| |
Efficacy and safety of sorafenib in patients in the Asia-Pacific region with advanced hepatocellular carcinoma: a phase III randomised, double-blind, placebo-controlled trial
|
| |
| |
The Lancet Oncology, Early Online Publication, 17 December 2008
Dr Ann-Lii Cheng MD a Corresponding AuthorEmail Address, Yoon-Koo Kang MD b, Zhendong Chen MD c, Chao-Jung Tsao MD d, Shukui Qin MD e, Jun Suk Kim MD f, Rongcheng Luo MD g, Jifeng Feng MD h, Shenglong Ye MD i, Tsai-Sheng Yang MD j, Jianming Xu MD k, Yan Sun MD l, Houjie Liang MD m, Jiwei Liu MD n, Jiejun Wang MD o, Won Young Tak MD p, Hongming Pan MD q, Karin Burock BS r, Jessie Zou MD s, Dimitris Voliotis MD t, Zhongzhen Guan MD u
Editors' note: Although the efficacy of sorafenib for the treatment of hepatocellular carcinoma has been established in a predominantly European population, its usefulness in the Asia-Pacific region-where the majority of cases of hepatocellular carcinoma occur, and where the main cause of this cancer is chronic infection with hepatitis B virus, rather than chronic hepatitis C infection-is less clear. In this phase III, randomised controlled trial, investigators assessed the efficacy and safety of the drug in an Asian-Pacific population. While the improvement in overall survival seen in this population was smaller than in the European population, these data are an encouraging sign in what has long been a difficult-to-treat cancer.
Conflicts of interest
KB is employed by Bayer Schering Pharma and owns stock in Bayer AG. DV is employed by Bayer Healthcare. A-LC has had a consultant/advisory role, received honoraria, and participated in an advisory board for Bayer Schering Pharma, Pfizer, and Merck Serono. Y-KK has had a consultant/advisory role, received honoraria, participated in an advisory board, and received research funding from Bayer Healthcare. ZG has received honoraria from Bayer. All other authors declared no conflicts of interest.
Summary
Background
Most cases of hepatocellular carcinoma occur in the Asia-Pacific region, where chronic hepatitis B infection is an important aetiological factor. Assessing the efficacy and safety of new therapeutic options in an Asia-Pacific population is thus important. We did a multinational phase III, randomised, double-blind, placebo-controlled trial to assess the efficacy and safety of sorafenib in patients from the Asia-Pacific region with advanced (unresectable or metastatic) hepatocellular carcinoma.
Methods
Between Sept 20, 2005, and Jan 31, 2007, patients with hepatocellular carcinoma who had not received previous systemic therapy and had Child-Pugh liver function class A, were randomly assigned to receive either oral sorafenib (400 mg) or placebo twice daily in 6-week cycles, with efficacy measured at the end of each 6-week period. Eligible patients were stratified by the presence or absence of macroscopic vascular invasion or extrahepatic spread (or both), Eastern Cooperative Oncology Group performance status, and geographical region. Randomisation was done centrally and in a 2:1 ratio by means of an interactive voice-response system. There was no predefined primary endpoint; overall survival, time to progression (TTP), time to symptomatic progression (TTSP), disease control rate (DCR), and safety were assessed. Efficacy analyses were done by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00492752.
Findings
271 patients from 23 centres in China, South Korea, and Taiwan were enrolled in the study. Of these, 226 patients were randomly assigned to the experimental group (n=150) or to the placebo group (n=76). Median overall survival was 6·5 months (95% CI 5·56-7·56) in patients treated with sorafenib, compared with 4·2 months (3·75-5·46) in those who received placebo (hazard ratio [HR] 0·68 [95% CI 0·50-0·93]; p=0·014). Median TTP was 2·8 months (2·63-3·58) in the sorafenib group compared with 1·4 months (1·35-1·55) in the placebo group (HR 0·57 [0·42-0·79]; p=0·0005). The most frequently reported grade 3/4 drug-related adverse events in the 149 assessable patients treated with sorafenib were hand-foot skin reaction (HFSR; 16 patients [10·7%]), diarrhoea (nine patients [6·0%]), and fatigue (five patients [3·4%]). The most common adverse events resulting in dose reductions were HFSR (17 patients [11·4%]) and diarrhoea (11 patients [7·4%]); these adverse events rarely led to discontinuation.
Interpretation
Sorafenib is effective for the treatment of advanced hepatocellular carcinoma in patients from the Asia-Pacific region, and is well tolerated. Taken together with data from the Sorafenib Hepatocellular Carcinoma Assessment Randomised Protocol (SHARP) trial, sorafenib seems to be an appropriate option for the treatment of advanced hepatocellular carcinoma.
Funding
Bayer HealthCare Pharmaceuticals and Onyx Pharmaceuticals, Inc.
Introduction
Hepatocellular carcinoma is the third most common cause of cancer mortality worldwide.1, 2 More than 75% of cases occur in the Asia-Pacific region, largely in association with chronic hepatitis B virus (HBV) infection.3, 4 More than 50% of cases of hepatocellular carcinoma occur in China alone, and an estimated 360 000 patients residing in east Asian countries, including China, Japan, Korea, and Taiwan, die from this disease each year.1, 2 Therefore, it is especially important to assess the efficacy and safety of new agents for hepatocellular carcinoma in these high-risk populations.
Surgical resection leads to a 60-70% 5-year survival for patients with hepatocellular carcinoma who present with solitary tumours and have excellent liver function. However, because most patients present with intermediate or advanced disease, surgical resection is only an option for less than 20% of patients.5 The findings of a meta-analysis6 of treatments used in randomised, controlled clinical trials, done before the availability of sorafenib, showed that a systemic therapy to improve survival is still needed. In fact, no systemic therapies in the meta-analysis were shown to confer a survival advantage over best supportive care.6 Unfortunately, few patients from the Asia-Pacific region present with early-stage disease.7
The aetiological factors of hepatocellular carcinoma vary by geographical region.8 Hepatitis virus infection is a significant risk factor in the Asia-Pacific region, with about 70% of patients with hepatocellular carcinoma presenting with chronic HBV infection and 20% presenting with chronic hepatitis C virus (HCV) infection. These rates are roughly reversed for European, North American, and Japanese patients;8 although Japan is geographically part of the Asia-Pacific region, HCV-related hepatocellular carcinoma represents about 75% of all cases of hepatocellular carcinoma.9 Because of the high prevalence of hepatocellular carcinoma in Asia, there is an urgent need for a systemic treatment that can extend survival. Currently, treatment options for patients in the Asia-Pacific region are limited, with practice guidelines recommending resection, ablation, chemoembolisation, radiotherapy, or chemotherapy, depending on liver function and tumour burden.10, 11 The prognosis of patients from the Asia-Pacific region is frequently worse than for those from other parts of the world, such as North America or Europe.12-14
Sorafenib is an oral multikinase inhibitor with antiproliferative and antiangiogenic effects. It has been shown to inhibit the activity of the serine/threonine kinases c-Raf (Raf-1) and B-Raf; the mitogen-activated protein kinases MEK and ERK; vascular endothelial growth factor receptors (VEGFR)-1, 2, and 3; platelet-derived growth factor receptors (PDGFR)-α and β the cytokine receptor c-KIT; the receptor tyrosine kinases Flt-3 and RET; and the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway.15-19 The intracellular signalling pathway Raf/MEK/ERK20, 21 and the extracellular receptors VEGFR and PDGFR22-24 have been implicated in the pathogenesis of hepatocellular carcinoma.
In the multicentre, double-blind, randomised phase III Sorafenib Hepatocellular Carcinoma Assessment Randomised Protocol (SHARP) study,25 sorafenib was shown to be efficacious and well-tolerated in patients with advanced hepatocellular carcinoma. Patients with advanced hepatocellular carcinoma who had not received previous systemic therapy were randomly assigned to placebo (n=303) or sorafenib 400 mg twice a day (n=299). The SHARP study mainly recruited patients from North America and Europe. The primary endpoints of the study were overall survival and time to symptomatic progression (TTSP), with secondary endpoints including time to radiological progression (TTP) and safety. The baseline aetiological factors for hepatocellular carcinoma in the SHARP trial reflected the geographical patient pool, and included HCV (29% vs 27%), alcohol-associated liver disease only (26% vs 26%), and HBV (19% vs 18%) for the sorafenib and placebo groups, respectively. Median overall survival was significantly longer in the sorafenib group (10·7 months [range 9·4-13·3]) than in the placebo group (7·9 months [6·8-9·1]; hazard ratio [HR] 0·69 [95% CI 0·55-0·87]; p<0·001), but median TTSP did not differ significantly between the study groups (4·1 months [3·5-4·8] vs 4·9 months [4·2-6·3]; HR 1·08 [0·88-1·31]; p=0·77). Median TTP was significantly longer in the sorafenib group (5·5 months [4·1-6·9]) than in the placebo group (2·8 months [2·7-3·9]; HR 0·58 [0·45-0·74]; p<0·001). The most common drug-related adverse events for all grades of severity in the sorafenib and placebo groups, respectively, included diarrhoea (39% [116 of 297] vs 11% [34 of 302]), fatigue (22% [64 of 297] vs 16% [47 of 302]), hand-foot skin reaction (HFSR; 21% [63 of 297] vs 3% [8 of 302]), and rash or desquamation (16% [47 of 297] vs 11% [34 of 302]).25
The SHARP study was designed for regulatory approval of sorafenib in the USA, Europe, and other geographical regions; subsequent to this study, sorafenib was approved for the treatment of unresectable hepatocellular carcinoma by the US Food and Drug Administration and for the treatment of hepatocellular carcinoma by the European Medicines Agency. To achieve regulatory approval in China, however, it was necessary to undertake a parallel study of sorafenib in about 200 patients from the Asia-Pacific region, the findings of which are reported here. In designing this study, we chose a 2:1 drug-to-placebo randomisation method to maximise the number of patients exposed to drug, but still providing a placebo comparator to confirm the efficacy of sorafenib in this population. By doing both the Asia-Pacific study and the SHARP study in parallel, a unique opportunity was presented to assess the efficacy and safety of sorafenib in patients with advanced hepatocellular carcinoma in different geographical regions and with varying underlying aetiological factors.
Discussion
The objective of this study was to assess the efficacy and safety of sorafenib in Asia-Pacific patients with advanced hepatocellular carcinoma. Although there was no primary endpoint for this trial, patients who were randomised to treatment with sorafenib had significantly longer overall survival than did those who received placebo. Additionally, sorafenib significantly prolonged TTP and improved DCR compared with placebo. No difference in the median TTSP, however, was detected between the study groups. In view of the significant differences in overall survival and TTP favouring the sorafenib group, the possibility exists that the instrument used for the TTSP endpoint (the FSHI-8 questionnaire) might not have been sensitive enough to capture small, but significant, differences in symptoms between the two groups. Furthermore, differences in TTSP might have been confounded by both adverse events associated with drug toxicity or ongoing cirrhotic liver impairment, despite tumour stabilisation with sorafenib treatment. Both treatment groups had similar total scores on the FSHI-8 questionnaire and scores with the FACT-HP questionnaire showed no difference in quality of life between the two groups (data not shown).
Sorafenib was generally well-tolerated and had manageable side-effects, with the most common drug-related adverse events including HFSR, diarrhoea, alopecia, fatigue, rash or desquamation, and hypertension. These adverse events were predominantly grade 1 or 2. Some grade 3/4 drug-related adverse events were reported more frequently in the sorafenib group than in the placebo group, most notably HFSR and diarrhoea.
The absolute risk of discontinuation due to adverse events was considered to be low, with a similar proportion of discontinuations reported for both groups (19·5% in the sorafenib group and 13·3% in the placebo group). The need for dose reductions or discontinuation of sorafenib might be mitigated by more aggressive monitoring for adverse events, such as HFSR, and initiating management strategies as soon as symptoms appear.29 Although no guidelines and limited data are available describing the pharmacological management of HFSR, the potential for optimising patient outcomes when treating with sorafenib makes this an important area for further study.30
The overall efficacy of sorafenib in patients with advanced hepatocellular carcinoma seen here was comparable with that reported in the SHARP trial. Although the absolute survival was greater in the SHARP trial for both study groups, the HRs for survival (ie, the reduction in the risk of death associated with sorafenib treatment) were comparable between the two studies (0·68 in our study vs 0·69 in the SHARP trial). HRs provide considerable statistical insight into the relative differences between the two study groups, and suggest comparable efficacy for sorafenib in both studies. Acknowledging that cross-study comparisons must be made with care, there are potential causes for the differences in absolute survival reported in the two studies. At baseline, more patients in our trial had extrahepatic spread, a greater number of hepatic tumour lesions, poorer ECOG PS, and increased concentrations of alpha-fetoprotein (AFP) compared with patients in the SHARP trial. These characteristics suggest that the patients enrolled in our study might have had more advanced disease than did those in the SHARP trial and some of these characteristics (eg, extrahepatic spread and increased concentrations of AFP) have been shown to be prognostic factors for mortality in patients with advanced hepatocellular carcinoma.12, 31
Subanalyses were done on the basis of various factors associated with the prognosis of patients with hepatocellular carcinoma, including age, macroscopic vascular invasion, extrahepatic spread, macroscopic vascular invasion or extrahepatic spread, ECOG PS, and HBV infection. Although not statistically significant, the HRs of these analyses were less than one, suggesting that sorafenib provided benefit to all the subpopulations analysed, including those patients who normally fare worse. However, one should note that there were wide 95% CI ranges for patients with an ECOG PS of 0 and for patients 65 years of age or greater, which might be attributed to a low number of patients in each subset (n=59 for ECOG PS 0 and n=32 for ≥65 years of age).
The safety profiles for sorafenib in our study and in the SHARP trial were similar; however, there were specific differences in the incidence of certain drug-related adverse events of any grade in the sorafenib groups of both studies-eg, the incidence of HFSR (any grade) was higher in our study compared with the SHARP trial (45·0% vs 21·2%), whereas the incidence of diarrhoea (any grade) was lower in our study (25·5% vs 39·1%). We do not believe that these differences were caused by physiological differences associated with ethnicity. The most commonly reported adverse events were similar for the two study populations, and few patients treated with sorafenib in either population had grade 3/4 drug-related adverse events. Furthermore, in a phase I study of sorafenib in Japanese patients, the drug-related adverse-event profile was consistent with that noted for other populations.32 Indeed, cultural differences between the patients enrolled in our study and those in the SHARP trial might have been responsible for the differences in adverse-event reporting.
Both our study and the SHARP trial restricted enrolment to patients with Child-Pugh class A disease, because impairment of liver function associated with Child-Pugh class B or C could have potentially confounded the results.7, 26 Nevertheless, we believe sorafenib might be efficacious and safe for patients with Child-Pugh class B or C liver function. In a phase I study in Japanese patients with hepatocellular carcinoma, no differences were noted in the efficacy, safety, or pharmacokinetic profile of sorafenib in patients with Child-Pugh class A or B disease.32 This consistency of pharmacokinetics across Child-Pugh class A and B patients was also seen in a phase II study of sorafenib in patients with advanced hepatocellular carcinoma.33 However, these data should be interpreted cautiously, because each study included a relatively small proportion of patients with Child-Pugh class B disease, and further study is merited.
A reason for doing a study in patients with hepatocellular carcinoma from the Asia-Pacific region was to assess the safety and efficacy of sorafenib in a patient population from a region of the world where the prevalence of hepatocellular carcinoma is greatest and aetiological factors for this disease might differ from those in other regions. For example, the incidence of HBV and HCV infection is different in the regions from which our study and the SHARP trial recruited patients.8 Indeed, 73·0% of patients in our study had baseline HBV infection compared with 12·0% in the SHARP study, whereas 8·4% of patients in our study had baseline HCV infection compared with 30% in the SHARP trial.25 There is some evidence that patients with HBV-associated hepatocellular carcinoma might have a worse prognosis than patients with HCV-related hepatocellular carcinoma. Cantarini and colleagues31 reported a case study comparing 102 patients with HBV-related or HCV-related hepatocellular carcinoma who were matched for sex, age, liver-disease status (cirrhosis vs chronic hepatitis), stage of hepatocellular carcinoma, and Child-Pugh class. In the overall population of patients with hepatocellular carcinoma, patients with HBV-related disease had a trend toward poorer survival (p=0·16); however, when the subset of patients with advanced hepatocellular carcinoma was assessed, the survival difference was statistically significant (p=0·025).
Although other studies suggest that sorafenib might be less efficacious in patients with HBV, we do not agree with these conclusions, and believe that the large proportion of patients with HBV enrolled in our study supports the efficacy of sorafenib in this important patient subset. Indeed, a subset analysis of patients infected with HBV in our trial showed that patients with HBV who were treated with sorafenib had a longer overall survival and longer TTP than patients who received placebo.34 Moreover, the safety profile of sorafenib in patients with HBV was comparable with that for the overall study population.34
Our findings conflict with those reported by Huitzil-Melendez and colleagues,35 which suggest that sorafenib might be less efficacious in HBV-positive patients with hepatocellular carcinoma. These findings, however, were based on a retrospective analysis of a phase II study that included a small population of HBV-positive patients. Nevertheless, additional studies are warranted to confirm the benefit of sorafenib in HBV-positive patients with hepatocellular carcinoma.
There was also a need to study sorafenib in Asia-Pacific patients with hepatocellular carcinoma because of the varying practices for surgical resection, transplantation, and other therapies for this disease in different geographical regions.8, 36 In many regions of Asia, patients who might otherwise be considered to have unresectable or non-embolisable disease, because of tumour size and number of tumours or partial portal venous thrombosis, might be accepted for surgical treatment or transarterial chemoembolisation. These differences in clinical practice reflect a more aggressive approach to the treatment of hepatocellular carcinoma in the Asia-Pacific region, which in turn might account for the inclusion of patients with more advanced or complex disease in our study. Finally, the effect of as yet unknown environmental factors (eg, socioeconomic status and exposure to aflatoxin) on prognosis in hepatocellular carcinoma in different patient populations remains unexplored.
Treatment with sorafenib has been deemed cost-effective compared with best supportive care in hepatocellular carcinoma.37 An economic assessment resulted in an incremental cost-effectiveness ratio of US$63 436 per life-years gained.38 This cost-effectiveness ratio is within the established threshold that society is willing to pay, and substantially lower than alternative thresholds suggested in recent years.39, 40
Our study shows that sorafenib is a well-tolerated treatment option with an acceptable safety profile for patients with advanced hepatocellular carcinoma from the Asia-Pacific region. The efficacy results of our study were consistent with those of the SHARP trial, despite the patients randomised in our study having more advanced disease and being more likely to have different disease aetiologies (eg, HBV infection) than the patients randomised in the SHARP trial. Furthermore, the primary findings and subanalyses of our study suggest that sorafenib could be used to treat patients with hepatocellular carcinoma with different prognostic factors.
Results
Between Sept 20, 2005, and Jan 31, 2007, 271 patients from China, South Korea, and Taiwan were enrolled and randomised in the study. Of the eligible patients, 226 were randomly assigned to receive either sorafenib (n=150) or placebo (n=76) and were included in the intention-to-treat population (figure 1).
The sorafenib and placebo groups were well balanced with regard to baseline demographic and disease characteristics (table 1). Most patients in both groups had ECOG PS 1 and extrahepatic spread, and 73·0% (165 of 226) of patients had HBV infection at baseline; the incidence of HBV infection was balanced between the sorafenib and placebo groups.
Patients receiving sorafenib had a median overall survival of 6·5 months (95% CI 5·56-7·56), compared with 4·2 months (3·75-5·46) for those receiving placebo (HR 0·68 [95% CI 0·50-0·93]; p=0·014; figure 2). 6-month overall survival was 53·3% in the sorafenib group and 36·7% in the placebo group (figure 2). Patients in the sorafenib group had a significantly longer median TTP (2·8 months [2·63-3·58]) than patients in the placebo group (1·4 months [1·35-1·55]; HR 0·57 [95% CI 0·42-0·79]; p=0·0005; figure 2). There was no meaningful difference in TTSP as assessed by the FHSI-8 between patients in the sorafenib and placebo groups: median TTSP was 3·5 months (2·80-4·24) in the sorafenib group and 3·4 months (2·40-4·08) in the placebo group (HR 0·90 [95% CI 0·67-1·22; p=0·50).
In the analysis for best response, five of 150 patients in the sorafenib group (3·3%) achieved a partial response and 81 of 150 patients (54·0%) had stable disease, whereas in the placebo group, one of 76 patients (1·3%) achieved a partial response and 21 of 76 patients (27·6%) had stable disease. The DCR was significantly greater in the sorafenib group (53 of 150 [35·3%]; 95% CI 27·7-43·6) than in the placebo group (12 of 76 [15·8%]; 95% CI 8·4-26·0; table 2; p=0·0019).
Sorafenib provided a clinical benefit in all preplanned subgroup analyses, despite some patients having characteristics associated with poor prognosis-including greater ECOG PS, extrahepatic spread, vascular invasion, older age (≥65 years), presence of macroscopic vascular invasion or extrahepatic spread (or both), and HBV infection (figure 3).
149 patients in the sorafenib group and 75 patients in the placebo group were included in the safety analysis. The incidence of treatment-emergent adverse events was high in both the sorafenib and placebo groups (146 of 149 patients [98·0%] vs 71 of 75 patients [94·7%], respectively). The overall incidence of drug-related adverse events of any grade was 81·9% (122 of 149 patients) in the sorafenib group compared with 38·7% (29 of 75 patients) in the placebo group. The most frequently reported drug-related adverse events in patients treated with sorafenib were HFSR (67 of 149 patients [45·0%]), diarrhoea (38 of 149 [25·5%]), alopecia (37 of 149 [24·8%]), fatigue (30 of 149 [20·1%]), rash or desquamation (30 of 149 [20·1%]), hypertension (28 of 149 [18·8%]), and anorexia (19 of 149 [12·8%]). All these drug-related adverse events occurred at a much higher incidence in patients in the sorafenib group than in those in the placebo group. Drug-related adverse events reported by patients receiving sorafenib were predominantly grade 1 or 2 (table 3).
The overall incidence of treatment-emergent serious adverse events (ie, those that were life-threatening, resulted in death, required patient hospitalisation or prolongation of hospitalisation, or resulted in a persistent or significant disability or incapacity) was similar in both the sorafenib and placebo groups (71 of 149 [47·7%] and 34 of 75 [45·3%], respectively). Drug-related serious adverse events were reported in 13 of 149 patients (8·7%) in the sorafenib group and one of 75 patients (1·3%) in the placebo group. Treatment discontinuation due to adverse events was similar in both groups (29 of 149 patients [19·5%] and ten of 75 patients [13·3%], respectively). Dose reductions due to adverse events were needed in 46 of 149 patients (30·9%) treated with sorafenib compared with two of 75 patients (2·7%) in those who received placebo. The most common adverse events resulting in dose reductions were HFSR (17 of 149 patients [11·4%] in the sorafenib group) and diarrhoea (11 of 149 patients [7·4%] in the sorafenib group). Deaths attributed to serious adverse events were reported for 46 patients in the sorafenib group and 24 patients in the placebo group, with no deaths being categorised as drug-related.
Methods
Patients
Patients at least 18 years old with advanced (unresectable or metastatic) hepatocellular carcinoma who had not received previous systemic therapy were eligible for this trial. Eligibility criteria also included histologically or cytologically proven hepatocellular carcinoma; Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0, 1, or 2; Child-Pugh liver function class A;26 and a life expectancy of at least 12 weeks. Patients were required to have adequate renal, haematological, and hepatic function, as indicated by a platelet count of 60_109/L or greater; haemoglobin concentration of 85 g/L or more; albumin concentration of at least 28 g/L; total bilirubin concentration of 51·3 _mol/L or less; alanine aminotransferase concentration of five-times the upper limit of normal (ULN) or less; serum creatinine concentration of 1·5-times the ULN or less; and a prothrombin time-international normalised ratio (INR) of 2·3 or less or prothrombin time less than or equal to 6 s above control. Additionally, patients considered for inclusion were required to have at least one tumour lesion (not previously treated with local therapy) that could be measured along one dimension according to Response Evaluation Criteria in Solid Tumors (RECIST).27 Patients who had received previous local therapy, such as surgery, radiotherapy, hepatic arterial embolisation, chemoembolisation, radiofrequency ablation, percutaneous injection, or cryoablation, were eligible for enrolment in the study, provided that either the target lesion increased in size by 25% or more, or the target lesion had not been treated with local therapy. Furthermore, the local therapy must have been stopped at least 4 weeks before study entry. Patients with recurrent disease after previous resection were considered eligible for the study. Exclusion criteria included: previous or concomitant systemic therapy (including new, molecularly targeted therapies); known history of HIV infection; clinically serious infections; known substance abuse; history of organ allograft; history of cardiac disease; known CNS tumour; known gastrointestinal bleeding up to 30 days before study entry; and pregnancy or breast-feeding.
All patients provided written informed consent before enrolment. The study was done in accordance with good clinical practice and Declaration of Helsinki guidelines. Documented approval from appropriate ethics committees and institutional review boards was obtained for all participating centres or countries before the start of the study, according to good clinical practice, local laws, regulations, and organisations. An independent Data Monitoring Committee (DMC) monitored safety and efficacy over the duration of the study. All study sites were compliant with local laws and regulations.
Procedures
The trial was done at 23 sites within the Asia-Pacific region, in China, Taiwan, and South Korea. Eligible patients were stratified by the presence or absence of macroscopic vascular invasion or extrahepatic spread (or both), ECOG PS (0, 1, or 2), and geographical region (China, Taiwan, or South Korea) and then randomised in a 2:1 ratio by means of an interactive voice-response system (IVRS) to receive either sorafenib 400 mg orally twice daily or placebo. Each investigator was provided with the code of the group assignment and randomisation schedule for each patient; in the event of an emergency, the code was to be used to identify the assigned group and randomisation schedule. However, during the course of the study, no investigators were unblinded to treatment assignment. The study drug was provided by Bayer (Leverkusen, Germany). Randomisation was done centrally through the IVRS. Treatment interruptions and dose reductions (first 200 mg twice daily, then 200 mg once daily) were allowed for drug-related toxicity. If further dose reductions were needed, the patient was removed from the study. Patients continued therapy until death or a criterion for stopping therapy was met. Criteria for stopping therapy included: adverse events that, in the opinion of the investigator, required termination of medication; progression of disease; pregnancy (as assessed by serum concentrations of _-human chorionic gonadotropin); deterioration of ECOG PS to 4; development of a second cancer; or withdrawal of consent. Other patients who met criteria for stopping therapy included patients who used an illicit drug that, in the opinion of the investigator, could contribute to toxicity, or if the patient was lost to follow-up.
The study was designed in parallel with the SHARP study to assess the efficacy and safety of sorafenib specifically in an Asia-Pacific population with advanced hepatocellular carcinoma as a supportive trial for regulatory submissions for hepatocellular carcinoma in Asia. Therefore, no primary efficacy endpoint was designated in the protocol. Overall survival was a primary endpoint in the SHARP trial and is also the most clinically relevant endpoint for measuring efficacy. For these reasons, overall survival was used to show the statistical power of this trial. Other efficacy variables included in this analysis were TTP, TTSP, disease control rate (DCR), and safety. Overall survival was measured from the date of randomisation until death from any cause. TTP or TTSP were measured from the date of randomisation until disease progression according to RECIST or symptomatic progression. Symptomatic progression was defined as deterioration to ECOG PS 4 status or a change from baseline score on the 8-item, symptom-focused Functional Assessment of Cancer Therapy-Hepatobiliary Symptom Index (FHSI-8) questionnaire, associated with a deterioration of symptoms. DCR was defined as the proportion of patients who had a best response rating of complete response, partial response, or stable disease, which was maintained for at least 4 weeks from the first manifestation of that rating. Patient reported outcome was measured with the Functional Assessment of Cancer Therapy-Hepatobiliary (FACT-HP) domain of physical wellbeing and additional concerns at baseline and at week 12 (or at the end of treatment visit if the patient discontinued before week 12).
Although sorafenib was taken orally on a continuous daily basis, the overall treatment period was divided into 6-week cycles for the purpose of the efficacy assessment. Tumour measurements were done at screening and every 6 weeks thereafter (within the period ranging from 10 days before each cycle to day 1 of each cycle). Patient visits were scheduled every 3 weeks to monitor safety and drug accountability and to administer the FHSI-8 questionnaire by study-site personnel. Patients were monitored for adverse events by use of the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 3·0.28
Statistical analysis
The trial was designed so that 165 deaths were needed to provide a high probability of recording a numerically positive treatment effect. We estimated that about 222 patients were needed (148 patients in the sorafenib group and 74 patients in the placebo group) to record 165 death events. Assuming a true underlying improvement in overall survival of 40% with 165 events (deaths), the probability of recording any numerically positive treatment effect on survival was about 98·5%; the probability of noting a greater than 20% improvement in overall survival with sorafenib compared with placebo was about 83·9%.
By the time 165 deaths had occurred, 191 progressions were also expected. With 191 progressions, and assuming a true underlying improvement in TTP of about two-thirds (67%), the probability of noting any numerically positive treatment effect on TTP (ie, an estimated HR between sorafenib and placebo of <1 for TTP) was around 99·9%, and this probability decreased to 83·9% if the true improvement in TTP with sorafenib was 40%. Although no formal interim analyses of efficacy were planned or done, in August, 2007, the independent DMC reviewed efficacy and safety data up to March, 2007, and preliminary data for overall survival up to August, 2007, which included 167 deaths and showed a one-sided p value of 0·014. The objectives of the DMC meeting included protection of patient welfare, monitoring interim data for patient safety, identification of issues, and recommendations about study design and conduct. Because patients in the sorafenib group had an overall survival and TTP that were statistically superior to those in the placebo group, and because the data were numerically similar to data noted in the SHARP study,25 the DMC recommended switching patients from the placebo group to receive sorafenib.
Preplanned subgroup exploratory analyses of median overall survival and HRs were done in patients by age group (<65 years vs ≥65 years), the presence or absence of macroscopic vascular invasion or extrahepatic spread (or both), ECOG PS, and HBV infection.
The efficacy analysis was based on the intention-to-treat population, which included all randomised patients. The population for safety analysis included all patients who received at least one dose of sorafenib. The Kaplan-Meier method was used to calculate survival curves for time-to-event variables, and the log-rank test was used for treatment comparisons. All reported p values were two-sided. As planned according to the study protocol, the main analyses were done using a non-stratified log-rank test. However, stratified log-rank tests were also done with the same stratification factors used for randomisation (ie, ECOG, tumour burden, and region), the results of which were consistent with the non-stratified test (data not shown). Furthermore, an exploratory Cox proportional hazard analysis was done to assess the effect of important baseline characteristics on overall survival. The results of these additional analyses were consistent with the primary analysis (data not shown). HR and 95% CIs were estimated by use of a non-parametric log-rank test, the Cox proportional hazards model. A proportional hazard was not assumed. A Cox proportional hazard model was only used as a convenient way to obtain an estimate of HRs, which can be interpreted as average relative risk over time even in the absence of proportional hazard. Treatment comparisons were done in multiple individual Cox models, separately for each factor, for subgroup analyses. DCR was calculated and a 95% CI was computed for each treatment group and compared by use of Fisher's exact test. Data analyses were done by use of SAS software (version 8·2). This trial is registered with ClinicalTrials.gov, number NCT00492752.
Role of the funding source
Bayer (Leverkusen, Germany) sponsored the study design, administration of treatment, and statistical analysis of the trial. Data were managed in parallel by both the sponsor and the principal investigators. The academic investigators (A-LC, Y-KK, ZC, C-JT, SQ, JSK, RL, JF, SY, T-SY, JX, YS, HL, JL, JW, WYT, HP, ZG) were responsible for the decision to publish, had unrestricted access to the final data, and vouch for the completeness and accuracy of the data and data analyses. Bayer funded writing assistance in the form of preparation of the manuscript, references, figures, tables, collation of author comments, editing, and administrative support. The corresponding author had full access to the data in the study, final responsibility for interpretation of the data, and the decision to submit for publication.
a National Taiwan University Hospital, Taipei, Taiwan
b ASAN Medical Centre, Songpa-gu, Seoul, Republic of Korea
c First Affiliated Hospital of Anhui Medical University, Hefei, China
d Chi Mei Medical Centre, Liou Ying Campus, Tainan, Taiwan
e Nanjing Bayi Hospital, Nanjing, China
f Korea University Guro Hospital, Seoul, Republic of Korea
g Nanfang Medical University Nanfang Hospital, Guangzhou, China
h Cancer Hospital of Jiangsu Province, Nanjing, China
i Zhongshan Hospital, Fudan University, Shanghai, China
j Chang Gung Memorial Hospital, Tao-Yuan, Taiwan
k Affiliated Hospital of Academy of Military Medical Sciences, Beijing, China
l Cancer Institute and Hospital of CAMS, Beijing, China
m Southwest Hospital of Third Military Medical University, Chongqing, China
n First Affiliated Hospital of Dalian Medical University, Dalian, China
o Shanghai Changzheng Hospital, Shanghai, China
p Kyungpook National University Hospital, Daegu, Republic of Korea
q Sir Run Run Shaw Hospital, Zhejiang, China
r Bayer Schering Pharma, Wuppertal, Germany
s Bayer Schering Pharma, Shanghai, China
t Bayer Healthcare, Montville, NJ, USA
u Sun Yat-sen University Cancer Centre, Guangzhou, China
|
|
| |
| |
|
|
|
