Microwave Ablation, Radiofrequency Ablation, Irreversible Electroporation, and Stereotactic Ablative Body Radiotherapy for Intermediate Size (3–5 cm) Unresectable Colorectal Liver Metastases: a Systematic Review and Meta-analysis

Purpose of Review Based on good local control rates and an excellent safety profile, guidelines consider thermal ablation the gold standard to eliminate small unresectable colorectal liver metastases (CRLM). However, efficacy decreases exponentially with increasing tumour size. The preferred treatment for intermediate-size unresectable CRLM remains uncertain. This systematic review and meta-analysis compare safety and efficacy of local ablative treatments for unresectable intermediate-size CRLM (3–5 cm). Recent Findings We systematically searched for publications reporting treatment outcomes of unresectable intermediate-size CRLM treated with thermal ablation, irreversible electroporation (IRE) or stereotactic ablative body-radiotherapy (SABR). No comparative studies or randomized trials were found. Literature to assess effectiveness was limited and there was substantial heterogeneity in outcomes and study populations. Per-patient local control ranged 22–90% for all techniques; 22–89% (8 series) for thermal ablation, 44% (1 series) for IRE, and 67–90% (1 series) for SABR depending on radiation dose. Summary Focal ablative therapy is safe and can induce long-term disease control, even for intermediate-size CRLM. Although SABR and tumuor-bracketing techniques such as IRE are suggested to be less susceptible to size, evidence to support any claims of superiority of one technique over the other is unsubstantiated by the available evidence. Future prospective comparative studies should address local-tumour-progression-free-survival, local control rate, overall survival, adverse events, and quality-of-life.


Introduction
Colorectal cancer (CRC) is one of the most common cancers worldwide and the second leading cause of cancer-related mortality, with almost 1.850.000 new cases worldwide and 881.000 deaths in 2018 [1]. Colorectal liver metastases (CRLM) will develop in 25-30% of these patients during the course of their disease and is the main cause of death in CRC patients [2][3][4][5]. When left untreated, the 5-year overall survival (OS) rate is dismal, with survival rates around 0-3% [6][7][8]. Although systemic therapy alone clearly improves survival, the only treatments that 1 3 can provide long-term disease control or in a subset of patients even cure, are local eradication of the tumour.
There is an ample amount of studies that have shown needle-guided thermal ablation to be effective and safe in the treatment of CRLM ≤ 3 cm [17]. After a median follow-up of 9.7 years, the EORTC-CLOCC trial reported a superior OS of RFA plus chemotherapy over chemotherapy alone (HR = 0.58; 95%CI 0.38-0.88) with an 8-year OS of 35.9% vs. 8.9% [25]. The efficacy of thermal ablation is even being compared to resection in CRLM < 3 cm to prove non-inferiority in the ongoing RCT COLLISION [26]. Conversely, for larger (> 3 cm) CRLM, the primary technique efficacy decreases exponentially, manifesting in higher rates of local tumour progression for all techniques [27][28][29][30][31][32][33].
The radiation oncology community has suggested SABR to represent a feasible alternative as local treatment option for a limited number of unresectable CRLM. Although SABR can be effective to establish local control, a tradeoff exists between tumour control and collateral damage to surrounding tissue and structures [34][35][36]. As the efficacy is unaffected by the proximity of large blood vessels and less affected by lesion size and a difficult-to-reach anatomical location, authors have suggested SABR as an alternative to thermal ablation for perivascular, sub-diaphragmatic, and larger CRLM [37,38].
IRE is a relatively new non-thermal ablative method, where cell death is caused by using high-voltage electric pulses that induce permanent disruption of the membrane [39]. It is thought to be a safe ablation method for tumours adjacent to vascular and biliary structures because it spares the extracellular matrix and as a result preserves critical tubular structures [40].
Extrapolating treatment results of small-sized CRLM, local ablative therapies are also often presumed to prolong survival for unresectable intermediate-size CRLM (3-5 cm). However, given the exponential decrease in local efficacy with increasing lesion size, this presumption requires validation. To ensure patients receive the optimal treatment method, knowledge about the preferred local ablative technique is indispensable. This multidisciplinary systematic review and meta-analysis critically assess and compare the outcomes of local treatment in patients with unresectable intermediate-size CRLM treated with the most widely used thermal and non-thermal ablation techniques.

Methods
This systematic review and meta-analysis was written according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and PICO (patients, interventions, comparisons, outcomes) protocol [41].

Search
A literature search was performed in the databases PubMed and Embase from January 1st 2008 till November 11th 2020. Keywords used in the search were as follows: colorectal liver metastases, microwave ablation, radiofrequency ablation, stereotactic body radiotherapy, and irreversible electroporation. The full search strategy is presented in appendix 1. The subsequent PICO question was used for the search strategy: P(population): patients with intermediate-size CRLM; intervention: RFA, MWA, IRE, and SABR with or without systemic therapy; comparison: systemic therapy alone; outcome: critical endpoints were local-tumour-progression-free survival/local control (LTPFS/LC), complications/toxicity, overall survival (OS), and important endpoints were diseasefree survival (DFS) and quality of life. The interventional oncology society prefers the use of the term LTPFS (to describe the time from the initial treatment to the first recurrence, regardless of whether the recurrence was reablated), where the radiation oncology society prefers the use of the term local control [42]. Conference abstracts, reviews, metaanalyses, and studies not concerning humans were excluded.

Study Selection
The abstracts retrieved by this literature search were independently screened by two authors (SN and RP). If the abstracts appeared to adhere to the in-and exclusion criteria, a full-text evaluation was performed. The references of relevant publications were reviewed. References appearing eligible were also submitted to a full-text evaluation. Manuscripts also 1 3 containing information on efficacy and safety of primary liver carcinoma and non-colorectal liver metastases were allowed if they reported their data on CRLM separately. Studies were excluded if they did not report on at least one of the abovementioned outcome measures distinctly for intermediate size CRLM and if the sample size was less than five. Discrepancies between authors were resolved by consensus.

Data Extraction
Two authors (SN and MD) extracted the data from the included studies. This concerned the following variables: name author, publication year, years of inclusion, total number of patients, and number of patients with CRLM 3-5 cm, whether patients received prior local treatment of the liver, presence of extrahepatic disease, size of CRLM, amount of CRLM 3-5 cm and/or ≥ 3 cm, treatment modality, and concomitant resections with thermal ablation. The collected data pertaining to study outcomes were for example median follow up, dose and fractions in SABR and biologically equivalent dose (BED10), local control, LTPFS, complications/toxicity, DFS, OS, and quality of life. This data was checked by a third author (RP). In case of discrepancies, these were discussed and resolved by consensus. Additional data of subgroups with intermediate size CRLM was requested and collected from authors that reported results of the comparison of SABR to thermal ablation.

Data Analysis
Quality assessment criteria per study were based on clinical criteria, such as the included number and specific reporting of intermediate-size CRLM, the population, and the outcome measures used. Pooled analyses were allowed if results from studies were sufficiently similar with regards to these criteria. Studies potentially sufficient to perform metaanalysis were assessed and a random effects model was used to account for statistical heterogeneity. Analysis with the Mantel-Haenszel method was performed to calculate risk ratios (RR) of local tumour progression. Review Manager 5.3 was used to perform the meta-analysis.

Guidelines
CRLM guidelines were searched using Guideline Central and Guidelines International Network databases.

Results
The search strategy yielded 1685 abstracts after removal of duplicates. After screening the abstracts for eligibility, 151 articles remained for full-text analysis, of which 124 were excluded. This left 27 articles that met our inclusion criteria for qualitative synthesis and 2 articles for quantitative synthesis with meta-analysis (see flowchart in Figs. 1 and 2). Very few publications reported on the outcomes of intermediate-size CRLM (3-5 cm) specifically. Therefore, we allowed publications reporting on the outcomes of CRLM ≥ 3 cm. Series that discontinued including patients before 2008 were excluded, due to the likelihood of outdated results.

Complications and Quality of Life
None of the studies reported the complication rate or the effect of thermal ablation on quality of life specifically for patients with CRLM > 3 cm. Irrespective of lesion size studies reported a major complication rate of 2-17% for percutaneous ablation [43,46,47,50,52,53,56]. Most reported major complications were: pleural effusion, pneumothorax, hepatic abscess, hepatic hematoma, perihepatic bleeding, or ileal perforation. Both Qin et al. and Veltri et al. did not find a correlation between the development of complications and lesion size [52,57]. Qin et al. found a mean lesion size of 1.8 cm vs 1.5 cm for patients with versus without complications (p = 0.101) [52]. Similarly, Veltri et al. found a mean size of 2.7 cm in both groups [57].

Colorectal Liver Metastases 3-5 cm
Two retrospective series reported DFS [45,60]. Gwak Table 3 for an overview of the efficacy of thermal ablation.

Patient and Lesion Characteristics
Strict adherence to the inclusion criteria resulted in two retrospective series, as most SABR series do not report separate results based on tumour type and tumour diameter > 3 cm [63,64]. Doi et al. compared SABR with a conventional fractionated schedule and included 24 patients in total, 15 patients with 21 CRLM > 3 cm and 16 patients (66.7%) with a history of focal hepatic resection(s) and/or thermal ablation(s) [63] (see Table 4). Joo et al. included 70 patients in total, half of the study population had received prior local hepatic treatment, and 19 patients (27%) presented with extrahepatic disease [64]. It was not stated how many patients had intermediate size CRLM.
To collect more data, one prospective phase II trial that studied the efficacy of SABR for 27 CRLM patients with a cumulative gross tumour volume (GTV) diameter > 3 cm unsuitable for surgery and thermal ablation was eventually added [35]. Cumulative GTV diameter here means either at least 1 CRLM > 3 cm or multiple smaller CRLM with a cumulative size > 3 cm. Twenty-four CRLM > 3 cm were included. In this study, 11 patients (26%) had extrahepatic disease (EHD) and half of the patients had undergone prior focal liver treatment(s).

Overall Survival
No study reported OS specifically for CRLM 3-5 cm. Doi et al. reported results both for SABR as for non-ablative radiotherapy and found a median OS of 45 months for patients with at least one CRLM > 3 cm [63]. Conversely, for patients with small-size CRLM ≤ 3 cm, they found a median OS of 27 months [63]. Scorsetti et al. reported a 1-, 2-, and 3-year OS from SABR of 68, 40, and 17%, respectively, for patients with CRLM > 3 cm [35] (Tables 5 and 6).

Toxicity and Quality of Life
No studies reported the complication rate or the effect of SABR on quality of life for patients with CRLM > 3 cm. Two studies reported no grade ≥ 3 toxicity [35,64]. Scorsetti et al. found grade 2 acute toxicity in 78% of the study population (55% fatigue, 25% transient hepatic transaminase increase, 12% nausea) [35]. One series reported 2/24 patients with grade 3 toxicity, 1 patient with grade 3 γ-glutamyl transpeptidase (GGT) elevation, and 1 patient with grade 3 GGT and blood bilirubin elevation presumably caused by cholangitis due to a recurrent tumour [63].     To extend data on IRE for CRLM > 3 cm we included the final results of an as of yet unpublished prospective multicentre phase IIb single-arm study (COLDFIRE-2 trial) where 51 patients were treated with IRE in 62 procedures. Although currently under review, the trial protocol was previously published [68], the results have been presented at ECIO 2019 in Amsterdam, and the outcomes are available as online abstract [69]. Twenty-one (27.6%) out of the 76 IRE-treated CRLM were 3-5 cm in size. Fruhling [67]. Subgroup analyses were performed for larger size lesions (> 3 cm) and additional data collection showed local tumour progression in 11/20 tumours following SABR and 22/41 tumours following thermal ablation with at least 1 year of follow-up.

Local Tumour Progression
Overall comparison of local tumour progression following SABR and thermal ablation showed no significant difference (p = 0.50).

Guidelines
Full-text analysis was performed for 12 guidelines [70][71][72][73][74][75][76][77][78][79][80][81]. One guideline included recommendations for CRLM > 3 cm: the UK National Institute for Health and Care Excellence (NICE) guideline stated that "there is controversy over the indication for RFA, most operators will no longer consider lesions > 4 cm in diameter for treatment" [71]. All other guidelines either did not report on RFA, MWA, SABR, or IRE at all, or they did not state recommendations for CRLM > 3 cm, or they did not state size limitations.

Discussion
Currently, the preferred treatment method for unresectable intermediate-size CLRM for patients, in whom downstaging or (further) downsizing systemic therapy failed, remains unknown. This systematic review and meta-analysis aimed to collect evidence regarding local ablative therapies to treat unresectable intermediate-size CRLM and to provide a comparison of the most well-known ablative techniques. Literature to reliably assess the oncological outcome was scarce for all treatment options. A substantial shortcoming was the lack of randomized controlled trials comparing treatment methods. In addition, apart from one prospective cohort [62] and one phase II trial [35], virtually, all included studies were retrospective series, with only two of the studies making a comparison between treatment options for intermediate-size CRLM. Furthermore, the reported oncological outcomes, the study population, and the timing of interventions with regard to periprocedural systemic chemotherapy were highly heterogeneous, making it impossible to draw any conclusion. The majority of publications on thermal ablation concerned RFA. However, for larger-size tumours, recently, preference has started to shift towards newer generation MWA systems or tumour-bracketing multiprobe ablation techniques as potentially superior alternatives to conventional RFA [82,83]. Presumed benefits of MWA over RFA are consistently higher intratumoural temperatures, faster heating, shorter procedure time, larger ablation volumes, and less susceptibility to the "heat-sink" effect at the cost of a somewhat higher biliary tract complication rate [84][85][86]. Although few studies compared RFA to MWA for patients with CRLM, several retrospective cohorts reported lower local recurrence rates following MWA compared to RFA, 6% vs. 20% (p < 0.01) [19], 10% vs. 20% (p = 0.02) [55], 8.6% vs. 20.3% (p = 0.07) [87], respectively. In this review, LTP rate at median follow-up after the first ablation ranged 11-78% for RFA [27,43,46,48,51,53,56,58] and 14-38% for MWA [44,52,61]. Although this seems to suggest a preference of MWA for CRLM > 3 cm, the number of MWA treated tumours was low (n = 41). A substantial part of the included publications on thermal ablation was relatively old. Consequently, recent advances in technique and improved awareness of the necessity to expand and confirm tumourfree margins following thermal ablation are inadequately represented [53].
For SABR, merely three articles met the inclusion criteria, and all reported different oncological outcome measures. Hence, no conclusions could be drawn regarding efficacy of SABR for intermediate-size CRLM. Many articles describing results for mixed disease and not for CRLM separately could not be included, because metastases deriving from different primary cancers or different organs containing colorectal metastases can have variable responses [88][89][90][91][92][93][94][95][96]. Several articles were excluded because they presented hazard ratios regarding small versus intermediate-size CRLM but did not report the actual outcomes per size-subgroup, or they reported on the size of CRLM in volumes and not diameter [34,90,97].
Two articles met the inclusion criteria for meta-analysis after additional data collection [66,67]. No difference in local tumour progression was found between SABR and thermal ablation. Two excluded publications compared SABR to thermal ablation for hepatic metastases [98,99],  [99]. SABR demonstrated a superior FFLP compared to RFA, especially for hepatic metastases ≥ 2 cm. There was no difference in median OS (25.9 months for RFA vs. 24.5 months for SABR). These studies, compared to the included studies in meta-analysis, imply a superior local control of SABR compared to thermal ablation for larger-size lesions. However, only comparing local control rates following one ablative procedure seems unjust when comparing a repeatable technique (RFA, MWA) with a technique that usually does not allow for retreatment (SABR). No studies reported a direct comparison of thermal ablation to SABR with regard to periprocedural complications and toxicity for intermediatesize CRLM, though both techniques are associated with an exceptionally low mortality and morbidity rate. Given the comparable overall reported mortality of 0.16% for thermal ablation [100] and 0.5% for SABR [101] (with 3/656 patients mistakenly published as 0.004%) and given the comparable serious adverse event rate of 4-5% for thermal ablation and 9% for SABR [100,101]. Because both ablative probes and ionizing radiation will potentially result in collateral morbidity by invading surrounding healthy tissue, we prefer to refrain from using the term non-invasive for SABR.
Only two studies concerning IRE were included in this review. This low number can be explained by the relative novelty of this technique and because it is generally a niche indication for CRLM unsuitable for resection and thermal ablation due to close proximity to biliary or vascular structures [40]. Interestingly, the results of the prospective phase II trial (COLDFIRE-2) did not reveal a difference in 1-year LTPFS for small-size versus intermediate-size CRLM, which may indicate that IRE, where electrodes bracket tumours, is less susceptible to differences in size [102].
A recent multidisciplinary consensus document concerning resectability and ablatability criteria for liver only colorectal metastases did not provide strict recommendations for unresectable intermediate-size CRLM due to a lack of evidence and also stated that the exact roles of SBRT and IRE in the treatment of unresectable CRLM need to be further investigated [103].
Although systematically acquired, the results of this systematic review and meta-analysis should be judged with restraint, as only a limited amount of studies could be 1 3 included, with poor quality and heterogeneous study populations. There is a high risk of publication bias due to the inclusion of mainly retrospective observational studies.

Conclusion
There are no randomized controlled trials or comparative studies on local treatment for patients with intermediatesize unresectable CRLM. Heterogeneity of the reported oncological outcomes and study populations reduced the amount of obtained data suitable for pooled assessment. Although long-term disease control was described in subsets of patients in all series, there is a lack of studies directly comparing RFA to MWA or to SABR or IRE. No hard conclusions or recommendations can be drawn and further prospective research is necessary to determine what local treatment option, if any, is preferable for intermediate-size unresectable CRLM, preferably in the setting of randomized controlled trials. Therefore, we strongly support the ongoing trials, the COLLISION-XL trial NCT04081168 (unresectable colorectal liver metastases: stereotactic body radiotherapy versus microwave ablation -a phase II randomized controlled trial for CRLM 3-5 cm), an RCT in Denmark for CRLM < 4 cm NCT03654131 (stereotactic body radiation therapy vs microwave ablation for colorectal cancer patients with metastatic disease in the liver), and an RCT in Italy for CRLM < 4 cm NCT02820194 (a trial on SABR versus MWA for inoperable colorectal liver metastases). Hopefully, the results of these trials will clarify and define the role of local ablative methods for the curative intent treatment of permanently unresectable intermediate-size CRLM.
Author Contribution SN and MRM contributed to the study design. Data collection and analysis were performed by SN, MD, RSP, and MRM. The first draft of the manuscript was written by SN and MD, and all authors commented on previous versions of the manuscript and all authors read and approved the final manuscript.

Data Availability Not applicable.
Code Availability Not applicable.

Conflict of Interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.