Feeding the tumor's own artery a lethal dose, then closing the tap
Transarterial chemoembolization (TACE) is an image-guided interventional radiology procedure that treats hepatocellular carcinoma (HCC) from the inside of its own blood supply. The first step is purely mechanical: threading a catheter from a peripheral artery, through the aorta, into the coeliac axis, and finally into the precise segmental hepatic artery branch that feeds the tumor — without ever opening the abdomen.
TACE begins with percutaneous puncture of the common femoral artery (traditional approach) or the radial artery at the wrist (increasingly preferred for patient comfort and lower bleeding risk). A vascular sheath is placed using the Seldinger technique, and a guidewire is advanced under fluoroscopy.
A 4–5 French catheter is steered retrograde (femoral) or antegrade (radial) through the iliac artery or subclavian/aortic arch into the abdominal aorta. From there it is directed into the coeliac trunk — the first major branch off the abdominal aorta, which in turn gives rise to the splenic artery, left gastric artery, and the common hepatic artery.
The common hepatic artery divides into the gastroduodenal artery and the proper hepatic artery, which itself splits into right and left hepatic arteries supplying the corresponding liver lobes. Correctly identifying this branching anatomy — which varies between patients in up to 45% of cases (replaced or accessory hepatic arteries) — is essential before any drug is delivered.
Roughly 20–45% of patients have a variant hepatic arterial anatomy (replaced right or left hepatic artery arising from the superior mesenteric or left gastric artery). Missing a variant vessel means missing part of the tumor's blood supply, so pre-procedural CT angiography is standard practice.
Once inside the hepatic arterial tree, a smaller coaxial microcatheter (as fine as 1.7–2.8 French, ~0.5–1 mm) is advanced through the guiding catheter into progressively smaller segmental and subsegmental branches — the vessels that supply only the tumor-bearing liver segment, not the whole lobe.
This "superselective" positioning is the anatomical foundation of TACE's therapeutic window: because HCC nodules recruit their own feeding arterioles, a microcatheter tip parked just proximal to the tumor blush can deliver drug and embolic material almost exclusively into the cancerous tissue, sparing the bulk of healthy hepatic parenchyma downstream of other branches.
The procedure is performed entirely under live digital subtraction fluoroscopy, with the interventional radiologist injecting small test boluses of contrast at each branch point to confirm the catheter is heading toward — and only toward — the tumor-feeding vessel before proceeding further.
Before any drug is infused, contrast dye is injected through the catheter tip to generate a real-time angiogram. This step exploits the single most important anatomical fact in liver oncology: the dual blood supply of the liver, and the fact that HCC hijacks one part of it almost exclusively.
The healthy human liver is unusual among organs in receiving blood from two independent sources: the portal vein, which carries nutrient-rich, partially deoxygenated blood from the gut and spleen and supplies roughly 70–75% of total hepatic blood flow, and the hepatic artery, which carries oxygenated blood from the aorta and supplies the remaining 25–30%.
As HCC develops, it undergoes a well-characterized process of neoangiogenesis: normal portal venules within the nodule regress while an abnormal, disorganized network of unpaired arterioles proliferates. By the time a nodule reaches the size of overt HCC (typically >1–2 cm), it derives essentially all of its blood supply from the hepatic artery — while the surrounding cirrhotic or normal liver tissue continues to rely predominantly on the portal vein.
This divergence is precisely what selective angiography visualizes: a bright, early-filling, densely staining "tumor blush" on hepatic arterial injection, sitting inside liver tissue that itself enhances far less intensely from the same arterial injection.
This dual blood supply is the entire physiological rationale for TACE: because HCC is fed almost exclusively by the hepatic artery while normal liver is fed mostly by the portal vein, occluding the hepatic artery starves the tumor while the surrounding liver survives on its portal inflow.
Digital subtraction angiography (DSA) subtracts a pre-contrast fluoroscopic "mask" image from post-contrast images, removing bone and soft tissue so that only the opacified blood vessels remain visible. Contrast is injected in small boluses (a few mL) at each catheter position, and the interventional radiologist reviews the arterial, capillary/parenchymal, and venous phases.
Key angiographic findings sought at this stage include: the number and course of tumor-feeding arteries (larger HCCs may recruit more than one), any arteriovenous shunting (tumor vessels draining directly into the portal or hepatic veins, which changes embolization strategy), and the precise segmental branch to cannulate for superselective delivery.
Cone-beam CT, now available on most modern angiography suites, can be fused with the live fluoroscopic image to provide 3D confirmation of tumor feeders that may be difficult to appreciate on 2D projection angiography alone — particularly for small or atypically located nodules.
With the microcatheter locked into the tumor-feeding branch, chemotherapy is infused directly into the vessel supplying the cancer. This intra-arterial route achieves a local drug concentration far beyond what could ever be delivered safely through a peripheral IV line, while sharply limiting the dose that reaches the rest of the body.
Two main technique variants exist. Conventional TACE (cTACE) mixes a chemotherapeutic agent — most commonly doxorubicin, sometimes cisplatin or epirubicin — with Lipiodol (an oily iodinated contrast agent) to form an emulsion. Lipiodol is selectively retained by hypervascular HCC tissue far longer than by normal liver, acting as both a drug carrier and a radio-opaque tracer, followed by a separate embolic agent (gelatin sponge particles) to complete vessel occlusion.
Drug-eluting bead TACE (DEB-TACE) instead uses calibrated, biocompatible polymer microspheres (e.g., 100–300, 300–500, or 500–700 µm) that are pre-loaded with doxorubicin through ionic exchange. The beads are infused directly into the tumor-feeding vessel, where they simultaneously embolize the vessel and slowly elute their drug payload over days, producing a more controlled, sustained-release pharmacokinetic profile and typically lower peak systemic doxorubicin levels than cTACE.
Systemic IV doxorubicin is limited by cumulative cardiotoxicity and dose-limiting myelosuppression, capping typical lifetime doses around 450–550 mg/m². Because TACE delivers the drug directly into the vessel feeding the tumor — and because the subsequent embolization traps it there rather than letting it wash out into systemic circulation — a given local dose achieves a much higher tumor tissue concentration for a much lower systemic exposure.
Studies measuring plasma doxorubicin levels after DEB-TACE report peak concentrations roughly an order of magnitude lower than after equivalent IV administration, which translates clinically into a substantially reduced incidence of alopecia, nausea, and cardiotoxicity compared to systemic chemotherapy regimens — while local tumor exposure is markedly increased.
Infusion is performed slowly and under intermittent fluoroscopic monitoring ("stasis" is the endpoint: injection continues until near-stagnation of contrast flow in the target vessel is observed), balancing maximal tumor drug delivery against the risk of reflux into non-target vessels.
The "embolization" half of chemoembolization is what gives the procedure its lasting effect: mechanically blocking the artery that was just used to deliver chemotherapy, cutting off oxygen and nutrients to the tumor and locking the drug payload in place for days to weeks of sustained local exposure.
Embolization uses particulate agents injected through the microcatheter to physically occlude the target artery. Options include gelatin sponge (Gelfoam) pledgets or slurry — historically common in cTACE, temporary/resorbable over weeks — polyvinyl alcohol (PVA) particles of calibrated size ranges, and, in DEB-TACE, the same drug-loaded microspheres that delivered the chemotherapy now serving a dual embolic role.
Particle size selection matters clinically: smaller particles (100–300 µm) penetrate deeper into the tumor's microvasculature, producing more thorough distal occlusion and tumor kill but carrying a higher risk of non-target embolization and tissue ischemia complications; larger particles (500–700 µm) occlude more proximally, which is safer near vessels with shunting but may leave some tumor microvasculature unembolized.
Injection continues under fluoroscopic monitoring until near-stasis — markedly slowed, near-stagnant contrast flow in the target vessel — signaling that the vascular bed has been adequately occluded without over-embolizing and risking reflux into the gastroduodenal or other non-target arteries.
Embolic particle size is a direct trade-off: smaller particles (100–300 µm) penetrate deeper into tumor microvessels for more complete ischemic kill, while larger particles (500–700 µm) occlude more proximally with a lower risk of non-target embolization — interventional radiologists titrate size to tumor vascularity and vessel anatomy.
Acute occlusion of a tumor-feeding artery predictably causes post-embolization syndrome: right upper quadrant pain, low-grade fever, nausea, and transient elevation of liver transaminases, occurring in the majority of patients within 24–72 hours and typically self-limited within a week with supportive care (analgesia, antiemetics, IV fluids).
Because total hepatic arterial occlusion in a single session risks liver failure — particularly in patients with compromised background liver function (cirrhosis, Child-Pugh B) — large or multifocal tumors are frequently treated in staged sessions, embolizing one hepatic segment at a time, spaced roughly 4–8 weeks apart, allowing the treated segment to recover collateral portal perfusion before the next segment is addressed.
Post-procedure imaging (contrast-enhanced CT or MRI, typically at 4–6 weeks) assesses treatment response using modified RECIST (mRECIST) criteria, which — unlike standard RECIST — measures only the viable, arterially enhancing portion of the tumor, since necrotic embolized tissue no longer enhances but may not immediately shrink in overall diameter.
The combined insult of arterial occlusion (ischemia) and locally concentrated chemotherapy drives coagulative necrosis of the tumor over subsequent days to weeks, while the surrounding cirrhotic or normal liver parenchyma — sustained by its still-patent portal venous inflow — is comparatively spared. This is where TACE's dual mechanism translates into measurable survival benefit.
TACE's place in HCC treatment was cemented by two pivotal randomized controlled trials published in 2002. Llovet et al. (Lancet, 2002) randomized unresectable HCC patients to TACE, transarterial embolization alone (no chemotherapy), or conservative treatment; the trial was stopped early after an interim analysis showed a clear survival advantage in the embolization arms. Lo et al. (Hepatology, 2002) independently randomized patients to cisplatin-Lipiodol TACE versus symptomatic treatment, reporting 1- and 2-year survival of 57% and 31% in the TACE group versus 32% and 11% in controls — a statistically significant benefit.
A subsequent meta-analysis (Llovet & Bruix, Hepatology 2003) pooling these and other randomized trials confirmed a significant 2-year survival benefit for arterial embolization/chemoembolization over best supportive care (odds ratio approximately 0.53), establishing TACE as an evidence-based, survival-prolonging therapy rather than a purely palliative procedure.
The Llovet (2002) and Lo (2002) trials, together with the 2003 meta-analysis, transformed TACE from a promising but unproven technique into a guideline-endorsed standard of care — the first interventional radiology procedure shown in randomized trials to extend survival in unresectable HCC.
The Barcelona Clinic Liver Cancer (BCLC) staging system, the most widely used HCC treatment algorithm, designates TACE as the recommended first-line therapy for BCLC Stage B — intermediate-stage HCC, defined as multinodular tumors confined to the liver in patients with preserved liver function (Child-Pugh A/B) and no vascular invasion or extrahepatic spread, who are not candidates for curative resection, ablation, or transplantation.
For comparison: BCLC Stage A (early, single small tumor) is preferentially treated with resection, ablation (radiofrequency or microwave), or transplantation, all of which offer curative intent; BCLC Stage C (advanced, with vascular invasion or extrahepatic spread) is treated with systemic therapy — tyrosine kinase inhibitors (sorafenib, lenvatinib) or immune checkpoint inhibitor combinations (atezolizumab-bevacizumab); BCLC Stage D (end-stage, poor liver function) receives best supportive care.
TACE is also frequently used outside its primary BCLC-B indication as a "bridge to transplant" — controlling tumor growth in patients awaiting liver transplantation to keep them within transplant eligibility criteria (e.g., Milan criteria) — and for downstaging tumors that initially exceed transplant or resection criteria.
A newer transarterial technique, radioembolization (also called selective internal radiation therapy, SIRT, or TARE), delivers microspheres loaded with the beta-emitting radioisotope Yttrium-90 (⁹⁰Y) instead of chemotherapy, via the same superselective catheter approach into the hepatic artery.
Unlike TACE, ⁹⁰Y-TARE microspheres are small enough (20–40 µm) and delivered at low enough particle load that they typically do not cause significant acute vessel occlusion — the tumor kill comes almost entirely from localized radiation dose (delivered over the ~64-hour half-life of ⁹⁰Y) rather than ischemia, allowing TARE to be used even in patients with portal vein thrombosis, where embolizing the hepatic artery in the setting of an already-occluded portal vein would risk global hepatic ischemia.
Head-to-head trials (e.g., SARAH, SIRveNIB for advanced disease) have generally shown similar overall survival between TARE and systemic therapy or TACE, with TARE offering a more favorable side-effect profile (less post-embolization syndrome) but requiring more complex pre-treatment dosimetry planning (including a technetium-99m-MAA "mapping" angiogram to exclude lung shunting before radioactive spheres are administered). TACE remains more widely available, faster to perform, and typically cheaper, keeping it the dominant intermediate-stage locoregional therapy worldwide.