High-intensity focused ultrasound versus transarterial chemoembolization for hepatocellular carcinoma: a meta-analysis

Abstract Purpose The application of high-intensity focused ultrasound (HIFU) in hepatocellular carcinoma (HCC) was promising. However, whether the effect of HIFU is comparable with that of transarterial chemoembolization (TACE) has not been determined. Materials and methods PubMed, Embase, Cochrane Library, Web of Science, WanFang Data, CqVip, CNKI, and CBM databases were searched for randomized controlled trials (RCTs), cohort studies, and case-control studies. The methodological quality of each study was evaluated. When there is no statistical heterogeneity, the fixed effect model would be used to merge data. Otherwise, the random effect model would be utilized. Sensitivity analyses were conducted by excluding one study each time. Subgroup analyses were conducted based on age, sex, tumor number, relative number of the patients with Child-Pugh C grade in each group, the percentage of patients with Child-Pugh C grade in the whole study, and tumor load. Publication bias was evaluated by Egger’s test and Begg’s test. Results Six cohort studies including 188 patients from HIFU group and 224 patients from TACE group were obtained for further analysis. The meta-analysis suggested HIFU and TACE showed no differences in postoperative 1-year overall survival (OS) rate, tumor response (including complete response, partial response, stable disease, and progressive disease), and postoperative complications. Moreover, compared with TACE, HIFU showed higher postoperative 6-month and 2-year OS rates. Subgroup analyses, meta regression analysis and sensitivity analyses indicated the findings above were reliable. Additionally, no potential publication bias was detected. Conclusion For HCC, when compared with TACE, HIFU might show comparable safety but better effect. Considering the limitations of current studies, more well-designed studies are needed to validate our conclusion.


Introduction
According to Global Cancer Incidence, Mortality, and Prevalence (GLOBOCAN) estimates in 2018, liver cancer was the sixth most commonly diagnosed cancer and the fourth leading cause of cancer death worldwide, with approximately 841,000 new cases and 782,000 deaths annually (Bray et al. 2018).Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer, accounting for 75-85% of cases.High-intensity focused ultrasound (HIFU) has been reported to be a novel treatment for HCC (Diana et al. 2016).The main trait of HIFU is its noninvasiveness.Therefore, in the early stages, the application of this technique was focused on some special conditions.For instance, it was used for patients who were found at an early stage but had inferior liver function and tolerance ability (She et al. 2016).In addition, it was also used as a local ablation treatment for patients who were diagnosed at a late stage (Rauch et al. 2015;Liu et al. 2019).HIFU can ablate tumors through a variety of complex mechanisms (Sadeghi-Goughari et al. 2020), including mechanical, thermal and cavitation.HIFU has been proved to be an effective and safe option for patients with liver cancer regardless of primary and metastatic lesions, and the 1-year survival rate was 70.69% and 48.00%, respectively (Ji et al. 2020).Therefore, the application of HIFU in the treatment of HCC was promising.
Transarterial chemoembolization (TACE) is a well-established and widely used method in treating HCC.Since the large number of HCC patients are diagnosed at a late clinical stage, they are unable to receive liver resection or transplantation.Currently, TACE is thought to be one of the most effective treatments for this population (Zhao et al. 2020).Therefore, in some cases, TACE and HIFU could both be selected.Although several studies compared the effect of HIFU versus TACE in HCC, the conclusions varied.Furthermore, as a novel treatment for HCC that was widely adopted after 2001, HIFU was rarely reported in terms of efficacy in treating HCC.Thus, original studies on HIFU-TACE comparison were limited.In our meta-analysis, after systematic searching, we found some published original studies comparing HIFU and TACE in HCC treatment.However, the conclusions of these studies were limited by small sample size (less than 50 cases in HIFU group).Considering this, through the method of meta-analysis, this study aimed to comprehensively evaluate whether HIFU is better than TACE in the treatment of HCC in the aspects of some vital outcomes, such as overall survival (OS) rate, tumor response, and postoperative complications.

Search strategy
The protocol of this meta-analysis was prospectively registered in the International Prospective Register of Systematic Reviews database (PROSPERO: CRD42020193484).This study was finished according to the guidelines for the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher et al. 2009).The PubMed, Embase, Cochrane Library, Web of Science, WanFang Data, CqVip, CNKI, and CBM databases were searched for original studies.The literature found was published before 21 June 2020.The search terms included 'HIFU', 'high-intensity focused ultrasound', 'high-intensity focused ultrasound ablation', 'focused ultrasound', 'focused ultrasound ablation', 'focused ultrasound surgery', 'focused ultrasound ablation surgery', 'FUAS', 'FUA', 'TACE', 'transarterial chemoembolization', 'chemoembolization, therapeutic', 'therapeutic chemoembolization', 'chemoembolizations, therapeutic', 'therapeutic chemoembolizations', 'embolization, therapeutic', 'HCC', 'hepatocellular carcinoma', 'carcinomas, hepatocellular', 'hepatocellular carcinomas', 'liver cell carcinoma, adult', 'liver cancer, adult', and 'adult liver cancer'.The search in the Web of Science database was conducted with the field of Subject but not with the field of Full text, because no more than 100 words could be searched with the field of Full text.In addition to that, the search in other databases was conducted with the field of Full text.No other limitations were set during the literature search.In the end, we searched the reference lists of the included articles and guidelines for extra studies.Institutional Review Board (IRB) approval and written consent was not required for conducting this meta-analysis.

Inclusion and exclusion criteria
The inclusion criteria were as follows: (1) Patients: The individuals were diagnosed with primary HCC.The diagnosis of HCC was based on the diagnostic criteria from the official guideline.In short, the diagnosis was made according to alpha fetoprotein (AFP) level and imaging findings.(2) Intervention and comparison: The patients in the HIFU group underwent treatment with HIFU, whereas the patients in the TACE group underwent treatment with TACE.For TACE technique, under X-ray fluoroscopy, puncture and intubation through femoral artery to tumor artery were finished.Then, chemotherapeutic drugs were injected via the catheter.According to the size of the tumor, iodized oil was injected to embolize the tumor vessels.For HIFU technique, the operation was finished by high-intensity focused ultrasound tumor treatment system with general anesthesia.Under the real-time monitoring of B-mode ultrasound or magnetic resonance imaging (MRI), the location, quantity and size of tumors were determined, and the treatment scope was also determined.From deep to shallow, the tumors in all levels are treated until the target area is completely ablated.The therapeutic dose was adjusted according to the gray value or signal intensity of the tissue after treatment on the image of ultrasound or MRI.(3) Outcomes: The outcomes of interest were postoperative OS rate, tumor response and postoperative complications.OS was defined as the period from the date of treatment (HIFU or TACE) to the date of death related to all causes.Tumor response was evaluated according to the WHO criteria or Response Evaluation Criteria in Solid Tumors (RECIST) criteria (Liti ere et al. 2019).In short, the tumor response included complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD).( 4) Study design: randomized controlled trials (RCTs), cohort studies, and case-control studies would all be considered.The exclusion criteria were as follows: (1) the study was an animal experiment; (2) the study was a case report, case series, or review; (3) the study was a conference abstract or book; (4) when the study reported repeat contents from another study, only the latest published study was selected; (5) the study included non-primary HCC and could not be separated from the primary HCC; (6) the study did not report any outcomes our meta-analysis needed; or (7) the study was not relevant to the theme of our study.

Study selection, data extraction, and quality assessment
Two reviewers (YBW and HSZ) independently screened the literature.After duplicates were excluded, the title, abstract and full text of each study was assessed for eligibility.Inconsistent results were resolved by consulting a third researcher.The following data was extracted: first author, publication year, study design, intervention method, sample size, and various outcomes.The Newcastle-Ottawa Scale (NOS) (Stang 2010) was used to evaluate the methodological quality of the included cohort and case-control studies.The scale is semiquantitative, and the full mark is nine stars.In this study, high quality was defined as a Newcastle-Ottawa Quality Assessment Scale score of 4 (McCarty et al. 2021).In this meta-analysis, we defined the vital confounding factors as liver function and clinical stage.In addition, other factors that existed in the original study were thought of as 'other confounding factors'.

Data analysis
With the DerSimonian-Laird method, the study-specific risk ratio (RR) for the categorical variables was utilized to calculate the pooled value.When there is no statistical heterogeneity, the fixed effect model would be used to merge data.Otherwise, the random effect model would be utilized.Stata 16.0 (StataCorp, College Station, TX) was used for data analysis.Heterogeneity between different studies was evaluated by I 2 statistics and corresponding p values.When p > 0.1 and I 2 < 50%, it meant low heterogeneity was found.If I 2 ranged from 50% to 75%, it suggested there was moderate heterogeneity.High heterogeneity was defined as I 2 > 75%.Sensitivity analyses were conducted by excluding one study each time in certain meta-analyses.Subgroup analyses were conducted based on age, sex, tumor number, relative number of the patients with Child-Pugh C grade in each group, the percentage of patients with Child-Pugh C grade in the whole study, and tumor load.Regarding the definition of tumor load, the tumor load of the Cheung et al. study (2013) was deemed to be low.Based on a comparison with this study, the tumor load of other studies was evaluated.The clinical stage, size, and quantity of tumors served as the reference base for evaluation.Univariate meta regression analyses were used to identify the sources of heterogeneity across studies.Publication bias would be evaluated by Egger's test and Begg's test.In this study, p < .05 was considered statistically significant.

Literature selection, characteristics, and methodological quality of the included studies
The literature selection process is shown in Figure 1.After the initial search, we obtained a total of 3990 citations.
Figure 1.Literature selection.Note.The PubMed, Embase, Cochrane Library, Web of Science, WanFang Data, CqVip, CNKI, and CBM databases were searched for original studies.Initially, we obtained 3990 citations.After excluding duplicates, there were 3418 citations.According to the title and abstract, we excluded 3399 unrelated records.Therefore, 19 records were then screened according to the full-text.In the end, we included six studies for quantitative synthesis.
More specifically, these citations originated from the PubMed (n ¼ 88), Embase (n ¼ 122), Cochrane Library (n ¼ 34), Web of Science (n ¼ 144), WanFang Data (n ¼ 278), CqVip (n ¼ 114), CNKI (n ¼ 3092), and CBM (n ¼ 118) databases.Subsequently, we started literature screening.Among them, 572 records were duplicates.Based on the title and abstract, we excluded 3,399 unrelated records.Therefore, 19 records were screened according to the full-text.Two records were the same, so we retained one record (Wu et al. 2012).Two studies included patients with TACE and Transcatheter arterial infusion (TAI) treatment in one group and the patients could not be separated, therefore they were excluded.One article (Cho et al. 2011) that belonged to a conference abstract was excluded.One citation (Facciorusso and Barone 2016) was excluded because it belonged to a book.One study (Xu 2014) included patients who accepted transcatheter arterial embolization (TAE) or TACE before HIFU treatment in the HIFU group.Four articles (Luo et al. 2015;Bao and Li 2008;Ma and Wang 2006;Qin 2003) belonged to a review.Two studies (Cheung et al. 2014;Xue 2013) included both patients with primary and recurrent HCC in one group, and these patients could not be separated.One study (Peng 2008) did not report any outcomes our meta-analysis needed.Additionally, the reference lists of the included articles, reviews and guidelines were screened, but no additional studies were found.In the end, our study included six studies (Huang et al. 2010;Ren et al. 2010;Zhang 2011;Wu et al. 2012;Cheung et al. 2013;Yang et al. 2014) for quantitative syntheses.
Our study included 188 patients from HIFU group and 224 patients from TACE group.The characteristics of the included studies are shown in Table 1.The included studies were published between 2010 and 2014.They belonged to cohort studies, so we used NOS to evaluate the methodological quality of the included studies (supplemental Table S1).The result indicated that two studies (Wu et al. 2012;Zhang 2011) obtained nine stars, one study (Yang et al. 2014) obtained eight stars, two studies (Huang et al. 2010;Ren et al. 2010) obtained seven stars, and one study (Cheung et al. 2013) obtained six stars.In more detail, two studies (Cheung et al. 2013;Yang et al. 2014) did not describe the source of the control group and lost one star.Three studies (Zhang 2011;Wu et al. 2012;Yang et al. 2014) controlled the vital confounding factors (liver function and clinical stage) and other confounding factors, so they got two stars.Three studies (Huang et al. 2010;Ren et al. 2010;Cheung et al. 2013) that were retained did not show they controlled the vital confounding factors, so they lost two stars.The scores of all studies were higher than four, so their methodological quality was considered high.

Postoperative complications
Three of six studies (Ren et al. 2010;Wu et al. 2012;Cheung et al. 2013) reported the outcomes of postoperative complications.However, none of them compared the incidence rate of postoperative complications.Meanwhile, none of these studies used a common complication classification system (such as the Clavien-Dindo classification system) to make a classification.We summarized the complications reported in each study and found the common postoperative complications were bleeding, skin burn, hollow viscera perforation, fever, aches, cardiopulmonary complications, bilirubin increase, transaminase increase, and loss of appetite.We found only three kinds of postoperative complications were reported by at least two studies.Therefore, we meta-analyzed these three kinds of postoperative complications.As a result, the meta-analysis indicated the two groups showed no difference in postoperative bleeding (RR ¼ 2.14; 95% CI ¼ 0.19-23.71;p ¼ .53;I 2 ¼ 0, P for heterogeneity ¼ .72; Figure 5(A)), skin burn (RR ¼ 2.14; 95% CI ¼ 0.19-23.71;

Subgroup analysis, heterogeneity analysis and meta regression analysis
According to information provided by the original studies, subgroup analysis was conducted based on age, sex, tumor number, relative number of the patients with Child-Pugh C grade in each group, the percentage of patients with Child-Pugh C grade in the whole study, and tumor load.The result showed that the comparison of the 6-month OS, 1year OS, 2-year OS, rate of CR, rate of PR, rate of SD, and rate of PD between TACE group and HIFU group in different subgroups showed consistent result with the meta-analysis including all studies (supplemental Table S2).In the subgroup analyses including patients with Child-Pugh grade C in the whole study < 8% or with low tumor load, HIFU group showed a higher rate of CR þ PR but a lower rate of SD þ PD (p < .05;supplemental Table S2).Besides, other  subgroup analyses showed consistent result with the metaanalysis including all studies (supplemental Table S2).
In our study, for the meta-analyses of 6-month OS, 1year OS rate, 2-year OS rate, CR, PR, postoperative complications (bleeding, skin burn, and transaminase increase), no statistical heterogeneities were found (All p values for heterogeneity > .05).As for the meta-analysis of SD or PD, there was a statistical heterogeneity.When we conducted meta-regression analyses for these two variables, it was indicated that insufficient observations (two studies) were detected, and meta-regression analyses could not continue and then were not conducted.
As for the meta-analyses of CR þ PR and SD þ PD, there were statistical heterogeneities and both meta-analyses included more than two studies.So, meta-regression analysis was then conducted.For the meta-analyses of CR þ PR and SD þ PD, same original studies were included.Through the observation of the forest plot and comparison of the basic traits of the patients between various studies, we identified one study conducted by Cheung TT et al (Cheung et al. 2013) presents an obvious heterogeneity.We then analyzed the possible origins of heterogeneity.
The origins of heterogeneity could be found mainly from two aspects.First, compared with the other two studies, the tumor load of patients included in the study conducted by Cheung TT et al is smaller.Second, compared with the other two studies, the liver function of patients reflected by Child-Pugh grades included in the study conducted by Cheung TT et al is better.Afterwards, for the variables of CR þ PR and SD þ PD, we conducted subgroup analysis and meta-regression analysis.In the subgroup analyses, after excluding the original studies with low tumor load or better liver function reflected by Child-Pugh grade, the meta-analyses of the remaining studies showed no statistical heterogeneities.Meta regression analyses based on these two variables furtherly confirmed that tumor load and the liver function of patients were two vital variables that affect the statistical heterogeneity.Therefore, tumor load and the liver function of patients were two vital variables that affect the statistical heterogeneity of the meta-analysis of the CR þ PR and SD þ PD.

Sensitivity analyses
In our study, we found the meta-analyses of 1-year OS rate, the rate of CR þ PR, and the rate of SD þ PD included more than two studies.For these outcomes, we conducted sensitivity analyses.As a result, by excluding one study each time and then conducting meta-analyses including the retained studies, the conclusions about the comparisons of 1-year OS rate, the rate of CR þ PR, and the rate of SD þ PD between groups had not changed.

Publication bias
Using the outcome of the 1-year OS, no publication bias was detected, as shown by Egger's test (p ¼ .862)and Begg's test (p ¼ .734).

Discussion
In this meta-analysis, we found that HIFU, compared with TACE, was associated with higher 6-month and 2-year OS rates.There was no significant difference in 1-year OS rate.Afterwards, we found there was no significant difference in the four kinds of tumor responses.In addition, we did not find any significant difference between HIFU and TACE groups in three major postoperative complications (postoperative bleeding, skin burn, and transaminase increase).
The OS rate and tumor response represent the long-term and short-term clinical benefits for patients, respectively.From the aspects of OS rate and tumor response, we found that the effect of HIFU was better than that of TACE in the treatment of HCC, although the conclusion can only be confined in some conditions.The patients' characteristics in those included trials might limit the validity of survival outcomes, and interpretation of such results needs to be cautious.As we found, the included studies mainly recruited patients who were not suitable for treatment with liver resection or liver transplantation, due to the inferior liver function, severe basic disease, mid-advanced liver cancer, and the patients' desire for treatment.Whether a better effect could be obtained in the HIFU group versus TACE group in other common conditions, such as good liver function and good general condition, has not been resolved by these studies.Therefore, more studies are still needed.
Surgical safety was another concern for clinical practitioners.To evaluate surgical safety, our study regarded postoperative complications as a vital indicator.We metaanalyzed three major postoperative complications.The result indicated HIFU and TACE showed no differences in postoperative bleeding, skin burn, and transaminase increase.In addition, Ren et al. (2010) found the HIFU group had no incidence of hollow viscera perforation or cardiopulmonary complications.They also found the patients in both groups had postoperative aches, and nearly all the patients had low appetite.Wu et al. (2012) found 28/40 patients in the HIFU group and 24/40 patients in the TACE group had a low fever postoperatively.Additionally, 3/40 patients in the HIFU group and 11/40 in the TACE group had an increase in bilirubin.Cheung et al. (2013) found there were no postoperative complications in the HIFU group.After immediate treatment, all of the complications reported in these studies finally resolved.Altogether, based on the current studies, we could conclude that, similar to TACE, HIFU would not lead to severe postoperative complications.The safety of HIFU in the treatment of HCC is considered to be reliable.However, it is important to highlight that the nonstandard report style of the postoperative complications might influence the evaluation of safety.In the near future, postoperative complications can hopefully be reported with the application of common complication classification systems.This would make treatment-related complications from various approaches comparable.
Subgroup analysis based on age, sex, tumor number, relative number of the patients with Child-Pugh C grade in each group, the percentage of patients with Child-Pugh C grade in the whole study, and tumor load supported the findings of each meta-analysis of the outcome in our study.Sensitivity analysis found any single study could not bias the conclusion of our meta-analysis.Additionally, no publication bias was detected.All of these extra analyses indicated the conclusion of this meta-analysis is reliable.However, no subgroup analysis was made according to tumor size and clinical stage.The reason for this is that the information provided in the original studies is limited, as shown in Table 1.Considering tumor size and clinical stage were key factors determining patients' prognosis, further original studies would be anticipated to report these results.In this way, the effect of HIFU and TACE could be evaluated more comprehensively.Aside from that, no subgroup analysis was performed based on different countries.That is because all the studies included were from China.Given that HCC is a significant health burden in China (Lin et al. 2020) and that HIFU was widely used there (Zhang and Wong 2020), the country accounted for the majority of HIFU clinical applications.Nonetheless, as the clinical application of HIFU expands, more studies from other countries are expected.
After reviewing the included studies, we found patients with Child-pugh grade C were included in both HIFU group and TACE group.Commonly, Child-pugh grade C is contradiction for patients to accept TACE treatment.We speculated that in most studies, the Child-Pugh grade of the patients would have been recorded when they were first admitted to the hospital.Usually, in clinical practice, the patients with poor liver function or Child-Pugh grade would receive supportive therapy before further treatments were considered.The poor status of the patients could be corrected before TACE or HIFU was conducted.Therefore, after receiving additional treatments, the patients' condition would be reevaluated, and they might not be in Child-pugh C but rather in Child-pugh B or A. In addition, some studies included patients who choose other treatments for patients with low Child-Pugh grades rather than TACE or HIFU.Furthermore, only small number of patients have Child-pugh grade C. Additionally, patients with Child-pugh grade C may consent to HIFU therapy.According to the study by Cheung et al. patients with decompensated cirrhosis or poor liver function, as demonstrated by the presence of gross ascites and disease at Child-Pugh B or above, may be eligible for HIFU ablation (Cheung et al. 2013).
In fact, HIFU system could be guided by ultrasound or MRI.We reviewed the included studies and analyzed how the HIFU system was guided in each study.For the included six studies, four studies (Ren et al. 2010;Wu et al. 2012;Cheung et al. 2013;Yang et al. 2014) reported the model of the HIFU system, and it was reported or could be speculated that HIFU treatment was guided by ultrasound.Another two studies (Huang et al. 2010;Zhang 2011) failed to mention the manufacturer and model of the HIFU system or the equipment used to provide treatment guidance.However, as these two studies (Huang et al. 2010;Zhang 2011) were conducted in China and HIFU in China is usually guided by ultrasound, we can only speculate that in these two studies also ultrasound was used for guidance.Since no study reported using MRI for guidance, the majority of the studies that were included in this analysis used ultrasound.
Considering all of the findings from our study, we found HIFU was effective and safe for the treatment of HCC after comparison with TACE.In fact, HIFU has been certificated to be an effective treatment for HCC by previous studies after selecting radiofrequency ablation (RFA) as a control (Cheung et al. 2013;Xiao and Liu 2015;Li et al. 2019).Recently, the application of HIFU in clinical practice has increased.In addition to HIFU's solitary application in HCC, HIFU was also reported to perform better when combined with other treatments, such as TACE (Zhang et al. 2019), ethanol ablation (Qiao et al. 2020), and various nanoparticles (Jiang et al. 2020).Ablative treatments, such as RFA, irreversible electroporation (IRE), and microwave ablation (MWA), could be palliative therapy in patients with tumor by obtaining a cytoreduction of the tumor and delaying the growth and spread of tumor (Paiella et al. 2018).Compared with TACE, RFA was reported to show comparable long-term OS rates (Yang et al. 2014;Martin et al. 2019;Oh et al. 2020).HIFU is also regarded as one of the ablative treatments.In our study, we certificated the reliable efficacy of HIFU in the treatment of HCC when compared with TACE.Therefore, ablative treatment could provide comparable effect for HCC when compared with TACE.Recent studies on the mechanisms of HIFU treatment have observed additional mechanisms, such as the immune effect, that are involved in the killing of tumors in addition to the conventional thermal and mechanical effects (van den Bijgaart et al. 2017).Therefore, the effect of HIFU on HCC is expected to be encouraging in the near future.
There are some limitations in this meta-analysis.First, the heterogeneity of the included studies, including difference in patient population and unclear tumor location, is of concern.Most of the included studies reported that they included patients who were not suitable for liver resection or liver transplantation.However, the reasons for these patients not being suitable for these treatments were not stated.As such when patients with different causes are included in the same study, heterogeneity in patient population emerges.The location of the tumor must also be evenly balanced in each study as it has a significant impact on how well HIFU works.Sadly, this factor was not taken into account as a confounding factor in any of the included studies.Secondly, we do not have any information about the energy and time spent on HIFU treatment.These factors are important characteristics in HIFU treatment and are directly associated with the effect of HIFU treatment.Therefore, they should be taken into consideration to evaluate the safety and efficacy of HIFU.Furthermore, only cohort studies but no randomized studies were included.Although this study provides a clear message about the comparable effect of HIFU and TACE, but the level of evidence still needs to be improved when further high-quality studies were released.In our study, we tried to search all of the well-recognized online database for original studies.In the end, we included only six studies.By searching and reviewing the previous literature, we found HIFU is one novel technique in the treatment of HCC.Currently, there may not be many studies investigating this field, and the studies included at this stage were observational studies.RCTs were expected to be published in the future.As more high-level evidencebased studies are possible, they could then be included in this meta-analysis to update it.Fourthly, part of the metaanalyses was performed by including only two studies.The conclusion might be not much valid.This condition might be led by inconsistent outcome reported in various studies.Considering all of these, more original studies are needed.
In spite of these limitations, our study had some strengths.This meta-analysis collected all the available controlled studies to evaluate the effect of HIFU versus TACE in the treatment of HCC.This meta-analysis was the first one to specially examine the effect of HIFU versus TACE for HCC.In addition, all of the included studies had a high methodological quality.In some vital variables, such as OS rate, complete tumor response and partial tumor response, there was no significant heterogeneity.Importantly, for some meta-analyses with obvious heterogeneity, we found the source of heterogeneity by using subgroup analysis and meta regression.We hope these findings would be helpful for patient counseling about the potential benefits and risks of HIFU and TACE.
In conclusion, based on current controlled studies, HIFU was associated with a higher OS rate and comparable tumor response compared with TACE.HIFU and TACE showed no significant difference in postoperative complications, including postoperative bleeding, skin burn, and transaminase increase.HIFU is a promising therapy, but more studies are needed to evaluate its safety and efficacy in the treatment of HCC.

Table 1 .
Characteristics of included studies.
a Represents the tumor size of the largest tumor.b Represents that it includes patients who accepted other treatments except for TACE and HIFU.HIFU: High-intensity focused ultrasound; TACE: Transarterial chemoembolization; NA: not available.