Evaluation of integrated color-coded perfusion analysis for contrast-enhanced ultrasound (CEUS) after percutaneous interventions for malignant liver lesions: First results

Abstract

BACKGROUND:

With the rising number of percutaneous ablation therapies in malignant liver lesions there is a need of reliable diagnostics after the intervention to differentiate between reactive changes and tumor.

PURPOSE:

To assess the success of percutaneous ablation therapies for malignant liver lesions using CEUS with perfusion analysis.

MATERIAL AND METHODS:

Retrospective analysis of perfusion analysis for 67 patients with 94 malignant liver lesions, treated with ablation therapies. The lesions were 70 hepatocellular carcinomas (HCC), 18 metastases, 4 cholangiocellular carcinomas (CCC), 2 lesions remained unclear. CEUS was performed after bolus injection of 1.6–2.4 ml of sulfur-hexafluoride microbubbles. The perfusion analysis was calculated using Peak, TTP, mTT and AUC with integrated software during the late arterial to early portal-venous phase for approximately 9 sec (5–15 sec). For the evaluation of the success after percutaneous treatment the perfusion results were compared to the follow-up control after 6 months with CT and MRI and CEUS.

RESULTS:

Perfusion analyses after percutaneous treatment of malignant liver lesions showed highly significant perfusion differences when comparing the center to the surrounding tissue and the margins (p<0.0001) for Peak and also for AUC. 62 lesions were successfully treated, meaning there was no local recurrence after 6 months. In cases of residual tumor CEUS showed a nodular marginal enhancement, the corresponding perfusion analyses showed nodular red and yellow pseudo-color shades.

CONCLUSIONS:

Using CEUS and perfusion analysis, a critical analysis of post-ablation defects in malignant liver lesions is possible. With the help of pseudo-colors, remaining tumor-vascularization can be detected.

Keywords

Ultrasound ablation procedures liver perfusion

1 Introduction

After the FDA approval, contrast enhanced ultrasound (CEUS) gains more and more importance especially in liver diagnostics [1]. CEUS shows a high diagnostic accuracy for characterization of liver lesions [2 –5]. Even lesions <10 mm can be classified [6]. The importance of CEUS with perfusion analysis for liver tumors has been shown in previous studies [7]. Not only in diagnosis of liver tumors but also during and as follow-up after percutaneous interventions, CEUS plays an important role [8 –11]. Percutaneous ablation therapies include radiofrequency ablation (RFA), microwave ablation (MWA) or irreversible electroporation. The aim of interventional procedures is a near to total devascularization of the tumor with sufficient safety margins (at least 5 mm). All ablation methods have in common that there are hyperemia and reactive changes 24 hrs after the intervention [12, 13]. Consequently it is difficult to differentiate between vital tumor parts, which appear more nodular, and reactive changes, when only the arterial phase is taken into consideration and there is no proper wash- in and outkinetics.

CEUS and perfusion imaging is a safe method for depicting the tumor’s mirco-vascularization and is even able to differentiate between different tumor entities [14, 15]. With an external perfusion tool, DICOM loops up to 1 minute can be analyzed with special software on a second computer. However, a disadvantage of this external tool is the high expenditure of time for data transfer and data analysis. Color-coded perfusion analysis software is more and more integrated in high-end ultrasound machines. This software can evaluate flow and volume parameters like TTP, RT, mTT, and Peak. Without the need of data transfer this data can be analyzed in a timely manner. The pseudo-colors used by both, integrated and external perfusion software, emphasize the effect of local perfusion differences [16].

The aim of this study was the analysis of the success after ablation therapy for malignant liver lesions with integrated perfusion software in the late arterial to early portal-venous phase.

2 Material and methods

This is a retrospective analysis of 67 patients after ablation therapy for 95 malignant liver lesions. There was an approval of the local ethical board for the retrospective reading. The patients gave their written informed consent for the i.v. contrast-media injection. There were no intolerances towards the i.v. ultrasound contrast media.

2.1 Methods imaging

Each patient received a 24 hrs post-interventional ultrasound performed on a high-end ultrasound machine (Accuson S3000, Siemens, Erlangen, Germany) using a 1–6 MHz convex abdominal probe. At first, the liver was scanned in B-Mode and CCDS (color-coded duplex sonography) of the portal vein, the hepatic artery and the hepatic veins. The whole liver was scanned for the localization of the post-ablation defect. After the intervention the local status of the ablation defect was assessed and abscesses were ruled out. Dual Mode was used with B-Mode and contrast harmonic imaging. The examinations were carried out with contrast harmonic imaging (CHI) and a low mechanical index (MI < 0.2) in regards to the EFSUMB guidelines for liver and non-hepatic indications [3, 9].

CEUS was performed after bolus injection of 1.6–2.4 ml of sulfur-hexafluoride microbubbles (SonoVue^®, Bracco) followed by a 10 ml bolus of NaCl via a cubital vein. All CEUS examinations were carried out by one experienced examiner (>3000 examinations/year; >10 years). 2.4 ml is the recommended dose of SonoVue ^® . But since in modern machines sufficient enhancement in liver tissue can be achieved also with lower doses (in particular in hypervasculated lesions) we decided to adapt the volume of injection according to the body weight. Patients >90 kg received 2.4 ml of SonoVue^®, patients from 70 to 90 kg 1.5–2.4 ml and patients <70 ml received 1.0–1.5 ml i.v. contrast media. To avoid repeated injections of i.v. ultrasound contrast-media the immediate (24 hrs after the ablation) follow-up examination was focused on the treated area for 1 minute in terms of suspicious remaining vascularization. With sweep technique the whole liver was scanned until the late phase (5 min after the injection). Single images were stored to exclude a late wash-out that hints toward tumor manifestation. The period of arterial enhancement is relatively short (only about 5–10 sec) so that only selected areas of the liver can be examined, while the duration of the portal and late phase is much longer (several minutes) allowing sweeps covering the whole liver tissue. Cine loops were stored digitally from the late arterial to the early portal- venous phase for 5 to 15 sec (mean 9 sec). The acquired loops were read with integrated perfusion software. Motion artifacts were balanced out by the integrated software itself after having placed the regions of interest (ROI) in the center of the ablation area, the margins and the surrounding tissue. After defining the regions, the perfusion parameters were calculated and are shown in pseudo-colors. Red and yellow meaning high perfusion, blue and green showing reduced perfusion. Time to Peak (TTP), Area under the Curve (AUC), Peak and mean Transit Time (mTT) were calculated. Those values were also documented in PACS. The results were also stored digitally in PACS. Besides the graphic display with TIC analyses the amount of perfusion was also shown with color-coded imaging.

For control after the intervention every patient received a post-interventional contrast-enhanced CT (ceCT) of the liver in arterial phase (25–35 sec) and portal-venous phase (70–90 sec) with bolus injection of iodic contrast-agent (100–130 ml Accupaque^TM, GE) (Siemens Definition Flash/Siemens Sensation 16, Siemens Healthcare, Erlangen Germany; collimation 5 mm with coronary and axial reconstructions) to rule out post-interventional complications 24 hrs after the intervention.

Furthermore a post-interventional contrast-enhanced MRI (ceMRI) was performed 1 day after the intervention (Skyra 3T, Siemens Healthcare, Erlangen, Germany), using T1/T2 sequences, diffusion imaging with ADC as well as contrast-enhanced sequences after bolus injection of 8–15 ml liver specific contrast agent (Gd-EOB-DTPA), (Primovist ^® , Bayer, Schering Pharma AG, Germany) using 3D vibe sequences from arterial phase (20–25 sec) up to late phase. After 3 and 6 months every patient received a follow-up MRI using the same protocol. If there was evidence of remaining tumor parts the intervention was classified unsuccessful. If there was no remaining tumor vascularization after 6 months, the intervention was classified as successful.

2.2 Method interventional treatments

All interventional treatments were carried out by an experienced interventional radiologist. The indication for tumor ablation was determined by an interdisciplinary tumor board. Exclusion criteria were resectability of the tumor or extrahepatic tumor manifestations (i.e. lymph node metastases or osseous metastases). The tumors had histopathologically been proven, or had clearly been classified as one tumor entity in 3 different imaging modalities (ceCT/ceMRI/CEUS).

RFA (Radiofrequency ablation): Necrosis is caused by heat generated from medium frequency alternating current. RFA has an advantage of predictable ablation zones. All RFA ablations were performed under general anesthesia using 3–5 cm electrodes (StarBurst Talon, AngioDynamics, Latham, NY).

MWA (Microwave ablation): Necrosis is caused by rapid oscillation of water molecules. MWA has an advantage of fast ablation times at the cost of difficultly predictable ablation zones. All MWA ablations were performed under general anesthesia using 2.45 GHz single applicators (Acculis MTA, AngioDynamics, Latham, NY).

IRE (irreversible electroporation): Is a non-thermal alternative to RFA and MWA. 2–6 IRE electrodes are placed parallel in and around the tumor area to deliver electrical pulses and provoke cell death by destroying the homeostasis of the cell membrane. In contrast to thermal methods, IRE protects heat-sensitive structures and also allows the ablation of tumors in the immediate vicinity of large vascular structures. All IRE ablations were performed under general anesthesia using the NanoKnife System (AngioDynamics, Latham, NY).

2.3 Statistics

All statistical analyses were performed with IBM SPSS Statistics (version 23, Chicago, IL, USA) and R 3.2.1. The data are presented as median, 25% quartile (Q1), 75% quartile (Q3) and range. Non-parametric Mann-Whitney U-test for independent variables and the Wilcoxon signed-rank test were used for dependent variables for comparisons between the groups. ROC analyses were performed to differentiate between the patient groups, and the optimal cut-offs were estimated according to the Youden indices. The estimates of the AUCs that corresponded to the 95% confidence intervals and the true classification rates are reported. All tests were two-sided, and values of p < 0.05 indicateda significant difference.

3 Results

There were 56 male patients and 11 female patients between 43–84 years (mean 62.9 years). The 67 patients in this study were treated with irreversible electroporation IRE in 18 cases, microwave ablation (MWA) in 75 cases and 1 radiofrequency ablation (RFA). The tumors were localized in segment I in 3 cases, in segment II in 10 cases, segment III in 4 case, segment IV in 13 cases, segment V in 12 cases, segment VI in 9 cases, segment VII in 21 cases and segment VIII in 22 cases. 62 lesions were successfully treated. The lesions were hepatocellular carcinomas (HCC) in 70 cases, 18 metastases (14 by colorectal cancer, 3 by breast cancer and 1 by pancreas cancer), and 4 cholangiocellular carcinomas (CCC). The lesions size ranged between 4–80 mm (mean 21.9 mm). The remained unclear tumors had been histopathologically proven in 44 patients and have clearly been identified with at least 2/3 imaging modalities (ceCT, ceMRI and CEUS) in all cases. 9 patients suffered from complications after the treatment. 1 patient showed a thrombosis of the portal vein, 1 an abscess, 1 cholecystitis,1 hematoma and 5 suffered from biliomas.

24 hrs after the interventional treatment of malignant liver lesions there were significant differences for the median Peak when comparing the center (3.1 rU) to the surroundings (11.9 rU) and the center to the marigns (13.0 rU) of the lesion (p < 0.0001). There was no significant difference for Peak when comparing the margins to the surroundings (p = 0.82).

For AUC there was also a significant difference (p < 0.0001) when comparing the center (8.9 rU) to the surroundings (37.4 rU) and the center to the margins (34.7 rU). However when comparing the margins to the surroundings there was no significant difference (p = 0.46).

For the relative perfusion parameters (TTP and mTT) there were no significant differences for any of the regions of interest (Table 1).

Table 1
Evaluation of the main perfusion parameters Peak (in relative Units, rU), Time to Peak (TTP, in sec), Area under the curve (AUC, in rU) and mean Transit Time (mTT, in sec) for the different regions of interest in the center, the surroundings, and the margins after interventional ablation therapy in malignant liver lesions

Center Margins Surroundings

Q1 Q2 Q3 Range Q1 Q2 Q3 Range Q1 Q2 Q3 Range

Peak 2.2 3.1 5.0 0.8–64.0 7.7 13.0 21.6 0.7–54.0 9.6 11.9 17.0 1.9–44.9

TTP 1.2 1.9 2.8 0.1–30.6 1.2 1.9 2.7 0.1–21.2 1.4 1.9 2.8 0.44–15.8

AUC 4.9 8.9 19.8 1.3–5251 18.7 34.7 73.3 2.9–1555 22.6 37.4 72.7 2.9–1628

mTT 2.1 3.0 4.2 1.1–60.9 1.9 2.9 3.9 0.8–28.1 2.1 2.9 4.2 1.0–27.6

	Center	Margins	Surroundings
Peak	2.2	3.1	5.0	0.8–64.0	7.7	13.0	21.6	0.7–54.0	9.6	11.9	17.0	1.9–44.9
TTP	1.2	1.9	2.8	0.1–30.6	1.2	1.9	2.7	0.1–21.2	1.4	1.9	2.8	0.44–15.8
AUC	4.9	8.9	19.8	1.3–5251	18.7	34.7	73.3	2.9–1555	22.6	37.4	72.7	2.9–1628
mTT	2.1	3.0	4.2	1.1–60.9	1.9	2.9	3.9	0.8–28.1	2.1	2.9	4.2	1.0–27.6

P values were calculated for the comparison between center and surroundings, center and margins as well as margins and surroundings. Lower quartile (Q1), median (Q2), upper quartile (Q3) and range are displayed.

After stratification for successfully and non-successfully treated lesions there were only significant differences when comparing the surroudings to the margins for relative Peak enhancement (p = 0.04). For AUC there were no significant differences at all. However, there was a tendency that relative AUC was higher in non-successfully treated lesions (Fig. 1).

Fig.1

When comparing the relative enhancement for Peak and AUC, there was only a significant difference when looking at the comparison for success vs. no success surroundings towards margins. For AUC there were no significant differences at all. M = Margins, S = Surroundings, C = Center, RE = relative Enhancement.

In cases of successful ablation, the defect appears in dark blue shades (Fig. 2). CEUS criteria for residual tumor in the ablation area are nodular irregular enhancement in the arterial phase with wash-out beginning in the portal-venous phase (60–90 sec) up to the late phase (3–5 min). In cases of reactive hyperemia there was a marginal hyper-enhancement without wash-out. Hyperemia is shown in pseudo-colors with a red/yellow margin, in comparison to a blue ablation defect (Fig. 3). In cases of IRE the hyperemia might also be found inside the ablation area.

In cases of residual tumor the changes of pseudo-colors perfusion evaluation by PEAK, AUC und mTT showed nodular changes in all cases (Fig. 4). For TTP there were no differences for hyperemia and residual tumor during the portal venous and late portal venous phase. During the possible short phase for tumor perfusion and post-ablation perfusion imaging up to 10 sec there were no significant differences for mTT, TTP and AUC or PEAK between residual tumor and hyperemia during the late arterial to early portal-venous phase.

Fig.2

The perfusion screen is divided into four quadrants. Up left the original CEUS images (left) and the original B-Mode image in the divided screen mode. Up right the perfusion images in pseudo-colors. Down right the numeric values of Peak (Spitze). Down left the duration of the cine loop. After ablation of a HCC lesion the ablation defect appears avascular. There is no remaining vascularization adjacent to the defect. No nodular or non-nodular hyperemia is visualized. The defect in the right liver lobe appears dark blue. The surrounding cirrhotic liver appears irregular in yellow, red and green shades. In B-Mode the ablation defect can not be fully visualized.

Fig.3

After ablation of a HCC in the right liver there is a more linear surrounding hyperemia, in the follow- up no remaining tumor was found so that this is more a reactive hyperemia. The ablation defect again appears in dark blue.

Fig.4

After ablation of a HCC lesion there is a nodular hyperemia at the right margin showing remaining vital tumor parts, which showed progressive growth during the follow-up. Dynamic CEUS in the early portal-venous phase (loop about 5 sec).

4 Discussion

Contrast– enhanced ultrasound (CEUS) allows a dynamic and outstanding assessment of capillary micro-vascularization. Before the intervention CEUS is able to characterize nodules in cirrhotic livers [17]. Furthermore, CEUS can detect (remaining) tumor vascularization after ablation of malignant liver lesions [18]. This is a requirement for evaluation of the success after ablation therapy in malignant liver lesions. It is vital to establish tools that are able to differentiate between reactive hyperemia and remaining tumor vascularization. This can be done with the help of dynamic DICOM loops for CEUS and pseudo-colors similar to perfusion analysis in CT and MRI. Pseudo-colors improve the assessment of post-interventional success. The expressiveness of those pseudo-color images can be reinforced with the help of perfusion analyses. With the help of integrated perfusion software, different phases of wash-in and wash-out kinetics can be evaluated with color-coded images and with curve analyses. Especially after defining Regions of Interest, parameters like TTP, Peak, mTT and AUC can be calculated. This is especially utile when assessing the success after percutaneous ablation treatments of malignant liver lesions. Using pseudo-colors, an increased microcirculation is displayed in shades of red and is represented by high values of AUC and Peak. Low tissue perfusion is coded in blue with corresponding low values for Peak and AUC.

Table 2

P values were calculated for the direct comparison between center and surroundings, center and margins as well as margins and surroundings for all perfusion parameters

	Center vs Surroundings	Center vs Margins	Margins vs Surroundings
	p	p	p
Peak	<0.0001	<0.0001	0.82
TTP	0.90	0.69	0.46
AUC	<0.0001	<0.0001	0.46
mTT	0.93	0.68	0.74

TTP and mTT are influenced by the contrast-media kinetics, and can also be influenced by the injection speed [19], since mTT corresponds to the average time that erythrocytes need in a determinate volume of capillary circulation. In this study there were no significant differences for mTT and TTP. Therefore it would be useful in follow-up studies to use an injection pump so the injection rate is standardized.

For the post-interventional changes of the center and the margins for RFA, MWA and IRE it must be remembered that only in RFA a track ablation is performed. By contrast, severe bleeding in the ablation area can be observed for MWA and IRE due to the missing ablation of the track. This is how apparently higher perfusion values can be measured. IRE spares larger penetrating vessels even in the former tumor area [20]. MWA leads to focused defects around the needle tip, an inhomogeneous defects at the margins. The perfusion imaging can be used for all ablation methods, however, defining the ablation areas is difficult and can only be performed manually. The goal of the interventions isa near to total devascularization.

This study emphasizes the role of CEUS for post-interventional control and follow-up. It can help to depict and evaluate the extent of devascularization, necrosis and vital tumor parts. A hypervascularization at the margins is a hint for reactive hyperemia or caused by inflammation after RFA [21, 22]. The criteria concerning remaining tumor have been discussed before [9, 23], which is irregular local nodular enhancement in yellow and red shades with wash-out.

Due to the rising number of ablative interventions, there is the need for a broad application to assess post-interventional results and the potential of CEUS. We consider Peak and AUC the most relevant perfusion parameters. Since only those parameters showed significant differences between the center and the surroundings/margins and are not so strongly influenced by the injection speed. AUC has already proven to be a powerful tool for follow-up in the French multicenter study [16].

Integrated perfusion analysis is not yet widely spread. The most critical point is that the perfusion analysis is limited to a 5–10 sec cine loop for this machine, so there is no proper dynamic wash-in/wash-out kinetics sine the loops are recorded during the late arterial to early portal-venous phase and only single images were recorded to the late phase, whereas external perfusion software is able to analyze loops up to 1 minute.

Another limitation is that the regions of interest are placed manually and visually, and are consecutively depending on the experience of the examiner. To avoid an inter-observer-bias, both the placing of the regions of interest and the examination itself were performed by one person each. However, compared to external perfusion software, the integrated perfusion tool is time-saving, since the DICOM loops do not have to be transferred to a second computer.

Successful application of perfusion analysis have among other things already shown the possibility of characterization of liver lesions [7], and the differentiation between HCC and CCC lesion [15]. Hence more studies are to be expected.

In the recent literature the diagnostic value of CEUS for the evaluation of post ablation defects has been proven to be high [1]. However an experienced examiner is needed to differentiate between post-interventional reactive changes and residual tumor. This pilot study aimed to show in how far color-coded perfusion analyses with new registration of Peak, TTP, mTT and AUC from the late arterial to the early portal-venous phase can help evaluate the success of treatment during follow-up. Ideally the recording of wash-in and wash-out kinetics should be performed between the arterial phase (15 sec) to the late phase (5 min). However, for this long data acquisition usually external perfusion software with data export is needed. In this study it was only possible to record loops between 5–10 sec (mean 9 sec). Even then the perfusion analysis takes between 7–10 minutes per patient. Even during this limited time areas with irregular increased micro-vascularization at the ablations margins can be displayed in pseudo-colors (red/yellow) whereas avascular areas appear blue. These underlying results show that the combination of loops, single images, and perfusion analyses are useful for the evaluation of post-interventional ablation areas.

References

Dietrich

, Lorentzen

, Appelbaum

, et al. EFSUMB Guidelines on Interventional Ultrasound (INVUS), Part III - Abdominal Treatment Procedures (Short Version). Ultraschall in der Medizin (Stuttgart, Germany: 1980). 2016;37(1):27–45.

Sporea

, Badea

, Popescu

, et al. Contrast-enhanced ultrasound (CEUS) for the evaluation of focal liver lesions - a prospective multicenter study of its usefulness in clinical practice. Ultraschall in der Medizin (Stuttgart, Germany: 1980). 2014;35(3):259–66.

Claudon

, Dietrich

, Choi

, et al. Guidelines and good clinical practice recommendations for Contrast Enhanced Ultrasound (CEUS) in the liver - update A WFUMB-EFSUMB initiative in cooperation with representatives of AFSUMB, AIUM, ASUM, FLAUS and ICUS. Ultrasound in medicine & biology. 2013;39(2):187–210.

Abraham-Nordling

, Oistamo

, Josephson

, Hjern

, Blomqvist

. The value of preoperative computed tomography combined with ultrasound in the investigation of small indeterminate liver lesions in patients with colorectal cancer. Acta radiologica (Stockholm, Sweden: 1987). 2017:284185117693461.

Rahbin

, Siosteen

, Elvin

, et al. Detection and characterization of focal liver lesions with contrast-enhanced ultrasonography in patients with hepatitis C-induced liver cirrhosis. Acta radiologica (Stockholm, Sweden: 1987). 2008;49(3):251–7.

Strobel

, Bernatik

, Blank

, et al. Diagnostic accuracy of CEUS in the differential diagnosis of small (< /=20mm) and subcentimetric (< /=10mm) focal liver lesions in comparison with histology. Results of the DEGUM multicenter trial. Ultraschall in der Medizin (Stuttgart, Germany. 2011;32(6):593–7.

Beyer

, Pregler

, Wiesinger

, Stroszczynski

, Wiggermann

, Jung

. Continuous dynamic registration of microvascularization of liver tumors with contrast-enhanced ultrasound. Radiology research and practice. 2014;2014:347416.

Pregler

, Beyer

, Wiesinger

, et al. Microwave ablation of large HCC lesions: Added value of CEUS examinations for ablation success control. Clinical hemorheology and microcirculation. 2016;64(3):483–90.

Solbiati

, Ierace

, Tonolini

, Cova

. Guidance and monitoring of radiofrequency liver tumor ablation with contrast-enhanced ultrasound. European Journal of Radiology. 2004;51(Suppl):S19–23.

10.

Dietrich

, Averkiou

, Correas

, Lassau

, Leen

, Piscaglia

. An EFSUMB introduction into Dynamic Contrast-Enhanced Ultrasound (DCE-US) for quantification of tumour perfusion. Ultraschall in der Medizin (Stuttgart, Germany: 1980). 2012;33(4):344–51.

11.

Clevert

, D’Anastasi

, Jung

. Contrast-enhanced ultrasound and microcirculation: Efficiency through dynamics–current developments. Clinical Hemorheology and Microcirculation. 2013;53(1-2):171–86.

12.

Cosgrove

, Piscaglia

, Bamber

, et al. EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography. Part Clinical applications. Ultraschall in der Medizin (Stuttgart, Germany: 1980). 2013;34(3):238–53.

13.

Gillams

, Goldberg

, Ahmed

, et al. Thermal ablation of colorectal liver metastases: A position paper by an international panel of ablation experts, The Interventional Oncology Sans Frontieres meeting 2013. European Radiology. 2015;25(12):3438–54.

14.

Han

, Liu

, Han

, et al. The Degree of Contrast Washout on Contrast-Enhanced Ultrasound in Distinguishing Intrahepatic Cholangiocarcinoma from Hepatocellular Carcinoma. Ultrasound in Medicine & Biology. 2015;41(12):3088–95.

15.

Wildner

, Pfeifer

, Goertz

, et al. Dynamic contrast-enhanced ultrasound (DCE-US) for the characterization of hepatocellular carcinoma and cholangiocellular carcinoma. Ultraschall in der Medizin (Stuttgart, Germany: 1980). 2014;35(6):522–7.

16.

Lassau

, Bonastre

, Kind

, et al. Validation of dynamic contrast-enhanced ultrasound in predicting outcomes of antiangiogenic therapy for solid tumors: The French multicenter support for innovative and expensive techniques study. Investigative Radiology. 2014;49(12):794–800.

17.

Kim

, Jang

. Contrast-enhanced ultrasound in the diagnosis of nodules in liver cirrhosis. World Journal of Gastroenterology. 2014;20(13):3590–6.

18.

Gao

, Zheng

, Cui

, Ma

, Liu

, Zhang

. Predictive value of quantitative contrast-enhanced ultrasound in hepatocellular carcinoma recurrence after ablation. World Journal of Gastroenterology. 2015;21(36):10418–26.

19.

Tang

, Mulvana

, Gauthier

, et al. Quantitative contrast-enhanced ultrasound imaging: A review of sources of variability. Interface Focus. 2011;1(4):520–39.

20.

Dollinger

, Muller-Wille

, Zeman

, et al. Irreversible electroporation of malignant hepatic tumors–alterations in venous structures at subacute follow-up and evolution at mid-term follow-up. PLoS One. 2015;10(8):e0135773.

21.

Sainani

, Gervais

, Mueller

, Arellano

. Imaging after percutaneous radiofrequency ablation of hepatic tumors: Part 1, Normal findings. AJR American Journal of Roentgenology. 2013;200(1):184–93.

22.