Ultrafast Doppler ultrasound (UFUS) with high spatial resolution (100 mm) and sensitivity represents a suitable noninvasive imaging modality for obtaining a comprehensive qualitative overview of the hepatic vasculature. UFUS also facilitates quantitative measurements of the microvasculature, aiming to enhance our understanding of vascular disease mechanisms.
Non-invasive in vivo imaging of the vasculature is a powerful tool for studying disease mechanisms in rodents. To achieve high sensitivity imaging of the microvasculature using Doppler ultrasound methods, imaging modalities employing the concept of ultrafast imaging are preferred. By increasing the frame rate of the ultrasound scanner to thousands of frames per second, it becomes possible to improve the sensitivity of the blood flow down to 2 mm/s and to obtain functional information about the microcirculation in comparison to a sensitivity of around 1 cm/s in conventional Doppler modes. While Ultrafast Doppler ultrasound (UFUS) imaging has become adopted in neuroscience, where it can capture brain activity through neurovascular coupling, it presents greater challenges when imaging the vasculature of abdominal organs due to larger motions linked to breathing. The liver, positioned anatomically under the diaphragm, is particularly susceptible to out-of-plane movement and oscillating respiratory motion. These artifacts not only adversely affect Doppler imaging but also complicate the anatomical analysis of vascular structures and the computation of vascular parameters. Here, we present a qualitative and quantitative imaging analysis of the hepatic vasculature in mice by UFUS. We identify major anatomical vascular structures and provide graphical illustrations of the hepatic macroscopical anatomy, comparing it to an in-depth anatomical assessment of the hepatic vasculature based on Doppler readouts. Additionally, we have developed a quantification protocol for robust measurements of hepatic blood volume of the microvasculature over time. To contemplate further research, qualitative vascular analysis provides a comprehensive overview and suggests a standardized terminology for researchers working with mouse models of liver disease. Furthermore, it offers the opportunity to apply ultrasound as a non-invasive complementary method to inspect hepatic vascular defects in vivo and measure functional microvascular alterations deep within the organ before unraveling blood vessel anomalies at the micron scale levels using ex vivo staining on tissue sections.
Preclinical mouse models are widely used to study the onset and progression of hepatic microvascular disease1,2. Longitudinal follow-up is advantageous, as blood capillary alterations occur at an early stage of the disease before inflammation and fibrosis appear, which may result in cirrhosis over time1. To investigate disease progression, powerful imaging tools are essential for visualizing and measuring blood vessel function in vivo. Ultrafast Doppler ultrasound (UFUS) represents a suitable modality due to its low cost, portability, simplicity, and real-time capability, but also because it is a non-invasive and time-saving procedure3. In comparison to conventional Doppler ultrasound, which may reach a similar spatial resolution but is restricted to large blood flow, UFUS allows the detection of slow blood flow down to 2 mm/s from 1 cm/s, allowing to probe the contribution of much smaller vessels4. UFUS improves the sensitivity, which is accomplished through plane wave transmissions with a large field of view in a single insonification to increase the frame rate and thus the number of ultrasound images available5. The high number of temporal samples improves the cancelation of the signal coming from the tissue and enhances the detection of the echoes reflected from the blood scatterers, boosting the sensitivity without requiring contrast agents6. By estimating the energy of blood scatterer echoes through the power Doppler estimation, a quantity is obtained that is proportional to the blood volume in each pixel even if the blood vessels are too small to be resolved in the image7.
UFUS offers a favorable combination of spatial and temporal resolution, making it a suitable tool for the qualitative assessment of the liver vascular architecture. While operating an ultrasound scanner is relatively straightforward, a comprehensive understanding of the liver anatomy is required to navigate through the hepatic vasculature of the different liver lobes in mice. Recent studies have focused on describing murine liver anatomy and the hepatic vascular tree using 3D micro-computed Tomography (µ-CT)8,9. Even though some of the morphological differences between human and mouse livers are well documented10, research performed in rodents often adopts anatomical terminologies relevant to humans, leading to some confusion in the context of quadruped animals. Even recognized standardized nomenclature, such as Nomina Anatomica Veterinaria (NVA), lacks sufficient information about consistent terminologies for rodent hepatic vascular structures11. Beyond the challenge of poorly documented anatomical structures in rodents, even greater difficulties arise in interpreting ultrasound readouts, particularly when a reproducible approach is anticipated.
Obtaining non-invasive functional information on living organs across various spatial scales poses challenges for in vivo imaging3. High-end liver imaging modalities such as CT, magnetic resonance imaging and angiography (MRI/MRA), or conventional Doppler imaging using focused waves are essential for evaluating disease progression, but they fail in capturing most blood flow dynamics with the exception of photoacoustic imaging12,13,14. UFUS has a high sensitivity and can detect slow blood flow in small vessels, which are not limited to flow larger than a centimeter per second4,15. Recently, we developed a customized imaging method based on anatomical considerations using UFUS, enabling reliable measurement of the hepatic blood volume in a mouse model with vascular expansion16.
In this protocol, we outline the procedure for non-invasive UFUS imaging of the liver using a commercially available ultrasound platform originally developed for functional brain imaging. To streamline the process and enhance information retrieval, we offer a comprehensive illustration of the mouse liver with anatomical nomenclature and a complete method of analysis. Additionally, we provide a detailed representation of the acquired imaging readouts as a reference, indicating common anatomical structures and landmarks to assist operator-dependent plane selection. Furthermore, we include a relevant illustration of both qualitative and quantitative analyses of the hepatic vasculature.
All animal experiments were performed under European legislation (Directive 2010/63/EU) and Dutch government guidelines and approved by the Animal Ethical Committee of Leiden University Medical Center.
1. Mouse preparation for UFUS imaging (Timing ~ 10 min)
2. Probe operation and plane selection (Timing ~ 10 min)
3. Doppler scan acquisition (Timing ~ 20 min)
4. Quantitative analysis (Timing ~ 1 h)
Following the protocol, imaging of the hepatic vasculature is facilitated based on the standardization of anatomical structures and corresponding terminology (Figure 1). To propose a standardized nomenclature for researchers using preclinical liver models, we schematically displayed the common liver positioning in the abdominal cavity of mice (Figure 1A), provided the structural arrangement of the different hepatic lobes (Figure 1B), and specified the arterial vascular network (Figure 1C). Equipped with anatomical expertise, in vivo UFUS imaging was performed by utilizing a 3D-printed water tank to reduce movement artifacts (Figure 2A). The acoustic impedance was corrected by introducing the probe into water and applying ultrasound transmission gel directly to the shaved skin of the animal. By adopting this setup, breathing and other movement artifacts are reduced since the probe is not in direct contact with the moving chest and abdomen of the animal. This way, we visualized the vascular structure of the right and left lobes, which are easily accessible by UFUS due to their superficial location (Figure 2B). Other liver lobes, such as the caudate and quadrate lobe, are situated deeper in the body, making it more challenging to anatomically identify the hepatic vasculature properly by UFUS.
Having demonstrated the importance of standardized and commonly accepted anatomical indications and interpretations of vascular structures, we performed multiple Doppler scans in various planes to provide a comprehensive overview of the hepatic vasculature imaged by contrast-enhanced UFUS in mice (Figure 3). We delineated the hepatic arterial network of the individual liver lobes, i.e., left lateral (LL), left medial (LM), right lateral (RL), and right medial (RM) lobes. Moreover, major arteries are labeled, such as the abdominal aorta (AA), hepatic artery (HA), left lateral hepatic artery (LLHA), and right medial hepatic artery (RMHA). It is important to note that liver lobes can be slightly differently positioned depending on the placement and peristaltic movement of the small intestine (SM), as well as the state of the stomach (S) filling. This contrast-enhanced overview of the entire mouse liver vasculature provides a comprehensive vascular atlas for in vivo liver imaging by UFUS.
An accurate selection of anatomical planes is crucial for obtaining a quantitative readout derived from one liver lobe. Robust blood volume measurements can be achieved by manually selecting 10 regions of interest (ROI) per plane (Figure 4A). For reliable blood volume quantification of the microcirculation, a stable acquisition is needed with low variation in mean image intensities. This is determined based on the mean image value (MIV) of the entire image over time (Figure 4B), in contrast to a recording that is susceptible to significant motion artifacts (Figure 4C). To handle motion artifacts, the automatic selection of low MIV was used for reliable quantification of hepatic blood volume measurements (Figure 4D) by plotting the standard deviation of all temporal low MIV of an entire image (Figure 4E). Doppler scans of multiple planes per animal show acceptable intra- and inter-variability within and between mice from 3 different cohorts after outlier treatment on 30 ROIs per animal with ROUT test 1% (Figure 4F). To improve the blood volume estimate per animal, it is recommended to record additional planes or increase the number of ROIs selected per plane. In addition, blood volume readings can be used to monitor drug treatment, specifically targeting microvasculature. By administrating the vehicle through two administration routes, we demonstrated that there is no effect on the hepatic blood volume measurements (Figure 4G). These results indicate that gene deletion or delivery of treatment is not affected by vehicle administration, ensuring a robust experimental design. Overall, an objective and reproducible quantitative protocol is provided for future research to explore hepatic vascular defects in various mouse models.
Figure 1: Anatomy of the murine liver. (A) Schematic overview of liver positioning in the abdominal cavity under the diaphragm of the mouse. (B) Schematic description of the murine liver with four liver lobes, i.e., left, quadrate, caudate, and right lobe, and the portal triad consisting of the portal vein, hepatic artery, and bile duct. (C) Schematic illustration of the arterial network of the haptic vasculature. Identification of hepatic arteries (HA) and branches (ramus) with Latin nomenclature provided. Abbreviation: HA = Hepatic artery. Please click here to view a larger version of this figure.
Figure 2: Ultrafast Doppler ultrasound (UFUS) imaging of the hepatic vasculature in mice. (A) Schematic overview of experimental setup for non-invasive in vivo imaging. (B) Schematic illustrationsof the imaging direction and associated representatives of the right and left medial lobe, as well as the left lateral lobe, imaged by UFUS. Abbreviation: AA = Aorta abdominals, LLL = Left lateral lobe, LML = Left medial lobe, RML = Right medial lobe. Scale bar = 1 mm. Please click here to view a larger version of this figure.
Figure 3: Representative contrast-enhanced ultrafast Doppler ultrasound (UFUS) of the murine hepatic vasculature. An overview of the entire hepatic vasculature of an adult mouse imaged in 3 positions, i.e., lateral right, medial, and lateral left, in 4 planes from cranial towards caudal. Reference is situated directly under the diaphragm imaged in medial and 1 cm lateral on the right and left, with the same planes in a caudal direction. Abbreviation: AA = Aorta abdominals, HA = Hepatic artery, LLL = Left lateral lobe, LLHA = Left lateral hepatic artery, LML = Left medial lobe, RLL = Right lateral lobe, RMHA = Right medial hepatic artery, RML = Right medial lobe, SI = Small intestine, S = Stomach. Scale bar 1 mm. Please click here to view a larger version of this figure.
Figure 4: Ultrafast Doppler ultrasound (UFUS) quantitative measurement of hepatic blood volume. (A) Representatives of the medial lobe (ML) of adult C57BL/6 wildtype control mice in plane 1 selected under the diaphragm, corresponding to 0 mm, plane 2 at -0.3 mm, and plane 3 at -0.6 mm with a selection of 10 regions of interest (ROIs). Scale bar 1 mm. Illustration of Power Doppler of 30s with (B) stable acquisition (grey) and (C) tissue motion (red) based on mean image values in arbitrary units (A.U.) of the whole image for quantitative assessment. (D) Automated recognition of low intensities (A.U.) for quantitative blood volume measurement over time. (E) Standard Deviation (St. Deviation) of low intensities plotted to identify unstable acquisitions (red) caused by motion artifacts indicated by a cut off of overall mean (µ) from number (N) of animals from 3 cohorts with 2 SD (2σ) assessment of standard deviation (SD) of low intensities (A.U.) in physiological condition (gray) in adult C57BL/6 wildtype control mice. (F) Quantification of in vivo measured blood volume changes over time averaging over 3 planes, i.e., 30 ROI in total per animal, are represented as pixel intensity (A.U.). Outlier detection by ROUT test 1%. (G) There is no difference between the type of delivery vehicle and the route of administration. Cohort 1 (n=6) received an intraperitoneal (ip) injection of saline, cohort 2 (n=3) ip injection of DMSO and Cohort 3 (n=6) oral administration (po) of cellulose. One-way ANOVA with Dunnett's multiple comparisons test, ns P > 0.05. Abbreviation: A.U. = Arbitrary units, DMSO = Dimethyl sulfoxide, ip = Intraperitoneal, po = Oral administration, ROI = Region of interest, St. Deviation = Standard Deviation, UFUS = Ultrafast Doppler ultrasound. Please click here to view a larger version of this figure.
Supplementary File 1: Ultrafast Doppler ultrasound (UFUS) quantitative script of hepatic blood volume measurements. Script gui_analysis.m provides a graphical interface for the automatic positioning of the region of interest (ROI) for blood volume measurements. Please click here to download this File.
In this protocol, we demonstrate how to perform non-invasive UFUS imaging of the mouse hepatic vasculature. We present a detailed description of the imaging procedure, encompassing animal handling, operation of the ultrasound scanner, and data analysis for computing blood volume changes in vivo. To facilitate ultrasound data interpretation, we provide a comprehensive overview of the anatomical structures that can be clearly visualized by UFUS. Furthermore, we describe a standardized quantification procedure for robust measurements of vascular parameters. Achieving a suitable resolution is possible without the use of a contrast agent, enabling a safe and straightforward quantitative assessment of the hepatic blood volume of the microcirculation. We emphasize the use of standardized anatomical nomenclature to prevent confusion and misinterpretation, particularly when assessing quantitative measurements of microvasculature.
UFUS imaging of murine vasculature has previously been reported, primarily in the field of neuroscience7. Brain imaging is not adversely affected by movement artifacts unlike the liver, which is highly susceptible to breathing motion. Due to its anatomical position under the diaphragm, hair removal in the abdominal region must be performed very gently to avoid introducing irregular breathing by minimizing physical forces on the chest and abdomen region. Moreover, achieving a proper degree of anesthesia is essential to ensure regular oscillating breathing, a factor that can be relatively easily corrected in post-processing steps. The sensitivity of UFUS has been enhanced here by applying a singular value decomposition (SVD) filter to remove motion artifacts6 and by further censoring high motion frames. The high frame rate Power Doppler movies are acquired here at 10 ms to further allow us to visualize and measure the effect of breathing motion on frames that generated high-value peaks due to motion artifacts. The breathing rate could then be identified from the peak in the Fast Fourier transform of the average image signal, and the corresponding peak frames corresponding to the high motion of the organ can be discarded.
Reproducible ultrasound Doppler imaging within a single animal over time, as well as between animals of different body weights, can still face several limitations. Firstly, liver positioning can be influenced by factors such as a highly filled stomach and peristaltic movement of the bowel. This influence is particularly pronounced when selecting the left lateral liver lobe as a reference. Therefore, we recommend choosing the left lateral lobe only in larger animals and prefer the right or left medial lobe in smaller animals. Secondly, heavy non-oscillating breathing motions can occur in distressed animals, resulting in motion artifacts that are challenging to correct and, consequently, cannot be used for blood volume computation. Thus, an appropriate choice of anesthesia, along with a dosage tailored to body weight, is crucial. Lastly, the resolution limit restricts the visualization of the hepatic microvasculature, i.e., blood vessels can be visualized with diameters ranging down to 100 mm3.
The use of contrast agents was recently proposed to acquire a detailed image of the mouse vasculature by tracking microbubbles. Such an approach can achieve high spatial resolution and provide quantitative blood velocity measurements17. However, tracking microbubbles remains highly challenging for moving internal organs especially when acquired in 2D.
Here, we showed that stable hepatic blood volume measurements obtained by ultrafast Doppler could be acquired in control mice from three cohorts receiving different vehicles administered by ip injections or oral gavage. Recently, such blood volume measurements were successfully used in a liver disease mouse model16. To elaborate, vascular expansion was detected on an electron microscopic level and confirmed by quantitative measurements using UFUS. This study emphasized the sensitivity of in vivo assessment to investigate vascular defects on functional and quantitative levels, even with a preserved hepatic angioarchitecture, which is promising for clinical diagnosis and monitoring.
In summary, this protocol serves as a guide for non-invasive qualitative and quantitative studies of the hepatic vasculature in mice. It can be also employed for longitudinal follow-up due to its non-invasive and straightforward procedure. Additionally, it holds the potential to provide functional assessments of blood vessels.
The authors have nothing to disclose.
We thank M.S. Zuurmond (Leiden University Medical Center) for preparing the graphical illustrations. This protocol was derived and adapted from our previously published original work de Haan et al. 202216. This project has received funding from the European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Grant agreement No. 813839; the Netherlands Organization for Health Research and Development (ZonMw) PTO 446002501 in collaboration with CVON-PHAEDRA Impact; Health Holland (PPS-Match call LUMC Top sector LSH 2020), project NV-STAB; The Novo Nordisk Foundation Centre for Stem Cell Medicine that is supported by a Novo Nordisk Foundation grant (NNF21CC0073729); by the European Innovation Council EIC Pathfinder Program Project MICROVASC n° 101070917; French National Research Agency ANR-17-T171105J-RHUS-0009 project RHU Quid-Nash; French National Research Agency ANR-22-CE19-0034-03, project Peri-fUS; and INSERM (Institut National de la Santé et de la Recherche Médicale) Accelerator of Technological Research in Biomedical Ultrasound.
(Hydroxypropyl)methyl cellulose HPMC | Sigma-Aldrich | H7509 | 0.5% HPMC |
Centrifuge ZentriMix 380 R | Hettich | 3200 | |
Centrifuge Tubes 50 mL conical Sterile Polypropylene | CELLSTAR tubes | 227 261 | |
Cotton tips | Meditip Clean | G80215 | |
Deionized water | in house | ||
Depilatory cream | nearby drugstore | ||
Dimethyl sulfoxide (DMSO) | Sigma | D2650-5X5ML | sterile |
Ethanol absolute | FRESENIUS KABI | 010-0622-1 | |
Feeding needle, reusable | FST Fine Science Tools | 18064-20 | |
Forceps with fine tip Dumont #5 forceps | FST Fine Science Tools | 11252-00 | |
Gauze pads 5 x 5 cm | Klinipress | 175 004 | |
GraphPad Prism | GraphPad Software, Inc | Version 9 | |
GUI analysis script | Inserm | gui_analysis.m | needs a 10 ms Doppler scan |
Heating pad 17 x 17 cm, 6 W | ThermoLux Acculux | 461265 | |
Insulin syringe 30G 0.30 x 8 mm, 0.3 mL and 0.5 mL | BD Microfine + Demi | 25636 | |
Ketamine Anesketin | Dechra | REG NL 111764 | 100 mg/mL |
Kimwipes KIMTECH low-lint wiper | ULINE | S-8115 | |
Lab tape 19 mm x 50 m x 0.125 mm | Stokvis | M451 | |
MATLAB software | MathWorks | Version R2022b | |
Moisturizing cream | nearby drugstore | ||
Needle yellow 30G 0.3 x 13 mm | BD Microlance | 304000 | |
Ophthalmic ointment Oculentum Simplex | Added Pharma | 220201 | |
Paper tissues | in house | ||
Permanent lab marker | VWR | 52877-310 | |
Physiological salt solution (NaCl 0.9%) | FRESENIUS KABI | B102986 | sterile |
Sharps disposable container Medical Dispo 2.5 L | APmedical | UN3291 | |
Sterile safe-lock 1.5 mL tubes | Eppendorf | 0030 120.086 | |
Syringe 1 mL | BD Plastipak | 303172 | |
Syringe 20 mL Luer-Lok | BD Plastipak | 300629 | |
Timer | West Bend | 40053 | |
Trimmer | Exacta Aesculap | GT416 | |
Ultrafast ultrasound platform | Iconeus & INSERM 'Biomedical Ultrasound' ART | Iconeus One prototype | with a 10ms sequence |
Ultrasound transmission gel Aquasonic 100 | Parker Laboratories INC. | BT-025-0037N | |
Water tank 3D printed 10 x 10 x 1.5 cm polymethylpentene (PMP) 0.125 mm | GoodFellow Cambridge limited | ME311100/2 | CAD file upon request |
Xylazine Sedamun | Dechra | REG NL 10084 | 20 mg/mL |
.