Color coded perfusion analysis and microcirculation imaging with contrast enhanced ultrasound (CEUS) for post-interventional success control following thermal ablative techniques of primary and secondary liver malignancies

AIM: Evaluation of the post-interventional success following ablative techniques (radiofrequency and microwave) using a new color coded perfusion quantification software with CEUS in patients with primary and secondary liver malignancies. MATERIAL AND METHODS: 75 patients (60 males, 15 females, age 24–84 years, mean 62.7 years) with 128 malignant liver lesions were included in this study. Between 01/2013 and 06/2018, the therapeutic interventional procedure in 88 lesions was MWA, in 40 lesions RFA. All patients underwent CEUS using a convex multifrequency probe (1–6 MHz) following application of 1–2.4 ml sulphur hexafluoride microbubbles, before and within 24 hours following RFA and MWA to detect residual tumor tissue. Postprocessing of the stored DICOM loops from 15 sec up to 1 min using a perfusion quantification software regarding peak enhancement (pE), time to peak (TTP), mean transit time (MTT), rise time (Ri) and Wash-in area under the curve (WiAUC) in the center of the lesion, the border area and periphery was performed. RESULTS: In patients treated with RFA, pE differences between center of the lesion vs. surrounding liver were found to be statistically extremely significant (p < 0.001), differences between center of the lesion and margin were also statistically significant (p < 0.01). mTT, TTP, WiAuC and Ri showed no significant difference between center, margin or surrounding liver. In patients treated with MWA, statistically significant differences (p < 0.05) were found for pE, Ri and mTT regarding the differences between center of lesion and surrounding tissue. WiAuC and TTP showed no significant differences between center, margin or surrounding liver. CONCLUSION: CEUS with perfusion imaging is a valuable supporting tool for post-interventional success control following RFA and MWA of primary and secondary liver maligancies. Focus should be placed upon pE following MWA and pE, Ri and mTT following RFA.

Keywords: Microwave ablation (MWA); radiofrequency ablation (RFA); contrast enhanced ultrasound (CEUS); perfusion analysis; primary and secondary liver malignancies

Due to vast improvements regarding efficacy, safety and wound recovery in recent years, thermal ablation procedures have become increasingly established in the clinical routine for the percutaneous treatment of non-resectable malignant liver lesions. In addition to microwave ablation (MWA), radiofrequency ablation (RFA) is the most commonly used technique and, according to current practice guidelines in the management of hepatocellular carcinoma (HCC), is now recommended as the standard of care for HCC patients in Barcelona Clinic Liver Cancer (BCLC) stage 0-A who are not suitable for surgery [[1]].

Percutaneous RFA is based on generation of an electric current (375 to 500 KHz) through an electrode tip inserted into the liver lesion that induces a local heat effect, in order to reach a temperature of from 60 to 100°C, necessary to obtain coagulation necrosis. The heat spreads in centrifugal direction from the energy source and decreases when blood flow is present [[4]]. Complete ablation rates in HCC less than 5 cm are reached in >95% of cases, the rate of local tumor recurrence ranges from 10 to 30% [[6]].

MWA is another thermal technique that creates an electromagnetic field around an electrode, thus inducing homogeneous heating and coagulation necrosis with a complete loss of microcirculation. The heat effect of MWA is faster, also, it reaches a higher temperature than RFA, and consequently has the potential advantage of simultaneously treating more lesions in a shorter time. Recent results show complete ablation rates of 95 to 100%, with local recurrence varying from 10 to 13% at 3 years [[9]].

However, there are various method-inherent limitations. Some hepatic tumors cannot be safely and effectively ablated thermally, due to their close proximity to large vascular structures [[11]]. Due to the so-called heat-sink effect, the heat is diverted by large vessels or biliary structures, thus leading to an incomplete ablation [[12]]. Also, the adjacency to heat sensitive neighboring organs (colon, gall bladder, heart) represents a relative contraindication for thermal ablation due to the risk of thermal damage [[13]].

Both ablative techniques are usually guided either by ultrasound (US) or computed tomography (CT). Due to its availability, convenience, low cost and real-time capability US is the most commonly used instrument in the guidance of RFA. Some studies recommend the use of CT for RFA guidance for its superior edge detection of ablated lesions, immediate coagulation evaluation and lesser artifacts [[14]]. However, disadvantages include radiation exposure, extended procedure time, possible contrast-induced nephropathy, and higher costs.

The aim of this study was to evaluate the post-interventional success following RFA and microwave ablation using a new color coded perfusion quantification software with CEUS in patients with HCC.

From January 2013 until June 2018, 75 patients (60 males, 15 females, age 24–84 years, mean 62.7 years) with 128 malignant liver lesions identified by characteristic imaging features or histopathology were included in this retrospective study. Each patient underwent pre-interventional CEUS and contrast-enhanced computed tomography (CT) or liver specific contrast enhanced magnetic resonance imaging (MRI) for detection and characterization of the liver lesions.

Approval for this study was obtained from the institutional review board in conformity with the Declaration of Helsinki (approval number 15-104-0115) [[15]]. All patients were informed about the procedure and gave their written informed consent. The authors comply with the Ethical Guidelines for Publication in Clinical Hemorheology and Microcirculation [[16]].

Microwave ablation (MWA) was performed in 88 lesions using the Acculis Microwave Tissue Ablation (MWA) System (Angiodynamics®, Latham, NY, USA), which operates at 2.45 GHz with a maximum power output of 140 W. In MWA necrosis is caused by rapid oscillation of water molecules. MWA has an advantage of fast ablation times at the cost of hardly predictable ablation zones. All MWA ablations were performed under general anesthesia using the standard Acculis microwave applicator with a 1.8 mm diameter and a 16 mm active tip.

The therapeutic interventional procedure in 40 lesions was a radiofrequency ablation (RFA). During RFA necrosis is caused by heat generated from medium frequency alternating current. RFA has an advantage of predictable ablation zones but takes longer for ablation (10–25 min) than microwave ablation. All RFA ablations were performed under general anesthesia using 3–5 cm electrodes (StarBurst Talon, AngioDynamics, Latham, NY).

All the interventions were performed under combined CT-fluoroscopic and CEUS-image guidance. In all patients, the treatment was performed under relaxation in intubation anaesthesia and with anaesthesia standby. The indication for local ablative therapy was provided in the scope of an interdisciplinary tumor board review.

Each patient was examined using fundamental B- mode, Color Coded Doppler Sonography (CCDS), Power Doppler and Contrast Enhanced Ultrasound (CEUS). The use of CEUS for this study was approved by the local ethical committee. Before the imaging procedures were conducted, written informed consent was obtained from each patient for MRI, CT and CEUS. Exclusion criteria of this study were contraindications for use of a contrast agent for CT or MRI including impaired renal function (creatinine >1.5 mg/dl, creatinine clearance <30 ml/min) or pre-existing strong allergic reactions for CT, MRI or CEUS. Acute thrombosis of the portal vein, acute inflammations, relevant ascites or decompensated cardiac failure were contraindications for the ablative techniques.

Within 24 hours following RFA/MWA ultrasound was performed by an experienced radiologist (more than 5000 examinations per year) using a multifrequency convex transducer (1–6 MHz, LOGIQ E9, GE Healthcare).

First, the whole liver was examined using B-mode sonography in sweep technique. Color Coded Doppler Sonography (CCDS) and Power Doppler (PD) ultrasound were used to evaluate native vascularization. Flow parameters were adjusted to the lowest possible pulse repetition frequency (PRF < 1000 Hz) and the best possible color imaging without blooming artifacts.

Contrast enhanced ultrasound (CEUS) was performed after bolus injections of 1–2.4 ml of sulphur hexafluoride microbubbles (SonoVue®, BRACCO, Italy) with a low mechanical index (MI < 0.16) applying CEUS with amplitude modulation and pulse inversion harmonic imaging (PIHI) technique [[17]]. The contrast harmonic imaging technique (CHI) uses a contrast-specific detection mode for real-time evaluation of the contrast-agent enhancement. The complete data of the contrast-agent examination was recorded up to 5 min. The liver microcirculation was evaluated continuously from an early arterial phase (beginning 15 sec. after contrast application) until a late parenchymal phase (>5 min.). The transducer was held steadily on the ablation defect for 1 minute to avoid repeated injection of contrast media, afterwards the whole liver was examined in sweep technique to look for wash-out. The first minute was documented as a video clip centering the ablation defect in sweep technique. Afterwards single images were stored until the late phase. All images were digitally stored in PACS.

An irregular enhancement pattern in the periphery of the ablation defect during early arterial phase, combined with a wash-out beginning in the portal venous phase was considered a characteristic criterion for residual HCC tissue. An evenly peripheral rim enhancement excluding wash-out was considered as physiological postablation reaction. A wedge-shaped, homogenous arterial enhancement peripheral to the ablation defect with cumulative enhancing portal-venous branches, but without wash-out was defined as arterio-portal-venous shunt.

The digitally stored ultrasound loops were evaluated using an external color coded perfusion quantification software (VueBox®, Bracco, Italy) [[18]]. Five parameters were calculated for each Region of interest (ROI) which included time to peak (TTP), mean transit time (mTT), peak enhancement (pE), Wash-in Area Under the Curve (WiAUC) and Rise time (Ri) [[19]]. Flow parameters such as regional blood volume and regional blood flow were calculated by the software and exported in a calculation protocol.

Regions of interest were manually placed in the center of the ablation defect, at the margins and the surrounding liver tissue.

CEUS examinations and the perfusion analysis were analyzed by two experienced independent radiologists in consensus. For each modality used, each observer recorded the diagnostic findings. Furthermore, the image quality was documented on a four points scale: 1 - excellent, 2 - minor diagnostic limitations, 3 – major diagnostic limitations, 4 - non-diagnostic.

Imaging modalities were evaluated using the data analysis hard-/software of the ultrasound system (LOGIQ E9, GE).

For calculation of the differences between the center of the lesion and the border area, the border area and the periphery and the center of the lesion and the periphery for each parameter repeated measures ANOVA with Bonferroni post test was performed using GraphPad Prism version 5.00 for Mac OS X, GraphPad Software, San Diego California USA with alpha <0.05 indicating statistically significant differences between groups.

Out of the 128 tumor lesions 96 (55 Patients) showed an enhancement pattern consistent with hepatocellular carcinoma (HCC) including arterial irregular hyperenhancement and portal venous to late washout on CEUS. Also, the pre-interventional imaging techniques showed characteristics compatible with HCC. 30 HCC lesions were treated using RFA, whereas 66 HCC lesions were treated by MWA. The underlying malignancies in the other 32 lesions were metastases caused by colorectal carcinoma in 12 patients (19 lesions), by pancreatic carcinoma in 1 patient (1 lesion), by neuroendocrine carcinoma in 2 patients (5 lesions), by duodenal carcinoma in 1 patient (1 lesion), by prostatic carcinoma in 1 patient (1 lesion), by Thymoma in 1 patient (2 lesions) and by cholangiocellular carcinoma (CCC) in 2 patients (3 lesions).

Out of the 88 lesions (55 patients) treated with MWA, in 73 lesions a complete ablation success could be verified in follow-up up to 6 month. Out of the 40 lesions (24 patients) treated with RFA, a complete ablation success was noted in 34 lesions by follow-up for up to 6 month. A partial ablation success following MWA was confirmed in 15 lesions (11 HCC and 4 metastases) and in 6 lesions (4 HCC and 2 metastases) following RFA. These patients were evaluated for a secondary procedure in the scope of an interdisciplinary tumor board review. No major complications were noted.

In 43 out of 55 patients with HCC in cirrhosis, the etiology was that of an alcoholic liver disease, 8 patients suffered from a chronic hepatitis C and 4 patients from primary biliary cirrhosis. Regarding the assessment of prognosis in liver cirrhosis, the Child-Pugh scores of our patients ranged from a (45 patients), b (8 patients) to c (2 patient). MELD (Model for end-stage liver disease) scores were within the range of 6–23 (6–18 patients, 7–15 patients, 8–7 patients, 9–5 patients, 12–5 patients, 15/16/19/22/23 – one patient each). The Barcelona Clinic Liver Cancer (BCLC) algorithm for HCC staging and treatment included stage 0 (22 patients), stage A (24 patients) and stage B (9 patients). In 35 out of 55 patients, a solitary HCC lesion was treated. In 11 patients 2, in 4 patients 3, in one patient 5, in 3 patients 6 and in one patient 8 HCC lesions were ablated, respectively. If three or more lesions per patient needed to be treated using RFA/MWA, the intervention was carried out over the course of several days.

In 11 patients, only one metastatic tumor lesion was treated using MWA/RFA. In 7 patients, the number of lesions treated was 2 and in 3 patients it was 3. The mean number of lesions treated in all patients was 1,6.

All 75 patients underwent pre-interventional MRI using a liver specific contrast agent (Primovist®, Bayer HealthCare AG, Germany; 0.1 ml/kg body weight) on a 3-T whole - body scanner (Skyra, Siemens Medical Solutions, Germany) and a pre-interventional CT based on a triphasic contrast enhanced protocol using a Dual Source scanner (SOMATOM Definition Flash, Siemens, Germany).

The pre-interventional tumor sizes of the 128 lesions ranged from 9 mm up to 58 mm with a mean size of 27 mm. The post-interventional defects sized from 11 mm to 73 mm with am mean size of 38 mm. In all lesions, a post-interventional reduction of the tumor microvascularization was observed.

In all 75 patients (100%) CEUS was feasible. The image quality in all examinations was excellent or had only minor diagnostic limitations (1–2 SD±0.362).

In patients treated with RFA, pE differences between center of the lesion vs. surrounding liver were found to be statistically extremely significant (p < 0.001), differences between center of the lesion and margin were also statistically significant (p < 0.01). The 95% confidence intervals (CI) of difference were 605.9 to 2697 for center vs. surrounding liver, –861.3 to 1230 for surrounding liver vs. margin and –2513 to –421.6 for center vs. margin. Other parameters including mTT, TTP, WiAuC and Ri showed no significant difference between center, margin or surrounding liver (Table 1, Fig. 1).

Graph: Fig.1 Box plots for all perfusion parameters following RFA. Highly significant differences were only observed for Peak (A) when comparing the center of the lesion (center column) versus surrounding liver (right column) (p < 0.001). Significant differences were also found for the comparison between center of the lesion (center colum) and margins (left colum) for pE (p < 0.01) (A). No significant differences were found regarding Rise, mTT, TTP or WiAUC (B–E).

Table 1 Mean±standard deviation (SD) of pE, WiAuC, mTT, TTP and Ri each using ROIs in the center of the lesion, the margins and the surrounding, acquired within 24 h following RFA

In patients treated with MWA, statistically significant differences (p < 0.05) were found for pE, Ri and mTT regarding the differences between center of lesion and surrounding. The 95% confidence intervals (CI) of difference regarding pE were 352.5 to 5881 for center vs. surrounding liver, –2268 to 3261 for surrounding liver vs. margin and –5384 to 144.3 for center vs. margin. For Ri, the 95% CI of difference were 1.4 to 18.8 for center vs. surrounding liver, –4.7 to 12.7 for surrounding liver vs. margin and –14.8 to 2.6 for center vs. margin. Regarding mTT, the 95% CI of difference was 12.1 to 125.6 for center vs. surrounding liver, –18.2 to 95.4 for surrounding liver vs. margin and –86.9 to 26.6 for center vs. margin. WiAuC and TTP showed no significant differences between center, margin or surrounding liver (Table 2, Fig. 2).

Table 2 Mean±standard deviation (SD) of pE, WiAuC, mTT, TTP and Ri each using ROIs in the center of the lesion, the margins and the surrounding, acquired within 24 h following MWA

Graph: Fig.2 Box plots for all perfusion parameters following MWA. Statistically significant differences were found for Peak (A), Rise (B) and mTT (C) when comparing the center of the lesion (center column) versus surrounding liver (right column) (p < 0.05). No significant differences were found for TTP and WiAUC (D–E).

The perfusion software uses pseudo-colors in order to display vascularization. Hypervascularization is shown in red and yellow shades. Devascularization is shown in blue and green colors. In all cases there was a profound visual reduction of vascularization displayed as blueish and greenish nuances.

Graph: Fig.3 Case of a partially successful MWA (display of peak enhancement). A. The post-embolization defect in CEUS original images shows residual contrast enhancement on the side, consistent with remaining tumor. B. In pseudo-colors the defect is shown in blue to yellow and red colors, proving an only partial devascularization. C/D. TIC analyses show that the perfusion curve for the center of the defect (green) is not quite close to the baseline confirming residual tumor.

Graph: Fig.4 Case of successful RFA (display of peak enhancement). A. After the intervention, the post-embolization defect in CEUS original images appears black, meaning nearly avascular whereas the margin appears partially vascularized. B. In pseudo-colors the defect is shown in blue colors, also showing a devascularization. C/D. TIC analyses show that the perfusion curve for the center of the defect (green) is close to the baseline.

Radiofrequency ablation is considered to be one of the most common thermal ablation modalities worldwide for HCC, with a 33–57% 5-year survival rate [[20]]. In comparison to that, Studies using microwave ablation for treatment of HCC lesions <5 cm have reported 29–68.6% 5-year survival rates [[21]].

Studies comparing HCC lesions ≤4 cm in maximal diameter treated by RFA or MWA have found no significant differences in local tumor control, complication rates, and long-term survival [[22]]. Also, in HCC lesions up to 5 cm maximal diameter treated by either RFA or MWA, no significant differences regarding 5 year survival and 5 year disease-free survival rates were found [[23]].

Assessment of treatment response, during and shortly after RFA and MWA, is crucial to determine treatment outcome and patient safety. An intervention can be classified as technically successful when the ablation zone completely overlaps or encompasses the target tumor plus an ablative safety margin [[24]]. For treatment of HCC lesions up to 5 cm diameter, technical success was achieved in 83.4% for RFA vs. in 86.7% for MWA [[23]].

Previously published studies [[25]] have shown a significant correlation between the presence of peritumoral vessels and poor local tumor control during thermal ablation for HCC. Following ablation, a central devascularized area is generally accepted to represent coagulated tissue. The surrounding tissue is characterized by profound hyperemia whereas on delayed contrast images, peripheral rim enhancement can be identified. Nodular enhancement and irregular regional enhancement pattern correlate with residual vital tumor. However, there are various reasons for contrast enhancement 24 hours after the ablation such as reactive changes in terms of immunoresponses [[27]] or remaining tumor.

In metastatic liver disease, the indication for MWA is the impossibility of a complete resection or a central position of metastases. Patients with up to 5 liver metastases and a size of ≤4 cm are eligible candidates for local ablation.

The color coded perfusion software (VueBox®) used in this study shows hyper enhancement in yellow and red shades, whereas devascularized areas appear blue. By using ROIs in the center of the lesion, the margins and the surrounding liver tissue the extent of devascularization could be analyzed. Previous studies have evaluated the perfusion parameters of gastrointestinal stoma tumors (GIST) [[29]] and prostate cancer [[18]]. However, this is the first time that a perfusion software was used for evaluation of the therapeutic success following RFA/MWA in primary and secondary liver malignancies.

MWA induces a homogeneous heating effect and coagulation necrosis with a complete loss of microcirculation. In RFA, a simultaneous application of a current to a cluster of three to five closely spaced electrodes, offers the potential of large volume coagulation necrosis for clinical tumor ablation therapy. However, fibrous modifications and hyperemia are seen in the margin areas of the ablation defect due to inflammatory changes [[30]]. A crucial point is that evaluation of the perfusion parameter often shows incomplete devascularization following ablative techniques, especially following RFA. Thus, for establishing the clinical success short-term follow-up is crucial. In this study, we were able to analyze and compare various perfusion parameters including peak enhancement, Wash-in area under the curve, mean transit time, rise and time to peak in the center of the lesion, its margins and the surrounding tissue for the first time. Parametric imaging using CEUS can suggest whether ablation therapy using MRA or RFA was successful or not. In lesions treated using RFA, perfusion analysis from arterial to portal venous phase shows highly significant differences regarding pE between center and surrounding liver as well as surrounding liver and margin. In lesions treated using MWA significant differences were detected regarding pE, Ri and mTT between center of the lesion and surrounding liver. However, perfusion evaluation can only be used as a secondary method since assessment of CEUS contrast dynamics concerning remaining nodular tumor lesions with possible early wash-out is crucial.

Our study had some limitations. The study population was heterogeneous with regard to the tumor sizes and the extent of liver cirrhosis. A further limitation is the still relatively small number of patients and lesions in the present study. Therefore, further studies with an increased number of patients are necessary. Evaluation of perfusion parameters using an external perfusion software prolongs the examination time for each patient for about 15–30 minutes, mean 20 minutes and is therefore a rather time consuming application. Image fusion of CEUS with CT or MRI is valuable imaging method following tumor ablation. One disadvantage is however that the fusion data cannot be further evaluated using the external perfusion software (VueBox®) as a different image format (DICOM) of the loops is required.

In summary our study indicated, that the contrast-enhanced ultrasound in combination with perfusion analysis is a valuable supporting tool for post-interventional success control following MWA/RFA of liver lesions.