Experimental study on monitoring microwave ablation efficacy by real-time shear wave elastography in ex vivo porcine brain

Abstract

Objective: Glioma constitutes the most common primary malignant tumor in the central nervous system. In recent years, microwave ablation (MWA) was expected to be applied in the minimally invasive treatment of brain tumors. This study aims to evaluate the feasibility and accuracy of microwave ablation in ex vivo brain tissue by Shear Wave Elastography (SWE) to explore the application value of real-time SWE in monitoring the process of MWA of brain tissue.

Methods: Thirty ex vivo brain tissues were treated with different microwave power and ablation duration. The morphologic and microscopic changes of MWA tissues were observed, and the diameter of the ablation areas was measured. In this experiment, SWE is used to quantitatively evaluate brain tissue’s degree of thermal injury immediately after ablation.

Results: This study It is found that the ablation range measured by SWE after ablation is in good consistency with the pathological range [ICCSWEL1-L1 = 0.975(95% CI:0.959 − 0.985), ICCSWEL2-L2 = 0.887(95% CI:0.779 − 0.938)]. At the same time, the SWE value after ablation is significantly higher than before (mean ± SD,9.88 ± 2.64 kPa vs.23.6 ± 13.75 kPa; p < 0.001). In this study, the SWE value of tissues in different pathological states was further analyzed by the ROC curve (AUC = 0.86), and the threshold for distinguishing normal tissue from tissue after ablation was 13.7 kPa. The accuracy of evaluating ablation tissue using SWE can reach 84.72%, providing data support for real-time quantitative observation of the ablation range.

Conclusion: In conclusion the accurate visualization and real-time evaluation of the organizational change range of the MWA process can be realized by real-time SWE.

Keywords:

• SWE
• MWA
• porcine brain
• US
• pathology

Introduction

Glioma constitutes the most common primary malignant tumor in the central nervous system [1]. According to the World Health Organization (WHO) shows that grade III and IV gliomas typically feature low differentiation, rapid growth, and robust invasiveness in the glioma grading system. These high-grade gliomas occur mainly in adults and account for over 90% of all gliomas, while the survival cycle of the patients averages about 15.6 months (95%CI 13.3–19.1 months) [2]. The conventional treatment for high-grade gliomas principally involves surgical resection, when possible, supplemented by extracorporeal radiotherapy and chemotherapy. Nevertheless, this often accompanies normal tissue damage and causes severe pain. Consequently, a new scheme for accurate treatment of gliomas is needed.

Microwave ablation generates high-frequency electromagnetic waves utilizing microwave antennas positioned in tumors, which instigates high-speed oscillation of ions and water molecules in the tissue and produces substantial heat, resulting in coagulative necrosis of proteins in tumor cells [3]. In comparison with existing ablation systems, MWA features a larger ablation area, shorter operation time, and a more effective generation of heat [4,5]. At present, as an emerging means of thermal ablation, MWA maintains a stable position within ablation surgery for the treatment of local tumors in various organ systems [3–7]. Nonetheless, in brain tumors, MWA has not attained wide use in clinical practice, one of the constraints being the capacity to monitor the degree of thermal coagulation. While enhanced computed tomography (CT) and magnetic resonance imaging (MRI) scans have been considered highly accurate detection methods for evaluating the efficacy of thermal ablation [8], they cannot be used for real-time intraoperative monitoring [9]. Ultrasound, which is able to produce real-time images without radiation damage, provides an essential image guidance tool for thermal ablation in tumor therapy. Prior to ablation, ultrasound may be used for differential diagnosis, preliminary localization, and treatment design before ablation, while also playing an important role in routine follow-up examination during and after ablation [10].

Conventional ultrasound (US) may be employed to guide real-time observation during puncture and surgery; nonetheless, in a conventional US examination, the artifact of bubble formation during ablation represents a considerable problem for visualization of the treatment area [11]. Ultrasound elastography is an imaging method based on differences in tissue hardness [12]. It has been proven to provide excellent contrast for thermal damage brought about by MWA and may be used to monitor and evaluate thermal ablation therapy [13,14]. Compared to conventional ultrasound, real-time ultrasound elastography can more accurately evaluate the size of the ablation area, according to reports [15,16].

By generating shear waves with acoustic radiation and tracking their velocity, shear-wave elastography (SWE) is capable of estimating mechanical properties within soft tissues [17], and quantitatively measuring tissue elasticity and hardness of tissues with a high degree of repeatability [13,18–21]. Currently, only limited systematic studies employ SWE to assess brain tissue hardness, mainly comprising SWE validation analysis of human brain diseases [22–24]; meanwhile, concerning animal models of brain diseases [25–27], the feasibility and accuracy of using SWE to evaluate and monitor the range and tissue changes of the ablation area during MWA of brain tissue remain unclear. Consequently, the central purpose of this study is to evaluate the feasibility and accuracy of MWA in ex vivo brain tissue by SWE, exploring the application value of real-time SWE in monitoring the process of MWA of brain tissue.

Methods

The ex vivo porcine brain model

On the day of experiment, thirty fresh ex vivo porcine brains were purchased from the pork processing facility. The laboratory temperature was maintained at 26.0 °C ± 2 °C. The brain tissue samples were placed flat on the operation bed and adjusted to a constant 36.0 °C ± 0.5 °C to simulate the internal environment of animals. Due to the brain tissue being loose, soft, and fragile, normal saline was continuously dripped on the surface of the brain tissue to keep fresh and increase close contact with the probe.

Microwave ablation

An MDW-A MWA system (Nanjing Ruibo Medical Instrument Co., Ltd., Nanjing, China) and microwave thermal ablation antenna (2450 MHz, diameter = 1.6 mm, length = 10 cm) were employed for MWA. Based on the conventional two-dimensional ultrasound guidance, the microwave antenna was horizontally inserted at 2.0 cm from the surface of the brain tissue, and ablated for 10, 20, 30, 40, 50 and 60 s at a different constant output power of 30, 35, 40, 45 and 50 W, respectively.

Conventional US and SWE examination

Before and during ablation, an Aplio i800 ultrasound machine (Canon, Japan) and an L18-5 linear array transducer (frequency range = 5–18 MHz) were employed for the ultrasound examination (Figure 1A). According to the US imaging in the ablation area, the boundary of the strong echo region of the ablation gas may be determined, including the maximum transverse diameter (USL1) in addition to the vertical diameter (USL2) of the ablation region (Figure 1B).

Figure 1. Ultrasound and pathological findings.

Note: A. Ultrasound guided microwave antenna insertion into ex vivo brain tissue； B. Hyperechoic (arrows) area gradually observed with slow enlargement over time. USL1 and USL2 are the maximum diameters that can be measured along the antenna insertion channel, and USL2 is perpendicular to USL1; C. SWE images of normal brain tissue; D. SWE image of microwave ablation area; E. Cross macroscopic images of ablation; F. Microscopic pathology after H&E staining. Section of with a well-demarcated area of coagulation necrosis; G. corresponded to green box 1 on Figure F; H. corresponded to green box 2 on Figure F

Thereafter a switch to SWE mode measurement was undertaken. In the associated SWE sampling frame (2.5 cm [transverse size] 1.5 cm [axial size]), the depth range: (2.0 ± 0.5 cm) comprises the ablation area and surrounding normal brain tissue (Figure 1C). In terms of the color scale, Young’s modulus was set to 0–60 kPa. When the elastic image became stabilized at five or six frames, it was frozen and stored. The SWE value of the ablation area, as shown in the Figure 1D, was measured immediately after conventional US exam following ablation. To facilitate comparison, the SWE diagram outlined the ablation area, with the electrode antenna insertion trace is used as the reference mark. Both the maximum diameter SWEL1 of the ablation area of the electrode antenna and the maximum short path SWEL2 perpendicular to the antenna insertion channel were measured. Under these conditions, when the quality control waves appear in parallel, the SWE measurement represents available data [28].

Histopathological examination

After ablation, a 0.4% trypan blue solution of approximately 0.03 ml was rapidly injected with a 1 ml syringe perpendicular to the end of the ablation antenna, in order that the trypan blue solution would be fully stained along the necrotic tissue around the ablation antenna path, thus rendering the area accessible to identification with the naked eye. The brain tissue was fixed with a 10% neutral paraformaldehyde fixation solution, and 72 h later, tissue samples were dissected along the blue staining position (Figure 1E). The unthawed area was then embedded in paraffin wax, while the section centered on the ablation antenna insertion tract was stained with hematoxylin and eosin and observed under a light microscope. Finally, all slides were digitalized employing Pannoramic Digital Slide Scanner (3DHISTECH, Sysmex, Budapest, Hungary) and different pathological regions were quantified on digitalized images (MRXS file extension dedicated for Case Viewer, 3DHISTECH, version 2.3.0., Figure 1F) and maximum transverse diameter (L1) and vertical diameter (L2) of the ablation area were measured by two independent pathologists.

Statistical analysis

Following normal distribution, the measurement data were presented as the mean ± standard deviation (SD), and measurement data with a nonnormal distribution were presented as the median (inter-quartile range, IQR). For use between groups, independent sample t-tests conducted comparisons of normally distributed data, while one-way ANOVA was employed for comparison of multiple groups. Furthermore, for comparison between the two groups, nonnormally distributed data were analyzed by the Mann-Whitney U test, while the Kruskal-Wallis test was harnessed for comparison of multiple groups. Pearson correlation analysis was used to analyze the correlation between the maximum area diameter of the MWA zone as measured by US and histology; interclass correlation coefficient (ICC) and Bland-Altman analysis was used to compare the consistency between SWE and the maximum diameter of ablation areas measured in the pathological specimens. To analyze the correlation between L1, L2, and microwave energy, correlation equation was derived through linear regression. Meanwhile, the receiver operating characteristic (ROC) curve was used to evaluate the ultrasonic elasticity in order to derive the evaluation of the ablation pathological type and calculation of the area under the ROC curve (AUC). Finally, statistical analyses and graphing were performed with SPSS (version 27, IBM, Armonk, NY) or GraphPad Prism (GraphPad Software, Inc.). All hypothesis tests were two-sided with a significance level of p < 0.05.

Results

The average weight among thirty ex vivo porcine brains was 101.91 ± 5.4 g, the mean length, 9.03 ± 0.78 cm; the width, 7.04 ± 0.71 cm; and the thickness, 2.12 ± 0.2 cm. Meanwhile, bilateral brain tissues were randomly ablated with variant ablation energy. Finally, sixty ablative foci were obtained after MWA from all ex vivo porcine brains, and satisfactory elastic images were obtained.

Pathology

Pathological manifestations

Following ablation, the experimenter applied a 1 ml syringe perpendicular to the tip of the vertical ablation antenna under the guidance of ultrasound, injecting about 0.03 ml of 0.4% trypan blue at the emission point of the ablation antenna to render the thermocoagulated tissue surrounding the antenna easily distinguished according to structural damage and coloring. The ablation focus profile exhibited an oval-like shape (Figure 1F–H) with the long axis of the antenna insertion channel as its long axis, which was divided into: (1) a brown-yellow area with a dark brown coking area as the center, centered in turn by a darker brown banded carbonized zone; under the light microscope, the cells were destroyed and disintegrated, showing a homogeneous and unstructured brown-yellow color (Figure 1G). (2) Deep staining area: adjacent to the carbonized zone, exhibiting hardness and density, attended by micropores; under the light microscope, brain cells exhibiting coagulated necrosis, brain nucleus dissolved to complete dissipation, cytoplasm exhibiting concentrated red staining, a cell outline unclear, and disintegrated, and cytoplasm showing diffuse light staining and decolorization (Figure 1H). (3) light staining area: under a light microscope, the morphology of brain cells appeared almost normal, exhibiting intercellular edema (Figure 1F–H).

Relationship between gross pathological range and time-intensity

With the increase of ablation intensity and time, the range of brain tissue ablation areas expanded, and the L1 and L2 of the ablation areas increased to varying degrees (Figure 2A and B), converting the combination of ablation intensity and time into energy, thus demonstrating that the ablation range expanded with the increase of microwave energy. The long diameter (L1) of the ablation range along the needle direction was linearly correlated with MWA energy (R² =0.862, p < 0.001), while the diameter (L2) of the ablation range perpendicular to the needle direction was also linearly correlated with MWA energy (R² =0.688, p < 0.001) (Figure 2C and D; Table 1).

Figure 2. Pathology measurements of the ex vivo porcine brain ablation areas.

Note: A, B The range of brain tissue ablation areas increases with the increase of ablation power and time, and the L1 and L2 of the ablation areas increase to varying degrees. C, D Establish a correlation equation through linear regression equation indicate the ablation range expanded with the increase of microwave energy.

Table 1. Ablation range with microwave energy.

Energy (J)	SWEL1 (Mean, cm)	SWEL2 (Mean, cm)	USL1 (Mean, cm)	USL2 (Mean, cm)	L1 (Mean, cm)	L2 (Mean, cm)
300	0.505	0.215	0.835	0.400	0.575	0.230
400	0.600	0.280	1.070	0.510	0.580	0.310
500	0.860	0.375	1.285	0.465	0.840	0.275
600	0.550	0.240	1.070	0.520	0.580	0.220
700	0.615	0.345	1.165	0.675	0.635	0.340
800	0.700	0.275	1.195	0.645	0.750	0.280
900	0.828	0.333	1.288	0.625	0.835	0.325
1000	0.950	0.485	1.255	0.590	0.985	0.425
1200	0.858	0.310	1.330	0.555	0.855	0.320
1400	0.890	0.375	1.540	0.805	0.885	0.365
1500	1.098	0.438	1.330	0.638	1.173	0.365
1600	1.030	0.465	1.695	0.705	1.165	0.380
1800	1.395	0.505	1.823	0.863	1.323	0.395
2000	1.460	0.550	1.803	0.738	1.393	0.453
2100	1.625	0.550	1.735	0.750	1.525	0.500
2400	1.940	0.570	2.060	0.820	1.605	0.535
2500	1.555	0.500	1.750	0.700	1.535	0.515
2700	1.885	0.820	2.035	0.895	1.635	0.550
3000	1.695	0.695	2.250	0.860	1.880	0.645

Ultrasonic findings

Conventional ultrasound

In the process of monitoring the MWA of ex vivo porcine brain by ultrasound, the researchers discovered a substantial echo gasification area around the central ablation antenna, which was oval, with a range that gradually expanded with the ablation time, diffusing and damping the ablation focus. With the cessation of ablation, the gasification range would reach the maximum, obscuring the edge of the ablation focus, while the rear would be accompanied by ill-defined sound shadow. Following the cessation of ablation, the gasification range would gradually decrease, accompanied by a patchy high echo in the focus, affected by undissipated microbubbles. The strong echo in the gasification portion of ablation focus area gradually disappeared, the center gradually exhibited low echo, and the periphery showed a annular strong echo. Finally, the strong echoes caused by gasification virtually disappeared after about 0.5–1 h.

Using pathological results as the gold standard, USL1 was positively correlated with L1 (r = 0.879(95% CI: 0.805 − 0.927), p < 0.001; Figure 2A), while USL2 was positively correlated with L2 (r = 0.524(95% CI: 0.309 − 0.688), p < 0.001; Figure 2B). The larger the gasification range following ablation, the larger the actual ablation range. Nonetheless, the gasification areas L1 and L2 following ablation proved larger than the actual ablation range and elastic measurement range, the difference being statistically significant (Figure 2E and F; p < 0.001).

SWE

Prior to ablation, ultrasound elastography revealed that the areas of interest in the ex vivo brain tissue were uniform blue, exhibiting no obvious boundary, with the mean SWE-value coming to about 9.88 ± 2.64 kPa. After ablation, ultrasonic SWE could be color-coded according to the difference in tissue hardness. As viewed by the naked eye, three quasi-oval areas with the antenna as the center had formed around the insertion channel of the ablation antenna, while the center presented a more uniform red area, and its surrounding width proved uniform and yellow. The outermost circle displays a clear boundary between lake blue and the surrounding blue normal tissue (Figure 1C). In vitro, in brain tissue at the same level and region, the SWE-value (mean, 23.6 ± 13.75 kPa) after ablation proved significantly higher than that of pre-ablation (Figure 4A, p < 0.001).

Using the ablation antenna insertion channel as a marker, the tissue elasticity values of the coking area, were recorded along with the measurements of the coagulative necrosis area, and the congestive edema area around the antenna insertion channel. Accordingly, the mean SWE-value in the pathological coking area was 48.53 ± 9.34 kPa; in the pathological coagulation necrosis area, the mean was 30.14 ± 7.01 kPa, and in the edema area, the mean was about 15.19 ± 6.03 kPa. The area under the ROC curve (AUC) using the SWE-value to identify normal tissue and ablation focus came to 0.86 (Figure 4B). According to the curve, the best cutoff value for the diagnosis of the ablation group was 13.7 kPa with the Youden index being 65.25%. The sensitivity of SWE with respect to determining ablation focus was 71.55%; the specificity came to 93.69%, and the identification accuracy could reach 84.72%.

Furthermore, SWEL1 was positively correlated with L1 (r = 0.962(95% CI: 0.937 − 0.977), p < 0.001; Figure 3C), while SWEL2 was positively correlated with L2 (r = 0.833(95% CI: 0.733 − 0.898), p < 0.001; Figure 3D), with no statistical difference revealed between the ablation area SWEL1 and SWEL2 and the actual ablation range (p > 0.05; Figure 3E and F). Referring to pathological results, with the ablation antenna as the center, SWE and histopathology exhibited positive consistency in term of measuring the maximum diameter of the microwave ablation areas [ICC_SWEL1-L1=0.975(95%CI:0.959 − 0.985), ICC_SWEL2-L2=0.887(95%CI:0.779 − 0.938)]. On average, SWEL1 measures proved 0.02 cm higher (95% confidence interval [CI]: −0.014 − 0.052) than those obtained with pathology L1. The upper limit of agreement (LOA) was 0.266 cm (95% CI: 0.210–0.323), and the lower LOA was −0.228 cm (95% CI: −0.285 to −0.172) (Figure 4C). SWEL2 measures were 0.03 cm higher (95% CI: 0.014 − 0.051) than those obtained with pathology L2, and the upper LOA was 0.175 cm (95% CI: 0.142 to −0.207), The lower LOA was −0.110 cm (95% CI: −0.142 to −0.077) (Figure 4D).

Figure 3. Comparison of ablation area measured by US, SWE and pathology.

Note: A, B. Correlation between USL1, USL2 and pathological results, USL1 was positively correlated with L1 (r = 0.879, P < 0.001), while USL2 was positively correlated with L2 (r = 0.524, P < 0.001); C, D. Correlation between SWEL1, SWEL2 and pathological results, SWEL1 was positively correlated with L1 (r = 0.962, P < 0.001), while SWEL2 was positively correlated with L2 (r = 0.833, P < 0.001). E, F. Comparison of ablation area measured by US, SWE, and pathology by Kruskal-Wallis test (****, P < 0.001);

Figure 4. SWE-value and measurements of the ex vivo porcine brain ablation areas.

Note: A: Comparison SWE-value of pre-MWA and post-MWA by Mann-Whitney U test (****, P < 0.001). B: Receiver operating characteristic curves demonstrate ability to use SWE to identify ablation area after MWA. AUC, area under receiver operating characteristic curve = 0.86; C,D. Bland-Altman analysis were used to compare the consistency of SWE and pathological specimens in measuring the maximum diameter of ablation area.

Discussion

MWA is included in the guidelines for the treatment variety of types of organ tumors. It features rapid heating rate, a high intertumoral temperature, a complete coagulative necrosis, a robust anti-heat settling effect, resistance to eschar and dryness, and a near absence of impact up on peripheral blood flow. In addition, MWA features less damage to the surrounding tissues [29], which provides a new possibility for the minimally invasive treatment of brain tumors. The correct judgment of the ablation range of the target area during and immediately after the operation proves one of the key factors in achieving safe and effective ablation. An accurate and reliable method for achieving near real-time imaging evaluation during ablation becomes a vital link in determining the quality of treatment to prevent injury from insufficient or excessive treatment [30]. In this study, ultrasound monitoring may be employed to monitor hyperechoic lesions formed by gases released from heated brain tissue during MWA, although it proves positively correlated with the pathological range of ablative foci. Nonetheless, this does not represent the actual pathological range of tissue coagulation necrosis following ablation. Typically, hyperechoic lesions overestimate the size of the ablation area [31], and with time, gradually dissipate while the ablation boundary gradually blurs [32]. In fact, within 1 h after ablation, it disappeared completely [33]. This renders measurement of the ablation range, particularly the upper and lower diameter of the ablation range, problematic and prone to error.

Previous studies have shown that the gas mass of the ablation zone has less effect on the elasticity and SWE is possible to judge the situation of the region in real time after the ablation in the short term [34]. This may be due to the fact that the velocity change of shear waves during the ablation process is highly dependent on temperature [35], while the size of bubbles during the ablation process is small [36] and the amount of microbubbles significantly decreases within 2–5 min after ablation [37]. Therefore, the impact on the propagation and detection of shear waves is relatively small, which is not enough to interfere with shear waves [36]. So in this experiment, SWE was used to quantitatively evaluate the degree of thermal damage to brain tissue after ablationfor the first time. Based on the physical properties of SWE, the Young’s modulus values of normal brain tissue measured in this study are basically consistent with previous studies [24]. After MWA, the Young’s modulus value of brain tissue in the ablated area measured by SWE was significantly higher than prior to ablation, and the ablation range measured by SWE following ablation provides a good correlation with the pathological range. In this study, the value of Young’s modulus of tissues in different pathological states was further analyzed by the ROC curve, and the threshold for distinguishing normal tissue from tissue after ablation came to 13.7 kPa. The accuracy of evaluating ablation tissue using SWE able to reach 84.72%, providing us with a new approach and data support for real-time quantitative observation of the ablation range. Consequently, the accurate online visualization of the tissue ablation process and real-time evaluation of the organizational change range can be realized by means of real-time monitoring in SWE.

This study still possesses several limitations. First, this is an in vitro study, and MWA may produce dissimilar effects on living tissue. Due to the influence of blood flow in living tissue, the temperature and hardness of ablation area in living tissue are not identical to those of ex vivo tissue. Previous studies have demonstrated that the volume of damaged tissue produced by MWA proves smaller in living tissue than in vitro tissue [38]. This could potentially also produce an impact on SWE imaging. As such, the next step which is also ongoing requires in vivo investigation in small animal and clinical intraoperative studies to validate the findings of this study. In addition, although this study studies reviews in detail the effectiveness and accuracy of SWE in the combination of common clinical ablation power and duration, the sample size remains small, and the next step is to consider a range situation that may be encountered in clinical application, including the combination of low, medium, high intensity and longer ablation duration, as well as elastic changes at different time periods after ablation, to provide more data support for clinical practice.

Conclusion

This study has the distinction of being the first to employ SWE to monitor the MWA process of ex vivo porcine brain tissue. It was found that, as compared with traditional US, SWE is able, in real-time, to delineate the boundaries of the brain tissue ablation area, to quantitatively analyze tissue hardness, and produce a strong correlation between the elastic imaging of the ablation area and pathological results. The usage of SWE to evaluate the thermal damage range of MWA shows excellent promise as a reliable new approach, providing data support for accurate, real-time visualization of clinical brain tumor MWA treatment.

Author contributions

R.L. wrote the manuscript; W.Z. and W.H. designed the research; R.L., HB.L. and YK.Z. performed the research; R.L., S.Z. and HX.Z. analyzed the data; LG.C. and LJ.D. contributed analytical tools.

Disclosure statement

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

Data availability

The raw data used to support the findings of this study are available from the corresponding author upon request.

Additional information

Funding

This study was funded by the State Key Program of National Natural Science of China (No. 82271995), Training Fund for Open Projects at Clinical Institutes and Departments of Capital Medical University (No.CCMU2023ZKYXY021), and Capital Medical University Research and Cultivation Fund (No. PYZ22123).