Intratumoral perfusion may affect microwave ablation area of hepatocellular carcinoma

Abstract Objectives We aimed to evaluate the effect of intratumoral perfusion on microwave ablation (MWA) area in hepatocellular carcinoma (HCC). Methods Patients who underwent curative MWA for HCC between October 2013 and May 2015 were enrolled. Three days before MWA, contrast-enhanced ultrasound (CEUS) was performed to illustrate the perfusion characteristics of the target lesion. Using the Sonoliver quantification software, time-intensity curves of dynamic CEUS were obtained, and quantitative parameters were extracted. Two microwave antennae were inserted into the center of the tumor and MWA was performed with a continuous power output of 50 W for 5 min. A second CEUS was performed to measure the size of the ablated region. Thereafter, an additional MWA procedure was performed until complete ablation with a 5–10-mm safety margin was achieved. Results A total of 38 patients who underwent curative MWA for 39 HCC nodules were enrolled. The mean age was 57 years (34–80 years), and the median maximum diameter of the HCC was 3.4 cm (interquartile range, 2–6.8 cm). Time-intensity curves were obtained and the area under the curve (AUC) was selected as a parameter for intratumoral perfusion. The AUC was inversely and linearly correlated with the size of the MWA area, including long- and short-axis diameters and ablation volume. A 1,000-dB·s change in the AUC produced an average change of 1.17 ± 0.44 mm, 0.725 ± 0.355 mm, and 2.4995 ± 0.6575 cm³ in the long- and short-axis diameters and ablation volume, respectively. Conclusions The intratumoral perfusion of HCC was inversely correlated with MWA area size.


Introduction
According to the latest global statistics, hepatocellular carcinoma (HCC) is the sixth most common cancer worldwide and fourth leading cause of cancer-related deaths [1].Ultrasonography-guided thermal ablation is a safe and effective therapy for HCC.Among various ablative strategies, microwave ablation (MWA) is regarded as a promising modality for HCC [2].However, incomplete MWA may occur in 5.9-20.0% of tumors, which is significantly associated with distant recurrence and survival during follow-up [3][4][5].Thus, complete tumor coverage during MWA is of utmost importance for patients with HCC.However, coagulated necrotic tissue after MWA is hyperechoic on grey-scale ultrasound, which makes it impossible to accurately evaluate the ablative margin, and may potentially lead to incomplete ablation [6].
A variety of factors may lead to the incomplete ablation of HCC, one of which is the heat-sink effect.The flowing fluid adjacent to the tumor can remove the local heat produced by thermal ablation, leading to incomplete ablation [7].Previous research has mainly focused on the heat-sink effect of large vessels adjacent to HCC during radiofrequency ablation (RFA), such as the portal vein [8,9].In contrast, only a few studies have assessed the heat-sink effect of the intratumoral microcirculation during MWA [10,11].
With the introduction of ultrasound contrast agents, contrast-enhanced ultrasound (CEUS) can dynamically display the intratumoral microcirculation of HCC in real-time, including the enhancement pattern of HCC during the arterial, portal venous, and late phases [12].Furthermore, the quantification software facilitates the quantitative analysis of microvascular perfusion in HCC.The obtained time-intensity curve (TIC) shows the wash-in and wash-out patterns of the contrast agent.The parameters extracted from the TIC characterize the blood flow and volume of the region of interest (ROI) [13].Several studies have demonstrated that parameters extracted from the TIC could reflect baseline perfusion characteristics and perfusion changes after antiangiogenic therapy and predict the response of HCC after treatment with lenvatinib, sorafenib, and bevacizumab [14][15][16].Among the different quantitative parameters, the area under the curve (AUC) reflects the blood volume in the ROI [16].Hence, if the entire tumor is considered as an ROI, the AUC reflects the internal blood flow of the entire tumor.
In this study, CEUS was performed before and after MWA to determine the intratumoral perfusion characteristics and size of the MWA area.The correlation between intratumoral perfusion of the HCC and MWA size was analyzed.Therefore, we investigated the influence of intratumoral perfusion on MWA in patients with HCC.

Participants
The study was approved by the Ethics Committee of Chinese PLA General Hospital and the approved protocol number is S2019-348-01.Written informed consent was waived owning to the retrospective design.Patients who underwent curative MWA for HCC between October 2013 and May 2015 were enrolled.The inclusion criteria were as follows: 1) the whole tumor could be clearly displayed on conventional ultrasound or CEUS; 2) the maximum diameter of the HCC did not exceed 7 cm; 3) there were no vessels with a diameter > 5 mm within 2 cm of the tumor margin; 4) the shape of the HCC was relatively regular; and 5) the Child-Pugh score was A or B. The exclusion criteria were as follows: 1) diffuse HCC and 2) severe coagulation abnormalities (prothrombin time >30 s, prothrombin activity < 40%, and platelet count < 40 � 10 9 /L).

Pre-MWA assessment
Before MWA, all patients underwent dynamic contrastenhanced computed tomography or magnetic resonance imaging to assess the enhancement patterns.Additionally, liver biopsy was performed under ultrasound guidance.Three days before the MWA, conventional grayscale ultrasonography and color Doppler flow imaging were performed to evaluate the location, size, shape, margin, echogenicity, and peripheral blood flow of the target lesion.CEUS was performed using the Acuson Sequoia 512 ultrasound system (Siemens Medical Solutions, Mountain View, CA) and a 4V1 vector transducer.A low mechanical index (<0.2) was selected to prevent microbubble disruption.A contrast agent (SonoVue, 2 ml; Bracco, Milan, Italy) was intravenously injected, followed by a flush with 5 ml sodium chloride solution (0.9%).The enhancement pattern of the target lesion was recorded during the arterial, portal venous, and late phases (Fig. 1).All CEUS procedures were performed and evaluated for consensus by two experts with at least 5 years of experience.

MWA procedure
The MWA system (KY-2000; Kangyou Medical, Nanjing, China) could produce 1-100 W of power at 2450 MHz.After the local injection of 1% lidocaine, two antennae were inserted into the tumor at a distance of 1.5 cm and placed in the target position under ultrasound guidance.For lesions that were inconspicuous on grey-scale ultrasound, the antennae were inserted under CEUS guidance.General anesthesia (propofol, 6-12 mg/kg per hour; ketamine, 1-2 mg/kg) was administered after insertion of the antennae [17].The power and duration of MWA per lesion were 50 W and 5 min, respectively.Subsequently, the microwave emission was paused for 5 min to dissipate the bubbles formed during ablation.CEUS was performed to calculate the MWA area.If the initial MWA was incomplete, an additional MWA procedure was performed until complete ablation was achieved with a 5-10mm safety margin.

Calculation of MWA area
Calculations of the MWA area included the maximum longaxis diameter (LAD), maximum short-axis diameter (SAD), Xaxis diameter (parallel to the intercostal space), Y-axis diameter (along the antenna axis), and Z-axis diameter (perpendicular to the X-Y plane) (Fig. 2).The volume of the MWA area was calculated using the formula for an ellipsoid: Volume ¼ 4/3p(X/2) � (Y/2) � (Z/2) [18].All measurements were based on the consensus of the two experts who performed CEUS.

Image analysis
The CEUS video was stored in DICOM format, and ROIs were marked on the CEUS images, including ROIs of the tumor (green line) and normal liver parenchyma (yellow line) at the same depth (Fig. 3a).Using the Sonoliver quantification software (version 1.1), the TIC of the HCC and normal liver parenchyma were obtained (Fig. 3b).Quantitative parameters extracted from the TIC included the AUC, maximum intensity within ROI (IMAX), rise time (RT, time taken from 10% IMAX to 90% IMAX), time to peak, mean transit time (time from wash-in to 50% wash-out of the contrast agent), and quality of fit between raw data and theoretical curve.

Statistical analysis
All statistical analyses were performed using the SPSS version 25.0 software (Chicago, IL).Normally distributed data were presented as mean ± SD, and skewed continuous data were presented as median (interquartile range).Categorical data were presented as frequency (%).The correlations between HCC size (maximum diameter of the HCC), LAD, SAD, and ablation volume with the AUC were analyzed using linear regression analysis.Statistical significance was set at p < 0.05.

Results
A total of 38 patients (33 men and 5 women) with 39 HCC nodules confirmed by the typical dynamic pattern on contrast-enhanced imaging and liver biopsy were enrolled in this study.Their age ranged from 34 to 80 years (mean age, 57 ± 11 years).The maximum diameter of the tumor ranged from 2 cm to 6.8 cm (median diameter, 3.3 cm).All the patients were diagnosed with HCC for the first time.The baseline characteristics of the enrolled patients and their tumors are shown in Table 1.
The quantitative CEUS parameters of the 39 tumors obtained from the TIC are shown in Table 2.Among the different quantitative parameters, the AUC, which is related to blood volume, was selected as the intratumoral perfusion parameter.The maximum diameter of HCC was inversely and linearly correlated with the AUC (r ¼ −0.434, p ¼ 0.007).The size of the MWA area after the 50 W-5 min protocol was calculated as described in the Materials and Methods section.The mean LAD, SAD, and volume of the MWA area were 4.8 ± 1.0 cm, 2.8 ± 0.7 cm, and 31.4 ± 17.8 cm 3 , respectively.
The results of the correlation analysis between the AUC and MWA area size are shown in Table 3.All three parameters of MWA size (LAD, SAD, and volume) showed inverse linear correlations with the AUC.When the AUC was included in the regression analysis as an independent variable and parameters of MWA size at 5 min were considered as dependent variables, the relationship between the AUC and MWA size parameters (LAD, SAD, volume, respectively) were as follows (Fig. 4): y ¼ −1.168 � 10 −3 x þ 6.028 (r ¼ −0.661, p < 0.001); y ¼ −7.250 � 10 −4 x þ 3.521 (r ¼ −0.566, p < 0.001); and y ¼ −2.499 � 10 −2 x þ 56.798 (r ¼ −0.785, p < 0.001).For every 1000-dB�s change in the AUC, there was an average change of 1.17 ± 0.44 mm, 0.725 ± 0.355 mm, and 2.4995 ± 0.6575 cm 3 in the LAD, SAD and volume of the MWA area, respectively, after the 50 W-5 min protocol.

Discussion
Image-guided thermal ablation is regarded as an effective therapy for small HCC.However, the heat-sink effect may lead to decreased thermal energy in the targeted area, resulting in incomplete ablation and tumor recurrence [7].Several animal studies have demonstrated that the heat-sink effect of the large blood vessels near HCC (portal vein, hepatic artery, hepatic vein and the main branches) may lead to a decreased ablative temperature and reduced ablation area [11,19].
To reduce the influence of the heat-sink effect, innovative techniques have been introduced to block the blood supply to the HCC before thermal ablation [20].These techniques include intravascular embolization and surgical or image-guided occlusion of large vessels adjacent to the targeted tumor [21,22].Thus, the peripheral blood supply to the HCC is reduced, and the ablation area is enlarged.However, such approaches are usually invasive and disrupt not only the blood flow near the tumor but also that of the surrounding normal tissues.
In addition to the large vessels near the tumor, intratumoral microvessels may affect the ablation area via the heatsink effect.Moussa et al. evaluated the effects of RFA combined with adjuvant intravenous liposomal doxorubicin on intratumoral microvascular morphology and patency [23].The results showed that RFA combined with liposomal  doxorubicin led to microvascular occlusion in the early stage of ablation compared with RFA without liposomal doxorubicin.Wu et al. found that, compared with vasoconstrictors (phenylephrine), a vasodilator (hydralazine) significantly decreased intratumoral blood flow and increased the coagulation size of RFA [24].Although the above results provide some insight, intravenous liposomal doxorubicin and vasodilators act not only on tumor vessels but also on the blood vessels of normal tissues, resulting in undesirable side effects.Therefore, the clinical applications of this medicine require further investigation.Previous research has shown that among the quantitative parameters obtained from CEUS analysis, the AUC is directly related to tumor blood flow; therefore, this was selected as the intratumoral perfusion parameter in this study [25,26].Postoperative CEUS was used to calculate the ablation size and its correlation with intratumoral perfusion was explored.
Our study revealed that all three quantitative parameters of the MWA area (LAD, SAD, and volume) after 5 min had inverse linear correlations with the AUC.Previous studies reported that only the SAD and volume had inverse linear correlations with blood perfusion, whereas the LAD had no correlation with blood perfusion [19,27].Several reasons may account for this difference in findings.First, previous research mainly focused on the heat-sink effect of normal blood vessels adjacent to the tumor in the ablative area, whereas this study mainly focused on the effect of internal perfusion of the tumor.Second, the energy sources, heat distribution, and ablation shapes of MWA and RFA differ.Third, the quantitative parameter, AUC, was selected to evaluate the in vivo intratumoral perfusion, whereas previous research mainly evaluated the blood perfusion of the target tumor based on the diameter of the blood vessels and distance from the tumor.Unlike previous research, to minimize the heat-sink effect of peripheral large blood vessels, we excluded patients with large vessels with a diameter > 5 mm within 2 cm of the tumor margin.
This study has several implications.First, although the heat-sink effect of MWA is less significant than that of RFA, it should not be ignored in clinical practice.Second, during MWA, in addition to the blood flow adjacent to the tumor, microvessels within the tumor can also remove the ablation heat and finally affect the ablative size.Third, CEUS can not only be used to evaluate the perfusion of tumors but can also be used to predict the ablation size.The formula obtained in this study can be used for the preliminary calculation of MWA size when using the 50 W-5 min protocol when there are no large vessels with a diameter > 5 mm within 2 cm of the tumor margin.
This study has some limitations.First, the sample size was small.Thus, further research with a larger sample size is warranted to evaluate the correlation between the intratumoral perfusion of HCC and MWA size.Another limitation is the variability of the tumor diameter, ranging from 2.0 to 6.8 cm.In previous studies, tumor diameter may influence HCC perfusion, although the results remain conflicting [28][29][30].In the present study, only four tumors were >5 cm in diameter, and their influence on intratumoral perfusion and ablation area could not be precisely determined.Future studies with a large sample size and subgroup analyses according to tumor size should be performed to verify the influence of tumor perfusion on ablation size by tumor diameter.Third, we used rigorous inclusion criteria.To eliminate the influence of peripheral blood vessels on thermal ablation, tumors with vessels >5 mm in diameter within 2 cm of the periphery were excluded.In addition, a strict ablation protocol was employed to ensure consistent heat distribution in different tumors.Future studies should be conducted on tumors with large peripheral vessels with subgroup analyses by tumor size and ablation protocol, in order to further explore the relationship between tumor perfusion and ablation volume.Fourth, shrinkage of the target tissue during MWA should not be ignored even though the same ablation protocol and temperature-monitoring system were used.Further studies are underway to better characterize the true ablation volume after MWA.
In conclusion, intratumoral perfusion of the HCC may affect the MWA size.The LAD, SAD, and volume of the MWA area was inversely and linearly correlated with the intratumoral perfusion of HCC.

Figure 1 .
Figure 1.Grey-scale and contrast-enhanced ultrasound of hepatocellular carcinoma in a 48-year-old male patient.(a) Grey-scale ultrasonography image showing a hypoechoic hepatic nodule.(b) Contrast-enhanced ultrasound in the arterial phase (21 s) showing hyperenhancement.(c) Contrast-enhanced ultrasonography image showing hypoenhancement in the late phase at 200 s.

Figure 2 .
Figure 2. Grey-scale and contrast-enhanced ultrasound of microwave ablation area after the 50 W-5 min protocol.(a) Grey-scale ultrasonography image showing a hyperechoic nodule.(b) Illustration of the X-and Y-axes of the contrast-enhanced ultrasonography image.(c) Illustration of Z-axis on a contrastenhanced ultrasound image.

Figure 3 .
Figure 3. Contrast-enhanced ultrasound and time-intensity curves of hepatocellular carcinoma (HCC) and normal liver parenchyma.(a) Contrast-enhanced ultrasound image of the entire HCC (green line) and normal liver parenchyma (yellow line).(b) Time-intensity curves of the HCC (green line) and normal liver parenchyma (yellow line) obtained using the Sonoliver quantification software.

Table 1 .
Baseline characteristics of enrolled patients and tumors.

Table 2 .
Quantitative CEUS parameters extracted from time-intensity curves.
Abbreviations: AUC, area under the curve; IMAX, maximum intensity; RT, rise time; TTP, time to peak; mTT, Mean transit time; QOF, quality of fit.

Table 3 .
Correlation analysis of AUC and size of MWA area.