resection hazardous; transplant or locoregional therapy preferred.
• Child-Pugh C (10 to 15) — decompensated; best supportive care or transplant only; high perioperative mortality. [1]

ALBI grade — the objective alternative

The ALBI (Albumin-Bilirubin) grade uses only bilirubin and albumin, making it more objective and reproducible than Child-Pugh (no subjective ascites/encephalopathy grading): [1]

• ALBI 1 — score under -2.60, best prognosis.
• ALBI 2 — -2.60 to -1.39, intermediate.
• ALBI 3 — over -1.39, worst prognosis. [1]

Formula: ALBI = (0.66 x log10 bilirubin in umol/L) + (-0.085 x albumin in g/L). [1]

MELD — transplant allocation

The MELD(-Na) score (bilirubin, INR, creatinine, sodium) governs liver transplant allocation; candidates within Milan criteria receive a MELD exception score because HCC risk climbs with waiting time. [1]

BCLC — staging and treatment in one algorithm (OABCD)

The Barcelona Clinic Liver Cancer (BCLC) algorithm links tumour burden, liver function and performance status to a first-line treatment — the framework examiners expect. [1]

BCLC is a guideline, not a law: many centres (especially in Asia) treat across boundaries — resecting multinodular disease or pushing TACE/systemic therapy past the lines. The up-to-7 criteria (sum of largest tumour diameter in cm plus number of tumours under 7) and other expanded criteria also shape transplant decisions. [1]

Treatment — curative, locoregional, systemic, supportive

Resection — can the liver survive the cut?

Consultant confession: in a cirrhotic, the question is never can I cut this tumour out — the tumour is almost always resectable. The real question is whether the remaining liver can survive the cut. That is why Child-Pugh, portal pressure, and the future liver remnant — not the tumour — decide resectability. [1]

• The resection sweet-spot: single HCC, Child-Pugh A, normal bilirubin (under 17 umol/L), platelets over 100, no clinically significant portal hypertension (HVPG under 10 mmHg, no varices).

• Non-cirrhotic HCC is the ideal candidate — no functional limit, so large resections are tolerated.

• Future liver remnant (FLR) is the linchpin: a minimum FLR of 40 percent is required in cirrhosis (versus 20 to 30 percent in a healthy liver). Below that, augment with portal vein embolisation (PVE) (contralateral hypertrophy over 4 to 6 weeks) or ALPPS (staged partition, hypertrophy in 1 to 2 weeks). [1]

• Anatomical resection follows Couinaud segments — preferred for HCC because it removes the tumour with its portal tributaries and intra-segmental satellites.

• Pringle manoeuvre — intermittent clamping of the hepatoduodenal ligament (hepatic artery + portal vein) to cut blood loss.

• CUSA or crush-clamp for parenchymal transection; intra-operative ultrasound is essential to map the tumour, vessels, and satellites. [1]

• 5-year survival 50 to 70 percent for single tumours without vascular invasion.

• Recurrence is the rule — 50 to 70 percent at 5 years — from intrahepatic metastases (within 2 years) or multicentric de novo tumours (later).

• Perioperative mortality 2 to 5 percent in high-volume centres, higher with portal hypertension. [1]

Transplant — Milan, the gatekeeper

Transplant is the only treatment that removes the tumour and the cirrhotic field together — abolishing recurrence risk in the native liver. It is reserved for patients within transplant criteria with decompensated cirrhosis or tumours not amenable to resection. [1]

Milan criteria (Mazzaferro, 1996): [1]

• Single tumour under 5 cm, OR up to 3 tumours each under 3 cm.
• No macrovascular invasion, no extrahepatic spread.
• Within these limits, 5-year survival is 70 to 80 percent — matching transplant for non-malignant disease. [1]

UCSF expanded criteria (Yao, 2001): single tumour under 6.5 cm, OR up to 3 with largest under 4.5 cm and total diameter under 8 cm — not universally adopted but approaching Milan outcomes in selected centres. [2]

While awaiting transplant, bridging/downstaging with TACE, RFA, MWA, TARE (Y-90), or SBRT controls progression and can bring beyond-criteria tumours back within Milan before listing. [1]

Post-transplant: 5-year survival 70 to 80 percent within Milan, HCC recurrence 10 to 15 percent; mTOR inhibitors (sirolimus, everolimus) may cut recurrence and are preferred immunosuppressants in some protocols. [1]

Thermal ablation (RFA / MWA)

Radiofrequency (RFA) and microwave (MWA) ablation are first-line curative options for small tumours (under 2 to 3 cm), achieving complete necrosis in 90 percent of under-2 cm lesions — comparable to resection in that range, especially when resection is contraindicated; 5-year survival 40 to 70 percent. Percutaneous ethanol injection (PEI) is now largely superseded but still useful in poor candidates and cystic lesions. [1]

Intermediate stage (BCLC B) — TACE

TACE is first-line for BCLC B — unresectable, multifocal HCC without portal vein invasion and preserved liver function (Child-Pugh A to B). [1]

HCC is arterially fed while the parenchyma is portal-fed, so delivering chemotherapy down the hepatic artery then embolising it concentrates drug in the tumour and starves it while sparing the liver. [1]

Via femoral or radial access under fluoroscopy, a catheter is advanced into the tumour-feeding arterial branch; chemotherapy (doxorubicin, cisplatin, or mitomycin C) is delivered emulsified in Lipiodol (selectively retained in HCC) or on drug-eluting beads (DEB-TACE), followed by embolisation with gelatin sponge or particles. [1]

• Absolute — portal vein tumour thrombus: the liver then depends entirely on arterial flow, so embolising it causes infarction. Also decompensated cirrhosis (Child-Pugh C), extensive bilobar disease, renal failure, encephalopathy.
• Relative — bilirubin over 2 to 3 mg/dL, significant ascites, hepatofugal flow. [1]

Post-embolisation syndrome (fever, pain, nausea — self-limiting over 1 to 3 days), liver abscess, hepatic failure, non-target embolisation (gallbladder, gut), contrast nephropathy. [1]

TACE beats best supportive care in BCLC B (median survival 20 to 30 months versus 16 untreated); repeat sessions are given for residual or recurrent disease. [1]

TARE (Y-90) — the option when the portal vein is thrombosed

Yttrium-90 microspheres deliver internal radiation into the tumour arterial bed and, unlike TACE, can be used in portal vein thrombosis because they cause minimal ischaemia, with less post-embolisation syndrome — used for downstaging, BCLC B, and selected BCLC C. Stereotactic body radiotherapy (SBRT) is an emerging option for tumours unsuitable for ablation or surgery. [1]

Advanced stage (BCLC C) — systemic therapy

Systemic therapy is for advanced HCC (BCLC C) — macrovascular invasion, extrahepatic spread, or progression after locoregional therapy — in patients with preserved liver function (Child-Pugh A, selected B7) and ECOG PS 0 to 2. [1]

First-line — atezolizumab + bevacizumab (IMbrave150): the anti-PD-L1 atezolizumab plus anti-VEGF bevacizumab combination is now standard first-line after IMbrave150, which gave median OS 19.2 months versus 13.4 for sorafenib (HR 0.66); dose atezolizumab 1200 mg + bevacizumab 15 mg/kg IV every 3 weeks. [4]

Everyone forgets: the patient passed for atezo+bev is the cirrhotic with portal hypertension — exactly the patient with occult varices. Do the OGD and band any varices before dose one, or the first cycle can be a catastrophic GI bleed; bevacizumab carries a real gastrointestinal bleeding risk. [4]

Adverse effects to counsel on: hypertension, proteinuria, arterial thromboembolism, bleeding, immune-related events (hepatitis, colitis, pneumonitis, endocrinopathies), infusion reactions. Contraindications: uncontrolled autoimmune disease, uncontrolled hypertension, recent haemorrhage or thrombosis, untreated varices. [1]

First-line alternatives — sorafenib and lenvatinib: for patients unsuitable for atezo+bev, the oral multi-kinase inhibitors remain first-line. Sorafenib (SHARP, 2008) targets Raf/MEK/ERK, VEGFR, PDGFR — dose 400 mg twice daily, median OS 10.7 versus 7.9 months for placebo; adverse effects are hand-foot skin reaction, diarrhoea, hypertension, fatigue, rash. [5]

Lenvatinib (REFLECT, 2018) targets VEGFR1-3, FGFR1-4, PDGFR, RET, KIT and is non-inferior to sorafenib (median OS 13.6 months) with higher response rates; it cannot be used with bile duct invasion or over 50 percent liver involvement. Adverse effects: hypertension, proteinuria, fatigue, weight loss, hand-foot reaction (less than sorafenib). [6]

Second-line options — after first-line TKI progression or intolerance, validated agents exist, a real advance in HCC: [1]

• Regorafenib (RESORCE) — for patients who tolerated sorafenib but progressed; 160 mg OD, 3 weeks on / 1 week off; OS 10.6 versus 7.8 months. [7]
• Cabozantinib (CELESTIAL) — multi-kinase (MET, VEGFR2, AXL) after up to two prior lines including sorafenib; 60 mg OD. [8]
• Ramucirumab (REACH-2) — anti-VEGFR2, the first biomarker-selected second-line, indicated only for AFP over 400 ng/mL; 8 mg/kg IV every 2 weeks. [9]
• Pembrolizumab (anti-PD-1) — for MSI-high or TMB-high tumours, or after atezo+bev in selected patients. [1]

[1]

BCLC D — best supportive care

For terminal HCC (PS over 2, Child-Pugh C, no transplant option) the goal is symptom control and dignity: [1]

• Pain — paracetamol (avoid high doses if decompensated), then weak then strong opioids (morphine); avoid NSAIDs (bleeding, renal risk).
• Ascites — spironolactone +/- furosemide; large-volume paracentesis with albumin 6 to 8 g per litre removed.
• Encephalopathy — lactulose + rifaximin.
• Nutritional support, palliative care referral, psychological and spiritual support, advance care planning. Median survival under 3 months. [1]

Complications and their management

• Tumour rupture with haemoperitoneum — emergency: resuscitate, then transarterial embolisation (TAE) to control bleeding, definitive treatment (resection if feasible) once stable.
• Portal vein tumour thrombosis — worsens portal hypertension and prognosis; contraindicates TACE; consider TARE (Y-90) or systemic therapy.
• Liver failure / decompensation — standard liver-failure measures; exclude precipitants (infection, bleeding, sedatives).
• Cancer cachexia — nutrition, megestrol, nausea control.
• Bone metastases — analgesia, palliative radiotherapy, bisphosphonates/denosumab.
• Spontaneous bacterial peritonitis in ascites — diagnostic paracentesis, third-generation cephalosporin (e.g. cefotaxime). [1]

Follow-up after treatment

After curative-intent treatment (resection, transplant, ablation): contrast CT or MRI every 3 to 6 months for 2 years, then 6 to 12 monthly, plus AFP every 3 months — recurrence is common after resection/ablation and de novo tumours keep coming, so manage the underlying liver disease (antivirals, abstinence, metabolic control) in parallel. [1]

After TACE/TARE, image (mRECIST) at 4 to 6 weeks to assess response, then repeat as needed; on systemic therapy, CT every 8 to 12 weeks (mRECIST) and monitor blood pressure, proteinuria, thyroid function, and hepatitis for checkpoint inhibitors. [1]

Prevention — the three levels (and why the HBV vaccine is the cheapest cure)

Prevention is the cheapest, highest-impact intervention in HCC, and it runs on three levels. [1]

Primary — prevent the cause

• Universal HBV vaccination — the cornerstone: a 3-dose schedule (birth, 1 month, 6 months). Taiwan was first (1984), after which childhood HCC fell by about 70 percent; WHO recommends universal vaccination globally and the birth dose is critical to block perinatal transmission.
• Antivirals for chronic HBV — entecavir 0.5 mg daily or tenofovir (TDF) 300 mg daily cut HCC risk by roughly 70 percent (5-year incidence from about 8 percent down to 2 to 3 percent); risk falls but is not abolished, because HBV DNA integration may already have occurred.
• HCV eradication — DAAs (e.g. sofosbuvir + velpatasvir for 12 weeks) achieve SVR over 95 percent; surveillance continues after SVR in cirrhosis.
• Alcohol cessation; weight and metabolic control for NASH; aflatoxin avoidance via improved grain storage in sub-Saharan Africa and Southeast Asia. [1][3]

Secondary — early detection

6-monthly liver ultrasound +/- AFP in the at-risk population detects HCC at a curable stage. [1]

Tertiary — prevent recurrence

Antiviral therapy after resection/ablation suppresses HBV and cuts de novo multicentric recurrence. [1]

Differential diagnosis

A liver lesion in a cirrhotic is HCC until proven otherwise, but the differential is broad: [1]

• Benign — haemangioma (peripheral nodular enhancement with centripetal fill-in), focal nodular hyperplasia (central scar), hepatic adenoma (OCP/anabolic steroids), regenerative/dysplastic nodules, focal fat sparing, simple cyst.
• Malignant — intrahepatic cholangiocarcinoma (delayed enhancement, capsular retraction, ductal dilation, raised CA 19-9), combined HCC-cholangiocarcinoma, metastases (hypovascular — breast, lung, colon; hypervascular — neuroendocrine, renal, melanoma), lymphoma, angiosarcoma.
• Infective — pyogenic/amoebic abscess (fever, pain, aspirate), hydatid cyst (Echinococcus), hepatic TB. [1]

The imaging signature (arterial hyperenhancement + washout) plus the cirrhotic setting is what makes LI-RADS LR-5 specific. [1]

Exam pearls — the high-yield spine

• HCC: 6th most common cancer, 3rd leading cause of cancer death; over 80 percent arise in cirrhotic livers. HBV is the global number one; NASH/MASLD is the fastest-growing in the West.
• Dual supply: portal vein ~75 percent, hepatic artery ~25 percent; HCC is arterial — the basis of arterial hyperenhancement, TACE, and anti-VEGF therapy.
• Couinaud: 8 segments, segment I = caudate (drains directly to IVC — hypertrophies in Budd-Chiari); resections follow segmental anatomy.
• Surveillance = 6-monthly USS +/- AFP in all cirrhotics and high-risk HBV carriers (Asian male over 40, Asian female over 50, African over 20, family history); any nodule over 1 cm gets multiphase CT/MRI.
• Diagnosis is non-invasive in cirrhosis: nodule over 1 cm with arterial hyperenhancement + washout = HCC (LI-RADS LR-5); do not biopsy (needle-track seeding).[4]
• Resection needs Child-Pugh A, no portal hypertension (HVPG under 10, platelets over 100, no varices), normal bilirubin, FLR at least 40 percent (augment with PVE/ALPPS).[1]
• Milan: single under 5 cm OR up to 3 under 3 cm, no vascular invasion, no extrahepatic spread; 5-year survival 70 to 80 percent.[1] UCSF: single under 6.5 cm, or 3 with largest under 4.5 cm and total under 8 cm.[2]
• BCLC: 0/A resect/transplant/ablate; B = TACE (contraindicated by portal vein thrombosis, Child-Pugh C); C = systemic therapy; D = best supportive care.
• First-line systemic: atezolizumab 1200 mg + bevacizumab 15 mg/kg q3w (OS 19.2, IMbrave150); pre-treatment OGD mandatory.[4] Alternatives: sorafenib 400 mg BD (OS 10.7, SHARP) and lenvatinib (OS 13.6, REFLECT).[5][6]
• Second-line: regorafenib (RESORCE), cabozantinib (CELESTIAL), ramucirumab only if AFP over 400 (REACH-2).[7][8][9]
• Prevention: universal HBV vaccination cut Taiwan childhood HCC by ~70 percent; entecavir/tenofovir cut HBV-HCC risk ~70 percent; DAAs cure HCV in over 95 percent.[3]

[1]

Ward-round test — two stems, thirty seconds each

Stem 1 — cirrhotic, 3 cm nodule, APHE + washout (answer)

A 62-year-old with HBV cirrhosis has a 3 cm right-lobe lesion on multiphase CT showing clear arterial phase hyperenhancement with portal venous washout; AFP is 280 ng/mL. The registrar wants to biopsy it 'to be sure'. What do you say? Model: This is LI-RADS LR-5 — diagnostic of HCC without biopsy. Do not biopsy an LR-5 lesion: the imaging signature (APHE + washout in a cirrhotic nodule over 1 cm) is specific, and percutaneous biopsy carries a 1 to 3 percent needle-track seeding risk. Instead, stage (Child-Pugh/ALBI, BCLC, CT chest) and take it to the hepatobiliary MDT — biopsy is reserved for atypical, non-cirrhotic, or LR-3/LR-4 observations. [1]

Stem 2 — Child-Pugh C, PS 3, multifocal (answer)

A 70-year-old with decompensated NASH cirrhosis (Child-Pugh C, albumin 24, bilirubin 70, refractory ascites, ECOG 3) has multifocal bilobar HCC. The team asks whether to start TACE or atezolizumab + bevacizumab. What is the call? Model: Neither — this is BCLC D, best supportive care. TACE is contraindicated (Child-Pugh C and the decompensated liver cannot tolerate ischaemia), and systemic therapy needs preserved liver function (Child-Pugh A, selected B7) and ECOG 0 to 2 — neither holds here. Focus on symptom control: analgesia, large-volume paracentesis with albumin (6 to 8 g per litre removed), lactulose + rifaximin for encephalopathy, nutrition, and palliative care. Median survival is under 3 months. [1]

References

1. [1]Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis N Engl J Med, 1996.PMID 8594428
2. [2]Yao FY, Ferrell L, Bass NM, et al. Liver transplantation for hepatocellular carcinoma: expansion of the tumor size limits does not adversely impact survival Hepatology, 2001.PMID 11391528
3. [3]Chang MH, Chen CJ, Lai MS, et al. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. Taiwan Childhood Hepatoma Study Group N Engl J Med, 1997.PMID 9197213
4. [4]Finn RS, Qin S, Ikeda M, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma N Engl J Med, 2020.PMID 32402160
5. [5]Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma N Engl J Med, 2008.PMID 18650514
6. [6]Kudo M, Finn RS, Qin S, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial Lancet, 2018.PMID 29433850
7. [7]Bruix J, Qin S, Merle P, et al. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial Lancet, 2017.PMID 27932229
8. [8]Abou-Alfa GK, Meyer T, Cheng AL, et al. Cabozantinib in Patients with Advanced and Progressing Hepatocellular Carcinoma N Engl J Med, 2018.PMID 29972759
9. [9]Zhu AX, Kang YK, Yen CJ, et al. Ramucirumab after sorafenib in patients with advanced hepatocellular carcinoma and increased α-fetoprotein concentrations (REACH-2): a randomised, double-blind, placebo-controlled, phase 3 trial Lancet Oncol, 2019.PMID 30665869