What this is
A working set of notes on liver tumour embolisation, the anatomy on the screen, how the doctor reaches the tumour, and the four flavours of the procedure.
Why This Procedure Works at All
The whole thing turns on a quirk of how the liver gets its blood. Picture a city block where every house gets water from two pipes, a big main running under the street, and a smaller side road. If you turn off the side road, the houses are fine. The main pipe still feeds them.
Healthy liver tissue is built like that. About three-quarters of its blood comes from the portal vein, the main pipe. Tumours in the liver, though, are greedy. They grow their own little blood supply almost entirely off the hepatic artery, the side road. Cut off the side road and the tumour starves while everything else keeps working.
The whole logic, in one line
Block the artery feeding the tumour, and the tumour dies while the rest of the liver carries on.
That single fact is what every part of this procedure exists to exploit. The catheters, the contrast, the agents, the screens, all of it is in service of finding the right side road and shutting only that one off.
The patients sent for this are usually in one of a handful of camps: liver cancer that can’t be cut out, cancer that started elsewhere (often bowel or neuroendocrine) and spread to the liver, people being kept stable while they wait for a transplant, or cases where a tumour is causing symptoms and the team wants to settle it down. Same procedure, broadly the same idea.
The Plumbing
The screen during the case is essentially a moving map of the body’s arterial highway. The doctor is driving up the aorta, taking the first big branch off the top (the coeliac trunk), then onto the artery heading to the liver, and from there into smaller and smaller branches until they reach the one feeding the tumour itself.
The textbook layout looks like this:
Aorta (the big pipe down the middle)
→ Coeliac Trunk (first big branch off the top)
→ Common Hepatic Artery (heading toward the liver)
→ Proper Hepatic Artery
→ Right Hepatic Artery → feeds the right side
→ Left Hepatic Artery → feeds the left side
The interesting thing is how often this isn’t what’s actually there.
Roughly 1 in 3 patients have unusual anatomy
About 36% of people have something different, a right hepatic artery coming off a completely different vessel, an extra branch nobody expected, a left hepatic supply hanging off the artery to the stomach. This is why pre-procedure CT review isn’t optional. If you see the doctor on the workstation in the morning studying images, this is what they’re working through.
The reason it matters: a hidden artery can supply the gallbladder, stomach, or bowel. If embolisation agent drifts into one of those by accident, those organs get injured. The morning’s CT review is the single biggest thing standing between the patient and that problem.
Where the Catheter Goes In
Two routes, same destination. The needle goes either in the crease at the top of the leg or at the wrist, and the catheter is threaded up through the body’s plumbing from there.
The traditional way is the femoral artery, the big one you can feel pulsing in the groin. It’s a generous-sized vessel, easy to work in, but it carries a tail of post-procedure care: 2 to 4 hours flat in bed, watching the puncture site for haematoma (a swelling collection of blood under the skin), checking distal pulses and the colour and warmth of the leg. Major bleeding is uncommon (around 1.8%) but when it happens it’s serious.
The newer way is the radial artery at the wrist, the same one you feel for a pulse. The patient sits up almost immediately, walks, often goes home the same day. A clear plastic TR Band is wrapped around the wrist afterwards and let down slowly over a couple of hours. Bleeding from the wrist is much less of a problem than from the groin, and the comfort difference for the patient is genuinely large. The catch is operator experience. The radial route has a steeper learning curve, so it’s not always offered.
Worth flagging at handover
If the access was radial, the TR Band is the thing to mention. It needs to be deflated gradually, usually a small amount of air let out every 15 minutes or so. Local protocol will spell out the cadence.
How the Case Unfolds
Once the patient is on the table with IV access and either light sedation or, less commonly, a general anaesthetic, the procedure has its own rhythm. The doctor numbs the access site, gets a needle into the artery, and slides in a small plastic tube called a sheath, usually a 5 or 6 French (French is just a measurement of how thick the tube is). The sheath becomes the doorway for everything that follows.
Then the mapping. A longer catheter goes up to the area near the liver and contrast (the iodinated dye, Omnipaque in our department) is injected. The arteries light up white on the screen and the doctor confirms what they saw on the morning’s CT, which branches feed the tumour, where the variant anatomy is, what to avoid. The catheter shapes you’ll hear named here are Cobra, Simmons, Shepherd’s Hook. They all refer to the bend at the catheter’s tip, different shapes hook into different branches.
After the diagnostic catheter is settled in the main hepatic artery comes the moment that really decides how safe the procedure will be. A much thinner tube called a microcatheter is threaded inside the bigger one (a smaller straw inside a larger straw). Names you’ll see on these are Progreat, Renegade, Maestro. The microcatheter can be steered all the way out into the tiny branch feeding just the tumour. A cone-beam CT scan on the table (the C-arm spinning a full circle around the patient) confirms the tip is sitting exactly where it needs to be.
This step matters more than any other
The closer the catheter sits to the tumour, the safer the embolisation. It’s what keeps the agent away from the gallbladder, the stomach, the duodenum. The whole rest of the procedure rests on this being done well.
Then the embolisation itself. The agent is delivered slowly while everyone watches the screen. What the doctor is waiting for is stasis, the moment when the contrast stops flowing because the artery is blocked. A final injection confirms the target vessel is closed and nothing’s gone where it shouldn’t. The sheath comes out. The groin gets either firm pressure for 15 to 20 minutes or a small closure plug called an Angioseal. The wrist gets the TR Band.
A few practical notes that show up around this rhythm. If the patient’s kidneys are struggling (eGFR below 30) the doctor sometimes uses CO₂ as the contrast instead of iodinated dye, because iodinated contrast can damage already-fragile kidneys. The setup is different and worth flagging. And anything involving doxorubicin (which shows up in the chemoembolisation flavours below) is cytotoxic, so the usual PPE and spill-kit protocol applies.
The Four Flavours
There are four main ways to do the actual blocking step. The first three use particles, the fourth uses tiny radioactive beads. The interesting question is what each one is trying to do. They’re not just variations on a theme, they’re different mechanisms doing different work.
TAE: pure blockage, no drug
Imagine pouring sand into a pipe until water can’t get through. That’s TAE. Particles get injected into the artery until it’s physically blocked off. No chemotherapy, no radiation. The tumour dies because it’s been starved of blood. The entire anti-tumour effect is ischaemic (tissue death from lack of blood supply).
The agents you’ll see are Gelfoam (a gelatin sponge that the body breaks down over 3 to 6 weeks, so it’s a temporary blockage), PVA particles (permanent plastic particles in a range of sizes), or calibrated microspheres like Embosphere or Bead Block (perfectly round, more predictable than PVA, less prone to clumping). For a deeper dive on these specifically, see Bland TAE — Embolisation Agents.
When this gets chosen
TAE often comes up for tumours that have spread to the liver from neuroendocrine cancers. These are exceptionally hypervascular, so pure ischaemia works well. Also for patients who can’t tolerate chemotherapy.
cTACE: blockage plus chemotherapy in oil
This one’s more interesting than it first sounds. The trick is Lipiodol, an oil made from poppy seeds with iodine in it. Lipiodol has a strange property. When it’s injected into the liver’s blood supply, it gets absorbed into tumour tissue and stays there for weeks, while normal liver washes it away. It’s effectively a tumour-seeking substance.
Imagine mixing medicine into a thick syrup that sticks to the inside of a pipe and slowly leaks out for weeks. The doctor mixes chemotherapy (usually doxorubicin, a red liquid, in 50 to 75 mg doses) into Lipiodol to form an emulsion (typically 1:1 or 1:2 ratio of drug to oil). The mixture is injected, then particles go in afterwards to block the artery and trap the drug inside the tumour.
What makes Lipiodol so useful
Lipiodol shows up bright white on x-ray and CT. After the procedure, a CT scan shows exactly where the Lipiodol settled, which tells the team where the drug actually went. Bright spots in the tumour mean a good treatment. Faded areas mean leftover viable tumour that may need another go. Built-in quality control.
DEB-TACE: beads that release drug slowly
Take the chemotherapy idea but solve the delivery problem differently. Instead of mixing the drug into oil, pre-made polymer beads are loaded with doxorubicin before the procedure starts. The beads get injected, lodge in the tumour’s blood supply, block it physically, and slowly release the drug over the next month or so. Like a time-release capsule sitting right where it’s needed.
The bead products you’ll see: DC Bead is the most common, then HepaSphere (a polymer that swells in saline to adapt to vessel size), LifePearl, and Embozene TANDEM. Some of the newer ones (LC Bead LUMI, Embozene TANDEM) are radiopaque, which addresses the one thing cTACE does that DEB-TACE doesn’t, showing up on follow-up imaging.
The honest comparison with cTACE: studies show no clear winner for survival. DEB-TACE has lower systemic chemotherapy levels (so fewer chemo side effects), but you don’t get the Lipiodol “map” on follow-up CT. The choice often comes down to what a centre is set up for.
SIRTEX: tiny radioactive beads
What "SIRTEX" actually means
SIRTEX is what you’ll hear in almost every Australian IR department for this procedure, but technically it’s a brand name. Sirtex Medical is the Sydney-based company that makes SIR-Spheres, the resin-bead product used to deliver the treatment. The procedure category is TARE (trans-arterial radio-embolisation). The radioactive material itself is Y90 (Yttrium-90). The other product, used less often in Australia, is TheraSphere, glass beads from Boston Scientific. So “SIRTEX,” “Y90,” and “TARE” all get used interchangeably in conversation, but they’re not the same thing. SIRTEX is the brand, Y90 is the isotope, TARE is the procedure.
This one’s a different beast. It’s not really an embolisation in the same sense. The point isn’t to block, it’s to deliver radiation from inside the tumour, with the blood supply just being the delivery mechanism.
Imagine sprinkling thousands of tiny grains of sand into a stream, but each grain is mildly radioactive, and only fires its radiation a few millimetres in any direction. The grains lodge in the tumour’s tiny blood vessels and zap the tumour from within. Healthy tissue right next door is barely affected because the radiation doesn’t travel far. The radioactive material is Yttrium-90, which has a half-life of about 64 hours. Most of the dose is delivered in the first six days, then it’s gone.
The product you’ll be working with is almost always SIR-Spheres: resin microspheres, about 40 to 60 million per dose, each one carrying a small amount of activity (~50 Bq). The competitor, TheraSphere, uses far fewer beads (about 1.2 million) but each carries much more activity (~2,500 Bq) because it’s glass rather than resin. They’re approved for slightly different indications: SIR-Spheres for bowel cancer that’s spread to the liver, TheraSphere for primary liver cancer that can’t be cut out. In practice the choice is often dictated by what a centre is set up for.
Y90 needs serious planning before procedure day
A week or two beforehand, the patient has a separate “test run” called a MAA scan. A small dose of a radioactive tracer is injected the same way the Y90 will be, then a nuclear medicine scan checks where it went. The key thing it’s looking for is lung shunting, how much of the injected material accidentally reached the lungs. If too much would go there, Y90 can cause lung damage, and the procedure is cancelled or modified. The cutoff is around 20%. Above that, Y90 doesn’t happen.
The Y90 itself comes to the room in a shielded container, and the patient is mildly radioactive for a short period afterwards. The nuclear medicine and radiation safety team will brief on the specific patient-care precautions. The embolisation tradition this comes from is similar to the ones above (it’s still the doctor in the angio suite, still the same arterial pathway), but the post-procedure side has a whole separate radiation-safety overlay.
Y90 tends to win the modality choice when the portal vein is blocked by clot (the other procedures get too risky because the artery becomes the only blood supply), when there’s a lot of tumour spread through both lobes, when prior TACE hasn’t worked, or for cancers that started in the bowel or are neuroendocrine.
What Follows the Procedure
The most common thing isn’t really a complication, it’s the procedure working as intended. Within 24 to 72 hours most patients develop post-embolisation syndrome: fever, pain in the upper-right belly, nausea. It’s the body’s response to a dying tumour, and it’s expected. Painkillers and antiemetics, and it settles.
The thing the team is genuinely watching for is liver decompensation. Some abnormality in the liver bloods after the procedure is normal. But jaundice (the patient turning yellow), a swelling abdomen (ascites, fluid building up in the belly), or new confusion, that’s the liver struggling. Patients whose liver was already poor going in are at highest risk, which is the whole reason patient selection is taken so seriously upfront.
Things going to the wrong place
Non-target embolisation, particles drifting into the wrong artery, can damage the gallbladder (causing inflammation and pain), the stomach or duodenum (causing ulcers), or the bile ducts. The superselective microcatheter technique and the cone-beam CT confirmation step exist specifically to prevent this. When they’re skipped, this is what shows up.
Y90-specific things
Lung damage if too much went to the lungs (which the MAA scan should have caught). A condition called RILD (radiation-induced liver disease). Inflammation of the gallbladder or stomach lining if beads drifted off-target. Almost all of these are pre-empted by the planning scan rather than caught afterwards.
Access-site complications follow the route. Groin: bruising, swelling, occasionally a false aneurysm (a leaky pouch of artery wall). Wrist: artery spasm during the procedure (the doctor may struggle to advance the catheter), or the artery slowly closing afterwards, usually with no symptoms because the hand has another artery.
Numbers Worth Knowing
| Stat | Value | Why it matters |
|---|---|---|
| Patients with unusual liver artery anatomy | About 1 in 3 (36%) | Why pre-procedure CT review is non-negotiable |
| Normal liver’s portal vein supply | 75% | Why blocking the artery doesn’t kill the rest of the liver |
| Major bleeding from groin access | ~1.8% | Almost zero with wrist access |
| Y90 radiation reach in tissue | 2.5 mm on average, 11 mm max | Why it spares nearby healthy tissue |
| Y90 half-life | 64 hours | Most radiation delivered in the first 6 days |
| DEB-TACE drug release period | ~1 month | Slow steady release vs one big hit |
| Y90 lung shunt cutoff | 20% | Above this, Y90 can’t be done, lung damage risk |
How the Modality Gets Chosen
No single answer
The choice between TAE, cTACE, DEB-TACE, and Y90 depends on liver function, how widespread the tumour is, whether the portal vein is patent, and what’s been tried before. There are patterns, but no algorithm.
A rough sketch of the decision space. Anatomy mapping comes first regardless, and cone-beam CT during the case is becoming the standard. Between cTACE and DEB-TACE the studies don’t show a clear winner, and the choice is usually centre-preference. Y90 wins when the portal vein is blocked, when there’s a lot of tumour, or when TACE has already been tried. And these are usually not one-and-done procedures. Most patients come back for repeat treatments based on how the tumour responds.
Related
- Bland TAE — Embolisation Agents, deep dive on the particles used in TAE and TACE
- Genicular Artery Embolisation, same arterial-catheter approach, different territory (knee osteoarthritis)
- Acetabulum Tumour Embolisation, another tumour embolisation, different anatomy
Sources
- Hepatic Chemoembolization — StatPearls (NIH)
- Transarterial Radioembolization for HCC — PMC
- Anatomy of Liver Arteries for IR — ScienceDirect
- Variations of Celiac Axis and CHA — PMC
- All You Need to Know About TACE — MDPI
- Distal Radial Access for TACE — PMC
- Radial vs Femoral TACE — SAGE Journals 2024
- TAE and TACE for HCC — PMC
- Drug-Eluting Beads: State-of-the-Art — PMC
- Intra-Arterial DEB Therapy — PMC
- Y90 Radioembolization of HCC and Mets — PMC
- TARE with Y90 for HCC — PMC
- Transcatheter Arterial Chemoembolisation — Radiopaedia
- Embolisation for Liver Cancer — Cancer Research UK
- CHA Catheter Insertion via Celiac Trunk Morphology — PMC
Last updated May 18, 2026.