Click document to read full article
SIRveNIB: Selective Internal Radiation Therapy Versus Sorafenib in Asia-Pacific Patients With Hepatocellular Carcinoma
Selective internal radiation therapy or radioembolization (RE) shows efficacy in unresectable hepatocellular carcinoma (HCC) limited to the liver. This study compared the safety and efficacy of RE and sorafenib in patients with locally advanced HCC.
Patients and Methods
SIRveNIB (selective internal radiation therapy v sorafenib), an open-label, investigator-initiated, phase III trial, compared yttrium-90 (90Y) resin microspheres RE with sorafenib 800 mg/d in patients with locally advanced HCC in a two-tailed study designed for superiority/detriment. Patients were randomly assigned 1:1 and stratified by center and presence of portal vein thrombosis. Primary end point was overall survival (OS). Efficacy analyses were performed in the intention-to-treat population and safety analyses in the treated population.
A total of 360 patients were randomly assigned (RE, 182; sorafenib, 178) from 11 countries in the Asia-Pacific region. In the RE and sorafenib groups, 28.6% and 9.0%, respectively, failed to receive assigned therapy without significant cross-over to either group. Median OS was 8.8 and 10.0 months with RE and sorafenib, respectively (hazard ratio, 1.1; 95% CI, 0.9 to 1.4; P = .36). A total of 1,468 treatment-emergent adverse events (AEs) were reported (RE, 437; sorafenib, 1,031). Significantly fewer patients in the RE than sorafenib group had grade > 3 AEs (36 of 130 [27.7%]) v 82 of 162 [50.6%]; P , .001). The most common grade > 3 AEs were ascites (five of 130 [3.8%] v four of 162 [2.5%] patients), abdominal pain (three [2.3%] v two [1.2%] patients), anemia (zero v four [2.5%] patients), and radiation hepatitis (two [1.5%] v zero [0%] patients). Fewer patients in the RE group (27 of 130 [20.8%]) than in the sorafenib group (57 of 162 [35.2%]) had serious AEs.
In patients with locally advanced HCC, OS did not differ significantly between RE and sorafenib. The improved toxicity profile of RE may inform treatment choice in selected patients.
The majority of patients with hepatocellular carcinoma (HCC) are found in the Asia-Pacific, where the main etiologies are different from those of the West.2 Comparative studies have suggested that treatment outcomes in HCC may differ between Asia-Pacific and Western patients and between different etiologies for the same therapeutic modalities. In both regions, early-stage HCC benefits from potentially curative therapies, but most patients are diagnosed with intermediate or advanced HCC, where treatment options are limited and the prognosis poor. Sorafenib, an oral multikinase inhibitor, is the reference treatment in locally advanced and metastatic HCC.
However, dose reduction or discontinuation of sorafenib is frequently necessary because of adverse events (AEs). Selective internal radiation therapy or radioembolization (RE) with yttrium-90 (90Y) microspheres delivers brachytherapy via the hepatic artery directly to liver tumors. In retrospective and cohort studies, patients with unresectable HCC treated with RE showed better survival than a matched control group and showed promising tumor regression in intermediate and advanced HCC limited to the liver.
Prospective randomized comparisons of efficacy and safety between RE and sorafenib in locally advanced HCC would address the clinical needs of this important segment of patients. The recently reported the SorAfenib versus Radioembolization in Advanced Hepatocellular carcinoma (SARAH) study, a randomized controlled trial comparing RE with sorafenib in patients from France with locally advanced nonmetastatic HCC, showed no significant difference in overall survival (OS). RE, however, delivered a better tumor response rate (TRR) than sorafenib and a significantly better health-related quality of life (HRQoL) and was better tolerated. In this, the Selective Internal Radiation Therapy Versus Sorafenib (SIRveNIB) study, we compared RE and sorafenib in an Asia-Pacific population with locally advanced HCC.
PATIENTS AND METHODS
The SIRveNIB study was conducted in accordance with the Declaration of Helsinki, approved by the Ethics Committee of each participating center, and complied with the provisions of the Good Clinical Practice Guidelines. All patients provided written informed consent. An independent Data Safety and Monitoring Board monitored safety and efficacy. This trial is registered with ClinicalTrials.gov (ClinicalTrials.gov identifier: NCT01135056). The study protocol has been published previously.
Eligible patients were > 18 years old with an unequivocal diagnosis of HCC (on the basis of the American Association for the Study of the Liver imaging criteria or biopsy), had locally advanced cancer (Barcelona Clinic Liver Cancer [BCLC] stage B or C without extrahepatic disease) with or without portal vein thrombosis (PVT), and were not amenable to curative treatment modalities. Patients were excluded if they had received more than two previous administrations of hepatic artery–directed therapy, hepatic artery–directed treatment within 4 weeks, previous treatment with sorafenib or vascular endothelial growth factor inhibitors, or previous radiotherapy. A complete list of inclusion and exclusion criteria is given in the Data Supplement.
Study Design and Interventions The SIRveNIB trial was a prospective, randomized, investigator-driven, open-label, multicenter, phase III trial conducted at centers in the Asia-Pacific region, comparing the efficacy and safety of a single delivery of RE versus continuous sorafenib dosing in patients with locally advanced HCC. Random assignment was performed through an internet application at the patient level. Eligible patients were randomly assigned in a 1:1 ratio to receive either RE or sorafenib and stratified according to center and the presence or absence of PVT. Oral sorafenib (Nexavar, Bayer HealthCare Pharmaceuticals, Berlin, Germany) treatment commenced at 400 mg twice daily in the week after random assignment. Treatment could be delayed, or the dose reduced or discontinued, in case of significant toxicities or treatment-related AEs.
Sorafenib treatment was continued until there was evidence of treatment failure (tumor progression at any site), complete response, the initiation of other HCC therapies, unacceptable toxicity, patient request to stop treatment, or death. Patients randomly assigned to RE underwent angiographic and technetium-99 macro-aggregated albumin assessment of suitability for RE as described previously. Eligible patients received a single delivery of RE ( 90Y loaded resin microspheres [SIR-Spheres; Sirtex Medical Limited, New South Wales, Australia]) within 35 days of random assignment. The activity and dose were calculated according to the body surface area model or the partition model. RE was not administered if the liver-to-lung shunt would result in > 20 Gy being delivered to the lungs.
The primary end point was OS, defined as the time from the date of random assignment to death as a result of any cause or the last follow-up date if the patient was alive. Secondary end points included: TRR, disease control rate (DCR), progression-free survival (PFS; according to Response Evaluation Criteria in Solid Tumors 1.1), and time to tumor progression (TTP) at any site and in the liver (Data Supplement). AEs were reported according to National Cancer Institute Common Terminology Criteria for AEs, version 4.02. HRQoL was measured using the EuroQol-5 Dimension (EQ-5D) questionnaire.
Tumor response was assessed by American Association for the Study of the Liver imaging criteria every 12 weeks from the date of random assignment to disease progression, through central review by radiologists at the National Cancer Center Singapore blinded to the treatment received. Patients were clinically assessed for safety every 4 weeks for the first 3 months, and every 3 months thereafter, until study conclusion. Patients completed EQ-5D questionnaires at each visit and were followed up until death (Data Supplement). AEs and serious AEs (SAEs) were recorded from the time of signing the written informed consent until 30 days after the final sorafenib dose, or until 30 days post RE regardless of causality, and for an additional 5 months post RE if judged by the investigator to be causally related to RE.
Sample size calculation was based on OS data reported by Kang et al25 and by Sangro et al, with assumed median OS of 9.35 months for patients treated with sorafenib and 14 months for patients treated with RE. The study was designed to detect superiority or detriment of RE compared with sorafenib with respect to OS. The sample size was determined using a two-sided log-rank test with type I error of 5% and power of 90%. A dropout rate of up to 20% was factored into the computations. The anticipated study duration was 5 years: 3 years of accrual and 2 years of follow-up, corresponding to an estimated hazard ratio (HR) of 0.67, with 266 deaths by study end. The number of patients required for random assignment was calculated as 360 (180 per group).
The primary analysis was conducted on the intention-to-treat (ITT) population (Data Supplement). Median OS, TTP, and PFS were estimated using Kaplan-Meier plots with corresponding two-sided 95% CIs. The two groups were compared using the Cox regression analysis on the basis of HR and two-sided 95% CI. Survival rates at 6, 12, and 18 months were estimated using Kaplan-Meier plots and compared between the two groups using the log-rank test. TRR and DCR were compared using the Fisher’s exact test. Considering the sizable proportion of patients not receiving the study treatments, sensitivity analyses were performed for OS, TTP, PFS, TRR, and DCR on the basis of the treated population, similar to the ITT population.
Subgroup analyses were performed using the Cox regression analysis to examine the effect on OS of the presence or absence of PVT, BCLC stage, hepatitis (B or C), Eastern Cooperative Oncology Group status, previous transarterial chemoembolization, Child-Pugh class, and the albumin-bilirubin score for the ITT and treated populations. Multivariable Cox regression analyses were performed to compare OS between the two groups adjusted for these covariates on the basis of the ITT and treated populations.
The EQ-5D index was derived using the Singaporean value set. HRQoL was analyzed using a linear mixed-effects model with patient-specific random intercepts and an unstructured covariance matrix, on the basis of the ITT and treated populations. AE rates were compared between the two groups using the Fisher’s exact test, on the basis of the treated population. Further analyses were performed to examine the differences between RE and sorafenib in terms of quality-adjusted time with grade > 2 toxicity (qTOX) and time without grade 2 to 4 toxicity of treatment (TWiST) before disease progression up to 2 years. Time spent with toxicity of > 2 (TOX) was weighted by mean EQ-5D utility during this state to obtain qTOX. Utility for TWiST state was set as 1. Restricted means and their bootstrap two-sided 95% CIs were derived to assess the statistical significance of treatment differences in TOX, qTOX, and TWiST, on the basis of the ITT and treated population. SAS software version 9.4 (SAS Institute, Cary, NC) was used. A P value < 0.05 was deemed to be statistically significant.
Between July 16, 2010 and May 25, 2016, 360 patients were recruited at 27 centers in 11 countries in the Asia-Pacific region, randomly assigned to RE (n = 182) or sorafenib (n = 178), and included in the ITT population (Fig 1). There were no clinically relevant differences between the two treatment groups at baseline (Table 1). Of 182 patients randomly assigned, 130 (71.4%) received a single administration of 90Y resin microspheres a median of 21 days after random assignment. The mean activity delivered to the liver was 1.8 GBq (range, 0.4 to 8.8 GBq; Data Supplement).
Overall, 162 (91.0%) of 178 patients received sorafenib; median time to start of treatment after random assignment was 3 days, and median treatment duration was 13.8 weeks (interquartile range, 19.3 weeks). The mean daily dose of sorafenib was 644.5 mg (SD, 203.5 mg). Overall, 52 patients (28.6%) and 16 patients (9.0%) randomly assigned to RE and sorafenib, respectively, did not receive the assigned treatment. In the ITT population, 33 patients (18.1%) and 42 patients (23.6%) in the RE and sorafenib groups, respectively, received subsequent treatment of HCC, with no significant crossover to either group (Data Supplement). The Data Supplement shows patient disposition at study conclusion.
A total of 266 patients had died at the time of analysis. Median OS in the ITT population was 8.8 and 10.0 months in the RE and sorafenib groups, respectively (HR, 1.12; 95% CI, 0.9 to 1.4; P = .36; Fig 2A). Differences in survival rates at 6, 12, and 18 months were 0.2% (P = .97), 29.8% (P = .07), and 24.2% (P = .42; Table 2). Median OS in the treated population was 11.3 and 10.4 months in the RE and sorafenib arms, respectively (HR, 0.86; 95% CI, 0.7 to 1.1; P = .27; Fig 2D). Differences in survival rates at 6, 12, and 18 months were 13.5% (P = .0081), 21.6% (P = .80), and 2.7% (P = .65; Table 2). OS at 6 months in the treated population was significantly different between RE and sorafenib: 81.0% (95% CI, 60.7 to 75.3) versus 68.0% (95% CI, 3.5 to 23.6; P = .0081), respectively. There was no statistically significant difference in OS between treatment groups in preplanned subgroups or in major risk factor–adjusted analyses (Data Supplement), except for patients with BCLC C disease in the treated population, with OS 9.2 versus 5.8 months (HR, 0.67; 95% CI, 0.4 to 1.0; P = .0475) in the RE and sorafenib groups, respectively.
In the ITT population, the TRR was significantly greater in the RE than in the sorafenib group: 16.5% versus 1.7% (P , .001), respectively (Fig 3A; Data Supplement). In the treated population, it was similarly significant: 23.1% versus 1.9%, respectively (P , .001; Fig 3A; Data Supplement). There was no statistically significant difference in DCR in either population (Fig 3B; Data Supplement).
For the RE and sorafenib groups, median PFS at any site was 5.8 versus 5.1 months (HR, 0.89; 95% CI, 0.7 to 1.1; P = .31; Fig 2B) for the ITT population and 6.3 versus 5.2 months (HR, 0.73; 95% CI, 0.6 to 0.9; P = .0128; Fig 2E) for the treated population. Median PFS in the liver was similar to PFS at any site in both treatment groups and populations (Data Supplement). Median TTP at any site in the RE and sorafenib groups was 6.1 versus 5.4 months (HR, 0.88; 95% CI, 0.7 to 1.1; P = .29; Fig 2C), respectively, in the ITT population and 6.4 versus 5.4 months (HR, 0.73; 95% CI, 0.6 to 0.9; P = .0188; Fig 2F), respectively, in the treated population.
In total, 1,468 treatment-emergent AEs were reported over the course of the study: 437 in the RE group and 1,031 in the sorafenib group. Fewer patients in the RE arm than in the sorafenib arm experienced one or more AEs: 78 (60.0%) of 130 patients and 137 (84.6%) of 162 patients; AEs of grade > 3: 36 (27.7%) and 82 (50.6%); or serious AEs (SAEs): 27 (20.8%) and 57 (35.2%), respectively (Data Supplement). Dermatologic events, including alopecia, palmar-plantar erythrodysesthesia syndrome, and rash, and GI events, including diarrhea and constipation, occurred at higher frequencies in the sorafenib group (Table 3).
In the ITT population, mean duration of TOX with and without HRQoL adjustment was significantly shorter and mean duration of TWiST significantly longer in the RE group compared with the sorafenib group. Similar results were found in the treated population (Data Supplement). There were no statistically significant differences in the EQ-5D index between the RE and sorafenib groups throughout the study in either the ITT or treated populations (Data Supplement).
In the SIRveNIB trial, RE failed to demonstrate any statistical difference in OS compared with sorafenib in the treatment of locally advanced HCC. RE was neither superior nor detrimental compared with sorafenib in locally advanced HCC. However, TRR and AE frequency were improved with RE over sorafenib. This is the first prospective study, to our knowledge, comparing these treatments in a population from the Asia-Pacific region and supports the evidence from SARAH that RE may offer a better-tolerated alternative to sorafenib for patients with HCC.
Median OSs in the sorafenib group of SIRveNIB for patients with BCLC B and C disease were similar to those reported in the subset analysis of the Asia-Pacific sorafenib trial11 (14.4 and 5.8 months and 14.8 and 5.8 months in the SIRveNIB ITT and treated populations, respectively, v 14.2 and 5.6 months in the sorafenib trial; Data Supplement). BCLC B is, however, a particularly heterogeneous group. Other staging systems, such as the Hong Kong Liver Cancer staging system, may provide more granularity, but this was available only after the trial commenced. The SARAH study had a similar design to SIRveNIB and was conducted in a European population, where the etiology of HCC differs from that in an Asia-Pacific population. The results of SARAH, similarly to SIRveNIB, failed to demonstrate a statistically significant difference in OS between RE and sorafenib groups (8.0 and 9.9 months, respectively; HR, 1.15; 95% CI, 0.9 to 1.4; P = .18), but RE was better tolerated with a significantly better HRQoL. TRRs were superior in the RE group in both SIRveNIB and SARAH. In the SIRveNIB-treated population, this manifested in better OS at 6 months with RE, compared with sorafenib (81.5% v 68.0%; P = .0081). No superiority in OS was observed at 12 and 18 months, consistent with the short TTPs in the ITT and treated populations. The SARAH study did not report OS for these time points.
In the ITT population, a trend toward longer TTP and PFS (at any site and in the liver) with RE was suggested and was significant in the treated population. The high proportion of patients in the ITT population assigned to the RE group who did not receive RE (28.6%) may provide a rationale for the failure to demonstrate the favorable effect of RE on secondary end points in the ITT population.
Tolerability was significantly better in the RE than in the sorafenib arm. Importantly, significantly fewer patients who received RE experienced one or more AEs, grade > 3 AEs, and SAEs. The high AE rate in the sorafenib group is consistent with treatment discontinuation rates observed in other studies. The safety profile of RE was as expected.
Although the SIRveNIB and SARAH trials had similar designs and outcomes, there were also significant differences. SIRveNIB only allowed a single delivery of RE, but SARAH allowed sequential or repeated RE delivery: in SARAH, 37% of patients receiving RE received more than one delivery. The exclusion criteria for serum bilirubin differed significantly: 32 mmol/L and # 50 mmol/L in SIRveNIB and SARAH, respectively. A higher proportion of patients in SARAH had BCLC C disease (68.4% and 67.1% for RE and sorafenib, respectively) than in SIRveNIB (54.9% and 61.2%, respectively). Patients with BCLC C disease showed better survival with RE than with sorafenib (9.2 v 5.8 months; P = .0475) in SIRveNIB, but not in SARAH, perhaps because of increased toxicity in patients with higher serum bilirubin levels.
The SIRveNIB study has some limitations. First, more patients randomly assigned to RE than to sorafenib did not receive the assigned treatment (28.6% and 9%, respectively). Sorafenib, a systemic oral treatment, can be initiated rapidly, whereas RE, a liver-directed treatment, requires several days of work-up and assessment of the patient’s suitability.
Consequently, 15.9% of patients allocated to RE were found to be ineligible but were included in the ITT analysis. The Asia-Pacific region is highly heterogeneous, and when SIRveNIB was conducted not all centers had facilities for RE. Patients from five of the 11 participating countries had to travel to Singapore for RE assessment and, if eligible, for RE, before returning to their home country’s center for follow-up. This delayed treatment initiation and, consequently, patients progressed, died, or withdrew consent before treatment. Second, HRQoL was measured using the EQ-5D questionnaire, which is not a specific index for liver disease. Although the EQ-5D index was similar between the RE and sorafenib groups, mean duration of TOX with and without HRQoL adjustment was significantly shorter and mean duration of TWiST significantly longer in the RE than in the sorafenib group. Last, the significantly better OS in the RE group compared with the sorafenib group for patients with BCLC C disease was based on a relatively small sample size, and the analysis was not adjusted for the multiplicity correction. This finding should be considered exploratory, requiring confirmation in a sufficiently powered study.
In conclusion, in this multicenter trial involving patients with intermediate or advanced HCC without extrahepatic metastases, OS did not differ significantly between the RE and sorafenib groups. RE was associated with a higher TRR and a significantly lower frequency of AEs than sorafenib.