Maarten Smits, UMC Utrecht

Electron microscope image of resin 90Y-microspheres




Radioembolization (RE) is a form of internal radiation therapy for tumors in the liver.

RE is performed with microspheres containing radioactive yttrium-90 or holmium-166 that are administered in the hepatic arteries from an endovascular catheter.

The liver has a dual blood supply. The liver parenchyma relies mostly on the portal vein for blood supply. In contrast, tumors arising in the liver are predominantly fed by the hepatic artery.1,2

Table 1. Tumor types


Table 2. Gross selection criteria

Patient selection

RE can be performed in patients with all kinds of tumors in the liver, ranging from primary tumors such as hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCC) to liver metastases from colorectal carcinoma, breast cancer, neuroendocrine tumors and melanoma, see table 1.




There is much discussion about the exact selection criteria for RE. Table 2 lists the gross selection criteria that many centers use.3

More information is available on the websites of the manufacturers of yttrium-90 microspheres: SIR-Spheres(R) and Therasphere(R). QuiremSpheres(R), containing holmium-166 are not commercially available yet.

Aberrant left hepatic artery arising from the left gastric artery (arrow)

Vascular anatomy

Before proceeding with a work-up angiography it is advisable to have a look at the patient’s vascular anatomy. The arterial anatomy of the liver region can vary widely between patients (see figure). A thin slice CT angiography viewed in MIP-mode can help to identify variant anatomy.4,5

Figures from ref 5 with permission from Springer

Work-up procedure in the angiosuite.

Digital subtraction angiography with contrast injection in the common hepatic artery. 
Injection position of 99mTc-MAA in the right hepatic artery after coil embolization of the gastroduodenal artery (long arrow) and the right gastric artery (short arrow). Subsequent injection position in the left hepatic artery. Images reproduced from Smits et al. Eur J Pharm 2013, with permission from Elsevier. 

Work-up angiography

  1. Perform a contrast run from the celiac axis and the superior mesenteric artery (for abberant vessels).
  2. Assess the vascular anatomy. The anatomy should have been assessed on CTA at this point.
  3. Embolization of non-target vessel (see subtopic)
  4. Perform C-arm CT (see subtopic)
  5. Inject 99mTc-MAA at the desired injection positions

Panel a displays a digital subtraction angiography with contrast injection in the common hepatic artery in a breast cancer liver metastases patient. Panel b: 1. common hepatic artery; 2. gastroduodenal artery; 3. right gastric artery; 4. left hepatic artery, 5. right hepatic artery. Tumorous areas are marked with ‘T’, catheter as a white dotted line.

Embolization of non-target vessels

Traditionally, all non-target vessels arising near to the injection position are coil-embolized. These vessels typically include the gastroduodenal artery (GDA) and right gastric artery (RGA). Currently, most centers decide whether or not to close these vessels based on the vascular anatomy (e.g. Where is the origin of the GDA? Is there retrograde flow in the GDA?) and intended injection position(s). Also, the use of an anti-reflux catheter may decrease the risk of back-flow and thus the need of closing the non-target vessels.


Cystic artery

The cystic artery can pose a dilemma. The cystic artery often arises from the right hepatic artery and closing this vessel to prevent radiation cholecystitis may be tempting. Closure may however cause ischemic cholecystitis. The consensus of the available literature on this topic seems to be that in case of visually (so no dosimetry needed) high 99mTc-MAA deposition in the gallbladder, one should seek a more distal injection position. If that is not possible, one can choose to (partially) close the cystic artery with coils or gelfoam, preferably at the time of the treatment angio to avoid formation of new collateral vessels.6,7

C-arm CT used to ensure complete tumor coverage. Depicted is a neuroendocrine tumor, supplied by the left- (in red) and right hepatic artery (in blue/green).

Cone-beam CT

One can use the C-arm at the angio suite to acquire cross-sectional CT images. This technique is called C-arm CT or cone-beam CT. The combination of CT with contrast injected via an endovascular catheter can be useful in several ways:

  1. Non-target embolization detection:
    Is there any enhancement of non-target organs?
  2. Selecting the target vessel supplying a tumor
    In case of multiple vessels projecting on top of each other on 2D-angiography, the target vessel can be identified using C-arm CT.
  3. Assessment of the target volume:
    What is the size of the volume that is reached from a certain injection position?
  4. Tumor coverage:
    Is the tumor completely covered? A tumor may be supplied by multiple vessels from within the liver or outside the liver A large variety of parasitic vessels have been described, particularly in HCC. Incomplete tumor coverage will lead to inadequate treatment.

99mTc-MAA SPECT and corresponding 90Y-SPECT. Images reproduced from Smits et al. Eur J Pharm 2013, with permission from Elsevier. 

99mTc-MAA imaging

After injection of 99mTc-MAA at the work-up angio, the distribution is assessed using nuclear imaging.


Non-target embolization

Distribution of 99mTc-MAA anywhere outside the liver can be detected with 99mTc-MAA-SPECT and should be resolved before treatment. Regions of interest are the duodenum, stomach, pancreas, falciform ligament and umbilicus. Free pertechnetate can cause diffuse high activity in the stomach simulating non-target embolization. This can be recognized by its characteristic pattern and accompanying high uptake in the thyroid gland. With the growing experience of many IR-teams and the growing use of C-arm CT, non-target embolization is now more and more frequently detected already during angio. Some experts have even gone as far as stating that 99mTc-MAA is only necessary for lung shunt detection.


Lung shunt

Hepatopulmonary shunting can be evaluated on planar imaging. In fact, the limit of a maximum 20% lung shunt fraction is based on studies using planar imaging. However, it is clear that attenuation-corrected Single Photon Emission Computed Tomography (SPECT) is far more accurate than planar imaging. Even when using SPECT, it seems that 99mTc-MAA tends to overestimate the amount of lungshunting when compared with the actual microspheres.8



a. Administration of 90Y-microspheres in the angio-suite. b. Therasphere administration system

Treatment angiography

After the work-up angiography and 99mTc-MAA imaging, the patient returns for the actual treatment procedure. This is generally performed 1 or 2 weeks after the work-up angiography.

In the angiosuite, a microcatheter is placed in the exact same position(s) as where the 99mTc-MAA has been injected. Next, the administration system holding a vial that contains the radioactive microspheres is attached to the microcatheter. The administration system consists of a perspex box for SirSpheres, Theraspheres, and Quiremspheres (perspex absorbes most beta radiation), see images.

The microspheres are brought into suspension and pushed forward through the tubing system and into the microcatheter. Depending on the type of microspheres and type of tumor, one should monitor the forward flow distal to the catheter to avoid reflux.

Post-treatment dosimetry

Post treatment imaging to visualize the distribution of the radioactive microspheres can be useful in order to confirm an adequate dose to the tumors and no non-target deposition.

Post treatment imaging for yttrium-90 microspheres can be performed with Bremsstrahlung SPECT (indirect gamma-emission from decelerating beta-particles) and with PET (32 positrons per million decays), see figure A.

Holmium-166 microspheres on the other hand can be visualized with SPECT (direct gamma-emission) and with MRI (holmium is a paramagnetic element). See figure B.


  • 1. Bierman HR, Byron RL, Jr., Kelley KH, et al. Studies on the blood supply of tumors in man. III. Vascular patterns of the liver by hepatic arteriography in vivo. Journal of the National Cancer Institute 1951; 12(1)
  • 2. Dezso K, Bugyik E, Papp V, et al. Development of arterial blood supply in experimental liver metastases. The American journal of pathology 2009; 175(2)
  • 3. Coldwell D, Sangro B, Wasan H, et al. General selection criteria of patients for radioembolization of liver tumors: an international working group report. American journal of clinical oncology 2011; 34(3)
  • 4. van den Hoven AF, Smits ML, de Keizer B, et al. Identifying aberrant hepatic arteries prior to intra-arterial radioembolization. Cardiovascular and interventional radiology 2014; 37(6)
  • 5. van den Hoven AF, van Leeuwen MS, Lam MG, et al. Hepatic arterial configuration in relation to the segmental anatomy of the liver; observations on MDCT and DSA relevant to radioembolization treatment. Cardiovascular and interventional radiology 2015; 38(1)
  • 6. Prince JF, van den Hoven AF, van den Bosch MA, et al. Radiation-induced cholecystitis after hepatic radioembolization: do we need to take precautionary measures? Journal of vascular and interventional radiology : JVIR 2014; 25(11)
  • 7. McWilliams JP, Kee ST, Loh CT, et al. Prophylactic embolization of the cystic artery before radioembolization: feasibility, safety, and outcomes. Cardiovascular and interventional radiology 2011; 34(4)
  • 8. Elschot M, Nijsen JF, Lam MG, et al. ((9)(9)m)Tc-MAA overestimates the absorbed dose to the lungs in radioembolization: a quantitative evaluation in patients treated with (1)(6)(6)Ho-microspheres. European journal of nuclear medicine and molecular imaging 2014; 41(10)

Keywords: subjects