Evaluation of dynamic effects of therapy-induced changes in microcirculation after percutaneous treatment of vascular malformations using contrast-enhanced ultrasound (CEUS) and time intensity curve (TIC) analyses

The aim of this follow-up study was to demonstrate the effect of percutaneous interventional treatment on local microcirculation of peripheral vascular malformations using CEUS and TIC analysis. MATERIAL AND METHODS: Retrospective analysis of 197 patients (136 female; 61 male; 3–86 years) with 135 venous (VM), 39 arterio-venous (AVM), 8 lymphatic and 15 veno-lymphatic peripheral vascular malformations before and after the first percutaneous treatment. CEUS was performed after i.v. injection of 1-2.4 ml of sulfur hexafluoride microbubbles (SonoVue^®) using a 6-9 MHz linear probe. Digitally stored cine loops (starting in the early arterial phase for 60 sec) were read by independent readers in consensus. Regions of interest (ROI) were defined in the center and at the margins of the malformation, as well as in the healthy surrounding tissue. TIC analyses with Time to Peak (TTP) and Area under the Curve (AUC) were calculated using integrated perfusion software. RESULTS: After the treatment there was a significant decrease for median AUC in VM in the center from 297.8 (14.5–2167.6) rU down to 243.3 (0.1–1678.8) rU (p = 0.043) and in the surrounding tissue down to 107.7 (20.2–660.2) rU (p = 0.018). For the other malformations AUC decreased in the center and the margins as well. TTP rose, however these changes did not reach the level of significance. CONCLUSION: Analyzing the capillary microcirculation TICs offer a possibility of monitoring therapy-induced capillary changes of vascular malformations.

CEUS; percutaneous interventional therapy; vascular malformation

The high-resolution ultrasound plays an important role in diagnosis of vascular diseases and structures. A multi-frequency linear probe for B-Mode and color-coded Doppler sonography (CCDS) can display morphological changes and hemodynamic effects of vascular malformations in arteries and arterioles as well as in venous structures, like shunts and fistulas depending on the angle of insonation and the speed of blood flow. Power Doppler can be used to help localize the malformation and is less dependent on the angle. However, only when using high-resolution CEUS, the capillary changes of the microcirculation can be shown and analyzed dynamically [[1] ]. CEUS is performed using CHI (contrast harmonic imaging), PIHI (pulse inversion harmonic imaging) and a low mechanical index (MI <0.2) [[3] ].

Extensive high-flow malformations with multiple shunts show a faster crossing of microbubbles from the arterial to the venous structures, which can be displayed dynamically, when using a high-resolution technique. Since the diameter of the microbubbles ranges between 3–10μm it is about the size of an erythrocyte [[4] ]. For veno-lymphatic malformations CEUS offers the possibility of showing structural differences compared to the surrounding tissue. And also in venous malformations, CEUS can depict the dynamic micro-vascularization compared to the surrounding tissue. Compared to ceCT and ceMRI, CEUS with second generation contrast agent (SonoVue®, Bracco) is a strictly vascular imaging [[3] ].

Percutaneous interventions are used to close fistulas, reduce blood flow, and in cases when further surgical treatment is needed, achieve best pre-operative circumstances. Vascular malformations lead to changes both in macro- and microcirculation including capillary changes [[5] ]. Depending on high- flow or slow-flow malformations, arterio-venous shunts and capillary hypercirculation are possible. But even in slow flow malformations, capillary feeding vessels can be visualized when using CEUS. As it was shown in our previous study, CEUS is a useful tool for dynamic imaging [[6] ]. By using time intensity curve analyses (TIC) in digitally stored cine sequences, perfusion and perfusion changes can be quantified [[6] ]. The extent of percutaneous interventions like embolization and sclerotherapy concerning microcirculation and tissue perfusion can be measured dynamically [[7] ].

The aim of this study was to quantify perfusion changes using TICs after the first percutaneous treatment in different types of vascular malformations.

Written informed consent for the CEUS examination was obtained in all cases by either the patients or in cases of children by their parents. The study was approved by the local review board and is in accordance with the ethical guidelines of the journal Clinical Hemorheology and Microcirculation [[10] ].

This is a retrospective study of 197 patients (136 female; 61male) between 3–86 years (mean 27.9 years). 135 venous malformations (VM), 39 arterio-venous malformations (AVM), 8 lymphatic malformations and 15 veno-lymphatic malformations were analyzed retrospectively by independent reading of digitally stored DICOM data.

The patients presented themselves to the interdisciplinary vascular anomalies center because of pain caused by the malformation. Indication for treatment was based on clinical findings and in interdisciplinary consensus. The classification of the malformation was based on the ISSVA classification for vascular anomalies [[11] ].

Every patient received CEUS before and 24 hours after the first percutaneous treatment. The examinations were performed using a high-end ultrasound machine (LOGIQ E9, GE) using a linear multi-frequency probe (6–9 MHz). Afterwards the examinations were stored digitally in PACS. Hence, an independent reading was possible.

In all cases color coded Doppler sonography (CCDS) and Power Doppler were performed for the arteries and veins in the affected extremity before the injection of contrast-media. Hemodynamic parameters like flow, shunts and fistulas were documented for the whole extremity. Pre-existing deep- vein thrombosis was excluded by B-Mode and CCDS. These images were also stored digitally in PACS.

All CEUS examinations were performed by one experienced examiner (>10 years, >3000 US/year). CEUS was performed after i.v. bolus injection of 1–2.4 ml sulphur hexafluoride microbubbles (SonoVue®, Bracco, Italy) followed by a bolus of 10 ml 0.9% NaCl injected through a 20-18 G peripheral cubital cannula ([NaN] , [NaN] ). 2.4 ml is the recommended dose of SonoVue [[12] ]. However, 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 1–2.4 ml and patients <70 ml received 1.0–1.5 ml i.v. contrast media depending on the localisation. In superficial malformations we injected less contrast media, than in deeper malformations. The amount was documented and the same amount was injected before and after the treatment. Cine loops were stored continuously for 60 sec after the arrival of the first microbubbles in the treated area. After having observed the center of the malformation (in correlation with the patients’ main complaints) for one minute, the whole extent of the malformation was examined in sweep technique and cine loops for up to 10 sec were stored digitally. The whole CEUS examination lasted up to 5 min after the injection. CEUS was performed by using a low mechanical index (MI <0.2), pulse inversion harmonic imaging technique (PIHI) and contrast harmonic imaging (CHI).

Because the malformations could cover the whole extremity the probe was set to a virtual convex mode so that an area with about 60×40×7 mm could be covered without moving the probe for the TIC analyses. Only those one-minute-cine-sequences, during which the probe was kept steadily in one place, were analyzed independently by two radiologists from the arterial to the venous phase (15 sec- 1 min) in consensus using raw data. For calculation of the TIC analyses the area of the malformation with the highest irregular enhancement was chosen before the treatment ([NaN] , [NaN] ) and marked for the post-interventional control with CEUS. TIC analyses were calculated after placing regions of interest (ROI) in the center and the margins of the malformations, as well as in the surrounding healthy tissue ([NaN] , [NaN] ). Time to Peak (TTP in sec) and Area under the Curve (AUC in rU = relative Units) were calculated for the regions before and after the treatment [[13] ]. For calculation algorithms integrated perfusion software (LOGIQ WORKS, GE) was used. The software calculates the average signal intensity as a function of time [[7] ]. TIC analyses calculate the parameters primarily descriptively as indirect parameters of local perfusion. The absolute values cannot be calculated. The average contrast intensity of the region of interest (ROI) is displayed as a function of time. The curves of all selected ROIs can be displayed simultaneously and can be compared to each other. The integrated perfusion software can calculate several perfusions parameters i.e. time to peak (time from the first image frame to peak intensity frame) or area under the curve.

Furthermore it must be kept in mind that intensities are compressed logarithmically by the ultrasound machine, meaning that double the visual signal intensity does not mean double the blood volume [[14] ]. This is why the perfusion quantification needs to be performed using the raw data.

Personal data like sex, age, localization, type of malformation, type of treatment and complications (DVT, abscess, necrosis) were extracted from the clinical information system [[15] ]. The post-interventional control was performed 24 hrs after the intervention after the compression draping had been removed ([NaN] , [NaN] ). The post-interventional scan was performed with the same protocol as the pre-interventional scan. Post-interventional TTP and AUC were calculated in the same way as pre-interventionally.

Besides descriptive analysis, an ANOVA analysis was performed. If a difference was found, and to exclude the effect of multiple testing, a Bonferoni correction was performed to determine which measurement had the most influence. Changes were considered to be significant for probabilities p < 0.05.

Of the 197 patients 39 patients received embolisations, 156 patients sclerotherapy and 1 patient each was treated with laser or coiling. 18 patients suffered from vascular malformation in the face/neck, 40 in the upper extremity, 105 in the lower extremity, and 34 in the torso.

Only three patients showed post-interventional complications 24 hours after the treatment. One patient had a partial deep vein thrombosis in the affected extremity, one partial necrosis in the treated area, and one patient suffered from pleural effusion after the treatment of a malformation on the chest wall.

Pre- and post-interventional TIC analyses could be performed in all cases.

For venous malformations there was a significant decrease in median AUC in the center (279.8 rU (14.5–2167.6) to 243.3 rU (0.1–1678.8), p = 0.043) and the surrounding tissue (129.4 rU (46.7–1182.6) to 107.0 rU (20.2–660.2), p = 0.018) after the first treatment. Furthermore there was an increase in median TTP that was significant for venous malformation in the surrounding tissue after the intervention (27.6 sec (0.9–90.2) to 33.4 sec (1.3–78.2), p = 0.022) ([NaN] ) ([NaN] ).

Table 1 Comparison of pre- and postinterventional TTP (time to peak) and AUC (area under the curve) for venous malformations in the center (C), the margins (M), and the surrounding tissue (S). Range, Median (Q2), 25 % quartile (Q1) and 75 % quartile (Q3) are displayed. P-values are calculated for the direct comparison between pre- and postintervention AUC or TTP respectively in the different regions

For arterio-venous malformations the median AUC decreased in all areas clearly, for example in the center from 590.9 rU (67.1–1531.9) to 490.4 rU (0.3–1503.9), however did not reach the level of significance (p = 0.222). TTP remained stable before and after the intervention ([NaN] ) ([NaN] ).

Table 2 Comparison of pre- and postinterventional TTP (time to peak) and AUC (area under the curve) for arterio-venous malformations in the center (C), the margins (M), and the surrounding tissue (S). Range, Median (Q2), 25% quartile (Q1) and 75% quartile (Q3) are displayed. P-values are calculated for the direct comparison between pre- and postintervention AUC or TTP respectively in the different regions

For the lymphatic and veno-lymphatic malformations AUC in the center and at the margins decreased. However these changes did not reach the level of significance either. TTP for these two types of malformations remained more or less stable after the first percutaneous treatment ([NaN] ).

Table 3 Comparison of pre- and postinterventional TTP (time to peak) and AUC (area under the curve) for lymphatic malformations in the center (C), the margins (M), and the surrounding tissue (S). Range, Median (Q2), 25 % quartile (Q1) and 75 % quartile (Q3) are displayed. P-values are calculated for the direct comparison between pre- and postintervention AUC or TTP respectively in the different regions

Table 4 Comparison of pre- and postinterventional TTP (time to peak) and AUC (area under the curve) for veno-lymphatic malformations in the center (C), the margins (M), and the surrounding tissue (S). Range, Median (Q2), 25 % quartile (Q1) and 75 % quartile (Q3) are displayed. P-values are calculated for the direct comparison between pre- and postintervention AUC or TTP respectively in the different regions

To sum it all up: after the first percutaneous treatment there were only significant changes for venous malformations.

To the present day CEUS is not yet fully integrated in the clinical routine. But as it has already been shown in other studies, that CEUS can visualize and quantify micro-vascularization in different organs [[16] ].

In high-flow malformations shunts can be shown using Duplex, hence CCDS can be used for the first “screening” in vascular malformations. In lymphatic malformations the CCDS does not reveal relevant changes. B-Mode by contrast can display the lymphatic edema. In venous malformations CCDS can show aneurysms, thrombosis, or stasis. However, the capillary microcirculation can only be displayed using CEUS.

By using TIC analyses capillary micro-vascularization can be detected. TIC has already been useful for postoperative control of free flaps. It was even possible to establish relative reference values. 300 dB show a sufficient blood supply, whereas necrotic areas only show about 10 dB [[23] ].

This study shows that CEUS also plays an important role as an additional imaging method in vascular malformations. The aim of the intervention is to close fistulas and shunts, or to reduce venous conglomerates, as well as arterial or venous aneurysms. The perfusion of the malformation should be adjusted to the surrounding tissue. Using CEUS, capillary microcirculation in different malformations (center and margins) can be evaluated in comparison to the surrounding tissue, since ultrasound contrast media is strictly intravascular [[3] ]. Using TIC analysis, the results can be displayed graphically. Still, a combination with other imaging methods is mandatory since MRA/CTA can depict the full extent of the malformation and CEUS can evaluate capillary changes. This study shows that one single percutaneous treatment can reduce blood flow and blood volume in the treated area of the malformation, as seen by an increase in TTP and a decrease in AUC. However one single percutaneous treatment is not sufficient in most cases to completely heal the malformation since not all the feeding vessels can be closed during one session because otherwise those parts might become necrotic. This is also consistent with our clinical experience, meaning that patients will be repeatedly treated for the same complaints.

The fact that TTP decreases in venous malformations after the intervention might be caused by reactive hyperemia due to the embolization. Not surprisingly, there were no significant changes for arterio-venous malformations. When closing shunts, other shunts and collateral circulation will be opened.

Using B-Mode alone, the determination of the malformation type is not always possible. Ectatic veins and aneurysms can be described. Thrombosis can be excluded by compression of the veins, which was done before and after the treatment. Other complications like necrosis or perfusion deficits can also be depicted quickly and easily.

Using CCDS, hemodynamic changes can be depicted, and are mandatory for planning and control of the intervention. But only with CEUS, low energy and a low mechanical index (MI <0.2) it is possible to judge capillary changes in malformations as well as feeding vessels and tissue perfusion using TIC analysis. This had been shown in our own previous work [[7] ].

To reduce the bias by inter-observer variability the patients in this study were only scanned by one examiner. The reading was always performed by the same two radiologists in consensus.

A limitation of this study is that only the nidus of the malformation was documented continuously for 1 minute after the arrival of the first bubbles and TIC analyses were only performed for the center (nidus) of the malformation. The rest of the malformation was only examined in sweep technique. Furthermore the regions of interest are placed visually and manually in the treated area and are therefore dependent on the experience of the examiner. Also the reading is based on the experience of the radiologist.

Another limitation of this study is that the nature of the surrounding tissue was different depending on the localization of the malformation, i.e. muscle or adipose tissue. Consequently, behavior of the contrast-media is different. Muscles with higher blood supply show a stronger contrast-media uptake than fat. Consequently there might be a bias, significantly or not, when comparing the center and margins to the surrounding tissue.

An advantage of using integrated perfusion software with TIC analyses is that the data can be analyzed in a timely manner. When using external perfusion software such as VueBox® (Bracco), that the data has to transferred to an external analysis computer and is thus more time consuming. A disadvantage of this integrated perfusion software is that only the original CEUS data and the TIC analysis are displayed as the results, whereas when using VueBox® the perfusion results are also displayed in parametric images using pseudo colors.

So far, there are no known contra-indications for repeated injections of ultrasound contrast media or an influence of the microbubbles on the perfusion of capillaries [[24] ]. Neither the thyroid gland, nor the kidneys are harmed. However, pseudo- allergic reactions can be evoked by the contrast-media even after the first i.v. injection [[12] ].

The number of lymphatic and veno-lymphatic malformations is too small to finally judge the impact of treatment. Further research with prospective studies for CEUS and perfusion analyses are needed to be able to evaluate the influence of percutaneous treatment in vascular malformations. Suggestions for prospective settings are for example using an injection pump instead of manual injection so that the contrast media is always injected with the same speed. As is has already be performed in this study the same examiner, the same machine and the same presets are mandatory.

This piece of work underlines the importance of capillary microcirculation and the research in vascular malformations. It points the important life’s work of Prof. Dr. Friedrich Jung, who should be honored in this way.

Graph: Fig.1a CEUS after i.v. injection of 1 ml SonoVue® of an AVM around the right knee before the treatment using the double view mode. On the left side the ectatic veins can be seen in B-Mode. On the right side the ectatic vessels pool the contrast media. The surrounding healthy tissue shows capillary soft tissue perfusion. b. Pre-interventional TIC analysis calculated by the integrated perfusion software (LOGIQ Works, GE). Up left the original CEUS images are shown with the manually placed regions of interest. Up right the corresponding TIC analysis is displayed and on the bottom the numeric values for AUC and TTP. The regions of interest (ROI) were placed in the center of the AVM around the right knee (top ROI), at the margins (middle ROI) and in the surrounding healthy tissue (bottom ROI). The software calculates the average signal intensity as a function of time. There is a fast inflow oft he contrast media. Peak is reached about 20 sec. after the injection. No relevant wash-out within 70 sec. c. Post-interventional TIC analysis calculated by the integrated perfusion software (LOGIQ Works). The regions of interest were placed in the center of the AVM around the right knee (top ROI), at the margins (middle ROI) and in the surrounding healthy tissue (bottom ROI). Compared to the pre-interventional TIC analysis (Image 2) there is a clear decrease in AUC in the center of the malformation (932 rU from 1397 rU), meaning there is a clear reduction of perfusion. Also at the margins the perfusion of the malformation was partially reduced but is still higher than in the surrounding tissue as shown by longer TTP and a decrease in AUC.

Graph: Fig.2a After i.v. application of 1.5 ml SonoVue (Bracco) there is a clearly higher capillary uptake in the venous malformation compared to the surrounding tissue. b. Case of a venous malformation at the right knee. The regions of interest were placed in the surrounding tissue (yellow, top), in the center (turquoise/red, middle) and at the margins (green, bottom). The corresponding TIC analyses and numeric values were calculated. c. After the first percutaneous intervention of the venous malformation around the knee the TTP (time to peak) is lower than before, most likely due to reactive hyperemia. The AUC (area under the curve) in the center, at the margins and in the surrounding tissue has also decreased.

Graph: Fig.3a Boxplot showing the results for TTP in venous malformation before (left three columns) and after the intervention in the center, the margins and the surrounding tissue. It is clearly visualized that there is a broad range of values. Only the increase of TTP pre- and post-interventionally was significant. b. Boxplot showing the results for AUC in venous malformation before (left three columns) and after the intervention (right three columns) in the center, the margins and the surrounding tissue. It is clearly visualized that there is a broad range of values.

Graph: Fig.4a Boxplot showing the results for TTP in AVM before (left three columns) and after the intervention (right three columns) in the center, the margins and the surrounding tissue. It is clearly visualized that there is a broad range of values. Only the decrease of TTP pre- an postinterventionally was significant. b. Boxplot showing the results for AUC in AVM before (left three columns) and after the intervention (right three columns) in the center, the margins and the surrounding tissue. It is clearly visualized that there is a broad range of values.