Effectiveness of microwave ablation for the treatment of hepatic hemangioma - meta-analysis and meta-regression

Abstract Background The treatment of hepatic hemangioma includes surgical resection, radiofrequency ablation and Transarterial embolization. However, complications, mortality and compromised effectiveness limit their applications. Microwaves with effective heating generation and short ablation time become a promising treatment. The aim of this study is to conduct systematic review and meta-analyses to evaluate the effectiveness of Microwave Ablation (MWA) for the treatment of hepatic hemangioma. Methods A systematic literature review was conducted in PubMed. Main outcomes were defined as hemangioma decreases in diameters and volume changes post-MWA. Conventional random-effect meta-analysis technique was applied to analyze the pooled data, and meta-regression model was established to explore the association among factors. Results There were nine studies with a total of 501 patients retrieved. The pooled estimate of mean differences and 95% CI of hemangioma decreases after MWA treatment in diameter and in volume change (%) were 3.009 cm and (1.856, 4.161), and 53.169% and (51.274, 55.065), respectively. The pooled estimates of liver enzyme, ALT and AST, elevation were 219.905 with 95%CI (160.860, 278.949) and 315.679 with 95%CI (226.961, 404.397), respectively. Major complications were defined as acute kidney injury (AKI), pleural effusion, diaphragmatic hernia, and jaundice that needed to be treated, and the pooled incidence was 0.017 with 95% CI of (0.006, 0.029). No mortality related to MWA was reported. Meta-regression showed ablation time was associated with pre-operative lesion size (p = .001). Conclusion MWA is effective and safe in treatment of hepatic hemangioma, and our study suggests that hemangioma size should be investigated in the future MWA pretreatment difficulty scoring system study.


Introduction
Hepatic hemangioma is the most common benign tumor of the liver with small capillary hemangiomata of no clinical significance.Larger cavernous hemangiomata more often come to the attention of the liver surgeon [1].Usually, patients present with asymptomatic and are diagnosed incidentally on imaging studies.Yet, when the masses keep growing, vague upper abdominal symptoms may manifest.Rapid enlargement or vessel thrombosis of the tumor can occasionally cause symptoms, and spontaneous rupture is exceedingly rare.The associated Kasabach-Merritt syndrome, however rare, is well defined.The prognosis of hepatic hemangioma is generally benign.In most cases, the lesions remain stable over long periods of time, and the constant rupture or large hemorrhage is seldom confronted in clinical practice.No malignant transformation of liver hemangioma has been reported before.Therefore, asymptomatic patients with definitive diagnosis can simply be observed, and symptomatic patients excluding the others causes should be candidates for treatment.
The treatment includes conventional operative resection, minimally invasive surgery of laparoscopic surgery, thermal ablation and arterial embolization.Indications of treatment include rupture, rapid expansion, and Kasabach-Merritt syndrome.Conventional hepatic resection has been considered the standard treatment, however, morbidity up to 25% has been reported in the previous studies [2,3].Transarterial embolization and radiation have been studied with conspicuous success and less morbidity, given the feature of minimally invasion [4][5][6][7].In thermal ablation, radiofrequency has been the most extensively ablation modality for treatment of hepatic hemangiomas.In hepatic hemangioma less than 7.5 cm, it is effective in ablating the lesion, yet limited power, small ablation areas and incomplete ablation require long ablation times, repeated operation and the unwanted result of perfusion-mediated tissue cooling (heat-sink), and limited power deposition together compromise the outcomes [8][9][10][11][12][13].In comparison, microwaves can produce hotter and larger ablation zones, and be faster and reproducible [14][15][16]; the principle of microwaves is to oscillate polar molecules of water producing thermal energy; since hemangioma is comprised of blood vessel and full of water, it is considered to be particularly suitable for microwaves treatment and can be effectively ablated by it.In the treatment of malignancies of liver, lung and kidney, microwave ablation has been increasingly used [17][18][19].The purpose of this study is to conduct a systematic review to further evaluate the current available evidences of microwave ablation in the treatment of hepatic hemangioma by using meta-analysis methods.

Literature search
A search was performed through PubMed using the terms of 'hepative hemangioma' or 'liver hemangioma' or 'liver benign tumor' from 1980 to 2022-08.Only articles in English were utilized.There were 2492 results on PubMed after initial searching.The title and abstract readings removed 2206 articles, which were either unrelated to microwaves ablation for treatment of hepatic hemangioma or of animal studies.Of these 286 articles remained for full examinations.The meta-analyses were done on nine studies with a total of 501 patients [14,[17][18][19][20][21][22][23][24] (Figure 1).

Including criteria
Studies were included, if they were original research articles and available as full text articles, and if they met the following criteria: clinical studies either retrospective study or controlled studies including microwaves ablation for treatment of hepatic hemangioma; studying subjects with hepatic hemangioma; microwaves ablation; measuring the outcomes of tumor decreases and functional parameters and major complications, and mortality.Studies that were excluded were animal studies, editorials, case reports, and conference reports.

Data extraction
The means and standard deviations of the measures of tumor decrease, ALT, AST, and ablation time were extracted.The mean differences of tumor decrease in diameter was computed and compared.The pooled estimation of ALT and AST elevation post-MWA were computed; the events of major complications defined as acute kidney injury (AKI), pleural effusion, diaphragmatic hernia, and jaundice that needed to be treated were extracted, the pooled incidence of major complications were calculated for the combining estimations.

Statistics
Meta-analyses were conducted through 9 pooled studies by using OpenMeta software.Inverse-variance methods were applied to combine mean differences for continuous data.The I2 statistic is calculated through synthesis to measure the heterogeneity among the studies' results to guide whether fixed-effect or random-effect model should be adopted.p Value <.05 was interpreted as showing significant difference in comparisons.

Meta-regression
In order to compute the coefficients of variables in metaregression models and estimate the correlation between potential affecting factors and pre-operative tumor size, meta-regression analyses were performed and scatterplots were drawn using OpenMeta software.The variable of preoperative tumor size and MW ablation time were added into the set of moderators.A random-effect meta-regression analysis between study variances was applied to evaluate whether MW ablation time was associated with pre-operative hemangioma size in the studies.

Summary of included studies
There are nine studies with a total of 501 patients, that met our study criteria, being pooled here.Tumor decreases in diameters and volumes, hepatic enzymes pre-and post-MWA, incidences of major complications and ablation times were extracted.Four studies presented the data of tumor decreases post-MWA in diameters, and the data in volumes were available in four studies.Six studies reported the hepatic enzyme elevation on day 1 post surgery.The major complications and ablation times reported in nine studies were pooled for incidence effect size estimation or meta-regression analysis.The details of the above data and the study design were summarized on Table 1.

Tumor decreases
Four studies with a total of 196 patients were included and reported the tumor shrinkage in diameters post-MWA.The heterogeneity test results revealed that the null hypothesis of pooled studies being homogeneity and measuring the same effect size could be rejected (Cochran's Q test, p < .001;I2 ¼ 88.88%).Thus random-effects model was adopted, and the pooled estimate of mean differences and 95% CI of tumor decreases after MWA treatment was 3.009 cm and (1.856, 4.161), which did not include 0; Five studies with a total of 155 patients reported tumor decreases in volume percentages.Fixed-effects model was applied (Cochran's Q test, p ¼ .123;I2 ¼ 0%) and the pooled estimate was 53.169% with 95% CI (51.274, 55.065), together indicating that MWA significantly reduced tumor sizes (Figure 2).Funnel plot was drawn to assess the publication bias in diameters by using Review Manager 5.4.The pooled studies visually spread evenly on both sides of the average, creating a roughly funnel-shaped distribution, indicating no publication bias by visual estimation (Figure 3).

Liver enzyme elevation
There are 6 studies with a total of 407 patients reporting the value of liver enzyme post surgery, and the pooled estimate of hepatic transferases of ALT and AST postoperation were 219.905 with 95%CI (160.860,278.949) and 315.679 with 95%CI (226.961,404.397), respectively.During the follow-up, hepatic enzymes returned to normal after three and sevendays in two studies (Study no. 1 and 2, Table 1).No study reported severe consequence attributed to the transient liver enzyme elevation (Figure 4).

Major complications post-MWA
Details of reported major complications are presented in Table 1.No mortality related to MWA was reported.Pooled incidence of major complications was 0.017 with 95% CI of (0.006, 0.029).The patients with the major complications were treated and fully recovered after the treatment (Figure 5).

Meta-regression
Due to different MW ablation times among the nine studies with 11 groups, the moderator of pre-operative hemangioma size in diameter was added into the model to evaluate whether MW ablation time was associated with pre-operative hemangioma size.The regression model gave the computing results being that the MW ablation time was significantly associated with the pre-operative hemangioma size with r ¼ 3.905, 95% CI [1.669, 6.141] and p ¼ .001,indicating a positive correlation between ablation time and baseline hemangioma size, meaning that if hemangioma size increases ablation time does as well (Figure 6).

Discussion
Although surgical resection or enucleation has become the standard treatment modality for symptomatic or enlarging or rapid growing hepatic hemangiomas [25], 0.5-2% mortality rate associated with operation restrain these procedures [26], besides, adverse event of enormous blood loss, elongated hospital stay and potential severe perioperative complications such as hepatic insufficiency, bile leakage and wound infection also prohibited it as the first line therapy in considerable cases [27].Laparoscopic resection followed by cool-tip cluster radio frequency (RF) ablation of hepatic hemangiomas had low incidence of massive blood loss, unbearable postoperative pain and complications [28].However, due to its high incidences of postoperative major complications including acute kidney injury, acute respiratory distress syndrome and about 40% technical failure rate, and complexity of the technique, it also remains controversial in dealing with those hemangiomas exceeding 10 cm [11,12,29,30].
The heating principle of microwaves is that electromagnetic field oscillating polar molecules can produce heat by turning electromagnetic energy into thermal energy [31].Considering the heat-sink effect and related complications of RF, MW has the potential to replace RF for the treatment of benign and malignant tumors in the liver, kidney, and lung [32][33][34].The working principle of MW devices is to make a heating energy conversion from electromagnetic energy produced by a uniform electromagnetic field surrounding antennas, that oscillates polar water molecules to generate heat and burn tissues.Thus, the charring of the tissue doesn't prohibit the effectiveness of MW, the area temperatures can be driven very high to efficiently ablate the lesion [35,36].Indeed, one study showed that the mean treatment time of RF was 39 min for tumors that ranged between 2.5 and 9.5 cm in diameter, and seven separate RF electrode insertions and 125 min of RF ablation were required eventually to treat the 9.5 cm tumor [8].Another presented that 40 min of RF were required to deal with a 7.5 cm hemangioma with three electrodes [10].By comparison, in MW study the mean treatment time was reported as 11.6 min.It was reported that only 24.5 min elapsed eventually before the largest hemangioma in this series exceeding 10 cm with mean diameter of 12.2 cm was effectively ablated.Therefore, the anesthesia and bleeding complications can be improved by the fast and short treatment time and less number of punctures.Hemangioma full of water content permits its ideal substrate for MW, and the vascular mediated cooling often limit other thermal ablation methods due to the very slow flow-through time but not MW [37].
In this meta-analyses, the pooled estimation of tumor shrinkage showed that microwaves were effective in treatment of hepatic hemangioma.Liver enzyme did increase post surgery, but was transient elevation and there was no  severe consequence caused by it.Major complications were reported in the pooled studies, yet patients with these were fully recovered without mortality cases being reported.The moderator of pre-operative lesion size was added into metaregression.A positive association between pre-operative hemangioma lesion size and MW ablation time indicates that the increase of pre-operative lesion size will lead to the elongation of MW ablation time, suggesting that tumor size could possibly add difficulty to hemangioma MWA treatment.More factors may also affect the procedure of operation, so we meant to add more modalities of risk site and the depth of tumors into the regression, however, few studies providing detailed information limited the analyses.Previous studies have established the surgical difficulty scoring system of open and laparoscopic hepatectomy in the treatment of Hepatocellular Carcinoma HCC [38].Yet, there is no study ever explored the difficulty scoring system evaluating MWA treatment.Considering hemangioma is benign, the difficulty scoring system of HCC may not be suitable to evaluate the treatment in hemangioma.Therefore, it is  essential to optimize the current difficulty scoring system to better adapt to it.The HCC scoring system did not include the tumor size, location and depth which, however, are considered to be affecting factors in hemangioma.Besides, hemangioma is often asymptomatic, without chronic liver disease background and has a better liver functional reserve, thus, Child-Pugh classification should be removed from the current scoring system.In this meta-regression, that only two studies reported the locations of tumor without detailed information of hepatic sections limited conduction of metaregression to explore the association.In the future studies, the depth and location near risk site of hemangioma should be explored more carefully in order to further evaluate the association of hemangioma locations and depth with the treatment outcomes.To validate the above proposal of establishing a more suitable difficulty scoring in hemangioma treatment, we believe, can enhance the effectiveness of WMA treatment and provide a more precise evaluation before operation.Besides, in malignant hepatic tumor, local thermal ablation has been showed effective, and a study has reported a novel prognostic biomarker of lymphocyte-tomonocyte ratio that could predict the survival and the tumor recurrence in colorectal liver metastasis patients before the local treatment [39,40].For hepatic hemangioma, it will be interesting that future studies can explore its potential prognostic biomarkers before local WMA for precision treatment and improving patient outcomes.

Limitations
The numbers of pooled studies are relatively small, and among them most lack comparison with a control group.Few studies giving details of hemangioma locations and depth that may enhance the difficulty of treatment and affect outcomes constrain a further analysis.Besides, subgroup studies of percutaneous and laparoscopic procedure, and 2 D vs. 3 D should also be conducted in the future when data are available.Therefore, further studies with larger samples sizes, control groups and detailed tumor sites will help improve the quality of evidence and provide more evidence to validate an optimized difficulty scoring system for the treatment of hemangioma.

Ethical approval
We declare that there are no studies with human participants or animals performed by any of the authors in this article.

Figure 2 .
Figure 2. Tumor shrinkage after MWA treatment IN diameter (A) and IN volume (B).(Data in parentheses demonstrate 95% confidence intervals (CIs).The horizontal lines represent the 95% CIs, and the diamonds show the pooled effect sizes.)

Figure 3 .
Figure 3. Funnel plot assessing the publication bias in diameters.

Figure 5 .
Figure 5.The pooled estimation of major complications incidences.

Table 1 .
Characteristics of pooled studies.
Note: N/A: not available; s: seconds.