Microwave Ablation for Colorectal Liver Metastases: A Systematic Review and Pooled Oncological Analyses

Simple Summary Liver resection for colorectal liver metastases (CRLM) represents the best curative option; however, few patients are candidates for surgery. Microwave ablation (MWA) can be a valid alternative in selected patients. This systematic review reports the oncological results of MWA for CRLM. The literature available on the Web was analyzed for reports concerning MWA for resectable CRLM, published before January 2021. Finally, 12 papers concerning MWA complications, recurrence-free (RF) cases, patients free from local recurrence (FFLR), and overall survival rates (OS) were selected. Global RF rates at 1, 3, and 5 years were 65.1%, 44.6%, and 34.3%, respectively. Global FFLR at 3, 6, and 12 months were 96.3%, 89.6%, and 83.7%, respectively. Global OS rates at 1, 3, and 5 years were 86.7%, 59.6%, and 44.8%, respectively. A better FFLR was achieved with an MWA surgical approach at 3, 6, and 12 months, with 97.1%, 92.7%, and 88.6%, respectively. Surgical MWA for CRLM smaller than 3 cm was a safe and valid option. MWA can be entered as part of the flowchart decision of CRLM curative treatment, especially for use in the parenchyma-sparing strategy and as a complement to surgery. Abstract (1) Background: colorectal liver metastases (CRLM) are the most common extra-lymphatic metastases in colorectal cancer; however, few patients are fit for curative surgery. Microwave ablation (MWA) showed promising outcomes in this cohort of patients. This systematic review and pooled analysis aimed to analyze the oncological results of MWA for CRLM. (2) Methods: Following PRISMA guidelines, PubMed, Scopus, EMBASE, Google Scholar, Science Direct, and the Wiley Online Library databases were searched for reports published before January 2021. We included papers assessing MWA, treating resectable CRLM with curative intention. We evaluated the reported MWA-related complications and oncological outcomes as being recurrence-free (RF), free from local recurrence (FFLR), and overall survival rates (OS). (3) Results: Twelve out of 4822 papers (395 patients) were finally included. Global RF rates at 1, 3, and 5 years were 65.1%, 44.6%, and 34.3%, respectively. Global FFLR rates at 3, 6, and 12 months were 96.3%, 89.6%, and 83.7%, respectively. Global OS at 1, 3, and 5 years were 86.7%, 59.6%, and 44.8%, respectively. A better FFLR was reached using the MWA surgical approach at 3, 6, and 12 months, with reported rates of 97.1%, 92.7%, and 88.6%, respectively. (4) Conclusions: Surgical MWA treatment for CRLM smaller than 3 cm is a safe and valid option. This approach can be safely included for selected patients in the curative intent approaches to treating CRLM.


Data Extraction and Outcomes of Interest
Once the articles meeting the specific inclusion criteria were identified, a series datasets were extracted independently by the two reviewers. These included: study t and design, patients' characteristics, the total number of MWA performed, mean/med lesion dimensions, the number of lesions per patient, synchronous/metachronous lesio treatment at recurrence, operation time, length of hospital stay, complications, follow

Data Extraction and Outcomes of Interest
Once the articles meeting the specific inclusion criteria were identified, a series of datasets were extracted independently by the two reviewers. These included: study type and design, patients' characteristics, the total number of MWA performed, mean/median lesion dimensions, the number of lesions per patient, synchronous/metachronous lesions, treatment at recurrence, operation time, length of hospital stay, complications, follow-up time, and the items defining the inclusion criteria (see below). All the selected articles Cancers 2022, 14, 1305 5 of 23 evaluated the postoperative complications using the Clavien-Dindo Score (CDs) [25]. Additional data, such as neoadjuvant and adjuvant treatments, specific postoperative complications, and types of MWA device and needle tip were registered as well.
Subsequently, and upon existing data, we stratified for lesions according to diameter (<30 mm vs. >30 mm), for percutaneous/surgical treatment procedures, and for the type of surgical approach (open, vs. laparoscopic, robotic) in the latter group.
The following primary outcome was extracted: rates of patients free from local recurrence (FFLR, at 3, 6, and 12 months at least). The following secondary outcomes were extracted: recurrence-free rates (RF, at 3, 6, and 12 months, at least) and overall survival rates (OS, at 3, 6, and 12 months, at least). We defined the overall recurrence rate as the appearance of new lesions during postoperative follow-up, irrespective of their location. Local recurrence rate was defined as a postoperative relapse occurring specifically on the ablated liver lesions.
Additional outcomes included the evaluation of recurrence-pattern patients with hepatic and extra-hepatic progression.

Quality Assessment
The quality of all included studies was assessed independently by A.M. and F.P. using the Newcastle-Ottawa scale (NOS) of quality assessment [26]. Any disagreements were resolved in consensus.

Statistical Analysis
Data of interest were collected as presented in the original manuscripts or were calculated from the reported raw data whenever possible. Quantitative data were presented descriptively as mean and standard deviations (SD). When continuous data were presented as medians and range, the method developed by Hozo et al. was used [27]. When continuous data were presented as medians and interquartile range, we used the method developed by Wan et al. [28]. Categorical variables were summarized as frequencies and percentages with 95% confidence intervals (CIs).
We extracted the recurrence and survival outcomes from the Kaplan-Meier curves and/or explicit data, when available. Subgroup survival analysis on patients with lesions of <30 mm, treated percutaneously or surgically, was performed as well as outcomes after MWA under laparoscopic versus open procedures. Finally, grouped analysis incorporated papers providing data on the local recurrence rates.

Literature Analysis
The literature search identified 4822 records. After excluding duplicates and citations, a pool of 3578 papers was screened. The first step of the literature analysis consisted of a screening of titles and abstracts to determine their eligibility and relevance to the topic: case reports, oral presentations, conferences, abstracts without a corresponding full text, reviews, meta-analyses, letters, editorials, books, non-English articles, reports describing study protocols, animal-model studies, and unrelated articles were excluded. After excluding 3469 papers, 109 articles were selected as being eligible for second-phase analysis.
The second step consisted of an analysis of the full text of articles meeting the specific inclusion criteria. Studies not including MWA as an exclusive treatment, with a sample of fewer than 10 patients, using MWA as a palliative treatment, not reporting the follow-up, combining MWA and other treatments on the same subjects, unrelated subjects, and those studies in which the patients who were affected by CRLM and/or treated with MWA could not be correctly isolated from those with other liver metastases and/or treated with other procedures (i.e., RFA) were excluded. Finally, 12 papers were included in the pooled analysis ( Figure 1).

Subgroup Analysis
To better understand the outcomes of MWA-treated patients, we sorted them into sub-categories and analyzed the outcomes of these subgroups.
The FFLR at 3, 6, and 12 months were 95.4%, 86.2%, and 78.5%, respectively. Overall, a recurrence was reported in 43.8% of patients, with a hepatic progression in 25.0%. The OS rates at 3 months, 6 months and 1, 3, and 5 years were 100%, 100%, 93.0%, 71.4%, and 53.6%, respectively. A graphic comparison between the two groups concerning RF, FFLR, and OS is depicted in Figure 3a-c.   (Tables 10 and 11). In this group, 153 lesions were treated in 77 patients. The mean lesion diameter was 22.81 mm (±7.79 mm). All patients presented with more than one lesion. Open surgery resulted in complications for 37.7% of patients, of which 14.3% were severe Three studies analyzed the data concerning laparotomic surgery for MWA [30,35,37] (Tables 10 and 11). In this group, 153 lesions were treated in 77 patients. The mean lesion diameter was 22.81 mm (±7.79 mm). All patients presented with more than one lesion. Open surgery resulted in complications for 37.7% of patients, of which 14.3% were severe (CDs ≥ 3). The RF rates at 3 and 12 months were 89.5% and 40.4%, respectively. The overall recurrence was 30.2%; the OS rates at 3 months, 6 months, and 1, 3, and 5 years were 100%, 98.7%, 81.8%, 42.9%, and 12.3%, respectively.  Only 2 studies reported data on laparoscopic MWA [38,39]. In this group, 242 lesions were treated in 122 patients. The mean lesion diameter was 10.63 mm (±1.57 mm). The majority (73.0%) had more than one lesion. Laparoscopically treated patients showed a complication rate of 13.1%; data concerning complications with CDs ≥ 3 were available for one study only, and no further analysis was possible. The RF rates at 3, 6, and 12 months were 99.2%, 95.1%, and 85.5%, respectively; the overall recurrence rate was 44.3%. Data concerning OS rates were available for one study only, and no further analysis was possible. A graphic comparison between two groups concerning OS is depicted in Figure 4.

Discussion
Surgical resection with a parenchymal sparing technique is the gold standard of care for CRLM [40,41]. Unfortunately, for oncological reasons and due to the patients' condition, few patients are candidates for curative-intent surgery. In this setting, when percutaneous thermal ablation techniques have been used in non-selected patients, oncological data from thermal ablation seemed to not be optimal when compared with a curative-intent surgical approach [32]. Modern CHT and target therapies for CRLM increased not only the resection rates but also the number of patients with complex CRLM and a high risk of recurrence in patients [32,40,42]. The main challenge is to define which patients can benefit most from surgery and thermal ablation techniques without increasing the morbidity and mortality rates [43].
Tabuse first developed the surgical technique of microwave coagulation in 1979 and applied it to the transection of hepatic parenchyma by coagulating the tumor tissue in many organs [44]. In the past, RFA and MWA have increased their usefulness in the context of CRLM treatment.
This study is the first systematic review focused only on the MWA of CRLM. In this review of 12 studies, we pooled 395 patients undergoing MWA for CRLM and reported the pooled analyses of OS, RF, and FFLR at 3 months, 6 months, and 1, 3, and 5 years.
A general observation of this systemic review was the performance of surgical MWA in 62.9% of patients. This underlines the need to explore other frontiers for CRLM treatment options, other than hepatic resection [20,29,30,32,35,[37][38][39]. It was also reported that the complication rate was 26.8%, of which cases only 8.5% were severe (CDs ≥ 3). These data were in concordance with the literature data concerning outcomes after liver resection for CRLM [45]. The pooled analyses of the oncological outcome of MWA showed OS survival rates at 1, 3, and 5 years of 86.7%, 59.6%, and 44.8%, respectively. This was also comparable with OS rates reported after the curative liver resection of CRLM [45]. The RF rates at 1, 3, and 5 years after MWA were 65.1%, 44.6%, and 34.3%, respectively, which were better than the previously reported data [45].
The present results, as well as the reported literature data, show a common intention

Discussion
Surgical resection with a parenchymal sparing technique is the gold standard of care for CRLM [40,41]. Unfortunately, for oncological reasons and due to the patients' condition, few patients are candidates for curative-intent surgery. In this setting, when percutaneous thermal ablation techniques have been used in non-selected patients, oncological data from thermal ablation seemed to not be optimal when compared with a curative-intent surgical approach [32]. Modern CHT and target therapies for CRLM increased not only the resection rates but also the number of patients with complex CRLM and a high risk of recurrence in patients [32,40,42]. The main challenge is to define which patients can benefit most from surgery and thermal ablation techniques without increasing the morbidity and mortality rates [43].
Tabuse first developed the surgical technique of microwave coagulation in 1979 and applied it to the transection of hepatic parenchyma by coagulating the tumor tissue in many organs [44]. In the past, RFA and MWA have increased their usefulness in the context of CRLM treatment.
This study is the first systematic review focused only on the MWA of CRLM. In this review of 12 studies, we pooled 395 patients undergoing MWA for CRLM and reported the pooled analyses of OS, RF, and FFLR at 3 months, 6 months, and 1, 3, and 5 years.
A general observation of this systemic review was the performance of surgical MWA in 62.9% of patients. This underlines the need to explore other frontiers for CRLM treatment options, other than hepatic resection [20,29,30,32,35,[37][38][39]. It was also reported that the complication rate was 26.8%, of which cases only 8.5% were severe (CDs ≥ 3). These data were in concordance with the literature data concerning outcomes after liver resection for CRLM [45]. The pooled analyses of the oncological outcome of MWA showed OS survival rates at 1, 3, and 5 years of 86.7%, 59.6%, and 44.8%, respectively. This was also comparable with OS rates reported after the curative liver resection of CRLM [45]. The RF rates at 1, 3, and 5 years after MWA were 65.1%, 44.6%, and 34.3%, respectively, which were better than the previously reported data [45].
The present results, as well as the reported literature data, show a common intention to treat CRLM that are less than 3 cm (91.6% of lesions) as a general principle of the advantageous influence of factors of the MWA [46,47]. In this subset of lesions, despite the presence of only 3 studies that analyzed oncological follow-up, 16.5% of patients experienced low-impact complications (CDs ≥ 3 = 0%). These results show a better safety profile than surgical resection and/or MWA from the Sweden Nationwide Registry, reporting severe complication rates of 16.4% and 7.0%, respectively [48].
Pooled analyses for RF rates and FFLR after MWA at 3, 6, and 12 months showed 91.8%, 83.5%, and 76.5% and 97.1%, 92.9%, and 88.6%, respectively, which was encouraging. In addition, hepatic progression rates of 32.4% and overall recurrence rates of 28.2% were in concordance with the data observed after resection. Interestingly, the OS rates after MWA for CRLM (100% at 3, 6, and 12 months) were encouraging and seemed better than surgical resection [48,49].
Analyzing the subgroup of surgical vs. radiological MWA approaches, no difference concerning complications was found. However, local hepatic control was more satisfactory after following the surgical approach, compared to radiological procedures. Indeed, the global RF and OS rates did not reflect the efficacy of treatment but, above all, the biological characteristics of CRLM. These results might be explained by the possible favorable tumor location for MWA and not for surgical resections, which are likely to be more complicated to treat in the same way as for MWA. As the rate of local recurrence seems dependent upon tumor location and its vessel proximity, in order to minimize the "heat sink effect", pedicle clamping during MWA has been used successfully [20,32]. In terms of comparing surgical MWA, only the laparoscopic approach ensured a good RF rate. We believe that these results might be due to the fact that smaller-sized CRLM were treated using the laparoscopic approach (10 mm vs. 22.8 mm in an open approach) [50,51].
Despite our efforts to create a homogeneous comparison group, we acknowledge that our systematic review suffers from several limitations. Only 4 studies were prospective; in addition, some papers reported a heterogeneous number and size of ablated lesions, showing a general lack of standardization and follow-up protocols. Therefore, these results should be interpreted with caution, considering that most of them were extracted from clinically heterogeneous studies.
Finally, we believe that MWA represents a promising technique in the treatment of CRLM. When used appropriately, especially in selected patients with tumors less than 3 cm in size, the oncological results are promising. RFA and MWA are not mutually exclusive but they are additional, with the advantage of being able to perform MWA near the large blood vessels. For the bigger biliary ducts, the use of thermal ablation remains absolutely contraindicated. This extends the use of intraoperative MWA, which can be performed safely using a laparoscopy. Resection with or without MWA can achieve encouraging oncological results, with low morbidity allowing patients to receive their systemic drugs more quickly; thus, their care is not interrupted. We are also impatient to discover the final results of the ongoing prospective randomized COLLISION trial (Colorectal liver metastases: surgery versus thermal ablation trial) [52].

Conclusions
Our findings indicate that MWA could be a valid tool for CRLM treatment, especially for deep lesions and those smaller than 3 cm. The surgical approach for MWA could improve local control and reduce complications.