Safety of Radioembolization with 90Yttrium Resin Microspheres Depending on Coiling or No-Coiling of Aberrant/High-Risk Vessels
- First Online:
- Cite this article as:
- Paprottka, P.M., Paprottka, K.J., Walter, A. et al. Cardiovasc Intervent Radiol (2015) 38: 946. doi:10.1007/s00270-015-1128-x
- 680 Downloads
To evaluate the safety of radioembolization (RE) with 90Yttrium (90Y) resin microspheres depending on coiling or no-coiling of aberrant/high-risk vessels.
Materials and Methods
Early and late toxicity after 566 RE procedures were analyzed retrospectively in accordance with the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE v3.0). For optimal safety, aberrant vessels were either coil embolized (n = 240/566, coiling group) or a more peripheral position of the catheter tip was chosen to treat right or left liver lobes (n = 326/566, no-coiling group).
Clinically relevant late toxicities (≥Grade 3) were observed in 1 % of our overall cohort. The no-coiling group had significantly less “any” (P = 0.0001) or “clinically relevant” (P = 0.0003) early toxicity. There was no significant difference (P > 0.05) in delayed toxicity in the coiling versus the no-coiling group. No RE-induced liver disease was noted after all 566 procedures.
RE with 90Y resin microspheres is a safe and effective treatment option. Performing RE without coil embolization of aberrant vessels prior to treatment could be an alternative for experienced centers.
KeywordsInterventional oncology Radioembolization Liver/hepatic Cancer
As the incidence of primary and metastatic liver cancer continues to increase [4, 12], the use of minimally invasive techniques as a treatment option is becoming more and more common. Radioembolization (RE) is a therapy method where particles of glass or resin, containing 90yttrium (90Y), are infused intra-arterial through a catheter directly into the hepatic arteries. RE is based on the two-vessel supply of the liver. Hepatic malignancies receive their blood supply mainly from the hepatic artery, whereas the normal liver parenchyma is perfused predominantly from the portal-venous circulation . Accordingly, a high-dose delivery to the tumor vasculature with relative sparing of normal liver tissue can be achieved, which has shown to be effective for primary and metastatic liver tumors [8, 10, 11, 14, 18, 31].
In the literature, several adverse events related to hepatic intra-arterial treatments and RE were reported, especially in the gastroduodenal region. Many of these complications arise from an unintended extra-hepatic deposition of 90Y particles (non-target embolization) via arteries originating from or close to hepatic arterial branches.
Therefore, it is necessary to know the patient’s individual anatomy including frequent variations such as a hepato-mesenteric trunk or gastro-hepatic trunk and to detect high-risk vessels such as the gastroduodenal artery, right and left gastric artery, cystic artery, inferior esophageal arteries, falciform artery, and supra-duodenal arteries [2, 17, 22].
Aberrant implantation of the radioactive particles may result in serious consequences. The reasons for aberrant implantation are often collaterals that are arising from the hepatic arteries distal to the therapy position or branching close to the therapy position with an extra-hepatic communication. Radiation and diminished blood supply due to non-target embolization may result in ischemia and inflammation leading to ulceration or even perforation of the stomach or duodenum, or in cholecystitis or pancreatitis [16, 32]. Therefore, up to now, it is recommended to protect potential aberrant vessels like the gastroduodenal artery by coil embolization even if they are proximal to the therapy position.
Nevertheless several experienced centers already perform RE without aggressive prior coiling of proximal at-risk vessels, if they deem the therapy position to be safe with regard to potential back-flow. Until now, no data exist to demonstrate whether this more deliberate coil embolization policy has an effect on the occurrence of early or late toxicity after RE.
In our study, we retrospectively compared patients who underwent RE with prior coiling with patients without prior coiling with regard to the appearance of early and late toxicities according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE v3.0).
Materials and Methods
The Institutional Review Board approved this retrospective analysis. 566 REs, performed using 90Y resin microspheres between 01/2009 and 12/2012, were retrospectively analyzed. We evaluated 240 RE procedures with coiling and 326 without coiling with regard to early and late toxicities.
Early (first week) and late toxicities (up to 6 months) were retrospectively categorized according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE v3.0). Patients were routinely followed-up in our institution 1–3 days after RE as well as 1, 6, 12, and 24 weeks after procedure. This follow-up included contact to the referring and family physicians to detect potentially missed complications. In accordance with CTCAE toxicities ≥Grade 3 were defined as clinically relevant.
Non-resectable primary or secondary liver cancer;
Absence of significant extrahepatic disease;
Failure to respond to other types of medical, surgical, or local ablative treatment modalities;
Adequate biochemical and hematological function and arteriovenous shunting ≤20 % to the lung vascular bed; and
Written informed consent.
90Y is a pure beta-emitter that decays to stable 90zirconium with an average energy of 0.94 MeV (half-life 2.67 days) with a maximum range of 11 mm in soft tissue (mean tissue penetration: 2.5 mm). The principle is the preferential tumor-related vascular distribution of the radioactive microspheres, which allows delivery of high doses within the tumor and relative sparing of normal liver tissue.
In our study, only 90Y resin microspheres (SIR-Spheres; Sirtex, Lane Cove, Australia) were used. The standard dose of 90Y resin microspheres, which contains approximately 35 million microspheres, is 1.5–2 GBq, (each microsphere contains 50 Bq activity).
The patients in this cohort were treated employing the partition modeling method because this model enables a more precise dose calculation. This more personalized dose calculation reduces the rare occurrence of radiation-induced liver disease, a late effect of excessive hepatic radiotherapy. To avoid an overestimating of the required activity in patients with prior surgical resections, we calculated the exact liver- and tumor-volume using a special software application (OncoTREAT, MeVis, Bremen, Germany).
Arteriography was performed via a transfemoral approach. The presence of angiographically occult afferent extrahepatic arteries and the magnitude of hepatopulmonary shunting were evaluated by hepatic arterial injection of approximately 100 MBq [99mtechnetium] macroaggregated albumin ([99mTc]-MAA) in the left, right, and sometimes segment 4 hepatic artery separately according to the planned catheter position for later treatment. In our institution, we do not perform any coil embolization during the [99mTc]-Angiography, because from our experience often new collaterals have developed 1–2 weeks later at the time of scheduled RE. [99mTc]-Angiography is followed by planar scintigraphy including the lung and liver to calculate the liver-to-lung shunt and single-photon emission-computed tomography (SPECT) or SPECT/CT to evaluate the distribution of the particles.
In case of non-target embolization, we tried to identify those vessels in the angiograms. If we could not find any vessels being responsible for these results, we performed a pre-treatment Dyna CT to identify those aberrant vessels.
When deemed a potential risk for extrahepatic deposition of microspheres and we could not place our catheter distal to the origin with a sufficient safety margin, vascular occlusion using microcoil embolization of the vessels was performed (“coiling group”), even if we did not see non-target embolization within the SPECT/CT. We did not perform a re-scan after adequate embolization of those vessels, if the non-target embolization seen in the SPECT/CT matched the vascular territory of the vessel in the angiograms/Dyna CT.
In patients where we considered it to be safe to deliver the microspheres without coil embolization by placing the microcatheter distal to the origin of the last visible extrahepatic arterial branch and when the SPECT/CT did not show non-target embolization, we performed the RE without further coil embolization (no-coiling group). The decision was at the discretion of the angiographer and included factors like distance from the catheter tip to the extrahepatic artery, quality of the antegrade flow in the treated hepatic artery and presence of potential “safety” vessels.
The 90Y resin microspheres, suspended in sterile water, were injected under intermittent fluoroscopic visualization, alternating with contrast medium, to assess antegrade hepatic arterial flow. 90Y-microsphere activity was pre-scribed according to the instructions for use, calculated with the partition modeling method, and administered as described elsewhere by way of either a one- or two-stage lobar treatment according to the tumor burden. If patients had only metastases in one lobe, they underwent an unilobar therapy. Patients with tumor distribution in the whole liver underwent sequential whole liver treatment within an interval of 4–6 weeks. In all cases the left, right, and if applicable the segment IV hepatic artery were catheterized separately and accordingly treated separately. We did not perform RE from the common or proper hepatic artery.
After RE, SPECT, or SPECT/CT scans were performed within 24 h of therapy delivery to confirm target deposition of the therapeutic material (so called Bremsstrahlung-SPECT).
Pretreatment and post-treatment laboratory tests included liver function tests, complete blood counts, and tumor markers. Patients resumed a routine schedule of laboratory tests and clinical examination at day 1–3, after 1, 6, 12, and 24 weeks. The most recent history was taken for side effects; these data were converted to a toxicity score according to CTCAE (version 3.0) and analyzed according to early (within the first week) and late toxicities (up to 6 months) retrospectively. Clinical relevant toxicities were defined as toxicities of level ≥3.
Early toxicities included nausea, vomiting, abdominal pain, and fever. Late toxicities included gastritis, pancreatitis, cholecystitis, pulmonary fibrosis, skin necrosis, and RE-induced liver disease (REILD). The quantitative data were compared with baseline results.
Statistical analyses were performed by using the Chi square test with a significance accepted at P ≤ 0.05. SPSS 20 for MAC (SPSS, Chicago, IL, USA) was used for data management and statistical analysis.
Significantly less “any” (Grade 1–3) early complications were seen for the no-coiling versus coiling group (P < 0.0001/contingency coeff. = 0.241)
144 (60 %)
96 (40 %)
114 (35 %)
212 (65 %)
31 (13 %)
14 (4 %)
312 (96 %)
There were significantly less “any” (Grade 1–3) early complications (P < 0.0001/Contingency coeff. = 0.241) and significantly less “clinical relevant” (≥Grade 3) early complications (P < 0.0003/contingency coeff. = 0.149) for the no-coiling group (Table 1).
No significant differences could be observed for “any” (P = 0.2698/contingency coefficient = 0.046) late complications
18 (8 %)
222 (92 %)
16 (5 %)
310 (95 %)
3 (1 %)
237 (99 %)
3 (1 %)
323 (99 %)
Comparing the coiling and the no-coiling group, no significant differences could be observed for “any” (P = 0.2698/contingency coeff. = 0.046) and “clinically relevant” (P = 0.9707/Contingency coeff. = 0.002) toxicities (Table 2).
After all 566 RE-procedures no REILD occurred.
Radioembolization with 90Y microspheres is known to be a safe and effective therapy for patients as long as the microspheres are delivered into the liver and nowhere else. There is a lot of data existing about the effectiveness of radioembolization therapy in patients suffering from hepatocellular carcinoma [6, 26] and intrahepatic cholangiocarcinoma  as well as patients with unresectable liver metastases of colorectal cancer [9, 27], breast carcinoma [7, 23], pancreatic carcinoma , and neuroendocrine tumors [8, 13, 31].
In these cases, the treating centers performed a “state of the art” RE in accordance to the guidelines with a coiling of the gastroduodenal artery (GDA), the cystic artery as well as all possible aberrant vessels being visible at the Tc99 m angiography. Regarding the coiling of the cystic artery McWilliams et al.  reported, that the embolization of the proximal cystic artery is a safe and feasible procedure and may be performed for liver-directed embolotherapy in order to minimize the exposure of the gallbladder to particulate, chemoembolic, or radioembolic agents.
Until now, no data exist on safety, success and toxicities post-radioembolization without prior coiling, especially comparing to state of the art radioembolization with pre-treatment coiling.
Our data show for the first time that radioembolization without prophylactic coiling leads to a significant reduction of “any” (Grade 1–2) and “clinically relevant” (≥Grade 3) early complications in the no-coiling group and comparable late toxicities between the two groups.
Nevertheless, in some cases unexpected complications can appear. These can be on the one hand complications due to the RE treatment [5, 15, 24, 29, 30] and on the other hand complications due to coil embolization associated with the therapy [1, 19, 25, 28].
Peterson et al.  performed a study in order to estimate the incidence of complications after 90Y microsphere RE for unresectable hepatic tumors. They could show that 78 patients (70 %) experienced a post-radioembolization syndrome (fatigue, abdominal pain, nausea, vomiting, anorexia or fever). 3 patients (3 %) suffered on a Grade 3 early complication. 2 % of the patients experienced clinically significant liver dysfunction; 11 patients (10 %) had gastrointestinal ulceration and 7 patients (6 %) suffered from a cholecystitis. Grade 2 pancreatitis occurred in 1 patient (1 %). After 12 months the cumulative incidence of late Grade 3 or 4 complications was 8 %. No Grade 5 toxicity occurred. These results are in correlation with our study where no Grade 4 or 5 early toxicities occurred. Peterson et al. postulated that radioembolization is a well-tolerated treatment for unresectable hepatic tumors with a low risk of Grade 3 or higher early or late toxicity.
These findings are consistent with our results were in 258 of 566 radioembolizations (45.6 %) early toxicities (7.9 % clinical relevant) and in 34 of 566 radioembolizations (6.0 %) late toxicities (1.06 % clinical relevant) were found.
In accordance to our results, Gil-Alzugaray  reported no relevant RILD in the treated patients with normal liver parenchyma, but in patients with cirrhosis or prior and subsequent chemotherapy.
Not only complications due to the procedure itself, but also due to coil embolization of aberrant vessels are possible. It is known that due to hemodynamic changes, coil embolization can result in a recruitment of previous existing but formerly not relevantly perfused hepato-intestinal collaterals (HICs), a reopening of coiled vessels or side branches, or even the development of new HICs, that can increase the risk of gastrointestinal complications following SIRT .
Schelhorn et al.  investigated the effect of the occlusion of the gastroduodenal artery (GDA) prior to RE with regard to arterial hepato-intestinal collateralization (HIC). They scheduled 606 patients, where the GDA was occluded in 86 patients (22 did not undergo RE due to their clinical performance). They found reopened or newly apparent HIC in 28 patients. In 25 of these 28 patients, it was possible to occlude these HIC or to choose another catheter position for a safe Y90 application. Although in the majority of the investigated cases RE was still feasible after recoiling or modification of the microcatheter position, in 3 (5 %) of patients, it was not possible and RE had to be abandoned. Hence, coiling of the GDA at the time of [99mTc]-Angiography with consecutively developing or newly forming HICs could be a problem for subsequent radioembolization. As a conclusion, RE without prior coiling of the GDA could be the safer treatment option.
Petroziello et al.  also investigated the rate of recanalization and collateral vessel formation after a side-branch embolization during Technetium angiography for planned 90Y RE. They reported that after side-branch embolization, eight of their 56 patients (14 %) presented later on with recanalized or newly developed HICs and seven out of 110 primarily coiled vessels (6 %) had newly developed collaterals.
As performing radioembolization without prior coiling is a possible and safe treatment option for experienced centers in accordance to our results with significant less early toxicities, we propose a step-by-step approach.
Our study has several limitations.
Data collection was carried out as a retrospective data analysis. Thus, we can not exclude a selection bias. Moreover, there have not been prospectively chosen, strict criteria whether to embolize or not. Secondly, our results represent only data of a single-center experience. Data of early and late toxicities after radioembolization without coiling from other centers are needed. Finally there might be a selection bias with regard to the classification of the patients in the “coiling” or “non coiling” group. We can assume that the latter had a more favorable and easier anatomy. Otherwise a coiling would have been performed. The results apply only to such a treatment approach as it is performed in our department. A possible bias could be the increasing experience of the interventional radiologist during the time course of this study. RE was done by a core team with already existing experience, as our center has performed a high number of RE since 2002. Therefore, we can rule out any bias regarding this issue.
In conclusion, radioembolization with 90Yttrium resin microspheres is a safe treatment option. Clinically relevant late toxicities appear only in 1 % of the treatments. We observed significantly less “any” (Grade 1–3) and “clinically relevant” (≥Grade 3) early complications in the no-coiling group. No significant differences in the late toxicities between the two groups were found. Nevertheless, a rigorous angiographic work up with identification of all potentially vessels at risk needs to be performed before RE. If a position of the catheter distal to the origin of aberrant vessels is not possible or if there are any doubts, coiling of these vessels is necessary and useful.
Conflict of interest
All authors have no conflict of interest.
Statement of Informed Consent
Consent was obtained from all individual participants included in the study.
Statement of Human and Animal Rights
For this type of study formal consent is not required.