Comparison of yttrium-90 radioembolization and transcatheter arterial chemoembolization for the treatment of unresectable hepatocellular carcinoma

Kooby DA, Egnatashvili V, Srinivasan S, et al. (2009) Comparison of yttrium-90 radioembolization and transcatheter arterial chemoembolization for the treatment of unresectable hepatocellular carcinoma. J Vasc Interv Radiol, 21(2):224-230

Introduction

Radioembolization is a form of brachytherapy that allows for concentrated beta-radiation administration to tumor tissue while minimizing damage to surrounding liver parenchyma (13,14). It appears to be somewhat tumor-selective based on natural disruptions to the micro-vasculature surrounding liver tumors (15,16) and can be delivered selectively with segmental, lobar, or whole-liver approaches (17). Radioembolization appears to rely less on static arterial embolization than chemoembolization and, as such, may induce less hepatocyte damage in patients with impaired baseline liver function and therefore be preferable in patients with portal vein tumor thrombus.

Materials and Methods

Patient Selection and Data Sources: As radioembolization was first used at our institution in January of 2003, data from a prospectively maintained database of patients treated with radioembolization (2003 to present) were compared with those from a retrospective database of patients treated with chemoembolization (1996 to present). Hospital and outpatient records for each identified patient were examined to verify disease status, treatment response, toxicity of therapy, and survival. Patients were contacted with letters and follow-up telephone calls, according to a scripted protocol approved by institutional review board.

Radioembolization: The use of SIR Spheres (SIRTex Medical, New South Wales, Australia) for HCC is off label by the U.S. Food and Drug Administration. All patients underwent superior mesenteric and celiac angiography to define the arterial supply of the liver, HCC, and stomach. Potential hepaticoenteric arterial communications were investigated (gastroduodenal, supraduodenal, retroduodenal, right or accessory gastric, falciform, and accessory or inferior phrenic arteries) and embolized with coils to prevent unintended radiation-induced injuries. The gastroduodenal artery was embolized in each case. Other vessels were embolized selectively if identified, but this was uncommonly performed. After protective coil embolization, 4–5 mCi (148–185 MBq) of technetium 99m-macroaggregated albumin was injected in the proper hepatic artery and scintigraphy performed to detect arteriovenous shunting and lung shunt fraction in all patients. The shunt fraction was defined as the proportion of total lung counts to total lung and total liver counts. A pulmonary shunt fraction of less than 20% was considered acceptable for treatment (19,20).

Therapy was administered as previously described (21). Briefly, radioembolization was performed via a 3-F microcatheter by using a lobar approach with 90Y resin-based microspheres (SIRSpheres). The 90Y dose was based on the extent of tumor involvement in the liver, which was calculated by using computed tomography (CT) or magnetic resonance volumetric imaging, adjusted by lung shunt fraction. The diameter of resin-based 90Y was about 29–35 m, with average activity of 40 Bq per microsphere. The actual dose was calculated by using the body surface area method, decreased by the degree of lung shunting. The body surface area was calculated in square meters as 0.20247 (height 0.725 m) (weight  0.425 kg). Tumor involvement was calculated as (tumor volume 100)/ (tumor volume liver volume). The actual dose (in gigabecquerels) was calculated as (body surface area – 0.2) (tumor involvement/100). Microsphere administration was performed in an angiography suite with a standard angiographic technique similar to that of chemoembolization. A microcatheter was placed in either the right or left hepatic artery before the administration of 90Y. The administration of 90Y was alternated with the infusion of sterile water and iodinated contrast medium. Contrast medium was used to visualize under fluoroscopy and monitor the status of flow to the treating tumor and feeding vessel. Sufficient pressure should be applied during the administration to keep 90Y adequately suspended in sterile water. The endpoint of the injection was either delivering the entire calculated and supplied dose (20–60 mCi[740–2,220 MBq] per lobe treated) or reaching the vascular flow stasis. Once vessel stasis was detected, no further injection of remaining 90Y dose was performed.

A radiation dose-rate meter was placed next to the source vial to assess the remaining activity within the vial during the procedure. The used vial and connecting catheters were placed in a 1-cm-thick Plexiglas jar to estimate residual activity. The administered activity was determined from the ratio of exposure from the Plexiglas jar to the exposure from the microsphere vial upon receiving the source (normalized to the same distance and corrected for the decay).

Treatment Effect: Patients were evaluated for treatment response with helical three-phase, thin-cut CT 4 weeks after treatment and then every 12 weeks if stable. Patients receiving incomplete treatmentat the first setting were re-treated at 4 weeks and then assessed 12 weeks later. Tumor response based on comparative evaluation of pre- and posttreatment axial imaging was graded by using Response Evaluation Criteria in Solid Tumors (RECIST). For patients with bilobar disease who received unilobar therapy, response was assessed according to the first side treated. 

Survival and Multivariate Analysis: Overall survival was calculated starting from the time of initial treatment up until the date of the last follow-up or death. Mortality data were collected from hospital medical records, from the Social Security Death Index interactive search engine, and through telephone contact. Because all patients had unresectable malignancy and/or unresectable cirrhosis, overall survival and disease-specific survival were similar; thus, overall survival is reported. Multivariate analysis of factors potentially bearing independent association with overall survival was performed, including dichotomous disease and treatment variables.

Toxicity: To standardize reporting, treatment complications were graded on a scale of 1 to 5 according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (version 3.0) (23). In articular, mild, moderate, severe, and life-threatening or disabling toxicity were represented by grades 1, 2, 3, and 4  respectively. Toxicity resulting in death was considered grade 5. For stratification, anything over a grade 2 event was assessed as a “major” toxicity. Specific categories of toxicities were also assessed. Treatment-related toxicity within 30 days of the first embolization is reported. 

Statisitical Analysis: Data were analyzed and outcomes evaluated for all patients in both treatment groups. The Fisher exact and Student t tests were used to compare categoric and continuous variables between the two groups, respectively. Overall survival curves for each variable were generated with the Kaplan-Meier technique. Differences in survival curves were assessed by using the log-rank test. Multivariable stratified Cox regression was used to evaluate the association between multiple variables and survival. A P value of less than .05 was considered to be statistically significant. The statistical analysis of results was performed with the aid of SPSS 15.0 software (SPSS,Chicago, Illinois).

Results

Patient Characteristics: Sixty-four of the 135 patients (47%) underwent additional therapy with partial hepatectomy, liver transplantation, or other liver directed treatment (radiofrequency ablation,cryoablation, or percutaneous ethanol injection) 

Treatment Effectiveness: Results of therapy are summarized in Table 2. On average, three posttherapy scans were obtained for each patient.

The median follow-up for all patients was 6 months, which was the same for each treatment group.

Table 3. Table 4

Discussion

Radioembolization is not a new concept as experimentation with regional infusion of radioisotopes for unresectable liver cancer dates back to the 1970s (25,26).Initial experience with radiation-induced hepatitis and pneumonitis tempered enthusiasm for this approach, especially as the therapeutic window was thought to be small (27). Better understanding of the principles of shunt fraction and administration made this form of therapy more accessible (28,29). More recently, one randomized trial showed a progression free benefit of radioembolization in patients with unresectable, liver-only colorectal cancer metastases (30), and radioembolization received Food and Drug Administration approval. To our knowledge, this comparative finding has not been reported. 

We acknowledge that our data and study design have limitations.Our study is a retrospective analysis with a relatively small number of patients. Because it is not a randomized trial, selection bias and late-look bias are inherent.Furthermore, the patients in this experience had advanced disease, rendering meaningful assessment

of survival with respect to therapy choice more questionable.

Questions/ Thoughts