In vivo evaluation of focused ultrasound ablation surgery (FUAS)-induced coagulation using echo amplitudes of the therapeutic focused ultrasound transducer

Abstract

Objective

Monitoring sensitivity of sonography in focused ultrasound ablation surgery (FUAS) is limited (no hyperechoes in ∼50% of successful coagulation in uterine fibroids). A more accurate and sensitive approach is required.

Method

The echo amplitudes of the focused ultrasound (FUS) transducer in a testing mode (short pulse duration and low power) were found to correlate with the ex vivo coagulation. To further evaluate its coagulation prediction capabilities, in vivo experiments were carried out. The liver, kidney, and leg muscles of three adult goats were treated using clinical FUAS settings, and the echo amplitude of the FUS transducer and grayscale in sonography before and after FUAS were collected. On day 7, animals were sacrificed humanely, and the treated tissues were dissected to expose the lesion. Echo amplitude changes and lesion areas were analyzed statistically, as were the coagulation prediction metrics.

Results

The echo amplitude changes of the FUS transducer correlate well with the lesion areas in the liver (R = 0.682). Its prediction in accuracy (94.4% vs. 50%), sensitivity (92.9% vs. 35.7%), and negative prediction (80% vs. 30.8%) is better than sonography, but similar in specificity (80% vs. 100%) and positive prediction (100% vs. 100%). In addition, the correlation between tissue depth and the lesion area is not good (|R| < 0.2). Prediction performances in kidney and leg muscles are similar.

Conclusion

The FUS echo amplitudes are sensitive to the tissue properties and their changes after FUAS. They are sensitive and reliable in evaluating and predicting FUAS outcomes.

Keywords:

• Focused ultrasound ablation surgery FUAS
• coagulative necrosis
• echo amplitude
• focused ultrasound (FUS) transducer
• sonography
• coagulation prediction

Introduction

The treatment of cancer is an important issue in both developing and developed countries. In the clinical practice of thermal ablation of cancer and solid tumors, ultrasound provides several benefits, such as enabling deeper tissue treatment, improved focus on the target tissue through its small wavelengths, precise control over the shape and location of energy deposition, and non-ionization. Over the past two decades, focused ultrasound ablation surgery (FUAS) has been developed into an effective and noninvasive ablation modality for soft tissues to treat thousands of patients worldwide [1–3]. The key to FUAS is that there is sufficient energy deposition in tissue to produce the temperature elevation to a cytotoxic level in a short time (i.e., >65 °C for a few seconds). Recent advances in FUAS have increased its popularity. Some promising results were achieved in managing various malignancies, including pancreas, prostate, liver, kidney, breast, and bone [4]. Other applications include brain tumor ablation and suppression of essential tremors [5].

To ensure the safety and effectiveness of clinical FUAS applications, imaging guidance plays an essential role in evaluating treatment outcomes [6]. Currently, single-modality imaging guidance, either ultrasound imaging or MR imaging, is used in clinics. B-mode sonography enables real-time anatomic imaging during FUAS, which is especially advantageous for moving organs. In contrast, respiration-induced liver motion may lead to noise in MR images [7]. In sonography guidance, FUAS-induced ablation is clearly shown in the B-mode ultrasound imaging using the integrated ultrasound probe, and the dimensions measured in the sonography were correlated to those measured in the pathology [8]. The changes to the grayscale level are usually applied to evaluate the ablation. According to clinical experience, the appearance of hyperechoes, with grayscale changes of more than 10, is a good indicator for the successful production of coagulative necrosis [9]. However, the generation of necrosis does not always correlate well with the appearance of hyperechoes [10]. It is reported that there were 66.7% of solitary uterine fibroids with the hyperechoic feature in a retrospective study from April 2015 to April 2019 [11]. In another multi-center retrospective study of 215 cases of FUAS for unresectable pancreatic cancers, the corresponding value was 41.9% (90 of 215 cases) [12]. Such a low possibility of hyperechoes significantly affects coagulation prediction and monitoring. Therefore, the investigation of accurate and reliable FUAS monitoring technology has great value.

Recent advances have already evidenced several new approaches. Target-specific techniques were developed for FUAS treatment planning using pretreatment imaging as well as controlling the irradiation and monitoring tissue lesions in 2D and 3D ultrasound imaging [13]. A tissue change monitoring (TCM) algorithm that quantitatively analyzes the backscatter signals pre- and immediately post-FUAS for each individual ablative site has been developed on the Sonablate^® device for estimating the temperatures in the treated prostate. Its performance has been validated in two clinical studies [14]. Pulse inversion ultrasonic imaging was proposed to eliminate the ultrasound interference and clearly show the echogenicity change by FUAS in the excised bovine liver, which is not affected by the organ motion in the abdomen due to respiration [15]. Acoustic radiation force impulse (ARFI) imaging could visualize thermally- and chemically-induced lesions in soft tissues. Agreement was found between lesion size in ARFI images and results from pathology [16]. Amplitude-modulated (AM) harmonic motion imaging (HMI) had the capability of following the protein-denatured lesion formation based on the variation of the HMI displacements, which is feasible for real-time monitoring of temperature-related stiffness changes in tissue [17]. Nakagami imaging is found to accurately evaluate the FUAS-induced thermal lesions in both axial and lateral directions with low computation complexity [18,19]. The entropy imaging has great potential for differentiating lesions in FUAS since its contrast-to-speckle ratio (CSR) is 3.4 times higher than that of B-mode sonography [20]. However, their performance in comparison to hyperecho changes in sonography has not been carried out to evaluate their translation potential for clinical application.

A commercial FUAS system was modified slightly to work in a testing mode, and echo amplitudes of the FUS transducer were collected for quantitative analysis. According to our previous ex vivo experiments in porcine tenderloin and bovine livers, it was found that echo amplitude and its changes could serve as monitoring parameters for successful coagulation. Its prediction metrics are higher in accuracy, sensitivity, and negative prediction but similar in specificity and positive prediction as conventional sonography under various conditions [21]. To further evaluate its performance and confirm its superiority to sonography monitoring at each treatment spot, in vivo experiments were carried out in this study. Three adult goats were used as the animal model, and their liver, kidney, and leg muscles underwent FUAS at various irradiation settings. Echos received by the FUS transducer and ultrasound imaging probe before and after FUAS were used for statistical analysis. Treated targets were dissected on day 7, and the produced lesion areas were recorded photographically and determined quantitatively. Coagulation prediction metrics (i.e., accuracy, sensitivity, specificity, positive rate, and negative rate) using these two methods were compared. This investigation may provide an alternative and effective approach for monitoring the ablation progress.

Materials and methods

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Chongqing Haifu Hospital (approval no. 2022-011, Chongqing, China), and all protocols are in accordance with the Principles of the Ethical Care and Use of Animals.

Animal model

Three healthy adult goats were used in our experiment at the age of 16.02 ± 4.33 months and a weight of 27.54 ± 1.88 kg and fed with standard nutrients. Fasting was performed 24 h before the experiment. Anesthesia was carried out through both abdominal and vein injections. 20% chloral hydrate (Huaxiang Biological Technologies, Wuhan, China) at a volume of 1–1.5 ml/kg was injected into the abdomen first, and then a vein injection of chloral hydrate was applied at a concentration of 1–2 ml/kg to maintain the anesthesia. If necessary, an injection of 1–2 ml of midazolam (Jiuxu Pharmaceutics, Zhejiang, China) at a concentration of 1 mg/mL was administered to the veins for calmness. After the anesthesia was confirmed, the hairs on the acoustic wave entry windows were removed by a clipper and razor, and the residual ones were removed using an 8% sodium sulfide solution (Jiuchong Chemicals, Shandong, China). Degreasing and degassing of the skin in the acoustic window were performed using 75% ethyl alcohol (Chinese Chemical Solution, Beijing, China) and a vacuum pump (Tengyuan, Taizhou, China) at a pressure of −87 kPa, respectively. After that, the goat was fixed to a custom-designed frame sitting on a Focused Ultrasound Tumor Therapeutic System (Model-JC200, outer diameter of 22 cm, inner diameter of 8 cm, focal length of 16.5 cm, central frequency of 0.97 MHz, Chongqing Haifu Medical Technologies Co., Ltd., Chongqing, China), and its position was adjusted for the optimal acoustic windows to the targets to be observed in the integrated B-mode sonography (Esonate, S.P.A., Genoa, Italy) as shown in Figure 1. The excitation frequency of sonography is 3.5 MHz, and the focus depth is adjusted according to the location of ablated targets.

Figure 1. Photo of experimental setup of the in vivo FUAS on a live goat.

FUAS and echo collection

In vivo experiments were carried out using the previously established protocols [22]. Firstly, the integrated B-mode sonography was applied to select the appropriate acoustic window by patiently adjusting the goat’s position. The degassed water was circulated inside the water cushion to maintain a gas concentration of < 1 mg/L and a temperature of 25 ± 2 °C. Secondly, in the testing mode, the emitted pulses were set to ten cycles at a driving voltage of 60 V and a pulse repetition frequency of 10 Hz. Because of the technical limitations of the power amplifier used in the system, the shortest pulse duration for the stable output is ten cycles. The received echos from the focal region were quantified via a digital oscilloscope (Picoscope 5000, Pico Technology, Cambridgeshire, UK) with a bandwidth of 100 MHz but without any post-processing. Five echos were collected for statistical analysis of their amplitudes. Thirdly, three targets (i.e., liver, kidney, and leg muscle) of the goat underwent FUAS. The therapeutic settings were similar to those in the clinical applications listed in Tables 1, 2, and 3, respectively. The acoustic power is set at 400 W because that is the pain threshold for most patients, according to our clinical experience. When treating the leg, fresh and degassed bovine livers obtained from a local slaughterhouse were put on the skin surface to increase the penetration depth of ultrasonic waves. Fourthly, after ultrasound ablation at each treatment spot, the echos of the FUS transducer were collected for the quantitative analysis. Grayscale changes in sonography in the region-of-interest (ROI) before and after FUAS were analyzed using GrayVal 1.0 software (Chongqing MicroSea Software Development Co., Ltd., Chongqing, China). According to our previous clinical experiences, a grayscale increase of 10 was set as the threshold for successful coagulative necrosis generation [23–26]. At each treatment spot, the focus position was fixed for both treatment and monitoring. After the FUAS session on all targets, the goat was sent back to the feeding room until spontaneous recovery, and its diet, daily activities, and vital signals were continuously monitored. Finally, seven days after the FUAS, the animal was sacrificed humanely using our established protocols [27]. The goat abdomen and leg were opened to resect the treated targets, which were cut axially for the maximum coagulative necrosis area. HIFU Measure Software (Chongqing MicroSea Software Development Co., Ltd., Chongqing, China) was applied to segment the lesion boundaries and calculate their areas quantitatively.

Table 1. Comparison of echo amplitude changes of the FUS transducer and gray levels in sonography and FUAS-induced lesion areas in the liver of goats.

Goat	Treatment spot	Acoustic window	Depth (mm)	Acoustic power (W)	Irradiation settings (m s on and n s off for k repeats)	Echo amplitude of the FUS transducer			Hyperecho	Lesion area (mm^2)
						Pre-operation (mV)	Post-operation (mV)	Changes (mV)
#1	1	between rib 2 and 3	62	400	m = 2, n = 2, k = 30	1070 ± 34	2203 ± 55	1133 ± 24	no	180 ± 2.01
	2		30	400	m = 2, n = 2, k = 30	329 ± 11	725 ± 23	396 ± 15	no	87 ± 1.37
	3		50	400	m = 2, n = 2, k = 30	1328 ± 46	3600 ± 112	2272 ± 86	yes	216 ± 2.26
	4	between rib 3 and 4	70	400	m = 2, n = 2, k = 30	730 ± 18	980 ± 13	250 ± 11	no	80 ± 1.54
	5		55	400	m = 2, n = 2, k = 30	850 ± 31	1389 ± 44	539 ± 25	no	84 ± 1.79
	6		35	400	m = 2, n = 2, k = 30	1427 ± 27	4250 ± 174	2823 ± 174	yes	237 ± 2.41
#2	1	between rib 2 and 3	50	400	m = 2, n = 2, k = 20	460 ± 25	646 ± 17	186 ± 18	no	na
	2		50	400	m = 2, n = 2, k = 20	784 ± 33	960 ± 21	176 ± 22	no	9.12 ± 0.97
	3		32	400	m = 2, n = 2, k = 20	820 ± 19	1025 ± 14	205 ± 15	no	14.32 ± 1.12
	4	between rib 3 and 4	57	400	m = 2, n = 2, k = 20	421 ± 8	432 ± 17	11 ± 4	no	na
	5		53	400	m = 2, n = 2, k = 20	498 ± 22	780 ± 11	282 ± 21	no	10.40 ± 0.65
	6		50	400	m = 2, n = 2, k = 20	978 ± 13	2478 ± 44	1500 ± 27	yes	18.33 ± 1.20
#3	1	between rib 1 and 2	45	400	m = 2, n = 2, k = 30	975 ± 18	1575 ± 22	590 ± 14	no	43.55 ± 2.12
	2		30	400	m = 2, n = 2, k = 30	1228 ± 25	3430 ± 112	2202 ± 107	yes	106.75 ± 2.43
	3	between rib 2 and 3	60	400	m = 2, n = 2, k = 30	880 ± 28	981 ± 24	101 ± 14	no	na
	4		40	400	m = 2, n = 2, k = 30	1044 ± 11	1577 ± 19	533 ± 18	no	22.36 ± 1.75
	5	between rib 3 and 4	54	400	m = 2, n = 2, k = 30	550 ± 14	730 ± 26	180 ± 16	no	na
	6		40	400	m = 2, n = 2, k = 30	1214 ± 21	3637 ± 107	2423 ± 96	yes	23.48 ± 2.22

Table 2. Comparison of echo amplitude changes of the FUS transducer and gray levels in sonography and FUAS-induced lesion areas in the kidney of goats.

Goat	Treatment spot	Depth (mm)	Acoustic power (W)	Irradiation settings (m s on and n s off for k repeats)	Echo amplitude of the FUS transducer			Hyperecho	Lesion area (mm^2)
					Pre-operation (mV)	Post-operation (mV)	Changes (mV)
#1	1	54	400	m = 2, n = 2, k = 20	1675 ± 45	2636 ± 56	961 ± 31	no	16.51 ± 1.32
	2	47	400	m = 2, n = 2, k = 20	794 ± 22	876 ± 19	82 ± 13	no	na
	3	50	400	m = 2, n = 2, k = 20	1402 ± 26	1421 ± 25	19 ± 7	no	na
#2	1	66	400	m = 2, n = 2, k = 20	965 ± 15	1383 ± 18	428 ± 17	no	12.12 ± 1.79
	2*	46	400	m = 2, n = 2, k = 20	608 ± 32	770 ± 26	162 ± 16	no	na
	3	36	400	m = 2, n = 2, k = 20	1801 ± 20	2945 ± 79	1144 ± 69	yes	54.45 ± 1.33
#3	1	53	400	m = 2, n = 2, k = 20	900 ± 14	950 ± 29	50 ± 18	no	na
	2	40	400	m = 2, n = 2, k = 20	923 ± 22	936 ± 27	13 ± 4	no	na
	3*	28	400	m = 2, n = 2, k = 20	1345 ± 31	1540 ± 17	195 ± 16	no	na

*: no coagulative necrosis, but apparent hemorrhage.

Table 3. Comparison of echo amplitude changes of the FUS transducer and gray levels in sonography and FUAS-induced lesion areas in the leg muscle of goats.

Goat leg	Treatment spot	Depth (mm)	Acoustic power (W)	Irradiation settings (m s on and n s off for k repeats)	Echo amplitude of the FUS transducer			Hyperecho	Lesion area (mm^2)
					Pre-operation (mV)	Post-operation (mV)	Changes (mV)
#1 right	1	48	300	m = 2, n = 2, k = 1	1956 ± 128	4796 ± 149	3200 ± 121	yes	265.72 ± 3.13
	2	45	300	m = 2, n = 2, k = 1	1807 ± 41	5943 ± 109	4136 ± 98	yes	217.90 ± 2.45
	3	52	300	m = 2, n = 2, k = 1	1644 ± 73	5243 ± 117	3599 ± 104	yes	156.40 ± 3.20
#2 left	1	52	250	m = 2, n = 2, k = 5	1991 ± 11	2522 ± 25	1031 ± 23	no	23.12 ± 1.11
	2	52	250	m = 2, n = 2, k = 5	2493 ± 42	3524 ± 39	531 ± 34	no	38.81 ± 1.34
	3	52	250	m = 2, n = 2, k = 5	1803 ± 94	2921 ± 169	1118 ± 95	no	28.22 ± 1.42
#2 right	1	91	400	m = 2, n = 2, k = 10	1202 ± 15	3697 ± 199	2495 ± 166	yes	44.19 ± 0.78
	2	88	300	m = 2, n = 2, k = 20	1261 ± 63	4861 ± 169	3600 ± 117	yes	145.30 ± 1.22
	3	83	250	m = 2, n = 2, k = 30	1502 ± 46	2990 ± 78	1488 ± 48	no	64.28 ± 1.56
#2 left	1	80	250	m = 2, n = 2, k = 30	2168 ± 110	3446 ± 193	1278 ± 143	yes	122.89 ± 2.45
	2	90	250	m = 2, n = 2, k = 30	1358 ± 76	2282 ± 82	924 ± 68	no	88.19 ± 2.10
	3	100	250	m = 2, n = 2, k = 30	1135 ± 42	2544 ± 71	1409 ± 44	no	72.44 ± 0.98
#3 right	1	100	300	m = 1, n = 2, k = 30	1067 ± 56	1699 ± 35	632 ± 38	no	114.30 ± 0.99
	2	92	300	m = 1, n = 2, k = 30	1349 ± 77	2920 ± 131	1517 ± 101	no	140.08 ± 1.67
	3	88	400	m = 1, n = 2, k = 30	1849 ± 47	2422 ± 92	573 ± 65	no	168.77 ± 2.31
#3 left	1	96	400	m = 1, n = 2, k = 20	1653 ± 53	3641 ± 180	1998 ± 154	yes	88.21 ± 0.67
	2*	90	400	m = 1, n = 2, k = 20	1856 ± 30	3859 ± 197	2003 ± 146	no	na

*: this data are excluded for analysis because of the significant motion.

Prediction performance

Prediction performance was evaluated using the following five metrics: (1) $$\mathrm{\text{accuracy}}=\frac{\text{accurate predictions}}{\text{total samples}}=\frac{\mathrm{\text{TP}}+\mathrm{\text{TN}}}{\mathrm{P}+\mathrm{N}}$$(1) (2) $$\mathrm{\text{sensitivity}}=\frac{\text{true positive}}{\text{true positive}+\text{false negative}}=\frac{\mathrm{\text{TP}}}{\mathrm{\text{TP}}+\mathrm{\text{FN}}}$$(2) (3) $$\mathrm{\text{specificity}}=\frac{\text{true negative}}{\text{true negative}+\text{false positive}}=\frac{\mathrm{\text{TN}}}{\mathrm{\text{TN}}+\mathrm{\text{FP}}}$$(3) (4) $$\text{positive predictive value}=\frac{\text{true postive}}{\text{true positive}+\text{false positive}}=\frac{\mathrm{\text{TP}}}{\mathrm{\text{TP}}+\mathrm{\text{FP}}}$$(4) (5) $$\text{negative predictive value}=\frac{\text{true negative}}{\text{true negative}+\text{false negative}}=\frac{\mathrm{\text{TN}}}{\mathrm{\text{TN}}+\mathrm{\text{FN}}}$$(5) where TP is the true positive, FP is the false positive, FN is the false negative, and TN is the true negative.

Statistical analysis

All experimental data were analyzed using SPSS 21.0 (IBM, Armonk, NY), and whether their distribution fits with the normal function was tested using a one-sample Kolmogorov-Smirnov test. The normal distribution data are presented as mean ± standard deviation, while the non-normal distribution data are in the format of median (quartile). A t-test was performed when comparing those groups with a normal distribution, while Mann-Whitney testing was performed for those with a non-normal distribution. p < 0.05 is considered statistically significant. The Pearson product moment was calculated for the correlation coefficient.

Results

Liver ablation

Because the liver lies behind the ribs, the FUS pulses need to propagate through the rib spaces for effective energy deposition. Table 1 lists the experimental conditions, FUAS parameters, echo amplitude changes, and therapeutic outcomes (i.e., lesion production). It is found that the echo amplitude changes of the FUS transducer correlate moderately with the coagulative necrosis areas (R = 0.682). In some cases, there are no grayscale changes in sonography but successful generation of coagulative necrosis in the liver (i.e., treatment spots of 1, 2, 4, and 5 in the #1 goat, see Figure 2b). Hyperechoes only show up at large lesion sizes (i.e., > 200 mm²). If the threshold of the FUS echo changes is set at 200 mV, its coagulation predictions are as follows: accuracy of 94.4%, sensitivity of 92.9%, specificity of 80%, positive predictive value of 100%, and negative predictive value of 80%. Such a threshold was determined from the Youden index with the maximum Kolmogorov-Smirnov (KS) distance of receiver operating characteristics (ROC) curve [21]. The corresponding prediction metrics based on sonography are 50%, 35.7%, 100%, 100%, and 30.8%, respectively. It is clear that the echo changes of the FUS transducer are more accurate and sensitive.

Figure 2. Comparison of sonography (a) before and (b) after FUAS to the liver of the #1 live goat at six different positions, and (c) the corresponding coagulative necrosis after the dissection of treated samples.

Kidney ablation

Every goat underwent FUAS in the kidney in three different positions. The experimental results of three goats are listed in Table 2. Three spots showed coagulative necrosis, but not in the remaining six spots, of which two had hemorrhage (see Figure 3). There is only one case of hyperechoes in sonography whose echo amplitude change of the FUS transducer is 1144 ± 69 mV. For the other two treatment spots with coagulative necrosis production but without the hyperechoes in sonography, the echo amplitude changes are 961 ± 31 mV and 28 ± 17 mV. The corresponding changes to the spots with hemorrhage are 162 ± 16 mV and 195 ± 16 mV. The others without hemorrhage and necrosis all have echo amplitude changes less than 100 mV. Coagulation prediction using echo amplitude changes has higher accuracy (100% vs. 77.8%), sensitivity (100% vs. 33.3%), and negative predictive value (100% vs. 75%) than using hyperechoes in sonography but the same specificity and positive predictive values (100%). The correlation coefficient between echo amplitude changes and lesion areas is 0.894.

Figure 3. Damage caused by FUAS in the kidney tissues of the #2 live goat (yellow arrow: coagulation necrosis, red arrow: congestion).

Leg muscle ablation

Representative coagulative lesions produced in the leg muscles of the #1 goat are shown in Figure 4. In order to increase the total attenuation in the acoustic wave pathway, a piece of ex vivo bovine liver at a thickness of about 4 cm was put in front of the legs of #2 and #3 goats. All experimental results are summarized and listed in Table 3. When treating the second spot in the left leg of the #3 goat, that goat regained consciousness with significant motions. No coagulative necrosis was found on that spot since the energy deposition is not constrained in a small region, but the echo amplitude changes are 2003 ± 146 mV, which may be due to the echos from the strong reflective interface. Thus, that data was excluded. The accuracy, sensitivity, and positive predictive value of echo amplitude changes of the FUS transducer and hyperechoes in sonography are 100% vs. 43.8%, 100% vs. 43.8%, and 100% vs. 100%, respectively. The correlation coefficient between echo amplitude changes and lesion areas is 0.594.

Figure 4. Coagulative necrosis in the leg muscles of the #1 live goat at six different positions after FUAS.

Discussion

Monitoring FUAS outcomes using sonography is one of the most popular approaches, but it has the shortcoming of limited sensitivity and accuracy. In this study, a new method based on the echo amplitude changes of the FUS transducer was proposed and evaluated in in vivo experiments in the goat’s liver, kidney, and leg muscle. Because of the piezoelectricity of the ultrasound transducer, it can work in both the therapeutic mode for high power output and the testing mode at lower power and short pulses as well. Thus, our proposed approach is not limited to the JC200 system used in this study. Preliminary results show that our proposed approach can accurately and noninvasively evaluate the total attenuation in the acoustic wave propagation pathway. If the total acoustic attenuation through the tissue layers is high, the echo amplitude is low with small coagulative necrosis. It is found that our approach is very sensitive to the production of coagulative necrosis (i.e., sensitivity of 92.9%, 100%, and 100% in liver, kidney, and leg muscle, respectively) with significant echo amplitude changes at the focus. The application of this approach may provide an alternative to sonography for monitoring FUAS progress. Although the number of animals involved in this study is limited, the promising results encourage us to pursue clinical trials in the near future. It is noted that the optimization of sonographic settings for different targets may change the ultrasound image quality but not the coagulation monitoring capabilities. Only a few modifications to the existing commercial system and working protocols may allow its quick translation into clinics. However, it will not replace sonography because echo information cannot provide anatomical structures for straightforward guidance. It is expected that multiple modalities, such as harmonic motion imaging (HMI) [28,29], manifold learning algorithm for backscattered RF signals [30], and backscattered energy (CBE) imaging [31], may work synergistically for FUAS monitoring in the future. In this study, the animals were sacrificed on day 7 to evaluate the coagulation outcomes. Although tissue changes, such as edema, rehydration, infiltration, and remodeling, may occur after the FUAS for the consequent changes to the shape and texture of the produced lesions, there was a sharp boundary between the HIFU necrosis and viable tissue for easy visualization, according to our experience [27]. A narrow region with inflammatory cell infiltration, consisting primarily of lymphocytes and monocytes sometimes with a small number of eosinocytes, was seen between the necrotic and normal zones.

The ribcage in the wave propagation pathway to the abdominal target will block the acoustic energy significantly and lead to much less energy deposition at the focus and the increased difficulty of producing coagulative necrosis [32,33]. The coagulation rate of 40% in the goat liver with the ribcage blockage is much less than that of 65% with the surgical removal of partial ribs [22]. Meanwhile, the energy required to ablate a unit volume was much higher (435.4 ± 201.8 J/mm³ vs. 76.8 ± 34.2 J/mm³). Because of the similar attenuation effects on sonography, its monitoring sensitivity and specificity also decrease [34]. Altogether, the FUAS of large hepatocellular carcinoma (HCC) partially obscured by the ribcage presents a challenge. However, there are certainly some available windows. Kennedy et al. adjusted the sound beam size and transmission location to allow as much acoustic energy as possible to penetrate through the intercostal space [35]. But it’s hard to quantitatively determine the quality of the acoustic wave pathway in sonography. So the shortest tissue penetration depth is applied as one of the criteria in selecting the acoustic window in practice [36]. According to our in vivo experimental results here, such a strategy is not a perfect solution. For example, when treating the liver through the 2^nd and 3^rd ribs, the 2^nd treatment spot in the #1 goat with the shortest penetration depth of 30 mm produced a lesion area of 87 mm², which is not the largest among all under the same irradiation parameters (Table 1). Our analysis of all experimental data shows that the correlation coefficient between penetration depth and lesion area is only −0.122, −0.14, and −0.194 for liver, kidney, and leg muscles, respectively. Therefore, a quantitative evaluation of the acoustic quality of the wave propagation pathway is important to make appropriate treatment plans. In our previous study, a boiling bubble was initially produced at the focus and then worked as a strong reflector. The total acoustic attenuation with the focusing angle in tissue was quantitatively determined using the echos, which provides an alternative for evaluating the acoustic window for clinical FUAS [3].

The echo amplitude of the FUS transducer is dependent on not only the acoustic attenuation in the wave propagation pathway but also the acoustic properties of the tissue at the focus. If the focus is close to a strong reflective interface, the echo amplitude is usually quite high. Thus, when using the echo amplitude approach to select the appropriate acoustic window, strong acoustic interfaces (i.e., gastric mural thickening, bone, and perinephric fat) [37] that can be determined in sonography should be avoided. In addition, a sufficient margin should be set between the ablation region and organ boundaries or vital tissues (i.e., large vessels and nerves) [38]. The influence of reflected waves from the interface or boundary on the FUAS outcomes should be investigated in future work.

If there are hyperechoes in sonography during FUAS, echo amplitude changes of the FUS transducer before and after the ablation provide a new metric for evaluating the tissue’s physiological variations. When hyperechoes occur in sonography, the corresponding echo amplitude changes and post-operative echoes’ standard deviations of the FUS transducer are significantly larger than those cases without hyperechoes (p < 0.001), as listed in Table 4, which may be due to the bubble cavitation in the focal region [10]. If there are no hyperechoes in sonography, the mechanism of FUAS is only the thermal effect. Physiological changes to the tissue in the focal region lead to a variation in the echo amplitude (see Figure 5). Because of the high acoustic intensity at the focus, bubble cavitation works together with the thermal effect to increase energy deposition and tissue damage [39,40]. Since the acoustic impedance of cavitation bubbles is much lower than that of surrounding tissue, strong reflection, scattering, and hyperechoes are produced [9]. Unstable bubble dynamics, especially those of inertial cavitation, may be the major reason for significant variations in the post-operative echo amplitudes. It is noted that our proposed monitoring approach is to evaluate the lesion production at each treatment spot, not the distribution of all lesions in the target, which is more preferred in evaluating the FUAS outcomes (i.e., contrast perfusion in contrast-enhanced sonography or MRI) but is usually infeasible for each spot because of the dose limit of the contrast agent in clinical use. Scanning the transducer’s focus over an area or a volume to estimate the lesion distribution may be an alternative solution and needs further investigation.

Figure 5. Comparison of sonograhy of treatment point ⑤ in the leg muscles in the #3 live goat (a) before and (b) after FUAS, and (c) the corresponding echo signals from the focal region.

Table 4. FUS echo difference of treatment points with/without grayscale changes in B-mode sonography.

organ	goat	echo amplitude changes (mV)		p value	post-operative echos’ standard deviation (mV)		p value
		with hyperechos	without hyperechos		with hyperechos	without hyperechos
liver	#1	2547 ± 390	468(360)	< 0.001	143 ± 44	33 ± 19	<0.001
	#2	1500 ± 27	186(176)	< 0.001	44 ± 12	16 ± 4	na
	#3	2313 ± 156	357(160)	< 0.001	109 ± 4	23 ± 3	<0.001
kidney	#1	1144 ± 69	122(42)	< 0.001	79 ± 22	27 ± 13	na
leg	#1	3525 ± 651	893 ± 317	< 0.001	125 ± 21	39(32)	<0.001
	#2	2458 ± 1161	1273 ± 305	< 0.001	187 ± 16	77 ± 6	<0.001
	#3	1988 ± 154	925 ± 560	< 0.001	180 ± 35	86 ± 48	na

Our animal experiment found that the target motion has a significant influence on the FUS echos. Increased frequency and range of the organ motion lead to the spread of the absorbed acoustic energy throughout a larger region and, subsequently, reduced temperature elevation for an insufficient thermal dose to achieve coagulation, especially in a highly perfused organ (i.e., the liver) [41]. Although the echo amplitudes and their changes are large, which may be due to the presence of an acoustic interface at the motion, no necrosis was found in the leg muscle. Respiration-induced intra-abdominal organ motion is significant in the liver and kidney and results in incorrect tumor targeting. Previous FUAS on the liver had required apnea during general anesthesia. However, techniques suitable for free breathing are desired in order to make the procedure less invasive. Mapping a signal from an external respiratory bellow to treatment locations within the liver allows the ultrasound transducer to be steered in real time, whose ablation performance is not significantly different from that with breath-holding [42]. A robotized FUAS system compensated the motion estimated by using a fast ultrasonic speckle tracking algorithm and reduced it by more than 80% using the ultrasonic visual servoing module [43,44].

Conclusion

The coagulation prediction performance by using the echo amplitudes of the FUS transducer has been preliminary evaluated in the in vivo experiments. The liver, kidney, and leg muscles of three adult goats were treated by FUAS in clinical settings. Echos of the FUS transducer and sonography were collected before and after the irradiation and then analyzed to compare their performance. It has been found that the echo amplitude changes of the FUS transducer are much more significant than those of sonography. With an appropriate threshold, the coagulation prediction of our approach outperforms that of sonography in accuracy, sensitivity, and positive prediction but works similarly in specificity and negative prediction. Coagulative lesion areas are found to correlate moderately with the echo amplitude changes of the FUS transducer, but not the tissue depth. Altogether, echos of the FUS transducer reflect the properties of tissue at the focus and its changes better than sonography, especially at high tissue depths and complicated anatomical structures. This approach has great clinical potential for selecting an appropriate acoustic window and evaluating the therapeutic outcomes in order to enhance the safety and effectiveness of FUAS.

Ethical approval

This study was approved by the Institutional Animal Care and Use Committee (IACUC) of Chongqing Haifu Hospital (approval no. 2022-011, Chongqing, China).

Acknowledgments

The authors thank Dr. Tianfeng Zhang for experimental work.

Disclosure statement

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

Data availability statement

The data presented in this study are available on request from the corresponding author.

Additional information

Funding

This work was supported by the Future Innovative Research Fund of Chongqing Medical University (2022-W0063).