Is transcatheter arterial chemoembolization plus sorafenib better than chemoembolization plus placebo in the treatment of hepatocellular carcinoma?

Abstract

Objective: To evaluate the efficacy and safety of transcatheter arterial chemoembolization (TACE) plus sorafenib compared with TACE plus placebo for hepatocellular carcinoma (HCC) using meta-analytical techniques.

Methods: A search of PubMed, EMBASE, and Cochrane Library databases were done from inception to December 27, 2019. Published trials including a treatment group receiving TACE + sorafenib and a control group receiving TACE + placebo with data for at least 1-year survival or tumor response or time to progression were included.

Results: Our study suggested that there was no evidence that TACE plus sorafenib was associated with a lower risk of disease progression compared with TACE plus placebo for treatment of HCC (hazard ratio 0.94 [95% confidence interval (CI), 0.84–1.05]), and no significant difference for treatment of HCC compared with TACE plus placebo in terms of 0.5-, 1-, 1.5-, and 2-year survival rates (risk ratio [RR] 1.01 [95% CI, 0.97–1.05]; RR 1.00 [95% CI, 0.92–1.08], RR 1.04 [95% CI, 0.89–1.23], RR 0.98 [95% CI, 0.72–1.34], respectively). The meta-analysis also showed that TACE + sorafenib
seemed to have no significant difference for treatment of HCC compared with TACE + placebo in terms of complete response, partial response, stable disease, progressive disease, overall response rate, and disease control rate. There was
an increased incidence of fatigue of grade 3/4 and elevation of aspartate aminotransferase and alanine aminotransferase of grade 3/4 in patients receiving TACE plus sorafenib compared with those receiving TACE plus placebo.
Conclusions: There is no additive benefit of TACE plus sorafenib compared to TACE plus placebo for HCC.

Keywords : Transcatheter arterial chemoembolization, hepatocellular carcinoma, sorafenib, placebo, meta-analysis

Introduction

Hepatocellular carcinoma (HCC) is the fifth most com- monly diagnosed cancer and the fourth leading cause of cancer-related death in the world.1 Over the past 25 years, the number of HCC cases has increased, with HCC becom- ing one of the fastest increasing causes of tumor-related deaths in Europe and the United States.2 It was pointed out that 650,000 individuals die from HCC every year, about three-fourths in East Asian countries.3,4 Nonradical treat- ment is available when some individuals with HCC are not diagnosed until the disease lesions are irremovable.5,6 Among carefully selected patients, surgical resection, liver transplantation, and local ablation are considered as pos- sible cures, with a 5-year survival rate of 40%–70% and
20% for untreated patients.7,8 Prevention of HCC progres- sion remains the most significant challenge for improving prognosis.

Transcatheter arterial chemoembolization (TACE), as a palliative treatment, has been proven to be effective for the treatment of patients with HCC who are not suitable for surgical resection. TACE has fewer adverse reactions than traditional systemic chemotherapy.9 Its mechanism is to temporarily prevent blood from flowing to the tumor sites to achieve the outcome of eliminating tumor lesions. In clinical practice, the application of TACE varies regarding embolizing agents, anticancer drugs, frequency, and sched- ule of administration10; the survival status and prognosis of patients with HCC may vary with the above factors.

Sorafenib, targeting multiple pathways, is a multiki- nase inhibitor11–13 and can inhibit tumor growth and neo- vascularization.14 It has been proved that sorafenib significantly improves overall survival and time to pro- gression (TTP) in patients with advanced HCC in two large, double-blind, randomized, placebo-controlled phase III clinical trials.15,16 Another two large, double- blind, randomized, placebo-controlled clinical trials found that the combination of sorafenib plus drug-eluting beads (DEB)–TACE did not improve TTP in a clinically mean- ingful manner compared with DEB-TACE alone,17 and that TTP is similar after neoadjuvant treatment with TACE + sorafenib before liver transplantation compared to TACE + placebo.18 These findings vary greatly among different studies, so it is a pending question whether TACE + sorafenib can achieve better outcomes compared to TACE + placebo. This meta-analysis aimed to analyze relevant clinical trials in recent years to evaluate the safety and efficacy of combination treatment used for HCC.

Methods

Study protocol

This systematic review and meta-analysis was performed by using the protocol specified by the Cochrane collabo- ration19 and reported according to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) (Supplementary Text 1, available online at www.tumorijournal.com).20 The review protocol has been submitted to PROSPERO. Two investigators inde- pendently searched the literature published in the PubMed, EMBASE, and Cochrane Library (CENTRAL) databases from inception to December 27, 2019. The main keywords used for the search were “hepatocellular carcinoma,” “HCC,” “transcatheter arterial chemoembo- lization,” “TACE,” and “sorafenib.” There was no lan- guage restriction in this search. We did not perform a search for unpublished literature.

Selection criteria

Acceptable works of literature complied with the follow- ing criteria: (1) study design: trials were described as randomized, controlled trials (RCTs); (2) subjects with HCC were confirmed pathologically or cytologically or diagnosed by computed tomography (CT) or magnetic resonance imaging (MRI); (3) the experimental groups received TACE plus sorafenib and the control groups received TACE plus placebo; (4) outcomes: reported at least one item of the results of TTP, survival, or tumor response, or adverse events/toxicity for the calculation of the hazard ratio (HR) or risk ratio (RR) at a 95% confi- dence interval (CI); (5) characteristics of patients: publica- tions were required to have similar baseline characteristics of registered patients such as age, liver function, percent- age of males, number of patients, tumor size, and tumor extent. Abstracts, letters, and case reports were excluded.

Data extraction

Using standardized forms for data extraction, we collected the following information: (1) first author’s name, year of publication, country, study type, study centers, characteris- tics of the study population, and tumor characteristics; (2) details of treatment; and (3) outcomes such as TTP, sur- vival, tumor response, and adverse events. Survival included 0.5- to 2-year survival rates. Tumor response was assessed by changes in tumor size on abdominal CT or MRI before and after treatment. The tumor response rate was recorded under the modified RECIST guidelines (Response Evaluation Criteria in Solid Tumours)21 for HCC in some included trials (complete response [CR; complete removal of the lesion after treatment]; partial response [PR; the sum of feasible target lesion diameters reduced by at least 30%]; progressive disease [PD; the sum of lesion diameters increased by at least 20% during treat- ment]; stable disease [SD]). The disease control rate (DCR) was defined as CR, PR, or SD, and the overall response rate (ORR) was defined as CR or PR.

Quality assessment

Using the Cochrane Risk of Bias Tool,22 we assessed the quality of each included RCT according to the Cochrane for- mat, which is based on the following: (1) random sequence generation; (2) allocation concealment; (3) blinding of par- ticipants and personnel and outcome assessment; (4) incom- plete outcome data; (5) selective outcome reporting; and (6) other sources of bias. The included trials were graded as low quality, high quality, or moderate quality based on the fol- lowing criteria23: (1) if randomization or allocation conceal- ment is evaluated as high risk of bias, regardless of the risk of other items, the trial is considered as low quality; (2) when randomization and allocation concealment are assessed as low bias risk in one trial, and other items are evaluated as low or unclear risk of bias, the trial is considered to be high qual- ity; (3) if the trial does not meet the high or low risk criteria, it is considered to be of moderate quality.

Statistical analysis

For TTP analysis, HRs with 95% CIs were combined using the inverse variance method, and the effect sizes of overall survival, tumor response, and adverse events were calculated by RR with 95% CI. HR or RR was used as the outcome measure for data synthesis. Statistical het- erogeneity between studies was examined using chi- square and I2 statistics were calculated to assess the homogeneity of the included trials. We presented the results of the fixed-effects model in the absence of statistical heterogeneity (p > 0.10). Otherwise, we reported the results from the random-effects model meta-analysis.

For I2 values below 25%, 25% to 50%, and above 50%, heterogeneity is considered low, moderate, or high, respectively.24 If there was a significant risk of bias, the source of bias had to be explored or discussed. For the exploration of heterogeneity, subgroup analysis was per- formed based on TACE type, anticancer drug, TACE and sorafenib/placebo interval, and cirrhosis. If necessary, sensitivity analysis would be carried out to assess the sta- bility of the results. Begg and Egger tests were employed to assess publication bias. All p values were two-sided, with p < 0.05 being considered statistically significant. All statistical analysis was performed using the statistical software Review Manager (version 5.3) from the Cochrane Collaboration (Oxford, UK) and Stata 12.0 (Stata Corporation, College Station, TX).

Results

Study selection and study characteristics

We identified 364 studies in our initial extensive search. After carefully reading titles and abstracts, we selected 18 articles for possible inclusion. After reading the full text, we excluded 12 for the reason that the same authors were involved in the respective sequential studies or unusable data. Ultimately, 6 trials17,18,25–28 (1416 patients) met our inclusion criteria, all RCTs (Supplementary Figure 1, available online at www.tumorijournal.com). We collected the research design and evaluated quality. The percentage of publications that described the mean age of the patients was 83% (5 of 6); Child-Pugh score, 67% (4 of 6); tumor diameter, 33% (2 of 6); etiology, 100% (6 of 6); and serum -fetoprotein, 67% (4 of 6). Some trials included patients with cirrhosis (4 of 6 [67%]) (Supplementary Table 1, available online at www.tumorijournal.com). Among these trials, 33% (2 of 6) used two or more anticancer drugs, and 3 of 6 (50%) used only one anticancer agent. The most common embolizing agent used was lipiodol (3of 6 [50%]). Generally, lipiodol was mixed with anticancer drugs at a dosage calculated according to tumor size or a uniform dosage. Intervals between TACE and sorafenib/ placebo ranged from 3 to 35 days (Supplementary Table 2, available online at www.tumorijournal.com).

Methodologic quality of RCTs

In terms of quality, only 1 RCT25 was of high quality, while others17,18,26–28 were of moderate quality. All of them were described with double-blinding. Four trials17,18,27,28 showed that random methods have been used, but none reported the details of the generation of random assignment sequences. Only 2 trials25,26 provided information on allocation con- cealment by use of tablets identical in appearance and num- bered bottles or sealed envelopes. Two trials17,25 performed intention-to-treat analysis. In 2 reports,18,26 dropout rates were stated, and corresponding explanations were given (Supplementary Figure 2 and Supplementary Table 3, available online at www.tumorijournal.com).

Meta-analysis outcomes

Time to progression. TTP was reported in 6 trials.17,18,25–28 Meta-analysis showed that there was no evidence that TACE + sorafenib was associated with a lower hazard of disease progression compared to TACE + placebo for treatment of HCC (pooled HR 0.94 [95% CI, 0.84–1.05]; p = 0.27) (Table 1 and Figure 1). Overall survival rate. The 0.5-, 1-, 1.5-, and 2-year survival rates were reported in 3,17,25,27 3,17,25,27 3,17,25,27 and 3 trials,17,25,27 respectively. Meta-analysis showed that TACE + sorafenib seemed to be of no significant difference for treatment of HCC compared with TACE + placebo in terms of 0.5-, 1-, 1.5-, and 2-year survival rates (RR 1.01 [95% CI, 0.97–1.05 ], RR 1.00 [95% CI, 0.92–1.08], RR 1.04 [95% CI, 0.89–1.23],
RR 0.98 [95% CI, 0.72–1.34], respectively) (Table 1).

Adverse events

Mortality. One study27 reported 84 deaths (43 in the sorafenib group and 41 in the placebo group) at the same cutoff date. One study18 reported four deaths that were not treatment-related during the observation period. Another study25 reported 3 deaths in each group owing to DEB- TACE, and 4 deaths attributed to therapeutic drugs, one of which was caused by massive variceal hemorrhage in the placebo group. The three treatment-related deaths were due to acute liver failure, infection, or hepatorenal failure in the sorafenib group.

Fatigue. Two18,25 and 3 trials17,25,26 were identified with outcome measurements of fatigue of grade 1/2 or grade 3/4, respectively. The pooled analysis suggested that there was no significant statistical difference between TACE + sorafenib and TACE + placebo for treatment of HCC in
terms of grade 1/2 fatigue (RR 0.97 [95% CI, 0.82–1.14]). However, the pooled analysis suggested that TACE + sorafenib greatly increased the occurrence of grade 3/4 fatigue compared to TACE + placebo (RR 1.55 [95% CI, 1.01–2.36]) (Table 2).

Abdominal pain. Two trials17, 25 were identified with out- come measurements of abdominal pain of grade 3/4. The pooled analysis suggested that there was no significant sta- tistical difference between TACE + sorafenib and TACE + placebo for the treatment of HCC in terms of abdominal
pain of grade 3/4 (RR 1.12 [95% CI, 0.51–2.46]) (Table 2).

Diarrhea. Two18,25 and 4 trials17,25,26,28 were identified with outcome measurements of diarrhea of grade 1/2 or grade 3/4, respectively. The pooled analysis suggested that there was no significant statistical difference between TACE + sorafenib and TACE + placebo for the treatment of HCC in terms of diarrhea of grade 1/2 (RR 1.74 [95% CI, 1.07–2.81]). TACE plus sorafenib seemed to be greatly increased the occurrence of diarrhea of grade 3/4 compared with TACE plus placebo (RR 4.82 [95% CI, 2.22–10.49]) (Table 2).

Nausea. Two18,25 and 3 trials17,25,26 were identified with out- come measurements of nausea of grade 1/2 or grade 3/4, respectively. The pooled analysis suggested that there was no significant statistical difference between TACE + sorafenib and TACE + placebo for treatment of HCC in
terms of nausea of grade 1/2 or grade 3/4 (RR 1.06 [95% CI,

Rash/desquamation. Four trials17,25–27 were identified, and the results of rash/desquamation were measured as grade 3/4. The pooled analysis suggested that compared with TACE plus placebo, TACE plus sorafenib greatly increased the occurrence of rash/desquamation of grade 3/4 (RR 8.24 [95% CI, 1.53–44.44]) (Table 2).

Hemorrhage/bleeding. Two trials17,25 were identified, and the results of hemorrhage/bleeding were measured as grade 3/4. The pooled analysis suggested that compared with TACE + placebo, TACE + sorafenib did not greatly increase the occurrence of hemorrhage/bleeding of grade 3/4 (RR 1.73 [95% CI, 0.74–4.08]) (Table 2).

Anorexia. Three trials17,25,26 were identified, and the results of anorexia were measured as grade 3/4. The pooled anal- ysis suggested that compared with TACE + placebo,TACE + sorafenib did not greatly increase the occurrence of anorexia of grade 3/4 (RR 1.39 [95% CI, 0.45–4.34]) (Table 2).

Vomiting. Two trials17,25 were identified, and the results of vomiting were measured as grade 3/4. The pooled analysis suggested that compared with TACE + placebo, TACE + sorafenib did not greatly increase the occurrence of vomit- ing of grade 3/4 (RR 0.50 [95% CI, 0.13–1.96]) (Table 2).

Alopecia. Two trials18,25 were identified, and the results of alopecia were measured as grade 1/2. The pooled analysis suggested that compared with TACE + placebo, TACE + sorafenib did not greatly increase the occurrence of alope- cia of grade 1/2 (RR 1.55 [95% CI, 0.83–2.89]) (Table 2).

Weight loss. Two trials17,25 were identified, and the results of weight loss were measured as grade 1/2. The pooled analysis suggested that compared with TACE + placebo, TACE + sorafenib did not greatly increase the occurrence of weight loss of grade 1/2 (RR 1.51 [95% CI, 0.79–2.91]) (Table 2).

Cirrhosis: present or absent. Subgroup analysis suggested that patients receiving TACE + sorafenib seemed to be of no significant difference for treatment of patients who had cirrhosis or not compared with TACE + placebo in terms of 0.5-, 1-, 1.5-, and 2-year survival rates, and CR, PR, SD,
ORR, and DCR. For the treatment of patients with cirrho- sis, TACE + sorafenib was associated with significantly less PD than TACE plus placebo, while TACE + sorafenib seemed to be of no significant difference for treatment of patients not having cirrhosis compared with TACE + pla- cebo in terms of PD (Table 3).

Discussion

This meta-analysis provides a comprehensive assessment of the efficacy and safety of TACE + sorafenib compared with TACE + placebo. Sorafenib is the sole anticancer drug that has been approved for systemic therapy for HCC.29 In clinical practice, the use of TACE varies sur- rounding embolizing agents, anticancer drugs, frequency, and schedule of administration10; survival status and prog- nosis of patients with HCC may vary owing to the above factors. Sorafenib can inhibit neoangiogenesis as well as tumor growth by acting on vascular endothelial growth factor receptor, Raf signaling, and platelet-derived growth factor receptor.14 The combination of sorafenib and TACE has been assessed in a series of single-arm phase I and II trials in which both concurrent sequential and sequential administration is feasible and effective.30–32

We applied meta-analytical techniques to assess the efficacy and safety of TACE plus sorafenib compared with TACE plus placebo for HCC, which provided conclusive evidence that there is no additive benefit of TACE plus sorafenib in the meta-analysis of the RCTs. Regarding adverse events, we found that compared with TACE plus placebo, TACE plus sorafenib was associ- ated with increased occurrence of grade 3/4 fatigue and grade 3/4 diarrhea, rash/desquamation of grade 3/4, hand- foot skin reaction, hypertension of grade 3/4, and elevation of AST and ALT of grade 3/4. One study27 in our included studies reported 84 deaths, 43 in the sorafenib group and 41 in the placebo group, at the end of the research. One abstracts of included trials, 18 articles were chosen for possible inclusion in the review. Ultimately, 6 trials involv- ing 1416 patients met our inclusion criteria, all of them RCTs, and the overall quality of the evidence was moder- ate. All of the RCTs were described as double-blind. Publication bias was not evaluated since the number of included trials concerning these outcomes was too small, therefore, the Egger or Begg test was inappropriate.

Subgroup analysis showed that there were no significant trends in which survival and CR, PR, SD, ORR, and DCR were greater for TACE plus sorafenib in patients who received either DEB-TACE or C-TACE. Similarly, no dif- ferences were seen in the interval between TACE and sorafenib/placebo, cirrhosis, or types of anticancer drugs used. Of note, subgroup analysis also showed that (1) patients who received DEB-TACE plus sorafenib had a lower PD than those who received DEB-TACE plus pla- cebo while conventional C-TACE plus sorafenib seemed to be of no significant difference for treatment of HCC com- pared with C-TACE plus placebo in terms of PD; (2) patients receiving TACE plus sorafenib seemed to have a lower PD for treatment of patients receiving doxorubicin compared with TACE plus placebo while TACE plus sorafenib seemed to be of no significant difference for treatment of HCC receiving doxorubicin ± other or carbo- platin compared with TACE plus placebo in terms of PD; (3) the experimental group was associated with signifi- cantly less PD than control group when the interval of TACE and sorafenib was 7 days or less, while TACE plus sorafenib seemed to be of no significant difference for treatment of HCC compared with TACE plus placebo in terms of PD when the interval of TACE and sorafenib was more than 7 days; (4) for the treatment of patients with cir- rhosis, the experimental group was associated with signifi- cantly less PD than the control group, while TACE plus sorafenib seemed to be of no significant difference for treatment of patients having no cirrhosis compared with TACE plus placebo in terms of PD. Several hypotheses may support evidence that the experimental group seemed to have a lower PD for treatment of patients receiving dox- orubicin compared with the control group: the activation of the multidrug resistance pathway may be blocked through the inhibition of the Raf/Ras/ERK/MEK pathway.33

For the treatment of patients with cirrhosis, the reasons why TACE + sorafenib was associated with significantly less PD than TACE plus placebo may be as follows: (1) the interaction between ELAVL1/HuR and beclin 1 mRNA is a key molecular event that triggers the activation of autophagy, and promotes the degradation of autophagic ferritin, which in turn leads to ferroptosis34; (2) sorafenib significantly decreases intrahepatic fibrogenesis, collagen deposition, as well as hydroxyproline accumulation35; (3) sorafenib ameliorates the proapoptotic and profibrotic effects of TGFB1 in hepatocytes, suggesting that this anti- cancer drug has therapeutic potential for the treatment of liver fibrosis35; (4) sorafenib significantly reduces histopathology and liver inflammation as well as decreases the expression of liver fibrosis–related genes, including matrix metallopeptidases, ACTA2, COL1A1, and TIMP metallopeptidases inhibitor 1.36

Although sorafenib is the standard of treatment strat- egy and the only anticancer drug for advanced HCC,29,37 through a comprehensive analysis of the included RCTs, our meta-analysis showed that there was no evidence that the experimental group was associated with a lower risk of disease progression compared to the control group for treatment of HCC (HR 0.94 [95% CI, 0.84–1.05]), and also showed that not only did TACE + sorafenib not improve the survival rate and tumor response rate, but caused a host of adverse events compared to TACE + placebo.

We included double-blind RCTs. All the examined studies were described with double-blinding and four stated that randomization had been used. None reported the details of the generation of a random allocation sequence. Only 2 trials25,26 provided information on allo- cation concealment by use of identical tablets in appear- ance and sealed envelopes or numbered bottles. If there is no double-blind, the result may produce selection bias, measurement bias, and performance bias. Second, the cri- teria for including patients with HCC vary from study to study. This may affect the consistency of the effects of these included studies and lead to heterogeneity between studies. Third, all of our literature originated from America, Europe, and Asia; the baseline characteristics of patients recruited in these studies are likely to be different in these regions, and we lacked subgroup analysis based on region owing to limited related studies. Finally, publication bias and sensitivity analysis could not be performed because our analyses included fewer than 10 trials.

Conclusions

The available evidence is sufficient to allow us to draw the following conclusions: (1) there was no evidence that TACE plus sorafenib was associated with a lower risk of disease progression compared to TACE + placebo for treatment of HCC, which was different from the previous meta-analysis38; (2) TACE plus sorafenib seemed to be of no significant difference for treatment of HCC compared to TACE + placebo in terms of survival and tumor
response; and (3) TACE + sorafenib was associated with increased occurrence of grade 3/4 fatigue and grade 3/4 diarrhea, rash/desquamation of grade 3/4, hand-foot skin reaction, hypertension of grade 3/4, and elevation of AST and ALT of grade 3/4 compared to TACE + placebo. There is no additive benefit of TACE plus sorafenib compared to TACE plus placebo for HCC.