Hepatic arterial 90Yttrium glass microspheres (Therasphere) for unresectable hepatocellular carcinoma: Interim safety and survival data on 65 patients


  • Brian I. Carr

    Corresponding author
    1. Liver Cancer Center, Starzl Transplantation Institute, Pittsburgh, PA
    • Liver Cancer Center, Starzl Transplantation Institute, E1552 Biomedical Science Tower, 200 Lothrop Street, Pittsburgh, PA 15213
    Search for more papers by this author
    • Telephone: 412-624-6672; FAX: 412-624-6666


Hepatocellular carcinoma (HCC) generally arises in a cirrhotic liver and, in most cases, is multifocal and bilobar. Although trans-hepatic artery chemoembolization (TACE) can be highly affective in shrinking tumors, it is limited by virtue of the damage that it can cause to the liver that is already damaged by chronic disease. A high priority in HCC research, after primary prevention and early detection, is to find new treatment modalities that are both effective and non-toxic to the underlying cirrhotic liver. A cohort of 65 patients with biopsy-proven unresectable HCC have been treated with hepatic arterial 90Yttrium microspheres (Therasphere), and the interim results are reported here. Only 1 cycle of Therasphere treatment ever was performed on 46 patients, 17 patients had 2 cycles, and 2 patients had 3 cycles of therapy. The median dose delivered was 134 Gy, typically as either 5 or 10 GBq (2–4 million microspheres). Clinical toxicities include 9 episodes of abdominal pain and 2 episodes of acute cholecystitis, requiring cholecystectomy. A main lab toxicity was elevated bilirubin which increased by more than 200 % in 25 patients (30.5 %) during 6 months of therapy, although 18 of these patients had only transient elevation. A prominent finding was prolonged and profound (> 70%) lymphopenia in more than 75% of the patients, but without clinical significance. Forty-two patients (64.6%) had a substantial decrease in tumor vascularity in response to therapy, and 25 patients (38.4%) had a partial response, by computed tomography scan. Median survival for Okuda stage I patients (n = 42) was 649 days (historical comparison 244) and for Okuda stage II patients (n = 23) was 302 days (historical comparison 64 days). All patients were followed after therapy for a minimum of 6 months. There were 42 deaths, 21 due to liver failure, 6 from HCC progression, and 3 from metastases. Therasphere appears to be a relatively safe and effective therapy for advanced-stage unresectable HCC. (Liver Transpl 2004;10:S107–S110.)


HCC, hepatocellular carcinoma; TACE, trans-hepatic artery chemoembolization; CT, computed tomography.


The traditional treatment for unresectable hepatocellular carcinoma (HCC) has been hepatic artery infusional chemotherapy or trans-hepatic artery chemoembolization (TACE). Approximately 95% of our HCC patients are considered unresectable, either due to their bilobar presentation, major branch portal vein invasion by tumor, or the severity of the associated cirrhosis. In those patients, a non-surgical approach to therapy is required. Although TACE is a widely used medical treatment of unresectable HCC, in which the embolization takes advantage of the vascular nature of this tumor type, it is limited by its toxicity to the underlying liver. Whereas the normal liver tolerates many forms of acute toxicity, including alcohol, hepatitis, and cancer chemotherapy, the liver that is already damaged by cirrhosis (most commonly from hepatitis B or C or chronic alcohol consumption) has poor reserve for tolerating further chemical insult. As a result, the cirrhotic liver can often not tolerate doses of chemotherapy that are usually given systemically for cancer, and when given as hepatic artery infusional chemotherapy or TACE, can result in irreversible liver failure. New treatments for HCC are, therefore, needed, that will be both effective and not toxic to the cirrhotic liver, in which approximately 80% of HCC arises. Our initial experiences with a new treatment, using 90Yttrium glass microspheres (Therasphere)1–6 are described.

Therapy Description

90Yttrium glass microspheres are 20 to 30 microns in diameter and made initially of silica containing 89Yttrium. The microspheres are bombarded by neutrons in a nuclear reactor and converted to 90Yttrium. This has a half-life of 62 hours, emits pure beta radiation, and has a maximum path length (radius of action) of 1 cm (average 2.5 mm). Thus, once inside the liver, neither medical personnel nor family members can be irradiated. It decays to inactive 90Zirconium. It is delivered from a nuclear reactor to institutions in vials containing 5, 10, or 20 GBq. The therapy is given into the hepatic artery with the intent of delivering 125–150 Gy (12,500–15,000 rads) of radiation to the target area of the liver.

Planning Procedure

Treatment planning includes a routine clinical history, physical examination, complete blood count, liver function, and renal function tests. Three Therasphere-specific tests are also performed before treatment. These are:

  • 1Computed tomography (CT) scan of chest and abdomen, with algorithm for calculation of the volume of each liver lobe, which is necessary for determination of the dose of Therasphere to be given. To date, we have excluded patients with extra-hepatic cancer.
  • 2Planning celiac and hepatic angiography, to define the vascular anatomy. Patients with arteries arising from the right or left hepatic arteries which supply the gallbladder, stomach, or intestine, and which cannot be embolized, are excluded from treatment to avoid radiation damage to these organs.
  • 3A 99mTc-macro-aggregated albumin scan is performed after injecting 99mTc- macro-aggregated albumin through the right or left hepatic artery at the time of the planning angiogram, to calculate the percent radiation that might go to the lungs, after a concomitant single photon emission tomography (SPECT) scan. Typically, patients with greater than 16% of the injected macro-aggregated albumin that gets shunted to the lungs (calculated lung dose > 16.5 mCi of 90Yttrium) are considered to be at high risk from radiation pneumonitis from this treatment approach and so are excluded from therapy.

Selection Criteria

After the above evaluation, patients are considered for treatment who have all of the following. Biopsy-proven HCC which is confined to the liver, bilirubin less than 2.0 mg/dL, creatinine less than 2.0 mg/dL, platelets > 60 K/μL, absolute granulocytes > 1,500/μL, conventional hepatic arterial anatomy, a lung shunt < 16%, and ECOG performance ≤ 2. Since Therasphere is approved by the Food and Drug Administration currently for the treatment of HCC in the U.S. under a Humanitarian Device Exemption, patients sign an institutional review board-approved consent form and all patients and their follow-up data are entered into our institutional review board approval protocol and database. We plan the treatment date, knowing the specific activity and delivery date and allowing the radioactivity to decay to the calculated delivery dose of 125–150 Gy, based upon the liver lobe volume to be treated, obtained from the CT scan volume calculations:

Dose Calculation

Amount of radioactivity required to deliver the dose to the selected liver target (whole liver or single lobe) is calculated using the following formula:

equation image

Calculation of the liver dose (Gy) delivered after injection is provided by the following formula:

equation image

* F is the fraction of injected activity localizing in the lungs, as measured by Tc-99m MAA scintigraphy. The upper allowable limit of injected activity shunted to the lungs is F × A = 0.61 GBq (16.5 mCi)

To minimize toxicity to the cirrhotic liver, we aim to treat only one liver lobe at any 1 treatment session. However, we have treated the whole liver through the proper hepatic artery in occasional patients with prior history of lobe resection.

Therapy Procedure

This is performed on an outpatient or on a 23-hour hospitalization basis. After a hepatic artery catheterization via the femoral artery (Seldinger technique), the catheter is placed in the major right or left hepatic artery and the total content of the Therasphere vial is rapidly injected via the manufacturer's administration set (Fig. 1). In attendance are the angiographer, nuclear medicine physician, radiation safety personnel, and the supervising oncologist (Occasional patients cannot get treated due to last minute surprises in the lung or liver shunt or vascular anatomy despite all the planning. The patients are then given TACE. Therefore, all our patients sign both Therasphere and TACE consent forms). Patients are monitored post-procedure for observation of the groin puncture site and vital signs and receive intravenous hydration, anti-emetics, and analgesics.

Figure 1.

90Yttrium administration set-up.


Patients are seen in clinic with physical examination, blood work, and CT scan each month for 3 months, then every 3 months post therapy. To minimize hepatic radiation toxicity, we have only given a re-treatment to the same lobe (maximum of 2 per lobe) if sequential CT scans show increasing tumor size.

Current Results

Sixty-five patients have been treated from August 2000 to August 2003. The longest survivor to date has been 32 months post-therapy.

Baseline Characteristics

The median patient age was 69 years. There were 47 males and 18 females. Underlying liver disease was as follows: cirrhosis, 48 patients (75%); HCV, 27 patients (45%); HBV, 15 patients (26.8%); alcohol, 28 patients (47.5%); intravenous drug abuse, 8 patients (16.3%); HIV, 2 patients. Baseline Okuda cirrhosis staging was as follows: stage I, 42 patients (64.6%); stage II, 21 (32.3%); and stage III, 2 patients (3.1%). Tumor location at baseline: bilobar = 32 patients (50.8%); right lobe = 28 patients (44.4%); and left lobe = 3 patients (4.8%). Tumor burden: 10 patients (15.4%) had > 50% tumor replacement of the liver and 18 patients (27.7%) had major portal vein thrombosis.


Forty-six patients had 1 cycle, 17 patients had 2 cycles, and 2 patients had 3 cycles of Therasphere treatment. Median time between repeat treatments was 90 days. The mean dose delivered was 145.7 Gy, and the median dose was 134.3 Gy (range: 61.1–280.9 Gy). A typical dose given was either 5 or 10 GBq (2 or 4 million microspheres).

Clinical Toxicities

Nine patients had abdominal pain, and 8 patients developed worsening ascites. Two patients developed acute cholecystitis, requiring cholecystectomy. Additional adverse events included 1 episode each of generalized pain and urinary electrolyte wasting; several patients had transient nausea and ankle edema.

Lab Toxicities

Twenty-nine patients (70.7%) had less than 100% increase in alanine aminotransferase (ALT) over baseline within 6 months; bilirubin was elevated by less than 100% in 14 patients (21.5%), by 100–150% in 16 patients (24.6%), by 150–200% in 10 patients (15.4%), and by more than 200% in 25 patients (38.5%) within 6 months after treatment (18 transiently). Prothrombin was elevated by less than 100% in 62 patients (96.9%) within 6 months of treatment. Decreases in platelets and absolute granulocyte count were minimal. However, a prominent finding was lymphopenia, which lasted for longer than 12 months in many patients. A more than 75% lymphocyte decrease occurred in 19 patients (33.3%); 50–75% in 28 patients (49.1%); 25–50% in 6 patients (10.5%); and less than 25% decrease in 4 patients (7.1%). There appeared to be no obvious clinical consequence of this profound and prolonged lymphopenia, which reversed in 2 transplanted patients.

Therapeutic Results of Treatment

CT changes post-therapy

42 patients (64.6 %) had a substantial decrease in tumor vascularity within 4 months of therapy that was persistent in most patients. Twenty-five patients (38.4%) had a formal partial response, as judged by a decrease in CT scan measurements, using the product of 2 perpendicular diameters.


Okuda stage I (n = 42) median survival was 649 days (95% confidence interval: 360–1012). Historical comparison median was 244 days. Okuda stage II (n = 23) median survival was 302 days (95% confidence interval: 166–621). Historical controls median survival was 64 days (Fig. 2).

Figure 2.

Survival from diagnosis by Okuda Stage I & II: Pittsburgh August 2000 through August 2003.

Causes of death

The cause of death (n = 42) was liver failure in 21 patients (1 definitely related, 2 probably, and 9 late); other causes, 6 (1 possibly related); unknown cause, 5; HCC progression, 6; metastatic cancer, 3; and road traffic accident, 1.


90Yttrium glass microspheres (Therasphere) represents promising new therapy for unresectable HCC confined to the liver with or without portal vein thrombosis, that combines effectiveness with many fewer toxicities than TACE. Furthermore, patients seem to like it because of the small number of treatments and thus hospitalizations needed and the fewer invasive procedures. Over a 12-month period, it is, therefore, much less expensive than repetitive TACE. All our patients were entered into a quality of life assessment study (QoL FACT-Hep score) and all scored higher than patients treated over the same time period with TACE,7 with fewer hospital days for either treatment or toxicities. In this initial U.S. experience, Therasphere appears to significantly extend survival in patients with unresectable HCC compared with historical controls that were matched for prognosis by Okuda staging.