I.  Radioembolization (SIRT) may be considered medically necessary to treat primary hepatocellular carcinoma that is unresectable and limited to the liver (see Policy Guidelines section).

II.  Radioembolization (SIRT) may be considered medically necessary in primary hepatocellular carcinoma as a bridge to liver transplantation. 

III.  Radioembolization (SIRT) may be considered medically necessary to treat hepatic metastases from neuroendocrine tumors (carcinoid and noncarcinoid) with diffuse and symptomatic disease when systemic therapy has failed to control symptoms. 

IV.  Radioembolization (SIRT) may be considered medically necessary to treat unresectable hepatic metastases from colorectal carcinoma, melanoma (ocular or cutaneous), or breast cancer that are both progressive and diffuse, in patients with liver-dominant disease who are refractory to chemotherapy or are not candidates for chemotherapy or other systemic therapies. 

V.  Radioembolization (SIRT) may be considered medically necessary to treat primary intrahepatic cholangiocarcinoma in patients with unresectable tumors.

VI.  Radioembolization (SIRT) is considered investigational for all other hepatic metastases except as noted above.

VII.  Radioembolization (SIRT) is considered investigational for all other indications not described above

VIII.  BCBSNE considers preoperative embolization of bone metastasis of Renal Cell Carcinoma, Thyroid cancer investigational because the effectiveness of this approach has not been established.

In general, radioembolization is used for unresectable HCC that is greater than 3 cm.

There is little information about the safety or efficacy of repeated RE treatments or about the number of treatments that should be administered.

Radioembolization should be reserved for patients with adequate functional status (Eastern Cooperative Oncology Group [ECOG] Performance Status 0-2), adequate liver function and reserve, Child-Pugh score A or B, and liver-dominant metastases.

Symptomatic disease from metastatic neuroendocrine tumors refers to symptoms related to excess hormone production.

The coding for radioembolization may depend on the medical specialty that is actually providing the therapy. The following CPT codes might possibly be used:

79445: Radiopharmaceutical therapy, by intra-arterial particulate administration;

77778: Interstitial radiation source application; complex;

75894: Transcatheter therapy, embolization, any method, radiological supervision and interpretation.

Beginning in 2014, the following code is available for the embolization procedure

37243: Vascular embolization or occlusion, inclusive of all radiological supervision and interpretation, intraprocedural roadmapping, and imaging guidance necessary to complete the intervention; for tumors, organ ischemia, or infarction.

Prior to 2014, the following code was available:

37204: Transcatheter occlusion or embolization (e.g., for tumor destruction, to achieve hemostasis, to occlude a vascular malformation), percutaneous, any method, non-central nervous system, non-head or neck; 

Since this therapy involves radiotherapy, a variety of radiotherapy planning codes may be a component of the overall procedure. For example, CPT code 77399 (unlisted procedure, medical radiation physics, dosimetry and treatment devices, and special services) may be used.  In 2004, the following HCPCS code was introduced for radioembolization: 

S2095: Transcatheter occlusion or embolization for tumor destruction, percutaneous, any method, using yttrium-90 microspheres.

The use of external beam radiotherapy and the application of more advanced radiotherapy approaches (eg, intensity-modulated radiotherapy) may be of limited use in patients with diffuse, multiple lesions due to the low tolerance of normal liver to radiation compared with the higher doses of radiation needed to kill the tumor. 

Various nonsurgical ablative techniques have been investigated that seek to cure or palliate unresectable hepatic tumors by improving locoregional control. These techniques rely on extreme temperature changes (cryosurgery or radiofrequency ablation [RFA]), particle and wave physics (microwave or laser ablation), or arterial embolization therapy including chemoembolization, bland embolization, or RE. 

RE, referred to as SIRT in older literature, is the intra-arterial delivery of small beads (microspheres) impregnated with yttrium-90 via the hepatic artery. The microspheres, which become permanently embedded, are delivered to tumor preferentially to normal liver, as the hepatic circulation is uniquely organized, whereby tumors greater than 0.5 cm rely on the hepatic artery for blood supply while normal liver is primarily perfused via the portal vein. Yttrium-90 is a pure beta-emitter with a relatively limited effective range and short half-life that helps focus the radiation and minimize its spread. Candidates for RE are initially examined by hepatic angiogram to identify and map the hepatic arterial system. At that time, a mixture of technetium 99-labelled albumin particles is delivered via the hepatic artery to simulate microspheres. Single photon emission computed tomography imaging is used to detect possible shunting of the albumin particles into gastrointestinal or pulmonary vasculature.

Currently, 2 commercial forms of yttrium-90 microspheres are available: a glass sphere, TheraSphere^®

(manufactured by Nordion, Ontario, Canada, under license by BTG International) and a resin sphere,

SIR-Spheres^® (Sirtex Medical, Lake Forest, IL). Noncommercial forms are mostly used outside the United States. While the commercial products use the same radioisotope (yttrium-90) and have the same target dose (100 Gy), they differ in microsphere size profile, base material (ie, resin vs glass), and size of commercially available doses. The physical characteristics of the active and inactive ingredients affect the flow of microspheres during injection, their retention at the tumor site, spread outside the therapeutic target region, and dosimetry calculations. FDA granted premarket approval (PMA) of SIR-Spheres^® for use in combination with 5-floxuridine chemotherapy by hepatic arterial infusion (HAI) to treat unresectable hepatic metastases from CRC. In contrast, TheraSphere^® was approved by humanitarian device exemption (HDE) for use as monotherapy to treat unresectable HCC. In January 2007, this HDE was expanded to include patients with HCC who have partial or branch portal vein thrombosis. For these reasons, results obtained with 1 product do not necessarily apply to other commercial (or noncommercial) products (see Regulatory Status section).

Unresectable Primary HCC

Most patients with HCC present with unresectable disease, and treatment options are limited secondary to the chemoresistance of HCC and the intolerance of normal liver parenchyma to tumoricidal radiation doses. Results of 2 (RCTs have shown a survival benefit for TACE therapy compared with supportive care in patients with unresectable HCC.^1,2 In 1 study, patients were randomly assigned to TACE, transarterial embolization (TAE), or supportive care. One-year survival rates for TACE, TAE, and supportive care were 82%, 75%, and 63%, respectively, and 2-year survival rates were 63%, 50%, and 27%, respectively. 

Targeted therapies have been investigated for HCC. For example, sorafenib was associated with improved OS in a randomized phase 3 trial with 602 patients.³ 

Unresectable Intrahepatic Cholangiocarcinoma

Cholangiocarcinomas are tumors that arise from the epithelium of the bile duct and are separated into intrahepatic and extrahepatic types. Intrahepatic cholangiocarcinomas appear in the hepatic parenchyma and are also known as peripheral cholangiocarcinomas. Resection is the only treatment with the potential for cure, and 5-year survival rates have been in the range of 20% to 43%.⁴ Patients with unresectable disease may select among fluoropyrimidine-based or gemcitabine-based chemotherapy, fluoropyrimidine chemoradiation or best supportive care.

Unresectable Metastatic CRC

Fifty to sixty percent of patients with CRC will develop metastases, either synchronously or metachronously. Select patients with liver-only metastases that are surgically resectable can be cured, with some reports showing 5-year survival rates exceeding 50%. Emphasis on treating these patients with potentially curable disease is on complete removal of all tumor with negative surgical margins. Most patients diagnosed with metastatic colorectal disease are initially classified as having unresectable disease. In patients with metastatic disease limited to the liver, preoperative chemotherapy is sometimes used in an attempt to downsize the metastases to convert the metastatic lesions to a resectable status (conversion chemotherapy).

In patients with unresectable disease that cannot be converted to resectable disease, the primary treatment goal is palliative, with survival benefit shown with both second- and third-line systemic chemotherapy.⁵ Recent advances in chemotherapy, including oxaliplatin, irinotecan, and targeted antibodies like cetuximab, have doubled the median survival in this population from less than 1 year to more than 2 years.⁵ Palliative chemotherapy by combined systemic and HAI may increase disease-free intervals for patients with unresectable hepatic metastases from CRC. 

RFA has been shown to be inferior to resection in local recurrence rates and 5-year OS and is generally reserved for patients with potentially resectable disease that cannot be completely resected due to patient comorbidities, location of metastases (ie, adjacent to a major vessel), or an estimate of inadequate liver reserve following resection. RFA is generally recommended to be used with the goal of complete resection with curative intent.⁶ The role of local (liver-directed) therapy (including RE, chemoembolization, and conformal radiotherapy) in debulking unresectable metastatic disease remains controversial.⁶

Unresectable Metastatic Neuroendocrine Tumors

Neuroendocrine tumors are an uncommon, heterogeneous group of mostly slow-growing, hormonesecreting malignancies, with an average patient age of 60 years. Primary neuroendocrine tumors vary in location, but most are either carcinoids (which most commonly arise in the midgut) or pancreatic islet cells. Carcinoid tumors, particularly if they metastasize to the liver, can result in excessive vasoactive amine secretion including serotonin and are commonly associated with the carcinoid syndrome (diarrhea, flush, bronchoconstriction, right valvular heart failure). 

Although they are considered to be indolent tumors, at the time of diagnosis, up to 75% of patients have liver metastases, and with metastases to the liver, 5-year survival rates are less than 20%. Surgical resection of the metastases is considered the only curative option; however, less than 10% of patients are eligible for resection, as most patients have diffuse, multiple lesions. 

Conventional therapy is largely considered to be palliative supportive care, to control, eradicate, or debulk hepatic metastases, often to palliate carcinoid syndrome or local pain from liver capsular stretching. Therapies for unresectable metastatic neuroendocrine tumors include medical (somatostatin analogs like octreotide), systemic chemotherapy, ablation (radiofrequency or cryotherapy), TAE or TACE, or radiation. Although patients often achieve symptom relief with octreotide, the disease eventually becomes refractory, with a median duration of symptom relief of approximately 13 months, with no known effect on survival. Systemic chemotherapy for these tumors has shown modest response rates of lcarcinoids, and is frequently associated with significant toxicity.⁷ Chemoembolization has shown response rates of nearly 80%, but the effect is of short duration and a survival benefit has not been demonstrated.⁷ 

Miscellaneous Metastatic Tumors

Small case reports have been published on the use of RE in many other types of cancer with hepatic metastases, including breast, melanoma, head, and neck (including parotid gland), pancreaticobiliary, anal, thymic, thyroid, endometrial, lung, kidney, gastric, small bowel, esophageal, ovarian, cervical, prostatic, bladder, and for sarcoma and lymphoma.⁸  

This policy was created in 2003, and has been updated annually with a MEDLINE search, most recently through January 30, 2015.

Unresectable Primary Hepatocellular Carcinoma

The evidence related to the use of radioembolization (RE) for unresectable primary hepatocellular carcinoma (HCC) 

Systematic Reviews

Vente et al (2009) conducted a meta-analysis evaluating tumor response and survival in patients who received yttrium-90 glass or resin microsphere RE for the treatment of primary liver cancer (HCC) or metastases from colorectal cancer (CRC).⁹ (Refer to Unresectable Metastatic CRC section for the data from the meta-analysis as pertains to that disease.) Included studies were from 1986 onward and presented tumor response measured by computed tomography (CT) scans and data on median survival times. To allow comparability of results with regard to tumor response, the category of “any response” was introduced and included complete response (CR), partial response (PR), and stable disease (SD). Overall tumor response could only be assessed as any response because response categories were not uniformly defined in the analyzed studies. 

In 14 articles, clinical data were presented on tumor response and survival for 425 patients with HCC who had received yttrium-90 RE. Treatment with resin microspheres was associated with a significantly higher proportion of any response than glass microsphere treatment (0.89 vs 0.78, respectively; p=0.02). Median survival was reported in 7 studies, in which survival time was defined as survival from microsphere treatment or from diagnosis or recurrence of HCC. Median survival from microsphere treatment varied between 7.1 and 21.0 months, and median survival from diagnosis or recurrence was 9.4 to 24.0 months.

The authors of the meta-analysis concluded that yttrium-90 RE is associated with high response rates, both in salvage and first-line settings, but that the true impact on survival will only become known after publication of several ongoing and/or to-be-initiated phase 3 studies, as well as the results of trials in which yttrium-90 RE and modern chemotherapy agents are combined with novel biologic agents.

Randomized Controlled Trials

In 2014, Kolligs et al reported results of a small pilot randomized controlled trial (RCT) comparing RE with transarterial chemoembolization (TACE) for the treatment of unresectable HCC, the SIR-TACE study.¹⁰ The study included 28 subjects with unresectable HCC, preserved liver function, and an Eastern Cooperative Oncology Group [ECOG] Performance Status of 2 or less, with no vascular invasion or extrahepatic spread, who had 5 or fewer liver lesions or a single lesion of 10 cm or less. Patients were randomized to RE (n=13) or TACE (n=15). Over posttreatment follow-up, PR rates were 13.3% for TACE and 30.8% for RE, with rates of disease control (CR, SD, PR) of 73.3% for TACE and 76.9% for RE. Median progression-free survival (PFS) was 3.6 months for RE and 3.7 months for RE.

Also in 2014, Pitton et al reported results from a small RCT comparing RE with TACE with drug eluting beads TACE (DEB-TACE) for the treatment of unresectable HCC.¹¹ The study included 24 patients, 12 randomized to each group. No deaths occurred within 30 days of the procedure for either group. There were no statistically significant differences between the groups in terms of in PFS (180 days for RE vs 216 days for TACE; p=0.619) and overall survival (OS; 592 days for RE vs 788 days for TACE; p=0.927). 

Prospective Noncomparative Studies

A comparison of tumor response and survival among subgroups of patients with and without portal vein thrombosis (PVT) was reported by Kulik et al in 2008.¹² Thirty-four percent of this phase 2 open-label cohort of 108 unresectable HCC patients treated with TheraSphere® had had either branch or main PVT. At 6 months, WHO criteria PR was observed in 42.2% of the overall cohort and in 34% and 66% of those with and without PVT, respectively. Kaplan-Meier survival was statistically longer in the PVT-free group (467 days) than the branch (304 days) and main PVT (133.5 days) groups. At baseline, the PVT groups had higher tumor burden, Okuda stage, pretreatment bilirubin concentrations, and proportion of patients with portal hypertension than the non-PVT groups. Adverse events for the PVT groups were presented among those with and without baseline cirrhosis. Cirrhotic patients with main PVT were more likely than those without PVT to have worsening of baseline ascites (55% and 15%, respectively) with yttrium-90 microsphere treatment; no difference was seen among those without cirrhosis, although the numbers were small.

In a 2002 review of the use of yttrium-90 RE for unresectable liver cancer, Salem et al described a previously-unpublished series of approximately 300 patients with liver carcinoma with selective internal radiotherapy (SIRT) under a humanitarian device exemption (HDE) at 8 unnamed institutions.¹³ The report provided no additional details on baseline characteristics of the patients and did not specify inclusion or exclusion criteria for treatment. Investigators only reported outcomes for a cohort of 54 HCC patients with Okuda stage I and II (median survival, 23 and 11 months, respectively; OS at 1 year, 68% and 37%, respectively). 

Observational Studies

A number of prospective and retrospective observational studies published since the 2009 Vente et al meta-analysis have reported outcomes for RE for HCC or have compared RE with other therapies for HCC.

In 2015, El Fouly et al reported results of a nonrandomized study comparing yttrium-90 RE with TACE among 86 patients with intermediate stage, nonresectable HCC.¹⁴ Sixty-three patients at 1institution were treated with TACE, while 53 patients at a second institution were treated with RE. Median OS in for TACE and RE did not differ significantly between groups (18 months for TACE vs 16.4 months for RE); similarly median time to progression (TTP) did not differ significantly between groups (6.8 months for TACE vs 13.3 months for RE). TACE patients had higher numbers of treatment sessions, hospital times, and rates of adverse events. Also in 2015, Gramenzi et al conducted a retrospective cohort study to compare RE with yttrium-90 with sorafenib for intermediate- or advanced-stage HCC.¹⁵ Patients with HCC refractory to other therapies and no metastases or systemic chemotherapy were included, 74 of whom were treated with sorafenib and 63 treated with RE. Median OS between groups was similar (14.4 months for sorafenib-treated patients vs 13.2 months for RE-treated patients). After propensity-score matching of 32 subjects in each group, there were no significant differences in median OS or 1-, 2-, and 3-year survival rates between groups.

Some of the larger prospective studies are described next.

In 2010, Salem et al reported the results of a single-center, prospective, longitudinal cohort study of 291 patients with HCC treated between January 2004 and December 2008.¹⁶ The patient population was heterogeneous and included patients with PVT (43%), advanced disease, and extrahepatic metastases (16%), which are usually exclusionary criteria for studies using locoregional therapy.Data were collected prospectively and included toxicity, imaging, and survival outcomes. Patients were staged by Child-Pugh scores. Eighty-seven percent of patients had received no prior therapy. A total of 526 treatments were administered (mean, 1.8; range, 1-5). Scans were performed 4 to 6 weeks after each treatment and then at 2 to 3 month intervals once all disease was treated. Median follow-up time was 30.9 months. Imaging follow-up was available in 273 patients, with an average of 4.3 scans per patient. By World Health Organization (WHO) criteria, response rates were 42%; by European Association for the Study of the liver (EASL) criteria, 57%, with 23% CR and 34% PR. Response rates were better in patients with Child- Pugh A disease (WHO=49%, EASL=66%) than those with Child-Pugh B disease (WHO=36%, EASL=51%), and WHO response rates varied by baseline largest tumor size: smaller than 5 cm, 44%; 5 to 10 cm, 42%; and larger than 10 cm, 33%. Survival for patients with Child-Pugh disease A and B was 17.2 months and 7.7 months, respectively (p=0.002). The authors concluded that patients with ChildPugh A disease, with or without PVT, benefitted most from the treatment but that the role of yttrium-90 in certain patients with HCC requires further exploration, including controlled studies comparing yttrium-90 with alternative locoregional therapies (radiofrequency ablation [RFA], TACE) and yttrium-90 in various combinations with systemic targeted therapies in advanced disease.

Carr et al (2010) reported on a consecutive series of patients with HCC who were seen at a single medical center and who were not candidates for surgical resection.¹⁷ Patients either received conventional cisplatin-TACE between the years 1992 and 2000 (n=691), yttrium-90 microspheres between 2000 and 2005 (n=99), or no treatment (n=142). Median OS for the yttrium-90 group versus the TACE group was 11.5 months (95% confidence interval [CI], 8 to 16 months) versus 8.5 months (95% CI, 8 to 10 months), respectively (p<0.05). Untreated patients had a median survival of 2 months. Although the authors felt there was a slight selection bias toward milder disease in the yttrium-90 group, they concluded that yttrium-90 and TACE appear to be equivalent regional therapies for patients with unresectable, nonmetastatic HCC. 

Additional case series describe outcomes for RE for HCC, with similar outcomes, including Kwok et al¹⁸ (N=46) and Saxena et al¹⁹ (N=45).

RE as a Bridge to Liver Transplantation for Primary HCC 

In 2014, Kulik et al reported results of a pilot RCT of yttrium-90 RE with or without sorafenib for patients with HCC awaiting liver transplantation.²¹ The study randomized 23 subjects; after accounting for losses due to self-withdrawal from the study, failure to confirm HCC, and death, the modified intention-to-treat (ITT) population included 10 subjects randomized to RE alone and 10 randomized to RE with sorafenib. Overall, 17 of 20 patients underwent liver transplantation, with no difference in median time-to-transplant between groups. However, the addition of sorafenib was associated with increased peritransplant biliary complications, and acute rejection.

In a 2013 retrospective review, Tohme et al reported on 20 consecutive HCC patients on liver transplant waiting lists who received RE as bridge therapy.²² When RE began, Milan criteria were met by 14 patients and sustained until transplantation. Of the 6 patients who did not meet Milan criteria initially, RE was able to downstage 2 patients to meet Milan criteria. After onset of RE, median time to liver transplant was 3.5 months. Complete or partial radiologic response to RE on modified Response Evaluation Criteria In Solid Tumors (RECIST) occurred in 9 patients. Additionally, on pathologic examination, 5 patients had no evidence of viable tumor whose disease met the Milan criteria. 

In 2014, Ramanathan et al reported on various therapies, including RE, for 715 HCC patients of whom 231 were intended for transplant.²³ In the ITT with transplantation arm, 60.2% were able to receive a transplant. Survival rates posttransplant were 97.1% and 72.5% at 1 and 5 years, respectively. Tumor recurrence rates were 2.4%, 6.2%, and 11.6% at 1, 3, and 5 years, respectively.

Lewandowski et al (2009) compared RE with chemoembolization in the efficacy of downstaging 86 patients with HCC from stage T3 to T2 (potentially making patients liver transplant candidates).²⁰ Patients were treated with either RE using yttrium-90 microspheres (n=43) or TACE (n=43). Median tumor size was similar between the 2 treatment groups (5.7 and 5.6 cm, for TACE vs RE, respectively.) PR rates were 61% versus 37% for RE versus TACE, respectively, with downstaging from T3 to T2 in 58% of patients treated with RE versus 31% with TACE (p<0.05).

Intrahepatic Cholangiocarcinoma 

The evidence related to the use of RE for intrahepatic cholangiocarcinoma (ICC) consists primarily of retrospective observational studies, along with a smaller number of prospective observational studies.

Systematic Reviews

In 2015, Al-Adra et al reported results from a systematic review of studies reporting outcomes for RE for ICC.²⁴ The review included 12 publications, 7 of which were published in abstract form only. Of the peerreviewed manuscripts, 3 are described as prospective cohort studies, which are described in more detail next (Mouli et al, Hoffman et al, and Saxena et al; of note, Hoffman et al is reported by the authors to be a retrospective study). The overall weighted median survival was 15.5 months (range, 7-22.2 months), based on 11 included studies. A weighted mean PR was seen in 28% of patients and SD was seen in 54% at 3 months posttreatment.

Also in 2015, Boehm et al conducted a systematic review to compare hepatic artery-based therapies, including hepatic arterial infusion (HAI), TACE, DEB-TACE, and yttrium-90 RE, for unresectable ICC.²⁵ Twenty studies that met the review’s inclusion criteria were included, 5 of which evaluated yttrium-90 RE.Median OS across studies was 22.8 months for HAI, 13.9 months for RE, 12.4 months for TACE, and 12.3 months for DEB-TACE. CR or PR occurred in 56.9% of patients treated with HAI, compared with 27.4% of those treated with RE and 17.3% of those treated with TACE.

Observational Studies

In 2013 Mouli et al reported on 46 patients treated with RE for ICC via a retrospective review of prospectively collected data from a single institution.²⁶ Survival varied depending on level of disease, multifocal, infiltrative and bilobar and ranged from 5.7 to 15.6 months. Five patients achieved resectable status and underwent curative resection. 

A retrospective study by Hoffman et al of RE with yttrium-90 resin microspheres included 24 patients with nonresectable chemorefractory intrahepatic ICC and no extrahepatic disease.²⁷ The mean age of the sample was 65.2 years, and the sample was 45.5% female. ECOG Performance Status was 0 in 51.5%, 1 in 21.2%, and 2 in 27.3%. Previous therapy included chemotherapy in 78.8%, surgery in 36.4%, TACE in 9.1%, RFA in 5.1%, and external beam radiotherapy (EBRT) in 3.0%. Tumor response was assessed by RECIST criteria. A CR was seen in 0%, PR, 36.4%, SD, 51.5%, and progressive disease (PD), 15.2%. Follow-up ranged between 3.1 and 44 months (median, 10 months). Median OS was 22 months and median TTP was 9.8 months. Favorable subgroups with respect to survival included those with ECOG Performance Status of 0, tumor burden as percentage of liver volume of 25% or less, response by CA-199 criterion and RECIST PR. The same subgroups except those with a CA-19-9 response had favorable TTP results. Data were collected retrospectively and no toxicity results were reported.

A 2011 study by Haug et al designed to evaluate prognostic factors of RE treatment included 26 consecutive patients with unresectable ICC who underwent RE with yttrium-90 glass microspheres.²⁸ All patients had a Karnofsky Performance Status of 60% or more. Mean age was 64.3 years, 31% had extrahepatic disease, and 42% were female. Treatment given previously included chemotherapy in 65%, surgery in 28%, local therapy in 20%, and none in 24%. Tumor response results according to RECIST criteria were: CR in 0%; PR, 22%; SD, 65%; and PD, 13%. Median OS was 51 weeks and multivariate analysis found that a PR from quantitative interpretation of positron emission tomography was a significant independent predictor of survival. The authors found no cases of grade 3 toxicity in transaminases or bilirubin.

In 2010, Saxena et al published results of a prospective evaluation of 25 patients with unresectable ICC who received RE with yttrium-90 resin microspheres.²⁹ Extrahepatic disease was present in 48%, mean age was 57 years, and 48% of patients were female. Prior treatment included surgery in 40%, chemotherapy in 72%, RFA in 6.1%, and EBRT in 3.0%. By RECIST tumor response criteria, CR was seen in 0%, PR, 24%, SD, 48%, and PD, 20%. Follow-up was collected between 0.4 and 55 months (median, 8.1 months). In the entire group, median OS was 9.3 months. Among subgroups, longer survival durat duration was seen in patients with peripheral tumors and those with ECOG Performance Status of 0. The proportion of patients with both grade 3 albumin toxicity and grade 3 bilirubin toxicity was 8%. Grade 3 alkaline phosphatase toxicity was observed in 4%. One patient (4%) experienced duodenal ulcer due to malperfusion of yttrium-90 microspheres.

A study by Ibrahim et al from 2008 reported results on RE with yttrium-90 glass microspheres among 24 patients with unresectable ICC.³⁰ The group was 33% female and had a median age of 68 years. Extrahepatic disease was present in 33%. ECOG Performance Status was 0 in 42%, 1 in 50%, and 2 in 8%. Prior chemotherapy had been used in 29%. Using the WHO tumor response criteria, CR was observed in 0%; CR, 27%; SD, 68%; and PD, 5%. Follow-up was collected over a median of 17.7 months and median OS was 14.9 months. Subgroups that had favorable survival results included those with ECOG Performance Status of 0, no previous chemotherapy and peripheral tumor. Grade 3 albumin toxicity was found in 17%, grade 3 bilirubin toxicity in 4%, and 1 patient (4%) developed a duodenal ulcer.

Rayar et al reported successful downstaging of unresectable ICCs after RE in 8 patients with initial unresectability due to involvement of hepatic veins or portal veins of the future liver remnant.³¹ After RE, all patients underwent successful R0 resection.

Metastatic Liver Tumors Unresectable Metastatic Neuroendocrine Tumors

The data on the use of RE for unresectable liver metastases from neuroendocrine tumors include 1 openlabel phase 2 study, retrospective reviews, and case series, which have been summarized in 1 systematic review.

Systematic Reviews

In 2014, Devic et al published results of a systematic review of studies evaluating RE for liver-dominant metastatic neuroendocrine tumors.³² The review included 12 studies that provided RECIST data for hepatic metastatic neuroendocrine tumors treated with RE. For yttrium-90 RE with resin microspheres only, objective radiographic response rates (CR or PR by RECIST) ranged from 12% to 80%, with a random-effects weighted average of 50% (95% CI, 38% to 62%). Disease control rates (CR, PR, SD) ranged from 62% to 100%, with a random-effects weighted average of 86% (95% CI, 78% to 92%).

Prospective, Noncomparative Trials 

Rhee et al reported the results of a multicenter, open-label, phase 2 study to assess the safety and efficacy of RE, using glass or resin microspheres, in 42 patients with metastatic neuroendocrine liver disease who had failed prior treatment(s), including medical (eg, octreotide), surgical resection, bland or chemoembolization, and RFA or cryoablation.³³ Mean patient age was 58±12 years for glass and 61±11 years for resin microspheres. RECIST criteria were used to assess tumor response, which showed 92% of glass patients and 94% of resin patients were partial responders or had SD at 6 months after treatment. Median survival was 22 and 28 months for glass and resin, respectively.  

Observational Studies

 In 2010, Cao et al reported the outcomes of 58 patients with unresectable neuroendocrine liver metastases from 2 different hospitals treated with yttrium-90 microspheres (SIR-Spheres) from 2003 to 2008. Data were examined retrospectively from a database.³⁴ Response was assessed with radiographic evidence before and after RE and measured RECIST guidelines. Patients typically had a CT scan within 3 months of treatment and every 3 to 6 months until disease progression or death. Systemic chemotherapy was routinely given at 1institution but not the other. Mean patient age at the time of RE was 61 (range, 29-84 years), and 67% of patients were men. Primary tumor site was variable and included small bowel, pancreas, colon, thyroid, lung, and unknown. Thirty-one patients underwent surgical resection of their primary tumor, which was classified as low grade in 15, intermediate grade in 7, and high grade in 7. Forty-three percent of patients had extrahepatic metastatic disease at study entry. Prior therapies before RE included liver resection in 19 patients, transarterial embolization (TAE) or TACE in 6, ablation or percutaneous ethanol injection in 10, previous chemotherapy in 20, concurrent chemotherapy in 34, and post-RE chemotherapy in 5 patients. Median follow-up was 21 months (range, 1-61 months). Fifty-one patients were evaluable, and 6 achieved a CR, 14 a PR, 14 had SD, and 17 had disease progression. OS rates at 1, 2, and 3 years were 86%, 58%, and 47%, respectively. Median survival was 36 months (range, 1-61 months). Prognostic factors for survival included extent of tumor involvement of the liver, radiographic response to treatment, presence of extrahepatic disease at the time of RE, histologic grade of tumor, and whether patients were responders (vs nonresponders) to RE. Factors that were not significant prognostic features included age, sex, ECOG Performance Status, and previous therapy. 

King et al reported outcomes in patients treated in a single-institution prospective study.⁷ Thirty-four patients with unresectable neuroendocrine liver metastases were given radioactive microspheres (SIRSpheres) and concomitant 7-day systemic infusion of 5-FU (fluorouracil), between 2003 and 2005. Mean patient age was 61 years (range, 32-79 years), and 65% were men. Mean follow-up was 35.2±3.2 months. The mean interval from diagnosis of hepatic metastases and treatment with SIR therapy was 36.6±6.7 months. Primary tumor sites were variable and included bronchus (n=1), thyroid (n=2), gastrointestinal (n=15), pancreas (n=8), and unknown (n=8). Subjective changes from baseline hormone symptoms were reported every 3 months. Twenty-four patients (71%) had, at baseline assessment, symptoms of carcinoid syndrome, including diarrhea, flushing, or rash. At 3 months, 18 of 33 patients (55%) reported improvement of symptoms, as did 16 of 32 (50%) at 6 months. Radiologic tumor response was observed in 50% of patients and included 6 CR (18%), and 11 PR (32%). Mean OS was 29.4±3.4 months.

Kennedy et al conducted a retrospective review of 148 patients from 10 institutions with unresectable hepatic metastases from neuroendocrine tumors who received resin microspheres.³⁵ All patients had completed treatment of the primary tumor and metastatic disease and were not excluded based on prior therapy. Total number of resin microsphere treatments was 185, with retreatment in 22.3% of patients (19.6% received 2 treatments, 2.7% received 3 treatments). All patients were followed with imaging studies at regular intervals to assess tumor response (using either WHO or RECIST criteria) until death, or they were censored if a different type of therapy was given after the microspheres. The male to female ratio was 49% to 51%, respectively, and median age was 58 years (range, 26-95 years). Median followup was 42 months. By imaging, response rates were SD, 22.7%; PR, 60.5%; CR, 2.7%; and progressive disease, 4.9%. Hepatic and extrahepatic metastases contributed to death in most patients, with 7% lost to follow-up. Median survival was 70 months. The authors conclude that RE can deliver high doses of radiation preferentially to hepatic metastases from neuroendocrine tumors with encouraging response rates by imaging and symptomatic improvement (although there were no data presented in the study regarding symptoms). 

Engelman et al retrospectively compared transarterial, liver-directed therapies, including RE, hepatic artery embolization (HAE), and hepatic artery chemoembolization (HACE), in 42 patients treated for metastatic neuroendocrine tumors.³⁶ Treatment decisions were at the discretion of the referring physician and interventional radiologist, but the decision to proceed with therapy was typically based on progression of symptoms nonresponsive to octreotide therapy or rapid progression of liver tumor burden on imaging. Seventeen patients had HACE, 13 had HAE, and 12 had RE. Among the 27 patients with symptoms from their liver metastases, there were no statistically significant differences in symptom improvement at 3 months after first liver-directed therapy across treatment modalities (6/13 for HACE; 4/8 for HAE; 5/6 for RE; p=0.265). There were no differences between treatment modalities in radiographic response at 6 months postprocedure (p=0.134), TTP (p=0.968), or OS (p=0.30).

Additional case series in patients with treatment-refractory, unresectable neuroendocrine hepatic metastases have shown good tumor response and improvement in clinical symptoms with RE.^37-39

Unresectable Metastatic CRC

The evidence related to the use of RE for metastatic CRC consists of several small- to moderate-sized RCTs, prospective trials, and retrospective studies using a variety of comparators, along with systematic reviews of these studies.

Systematic Reviews

In a 2013 (preprint) systematic review, Saxena et al evaluated 20 experimental and observational studies on RE for chemoresistant, unresectable CRC liver metastasis, which included a total of 979 patients.⁴⁰ The review included 2 RCTs (Gray et al [2001] and Hendlisz After RE, the average reported CRs and PRs from 16 studies was 0% (range, 0-6%) and 31% (range, 0-73%), respectively. Nine months was the median time to intrahepatic progression (range, 6-16 months). Eleven studies reported OS rates and 12 months was the median survival time (range, 8.3-3.6). 

In another 2013 systematic review, Rosenbaum et al evaluated 13 relevant articles on RE as monotherapy and 13 studies on RE combined with chemotherapy for chemoresistant, unresectable CRC liver metastasis.⁴¹ CR, PR, and SD rates ranged from 29% to 90% with only RE and from 59% to 100% for RE with chemotherapy. At 12 months, survival ranged from 37% to 59% with only RE and from 43% to 74% for RE combined with chemotherapy.

A technology assessment (2010) from the California Technology Assessment Forum (CTAF) assessed 25 studies on the use of RE and inoperable metastatic CRC to the liver, including 2 RCTs (Gray et al [2001]⁴² and Van Hazel et al [2004],⁴³ described next), 1 small retrospective study comparing SIRT with chemoembolization (N=36), and 21 case series.⁵ The assessment concluded that the 3 comparative studies all used different control interventions and that the nonrandomized study did not show any convincing improvements over chemoembolization. The author stated that the assessment showed it is feasible to deliver radiotherapy to liver metastases and achieve at least PR in a substantial portion of patients with relatively few serious adverse events and that the results of the 2 randomized studies were encouraging but not definitive, as the trials were very small, response rates in the control groups were lower than expected, and control groups were not given what is currently considered standard first-line chemotherapy for metastatic CRC. The assessment concluded that the use of SIRT for unresectable CRC did not meet any of the CTAF technology assessment criteria, with the exception of criterion number 1 (ie, the technology has final approval from the appropriate government regulatory bodies).

A 2009 Cochrane review attempted to assess the efficacy and toxicity of RE, alone or with systemic or regional hepatic artery chemotherapy, in the treatment of metastatic CRC liver metastases ⁴⁴ Two articles which met the authors’ inclusion criteria were included, Gray et al (2001)42 and van Hazel et al (2004),⁴³ described next. The authors concluded that there was a lack of evidence that SIRT improves survival or quality of life (QOL) in patients with metastatic CRC, whether it is given alone or with chemotherapy, and that there is a need for well-designed, adequately powered phase 3 trials assessing the effect of SIRT when used with modern combination chemotherapy regimens.

The 2009 meta-analysis by Vente et al,⁹ previously described, included 19 studies with a total of 792 patients with metastatic CRC who had undergone yttrium-90 RE.⁹ Included in the meta-analysis were 2 RCTs (Gray et al [2001]⁴² and van Hazel et al [2004]⁴³). Two covariates were included in the metaregression model: (1) whether an older generation of cytostatic agents (5-FU/LV [leucovorin or floxuridine) or a newer generation (5-FU/LV plus oxaliplatin [FOLFOX] or 5-FU/LV plus irinotecan [FOLFIRI]) was used, and (2) whether yttrium-90 RE was given as salvage therapy or as first-line treatment with adjuvant chemotherapy. The specific cytostatic agent(s) that were used did not affect response (p=0.96). Tumor response to yttrium-90 RE was high, with any response rates of approximately 80% in a salvage setting, and more than 90% when used as first-line treatment as neoadjuvant to chemotherapy, regardless of the chemotherapy regimen used. Median survival after yttrium-90 RE, irrespective of differences in determinants (microspheres type, chemotherapy protocol, salvage or first line) varied from 6.7 to 17.0 months. 

Randomized Controlled Trials

A study was published by Gray et al in 2001 and randomly assigned 74 patients with bilobar unresectable liver metastases to monthly HAI with 5-FU alone or with the same chemotherapy plus a single infusion of yttrium-90 microspheres.⁴² The investigators closed the study after entering 74 patients (n=70 eligible for randomization). The original goal was 95 patients. Reasons cited for the early closure included: (1) increasing patient and physician reluctance to participate; (2) decision by FDA to accept intermediate end points to support applications for premarket application approval; and (3) lack of funding to complete the study. The smaller study population was adequate to detect increases in response rate (from 20% to 55%) and median disease TTP (by 32% from 4.5 months), with 80% power and 95% confidence, but lacked sufficient statistical power to detect changes in survival. To monitor responses to therapy, investigators serially measured serum levels of carcinoembryonic antigen (CEA) and estimated tumor cross-sectional area and volume from repeated CT scans read by physicians blinded et al [2010], described next), 5 studies classified as non-RCTs or well-designed cohort studies, and 13 observational studies to treatment assignment. They reported increased overall responses (CR plus PR) measured by area (44% vs 18%, p=0.01; HAI plus SIRT vs HAI, respectively) and volume (50% vs 24%, respectively; p=0.03), or by serum CEA levels (72% vs 47%, respectively; p=0.004). They also reported increased TTP detected by increased area (9.7 vs 15.9 months, respectively; p=0.001) or volume (7.6 vs 12.0 months, respectively; p=0.04). However, there were no significant differences between treatment arms in actuarial survival rates (log-rank test, p=0.18) or in 11 QOL measures. Treatment-related complications (grades 3-4) included 23 events in each arm (primarily changes in liver function tests). Nevertheless, investigators concluded that a “single injection of SIR-Spheres^® plus HAI is substantially more effective” than the same HAI regimen delivered alone. 

 Despite the investigators’ assertions, these results are inadequate to support their conclusions for the following reasons: 

1.     Accrual was halted early, leaving the study underpowered. 

2.     Although the study involved oversight by an institutional review board, the report suggests early closure was at the sole discretion of the principal investigator without independent review or prospectively designed data monitoring procedures and stopping rules. 

3.     While in this study, response rate and TTP after SIRT plus HAI appeared superior to the same outcomes after HAI alone, results for the SIRT plus HAI group are within the range reported by other randomized trials of HAI in comparable patients.^16,17 

4.     Results of this study may reflect use of a shorter-than-standard duration of HAI therapy and are confounded by administration of nonprotocol chemotherapy before and after SIRT.  

5.     The reported increases in response rates and TTP improved neither duration of survival nor QOL. 

Another randomized trial was a phase 2 study published in 2004 by the same research group as the phase 3 trial.⁴³ The study involved 21 patients with advanced colorectal liver metastases; a total of 11 patients received SIR-Spheres plus systemic chemotherapy (fluorouracil and leucovorin), and 10 received the same systemic chemotherapy alone. While the time to progressive disease was greater in those receiving combination therapy (18.6 vs 3.6 months, respectively; p<0.001), the small size of the study limits any conclusions. 

A phase 3 study, 46 patients, compared intravenous 5-FU with hepatic intra-arterial injection of yttrium-90 microspheres (SIR-Spheres) with intravenous 5-FU in CRC metastatic only to the liver and refractory to standard chemotherapy.⁴⁵ The time to liver progression, the primary outcome, was significantly improved in the group receiving SIR-Spheres, 2.1 versus 5.5 months, respectively (p=0.003). However, there was no difference in the more important outcome of median survival, which was 7.3 and 10.0 months, respectively (p=0.80). 

Prospective Noncomparative Trials 

A single-arm, open-label study was reported by Mulcahy et al (2009) and involved 72 patients with unresectable hepatic colorectal metastases treated with yttrium-90 microspheres (TheraSphere®).⁴⁶ To determine response, 128 lesions were used. A PR rate using WHO criteria were noted in 29 of 72 patients (40.3%), and at the lesional level, the response rate was 40.6% (PR, 37.5%; CR, 3.1%). SD was observed in 44.5% of patients, and disease progression was found in 14.8% of patients. Median follow-up was 26.2 months. Median OS was 40.3 months (95% CI, 29.0 to 51.6 months) for all patients from the time of cancer diagnosis, 34.6 months (95% CI, 24.4 to 41.8 months) from the time liver metastases were diagnosed, and 14.5 months (95% CI, 9.6 to 21.9) from the time of yttrium-90 therapy. A substratification analysis was performed, and favorable prognostic factors that indicated a benefit from yttrium-90 therapy included an ECOG Performance Status of 0, a liver tumor burden of 25% or less, and the absence of extrahepatic disease. For the patients with an ECOG Performance Status of 0 at the time of yttrium-90 treatment, the overall median survival from the onset of liver metastases was 42.8 months, or a 5-year survival rate of 25.9%, which are comparable outcomes with survival data for patients treated with primary resection, chemotherapy followed by resection, or RFA.  for patients with CRC liver metastases who received RE after failure of chemotherapy. The largest study identified was by Saxena et al, who retrospectively reviewed outcomes for 302 patients with unresectable, chemotherapy-resistant CRC liver metastases treated with yittrium-90 RE at a single institution from 2006 to 2012.⁴⁷ One hundred fifteen subjects (38%) developed clinical toxicity after treatment, most of which (5%) was considered minor and self-resolved. Five patients (2%) died within 30 days of treatment, 2 from suspected pulmonary embolus, 2 from clinical disease progression, and 1 from radiation hepatitis. Two hundred ninety-three patients (97%) had follow-up beyond 2 months post-RE therapy with CT imaging. Of these, 2 patients (1%) had CR, 111 (37%) had PR, 96 (32%) had SD, and 84 (28%) had PD. The median OS after RE treatment was 10.5 months.

In another large prospective series, Lewandowski et al reported outcomes from a single-center review of 214 patients treated with yttrium-90 RE for CRC liver metastases from 2001 to 2013.⁴⁸ Median OS was 43 months from the time of primary cancer diagnosis, 34.6 months from the time of hepatic metastases diagnosis, and 10.6 months from the time of yttrium-90 initiation. Grade 3 absolute lymphocyte, bilirubin, albumin, alkaline phosphatase, and aspartate aminotransferase toxicities were observed in 39%, 11%, 10%, 8%, and 4% of patients, respectively; while grade 4 absolute lymphocyte and alkaline phosphatase toxicities were observed in 5% and 3% of patients, respectively. In multivariable models, a number of factors were independently associated with OS, including receiving fewer than 3 cytotoxic drugs before RE (OS, 15.2 vs 7.5 months; hazard ratio [HR], 0.67; 95% CI, 0.46 to 0.98; p=0.042), receiving no biologic agents before RE (OS, 18.6 vs 9.4 months; HR=0.56; 95% CI, 0.36 to 0.88; p=0.012), and having no extrahepatic disease (HR=0.58; 95% CI, 0.41 to 0.83; p=0.002).

In another relatively large review, Kennedy et al reported results for use of resin microspheres in 208 patients with liver metastases from CRC who had failed or were not candidates for standard chemotherapy.⁵⁰ There were no CRs but were 35% PRs by CT, as determined by a 50% decrease in 1 tumor measure at 12 weeks. Median survival was 10.5 months for responders but 4.5 months for nonresponders. The authors noted that most patients died with persistent liver disease and had uncontrolled systemic metastases. No QOL functional status measures were reported. In addition, the authors noted that their report was a retrospective review with associated problems of a mixture of patients and a lack of a controlled treatment protocol. 

Kalva et al reported results of a retrospective study of 45 patients with CRC liver metastases who failed systemic chemotherapy and were treated with yttrium-90 RE at a single center from 2005 to 2011.⁵¹ Twenty-three patients (51%) had no toxicities and 6 patients (13%) had grade 3 toxicities. One patient (2%) had a PR, 34 (71%) had SD, and 6 (13%) had PD. The median survival was 186 days (95% CI, 149 to 277 days). 

Jakobs et al retrospectively reviewed patients with CRC liver metastases in whom chemotherapy had failed and who therefore received a single-session, whole-liver treatment with yttrium-90 RE (N=41).⁴⁹ Response was partial in 7 patients (17%). Twenty-five patients (61%) had SD, and 4 (10%) had PD. Median OS was 10.5 months. Median survivals for patients with PR, SD, and PD were 29.3 months, 10.9 months, and 4.3 months, respectively. No severe toxicities were observed.

A retrospective, matched-pair comparison of RE and best supportive care (n=29) versus best supportive care alone (n>500) for chemorefractory, liver-dominant colorectal metastases showed prolongation of survival in the group of patients who received RE.⁵²

Miscellaneous Metastatic Tumors

Breast Cancer

Most studies on the use of RE for metastatic breast cancer evaluate the use of RE alone (ie, not in combination with chemotherapy) either during a hiatus between lines of chemotherapy or in patients refractory to standard of care chemotherapy.⁸ In 2013, Smits et al reviewed 6 studies on RE for metastatic breast cancer with a total of 198 study participants.⁵³ CR, PR, and SD control rates at 2 to 4 months after treatment varied from 78% to 96%. In 4 studies, the median survival ranged from 10.8 to 20.9 months. Ten patients had gastric ulceration, a

Observational Studies

A number of retrospective and prospective observational studies have reported outcomes and 3 patients had mortality related to treatment. 

In 2014, Gordon et al reported results of a retrospective review of prospectively collected data for 75 patients with breast cancer liver metastases and stable extrahepatic disease treated with yttrium-90 RE at a single center.⁵⁴ Included patients had hepatic tumor progression after cytotoxic systemic chemotherapy. For the 48 patients for whom data were available who had systemic chemotherapy with RE, 32 (66.7%) received additional systemic chemotherapy. The 30-day mortality rate after RE was 4% (n=3, including 1 case of sepsis and 2 cases of hepatic decompensation). Clinical grade 3 toxicity occurred in 5 patients (7.6%) and biochemical grade 3 toxicity occurred in 37 patients (54.4%). The median OS was 6.6 months (95% CI, 5.0 to 9.2 months). Post-RE imaging was available for 68/73 living patients (93.2%). The median time to hepatic progression was 3.2 months (95% CI, 1.2 to 8.5 months), while the median distant TTP was 4.1 months (95% CI, 2.7% to 7.0%).

In 2014, Saxena et al reported results of a retrospective analysis of 40 subjects who underwent yttrium-90 RE treatment for unresectable, chemoresistant breast cancer liver metastases treated from 2006 to 2012 at a single institution.⁵⁵ At study entry, all patients had received at least 1 line of systemic chemotherapy. Twenty-four patients (60%) had evidence of limited extrahepatic disease; for the 16 with liver-exclusive disease, 14 patients underwent RE for progression of disease, while 2 patients underwent RE for increasing symptoms. For patients with extrahepatic disease, the indication for RE was discordant liver progression in 20 patients and symptoms in the remaining 4 patients. Sixteen patients (40%) had clinical toxicity after treatment, all of which were grade 1-2. Thirty-eight patients (95%) had follow-up beyond 1 month post-RE with CT imaging. Overall, 2 of 38 patients (5%) had a CR, 10 of 38 patients (26%) had PR, 15 of 38 patients (39%) had SD, and 11 of 38 patients (29%) had PD. The median survival after the first RE treatment was 13.6 months.

A 2013 study by Cianni et al included 52 women with chemotherapy-refractory breast cancer and inoperable liver metastases.⁵⁶ RE treatment entailed yttrium-90 resin microspheres. The median age was reported as 57.5 years. ECOG Performance Status was 0 in 55.7%, 1 in 26.9% and 2 in 17.3%. Extrahepatic disease was present in 46.1%. Chemotherapy had been administered previously in all patients, surgery in 17.3%, TACE in 3.8%, and RFA in 3.8%. Tumor response results by RECIST criteria were: CR, 0%; PR, 56%; SD, 35%; and PD, 10%. Median OS was 11.5 months. Patients were retrospectively divided into 2 risk groups based on ECOG Performance Status, degree of liver tumor burden and whether extrahepatic disease was present. Median survival in the low-risk group was 14.3 months, significantly better than in the high-risk group (8.2 months). Grade 3 gastritis was seen in 2 patients (4%). 

Haug et al published a case series of 58 women with chemotherapy-refractory breast cancer and unresectable hepatic metastases.⁵⁷ The patients received RE with yttrium-90 resin microspheres. The mean age was 58 years, and all patients had a Karnofsky Performance Status of 60% or higher. Extrahepatic disease was present in 66%. Prior treatments were not mentioned. By RECIST criteria, a CR was seen in 0%; PR, 25.6%; SD, 62.8%; PD in 11.6%. Mean follow-up covered 27.5 weeks. The median OS for the sample was 47 weeks. Two indices derived from quantitative interpretation of positron emission tomography were significant predictors of survival. Bilirubin toxicity was at grade 3 in 3% and grade 4 in 2%. Transaminase toxicity was grade 3 in 5% and grade 4 in 2%.

Jakobs et al reported on the safety and survival of 30 patients (29 women, 1 man) who underwent RE with resin microspheres in a single-session, whole-liver treatment for breast cancer metastases.⁵⁸ All patients had failed prior polychemotherapy regimens (including at least anthracyclines and taxanes, hormonal therapy, and trastuzumab, when applicable). Twenty-three patients had follow-up data. At median follow-up of 4.2 months, PR, SD, and PD was observed in 61%, 35%, and 4% of patients, respectively. Clinically significant toxicities were observed in 8 of 30 patients and included increasing liver enzymes and bilirubin levels, nausea and vomiting, gastric ulcers and ascites; 1 death was due to treatment-related hepatic toxicity. Median follow-up was 14.2 months, with a median OS of 11.7 months. Median survival of responders versus nonresponders was 23.6 and 5.7 months, respectively. Median survival of patients with and without extrahepatic disease was 9.6 versus 16 months, respectively.

Bangash et al reported on the safety and efficacy of the use of RE with glass microspheres in 27 female patients with progressing liver metastases from breast cancer while on polychemotherapy.⁵⁹ Seventeen patients received 20 left lobe of liver treatments, and 20 received 22 right lobe of liver treatments. At the 90-day follow-up CT, CR and PR was observed in 9 patients (39%); SD, 12 (52%); and PD, 2 (9%). Median survival for ECOG 0 versus 1 to 3 was 6.8 versus 2.6 months, respectively, and for patients with tumor burden less than 25% versus greater than 25% was 9.4 and 2.0 months, respectively.

Hepatic metastases from breast cancer in 44 patients at 3 hospitals were retrospectively reviewed by Coldwell et al.⁶⁰ Patients had failed first-, second-, or third-line treatment for their primary tumor and were not candidates for RFA, TACE, resection, intensity-modulated radiotherapy, or stereotactic radiotherapy. At 12 weeks, a PR (using WHO criteria, at least 50% reduction in the cross-product of the tumor dimensions) to SIR-Spheres was observed by CT in 47% of patients with recorded follow-up (82% of the total). Symptoms were reported to improve, although no specifics were provided. There were no radiation-related liver failures observed, and, at a median follow-up of 14 months, the cohort had not yet reached its expected median survival of 14 months. 

Melanoma

The evidence related to the use of RE for melanoma consists of relatively small observational studies, many of which focus on patients with uveal melanoma, in whom the liver is the most common site of metastatic disease.

In 2014, Xing et al conducted a retrospective observational study to compare outcomes for patients with unresectable melanoma (both uveal and cutaneous) liver metastases refractory to standard chemotherapy treated with either yttrium-90 RE (n=28) or best supportive care (n=30).⁶¹ The groups were similar at baseline in terms of Child-Pugh class, ECOG Performance Status scores, age, sex, and race. Patients treated with RE had larger tumor size at baseline than those treated with best supportive care (mean, 7.28 cm vs 4.19 cm; p=0.02). Median OS from diagnosis of melanoma liver metastases was longer in RE-treated subjects (19.9 months vs 4.8 months; p<0.000), as was the median OS from diagnosis of the primary melanoma (119.9 months vs 26.1 months; p<0.001). Pre- and posttreatment imaging studies were available for 24 of 28 (85.7%) of those treated with RE. Of those, no patients had a CR; 5 patients (17.9%) had PR, 9 patients (32.1%) had SD, and 10 patients (35.7%) had PD. Two patients receiving RE had major (grade 5) clinical toxicities (ascites and hepatic encephalopathy and eventual mortality).

Also in 2014, Eldredge-Hindy et al retrospectively evaluated outcomes for the use of yttrium-90 RE in 71 patients with biopsy-confirmed uveal melanoma liver metastases.⁶² The median time from the diagnosis of liver metastases to RE was 9.8 months (95% CI, 7.4 to 12.2 months), and 82% of patients had received prior liver-directed therapies. Sixty-one patients (86%) had CT or magnetic resonance imaging evaluation of treatment response at 3 months post-RE. Of those, 5 patients (8%) had a PR, 32 patients (52%) had SD, and 24 patients (39%) had DP. Median OS RE was 12.3 months (range, 1.9-49.3 months).

Several smaller studies published from 2009 to 2012 reported on the use of RE in patients with hepatic metastases from melanoma.^63-66 Three studies included only patients with ocular melanoma,^63-65 and the fourth included patients with either ocular or cutaneous melanoma.⁶⁶ Sample sizes ranged between 11 and 32 patients. Three studies excluded those with poor performance status. Median age was in the 50s for 3 studies and 61 in the fourth. One article did not describe any previous treatment, and one described it incompletely. Three studies reported tumor response data, by RECIST criteria. Among 32 patients in the study by Gonsalves et al (2011), 1 patient had a CR (3%), 1 had a PR; 18, SD (56%); and 12, PD (38%).⁶³ In the study of 13 patients published by Klingenstein et al (2013), none had a CR; 8, PR (62%); 2, SD (15%); and 3, PD (23%).⁶⁵ Nine of 11 patients in the article by Kennedy et al (2009) provided response data: 1 had CR; 6, PR; 1, SD; and 1, PD.⁶⁴ Median survival in Gonsalves, Klingenstein, and Kennedy were 10.0 months, 19 months and not yet reached, respectively. Gonsalves reported 4 patients (12.5%) with grade 3-4 liver toxicity. Klingenstein observed 1 patient with marked hepatomegaly. Kennedy described 1 grade 3 gastric ulcer. The fourth study (Piduru et al [2012],⁶⁶ N=12) did not include 

Pancreatic Cancer

Michl et al reported on RE for pancreatic cancer in 2014.⁶⁷ Response was seen in 47% with median local PFS in the liver of 3.4 months (range, 0.9-45.0). Median OS was 9.0 months (range, 0.9-53.0), and 1-year survival was 24%.

Data on the use of RE in other tumors metastatic to the liver are limited and are composed of patient numbers too small to draw meaningful conclusions.

RE (Radiation Lobectomy) as a Bridge to Hepatic Resection

In 2013, Vouche et al reported on 83 patients treated with RE with (90)Y-loaded microspheres as a technique to control or limit tumor progression in unresectable, unilobar hepatic disease and to hypertrophy a small future liver remnant.⁶⁸ Patients included in the study had right unilobar disease with HCC (n=67), cholangiocarcinoma (n=8), or metastatic CRC (n=8). One month after RE, right lobe atrophy (p=0.003), left lobe hypertrophy (p<0.001), and future liver remnant hypertrophy (p<0.001) were observed and remained during follow-up. Successful right lobectomy was later performed in 5 patients, and 6 patients received liver transplants. However, further studies are needed to assess RE as a bridge to hepatic resection.

The available evidence for the use of radioembolization (RE) for the treatment of primary and metastatic liver tumors varies depending on the tumor type. 

For the use of RE in the treatment of hepatocellular carcinoma (HCC), the evidence consists primarily of retrospective and prospective observational studies, with limited evidence from randomized controlled trials (RCTs). Observational studies suggest that RE has high response rates compared with historical controls. Two small pilot RCTs compared RE with alternative therapies for HCC, including transarterial chemoembolization (TACE) and TACE with drug-eluting beads, both of which demonstrated similar outcomes for RE. Evidence from observational studies demonstrates that RE can allow successful liver transplantation in certain patients. The available evidence, including clinical input, is sufficient to draw conclusions and to determine that outcomes are improved for the use of RE for the treatment of primary HCC that is unresectable and limited to the liver or as a bridge to liver transplantation.

For the use of RE in the treatment of hepatic metastases from neuroendocrine tumors, the evidence consists of 1 open-label phase 2 study, retrospective reviews and case series, some of which compare RE with other transarterial liver-directed therapies. This evidence suggests that RE has similar outcomes to standard therapies and historical controls for patients with neuroendocrine tumor-related symptoms or progression of liver tumor burden. There was support from clinical input for the use of RE for the treatment of hepatic metastases from neuroendocrine tumors. Therefore, the available evidence is sufficient to determine that RE is associated with improved outcomes for the treatment of hepatic metastases from neuroendocrine tumors.

A major cause of morbidity and mortality in patients with colorectal disease metastatic to the liver is liver failure, as this disease tends to progress to diffuse, liver-dominant involvement. For the use of RE in the treatment of unresectable metastases from colorectal carcinoma, the evidence consists of several small- to moderate-sized RCTs, prospective trials, and retrospective studies using a variety of comparators, along with systematic reviews of these studies. Although this evidence describes wide ranges of clinical response to therapy, there was strong support from clinical input for the use of RE for the treatment of hepatic metastases from colorectal cancer; the use of RE to decrease tumor bulk, and/or halt the time to tumor progression and liver failure, may lead to prolonged progression-free and overall survival in patients with no other treatment options (ie, those with chemotherapy refractory liver-dominant disease). Other uses include palliation of symptoms from tumor bulk. Therefore, the available evidence is sufficient to determine that RE is associated with improved outcomes for the treatment of colorectal carcinoma liver metastases with liver-dominant disease.

For the use of RE for the treatment of intrahepatic cholangiocarcinoma, the evidence consists of retrospective and prospective observational studies, some of which compare RE with alternative therapies. Although no randomized trials are available, there is some suggestion that RE for primary intrahepatic cholangiocarcinoma has response rates similar to those seen with standard chemotherapy. RE may play a role patients with unresectable tumors that are chemorefractory or unable to tolerate systemic chemotherapy. Clinical input in 2015 supported the use of RE for intrahepatic

cholangiocarcinoma. Given the low likelihood of large-scale clinical trials for this rare tumor, the available evidence is sufficient to conclude that RE is associated with improved outcomes for patients with primary intrahepatic cholangiocarcinoma.

Similarly, for other tumors metastatic to the liver, including breast cancer and melanoma, the evidence consists of observational studies. In 2015, clinical input supported the use of RE for the treatment of liverdominant metastases from breast cancer and melanoma in patients who are not candidates for or who have not responded to systemic therapies. Given the clinical input, the available evidence is sufficient to conclude that RE is associated with improved outcomes for patients with hepatic metastases from breast cancer and melanoma with liver-dominant disease.