Percutaneous ablation of colorectal liver metastases: a comparison between the outcomes of grayscale US guidance and Sonazoid CEUS Kupffer phase guidance using propensity score matching

Abstract Purpose To assess the utility of Sonazoid contrast-enhanced ultrasound (CEUS) for guiding percutaneous microwave ablation (MWA) for colorectal liver metastases (CRLMs). Materials and Methods The medical records of patients who had undergone ultrasound (US)-guided percutaneous MWA between July 2020 and June 2022, were reviewed. Propensity score matching (PSM) with a ratio of 1:1 was used to balance the potential bias between the grayscale US-guided and Sonazoid CEUS-guided groups. Local tumor progression (LTP), intrahepatic recurrence (IR), and complication rates were compared between the two groups. Results Of 252 patients enrolled, 247 achieved complete ablation, and the technical effectiveness was 98.0% (247/252). Of these 247 patients, 158 were in the grayscale US-guided group and 89 in the Sonazoid CEUS-guided group. The median follow-up period was 14.6 months. After PSM, there were no significant differences in LTP, IR, or complication rates between the two groups (p = 0.100, p = 0.511, p > 0.99, respectively). Multivariate analysis identified tumor size ≥ 3 cm (hazard ratio [HR], 7.945; 95% CI, 2.591-24.370; p < 0.001), perivascular (HR, 2.331; 95% CI, 1.068-5.087; p = 0.034), and tumor depth > 8 cm (HR, 3.194; 95% CI, 1.439-7.091; p = 0.004) as significant factors associated with LTP. For tumors with poor vision on grayscale US, Sonazoid CEUS-guided ablation achieved a better LTP rate than grayscale US-guided ablation (3.7% vs.14.8%, p = 0.032). Conclusion For tumors with poor vision on grayscale US, Sonazoid CEUS guidance is recommended for better local tumor control.


Introduction
Colorectal cancer (CRC) is the second most common cancer and the third leading cause of cancer-related death worldwide [1,2].Approximately 20% of patients with CRC have liver metastases at diagnosis, and more than 50% will develop liver metastases during the disease course [3,4].Although surgical resection is the first choice for patients with liver metastases, 80% cannot undergo surgery because of heavy tumor burden, unfavorable tumor location, or severe comorbidities [5].Ablation offers a treatment option for patients without surgical options [6].Compared to surgery, thermal ablation has the advantages of being minimally invasive, having a short hospital stay, and fewer complications [7,8].
Computed tomography (CT) and ultrasound (US) are the most widely used imaging modalities for guiding ablation.Compared to CT, US has the advantages of facilitation, occurring in real-time, being radiation-free, and flexibility [9].Unfortunately, some colorectal liver metastases (CRLMs) are invisible or have poor visibility on grayscale US [10].This problem can be resolved by contrast-enhanced ultrasound (CEUS).CEUS with SonoVue (sulfur hexafluoride microbubbles; Bracco, Milan, Italy) is more sensitive than grayscale US in detecting CRLM [11].However, when guiding ablation in patients with multiple CRLMs, the role of SonoVue is limited owing to short-lasting enhancement.Sonazoid (perfluorobutane microbubbles; Daiichi-Sankyo, Tokyo, Japan; GE Healthcare, Milwaukee, WI, USA) is a perfluorobutane microbubble that can be phagocytosed by hepatic Kupffer cells, allowing enhancement of the liver parenchyma for more than 60 min [12].This specific Kupffer imaging method has realized the operating window required to perform the procedure and facilitated precise applicator placement during the procedure [13].Studies have shown that compared to grayscale US-guided radiofrequency ablation (RFA), Sonazoid CEUS-guided RFA decreases the number of sessions in hepatocellular carcinoma (HCC) [14,15].Ultimately, compared with grayscale US guidance, Sonazoid-CEUS guidance can lead to superior local tumor control in HCC [16].However, few studies have investigated the utility of Sonazoid CEUS in guiding the percutaneous ablation of CRLM.This study aimed to evaluate the utility of Sonazoid for guiding percutaneous microwave ablation (MWA) in CRLM.

Patients
The Institutional Review Board of the Sixth Affiliated Hospital of Sun Yat-sen University approved this retrospective study (approval protocol number: 2021ZSLYEC-340) and informed consent for the treatment was obtained from all patients.The requirement for informed consent to publish these results was waived because of the retrospective nature of this study.Patients diagnosed with CRLM who underwent percutaneous US-guided microwave liver ablation between July 2020 and June 2022 were included.The inclusion criteria were [1]: patients with pathology confirmed CRC [2], number of CRLMs <9 [3], the maximal size of CRLMs � 5 cm, and [4] all tumors detectable by preoperative US/CEUS.The exclusion criteria included [1]: patients without preoperative contrast-enhanced CT/MRI imaging one month before ablation [2], a history of other malignant tumors [3], prior non-R0 resection of CRLMs [4]; insufficient follow-up time (survival status ¼ "Local tumor progression-free survival" and "intrahepatic recurrence-free survival" & survival time < 6 months), and [5] inability to ablate the lesions because of proximity to major biliary structures or adherence to the gastrointestinal tract.The patient enrollment flowchart is shown in Figure 1.

Ablation
A 2450-MHz microwave ablation system (KY2000; Nanjing Kangyou Biological Energy Co., Ltd., Nanjing, China) was employed.This system consisted of a microwave generator with an output power range of 1-100 W, a flexible coaxial cable, and a cooled-shaft antenna (KY-2450-b; Nanjing Kangyou Biological Energy Co., Ltd., Nanjing, China).The antenna was 18 cm long and 1.9 mm in diameter (15 G), and the temperature was maintained below 40 � C by adjusting the flow of the cold water.Based on the number, size, and visualization of the tumor on grayscale US, the radiologist decided to perform the procedure under the guidance of grayscale US or Sonazoid CEUS.Before puncture, 2% lidocaine was administered into the liver capsule using a 24-G needle 10 cm in length to provide local anesthesia.The key MWA parameters, including the number of antennas, number of insertions, ablation time, and output energy, were determined according to the tumor size, shape, location, and adjoining structures and organs.Generally, 1-2 antennas and 1-2 insertions are used for tumors <3.0 cm in diameter, and two antennas and multiple insertions are used for tumors � 3.0 cm.After the antenna was inserted into the center of the tumor under the guidance of grayscale US or Sonazoid CEUS, an energy output of 40-60 W was applied for 3-10 min.A radiologist with more than 20 years of experience in liver tumor ablation performed all the procedures.

Definitions
Technical success, local tumor progression (LTP), intrahepatic recurrence (IR), and complications were evaluated according to terminology standardization guidelines [17].Technical success was defined as the treatment of the tumor by following the plan and ensuring that the ablation zone completely covered the tumor when the procedure was completed.Complete ablation was defined as the absence of enhancement on follow-up contrast-enhanced computed tomography (CECT) or contrast-enhanced magnetic resonance imaging (CEMRI) 1 month after ablation.Technical effectiveness was defined as the complete ablation of the tumor at the imaging follow-up 1 month after MWA [18].LTP was defined as the appearance of new nodular enhancement in the marginal area of the ablative zone after complete ablation.IR refers to the appearance of new tumors within the liver without connecting to the ablation zone or the resection margin [19].Major complications refer to events that require additional therapy and hospital admission, whereas minor complications refer to events with no severe consequences and requiring no therapy.LTP, IR, and complication rates were compared between the grayscale US-guided-and Sonazoid CEUS-guided groups.Procedure time was defined as the duration from the start of local anesthesia to the time the patient left the operating room for ultrasound intervention.The Tumor Burden Score (TBS) was estimated by incorporating the maximum tumor size and number of tumors and calculated according to the following mathematical equation: [20].
TBS ¼ ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi The location of each tumor was classified as perivascular or subcapsular.A perivascular tumor was defined as a tumor margin less than 1 cm from the first-or second-degree branch of a portal or hepatic vein with a diameter greater than 3 mm [21].A subcapsular tumor was defined as a tumor with a distance between the tumor margin and liver capsule of less than 1 cm in diameter [22].A five-point confidence scale was used to grade the visibility of the target tumor compared to that of the adjacent liver parenchyma, with 1 denoting invisibility and 5 denoting good visibility [23].Scales 1 and 2 are generally classified as poor vision and scales 3-5 are generally classified as clear vision.CEUS was performed 30 min and 24 h after ablation.CEUS and CECT/MRI were performed 1 month postoperatively.CEUS was performed every 3 months.CECT (Aquilion One, Toshiba Medical System, Tokyo, Japan) or CEMRI (Optima MR360, GE Healthcare, Milwaukee, WI, USA) was performed every 3 months for half a year, and every 6 months thereafter for follow-up.

Statistics
Using propensity score matching (PSM) to balance the potential bias of baseline information between the two groups before analysis, the propensity score was calculated using a logistic regression model.Patients in the two groups were matched using the 1:1 nearest-neighbor matching method with a match tolerance of 0.1.An absolute standard difference lower than 0.2 were used to assess the balance of variables used for matching.The variables included in the propensity score model were as follows: age, sex, comorbidities, body mass index (BMI), carcinoembryonic antigen (CEA) levels, primary tumor location, T and N categories, extrahepatic metastasis, prior chemotherapy, after chemotherapy, therapeutic purpose, synchronous or metachronous liver metastases, disease-free interval (DFI), number and maximal size of metastases, and tumor burden score (TBS).
Continuous variables satisfying a normal distribution were expressed as mean ± standard deviation (SD) and variables that did not satisfy a normal distribution were expressed as median (range).Continuous variables were compared using

Efficacy of MWA
In total, 252 patients with 558 CRLMs were included in the (Figure 1).

Propensity score matching
A total of 247 patients (167 males and 80 females; mean age, 55.5 ± 10.9 years; range, 25-79 years) were included for the analysis of LTP, IR, and complications.Ablation was performed in 158 patients in the grayscale US-guided group and 89 patients in the Sonazoid CEUS-guided group.The median follow-up period was 14.6 months (range, 1.8-31.3months).Before PSM, the mean number of ablated tumors in the grayscale US-guided group and Sonazoid CEUS-guided group per patient was 1.96 ± 0.11 and 2.70 ± 0.15, respectively.The median number of ablated tumors was 1 (range, 1-8) and 1 (range, 1-7) in the grayscale US-guided and Sonazoid CEUS-guided groups, respectively.Patients in the Sonazoid CEUS-guided group had more CRLMs than those in the gray scale US-guided group (p < 0.001).Additionally, the number of liver metastases, TBS, extrahepatic metastases, and therapeutic purposes differed between the two groups.After PSM, 144 patients were matched and baseline information showed no significant differences between the two groups.The median number of ablated tumors was 2 (range, 1-8) and 2 (range, 1-6) in the grayscale US-guided and Sonazoid CEUS-guided groups, respectively.Baseline information before and after PSM is summarized in Table 1.The characteristics of the CRLMs in the grayscale USguided and Sonazoid CEUS-guided groups are summarized in Table 2. Grayscale US vision and CRLM depth significantly differed between the two groups.Tumors with poor vision in US were usually ablated under the guidance of Sonazoid CEUS, and tumors with a depth > 8 cm were more often ablated under the guidance of grayscale US.There were no significant differences in other characteristics between the two groups (Table 2).

Local tumor progression
Before PSM, 27 of 158 (17.1%) and 13 of 89 (14.6%) patients developed LTP in the grayscale US-guided and Sonazoid CEUS-guided groups, respectively.After PSM, 16 of 72 (22.2%) and 11 of 72 (15.2%) patients developed LTP in the grayscale US-guided and Sonazoid CEUS-guided groups, respectively.There was no significant difference in the LTP rates before and after PSM between the two groups (Figure 2(A,B)).
The subgroup analyses of LTP are summarized in Figure 3.In the subgroup of tumors with poor US visibility, 12 of 81 (14.8%) and 4 of 108 (3.7%) tumors developed LTP in the grayscale US-guided and Sonazoid CEUS-guided groups,  respectively (p ¼ 0.032) (Figure 4(A)).In the subgroup of tumors with a clear vision, no significant difference was found between the two groups (p ¼ 0.559) (Figure 4(B)).In univariate analyses, extrahepatic disease (EDH) and nodal status of the primary tumor were factors associated with IR (Supplementary material, Table S1).In multivariate analysis, EDH (HR, 2.070; 95% CI, 1.202-3.565;p ¼ 0.009) was a significant factor associated with IR (Supplementary material, Table S2).

Complications and procedure time
Before PSM, the incidence of complications was 6.32% (10/158) in the grayscale US-guided group and 10.11% (9/89) in the Sonazoid CEUS-guided group (p ¼ 0.284).The minor and major complication rates were similar between the two groups.After PSM, there was no significant difference in the rates of major (5.6% vs. 6.9%,p ¼ 0.374) or minor (4.2% vs. 2.8%, p > 0.99) complications related to ablation between the two groups.
Before PSM, the procedure time was significantly longer in the Sonazoid CEUS-guided group than in the gray scaleguided group [50.0 (range, 15-120) vs. 40.0(range, 13-110), p < 0.001)].After PSM, there was no statistically significant difference in procedure time between the two groups [42.5 (range, 15-110) vs. 45.0 (range, 13-120), p ¼ 0.113].summarizes the comparison of the procedure time and incidence of complications between the two groups before and after PSM.

Discussion
Sonazoid CEUS-guided ablation achieved a non-inferior LTP and IR rate and did not increase the procedure time or complication rate compared to grayscale-guided ablation.Sonazoid CEUS with an extended window is a powerful adjunct for targeting liver metastases that are poorly visualized on grayscale US.For tumors with the poor vision on grayscale US, Sonazoid CEUS guidance helps obtain better local tumor control.
According to previous reports, the incidence of LTP after percutaneous thermal ablation for CRLM ranges from 4.1% to 32.6% [18,24,25].In this study, the LTP rate was 18.7% (27/144), which is consistent with the results of previous studies.In multivariate analysis, a tumor diameter � 3 cm and perivascular location were independent factors  associated with LTP, which agrees with previous studies [22,26].In subgroup analysis, ablated tumors with poor vision guided by Sonazoid CEUS achieved a lower LTP rate than those guided by grayscale US (3.7% vs. 14.8%;p ¼ 0.032).For tumors with a poor vision on grayscale US, CEUS improves the visualization of tumors and thus helps precisely place the ablation needle [25].Therefore, we suggest that when liver metastases have poor vision on grayscale US, Sonazoid CEUS-guided ablation is a better choice, helping reduce the LTP rate.After chemotherapy, treatment-related changes may lead to poor definition of tumor margins on imaging [27,28].In the subgroup analysis of patients who received chemotherapy before ablation, patients in the Sonazoid CEUS-guided group achieved a lower LTP rate than those in the grayscale US-guided group; however, the difference was not significant (18.0% vs. 28.6%,p ¼ 0.136).Additionally, LTP is closely related to post-ablation safety margins.A commonly used method for measuring safe margins is the use of anatomical landmarks on pre-and post-ablation CT imaging [29].Recently, image data fusion imaging has been developed [30,31].A clear display of the lesion boundary facilitates safe margin measurement.In our experience, Sonazoid CEUS helps better design the ablation procedure by improving the visualization of lesions.
Compared with grayscale US, CEUS is more sensitive for detecting malignant liver tumors, particularly smaller tumors [32,33].In this study, the number of ablated tumors in the Sonazoid CEUS-guided group was significantly higher than that in the grayscale US-guided group (2.70 ± 0.15 vs. 1.96 ± 0.11, p < 0.001).SonoVue and Sonazoid are commonly used contrast agents; one study showed that Sonazoid was capable of detecting more lesions than SonoVue [34].Sonazoid, with a stable Kupffer phase, provides sufficient time for full detection of the liver and radical ablation of all tumors through a single injection.Comprehensive treatments, including radical ablation of liver metastases with no evidence of disease (NED) status, help achieve a better prognosis [35,36].In this study, Sonazoid CEUS-guided ablation helped 97.8% (89/91) of the patients who underwent radical ablation.According to a meta-analysis, multiple tumors were independent factors of recurrence-free survival (RFS) and overall survival (OS) [37].In this study, Sonazoid was often used to detect tumors and perform ablation in patients with multiple tumors, resulting in a higher IR rate in the Sonazoid CEUS-guided group than in the grayscale US-guided group.After PSM, the tumor burden was balanced between the two groups and Sonazoid CEUS guidance did not increase the incidence of IR (p ¼ 0.511).
A systematic review of 32 studies reported major complication rates of 4.6%, minor complication rates of 5.7%, and mortality rates of 0.15% in MWA [38].Thus, the 6.3% (9/144) major complication rate, 3.5% (5/144) minor complication rate, and 0% mortality rate in this study were within the tolerable risk limits.Moreover, there were no differences in the major or minor complication rates between the grayscale USguided and Sonazoid CEUS-guided groups, indicating that Sonazoid CEUS guidance did not increase the risk of MWA.
This study has some limitations.First, this was a retrospective study, and bias in the selection of different guidance approaches was inevitable.Although baseline information was balanced after PSM, the raw data were damaged because some patients were excluded from the final analysis.Second, the evaluation of tumor visibility based on the decisions of radiologists is subjective, which may be inevitable because of the subjective nature of US examinations.Third, 1-month follow-up may not be sufficient for detecting residual tumors at the ablation site, and the definition of complete ablation is a relative concept.

Conclusion
In conclusion, grayscale US or Sonazoid CEUS guidance for the percutaneous ablation of CRLM does not appear to affect LTP, IR, or complications.The use of Sonazoid CEUS guidance was conducive to decreasing the LTP rate for tumors with poor vision on grayscale US and detecting more tumors for curative ablation.Thus, we recommend Sonazoid CEUS as a better choice for tumors with poor vision on grayscale US.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 3 .
Figure 3. Subgroup analyses of comparison of grayscale US-guided and Sonazoid CEUS-guided ablation for predicting local tumor progression-free survival (LTPFS) based on study variables among matched patients after propensity-score matching.BMI: body mass index; CEA: carcinoembryonic antigen; EHD: extrahepatic disease; DFI: disease-free interval; US: ultrasound.

Figure 4 .
Figure 4. Kaplan-Meier curves of subgroup analysis for lesions with poor vision and clear vision on grayscale ultrasound (US).(A) local tumor progression-free survival (LTPFS) curve for lesions with poor vision on grayscale US. (B) LTPFS curve for lesions with clear vision on grayscale US.

Table 1 .
Baseline information of patients before and after PSM (N ¼ the Student's t-test or Mann-Whitney U test, and categorical variables were compared using the chi-square test.Survival outcomes were calculated using the Kaplan-Meier method and compared using the log-rank test.Cox univariate and multivariate regression analyses were used to analyze the significant factors for LTP and IR.Factors with p < 0.05 in univariable analysis were further entered into multivariable Cox regression.Subgroup analysis used Cox regression and was performed on matched patients after PSM.All statistical analyses were performed using SPSS (version 25.0) and MedCalc (version 11.2; 2011 MedCalc Software bvba, Mariakerke, Belgium) software.All statistical tests were two-sided, and p < 0.05 indicated statistical significance.
Technical success was achieved in 100% (558/558) of the tumors, according to CEUS performed 30 min and 24 h after ablation.Incomplete ablation occurred in 6 CRLMs of 5 patients on contrast-enhanced imaging acquired 1 month after ablation, and technical effectiveness was achieved in 98.9% (552/558) of tumors and 98.0% (247/252) of patients.

Table 2 .
The characteristics of CRLMs in the grayscale US-guided group and Sonazoid CEUS-guided group (N ¼ 366).

Table 3 .
Results of univariate and multivariate Cox regression analyses for evaluating factors associated with local tumor progression-free survival.

Table 4 .
the comparison of procedure time and incidence of complications in the two groups before and after PSM.PSM: propensity score matching; US: ultrasound; CEUS: contrast-enhanced ultrasound.� p < 0.05 was considered a significant difference.