Introduction

Lymph nodes are common sites of cancer metastasis, and the extent of lymph node involvement is crucial for patient prognosis and determining treatment strategies1,2. Lymph nodes have long been a target of ultrasound imaging, as many are located in superficial regions such as the cervical, axillary, and inguinal regions, where high-frequency probes can capture real-time images with high spatial resolution and without ionizing radiation. Their shape, size, echogenicity, and blood flow can be easily detected and characterized using B-mode and Doppler ultrasound, making them frequent targets for ultrasound-guided fine-needle aspiration3,4. Ultrasound-guided fine-needle aspiration cytology has demonstrated superior diagnostic performance compared to MRI and CT in detecting cervical lymph node metastases5. However, ultrasound-guided fine-needle aspiration does not always provide a definitive diagnosis. When lymph node metastases or lymphoma are suspected, one or more lymph nodes or lymph node stations are typically excised for histopathological examination. However, some excised lymph nodes may be benign. Lymph node excision can cause side effects, including lymphedema, pain, or nerve damage6,7. Developing an in-vivo technique to visualize and quantify the microvasculature to potentially differentiate between normal and malignant lymph nodes could optimize lymph node excision, thereby improving quality of life and reducing morbidity.

Super-resolution ultrasound imaging (SRUS) results in high-resolution microvascular imaging that exceeds the diffraction limit of ultrasound, allowing visualization of vessels with diameters in the tens of micrometers8. In a subset of SRUS known as ultrasound localization microscopy (ULM), microbubbles (MB) are injected intravenously and then detected, isolated, localized, and tracked to create microvascular density and velocity images. SRUS, including ULM, has been investigated for microvascular imaging in various organs and tissues in both animal and human studies9,10,11,12,13,14,15,16,17,18,19,20,21,22,23. SRUS has great potential for the diagnosis of diseases that affect the microvasculature, including cancer. Blood vessels undergo structural and functional alterations in various types of cancer, leading to a disorganized architecture of tortuous and leaking vessels with variable blood flow24. If these microvascular changes can be visualized and quantified using SRUS, it could improve the diagnosis and monitoring of various cancer diseases, including cancer with lymph node involvement.

SRUS/ULM has previously been applied to human lymph nodes13,20,21. In 2022, Zhu et al.20used ULM to create images of the microvasculature and MB flow direction in reactive and metastatic human lymph nodes. They used a clinical scanner (Resona 7S, Mindray) and an L11-3U Mindray transducer at a 5.6 MHz central frequency. A 1.2 ml bolus of SonoVue microbubbles (Bracco, Milan, Italy) was injected intravenously for 2 seconds, followed by a saline flush. Contrast-enhanced ultrasound (CEUS) data were recorded for at least 80 seconds with a mechanical index (MI) of 0.085. At 40 seconds, a flash sequence at an MI of 0.553 was used to adjust bubble concentration, and a second flash was given at 60 seconds if individual bubbles remained indistinct based on visual assessment. If the MB concentration was appropriate, the minimum data length was 30 seconds. In 2022, using the same scanner, transducer, and MB injection procedure, Yan et al13. demonstrated ULM velocity maps of the microvasculature of one cervical lymph node. These ULM approaches require an intravenous microbubble injection and monitoring to ensure optimal imaging conditions, which involves approximately 1.5 minutes for the injection and scanning procedure. These reasons make it less suitable for repeated clinical studies. In 2023, Ferroni et al.21 used high-definition microvasculature imaging (HDMI) to visualize the microvasculature of suspicious axillary lymph nodes in breast cancer patients in 3 seconds. They used an Alpinion Ecube12-R ultrasound machine (ALPINION Medical Systems, Seoul, Korea) and an L3-12H ALPINION transducer at a central frequency of 8.5 MHz. This technique could only visualize submillimeter vessels down to 300 \(\mu\)m, which is 1.6 times the imaging wavelength, and it cannot provide velocity maps or flow directions of the microvasculature.

A promising solution to the existing challenges in microvascular imaging of human lymph nodes using ultrasound is a fast contrast-free SRUS approach. Super-resolution ultrasound imaging using erythrocytes (SURE) has been introduced as a fast contrast-free technique for visualizing the microvasculature25,26,27,28. SURE enables tracking of the peak positions of erythrocyte scatterers to estimate microvascular blood flow direction and velocity. Only a few seconds of data acquisition is required and the MI and acoustic pressure do not need to be reduced to prevent MB rupture. SURE has been validated through simulations, phantoms, and studies in rat and rabbit kidneys25,26,29. In this study, we hypothesize that the SURE pipeline can visualize the microvasculature and flow direction and estimate the blood flow velocity in human lymph nodes in just 2 seconds after acquisition without contrast injection.

Results

The SURE images of four normal axillary and inguinal lymph nodes from healthy volunteers are presented.

Axillary lymph nodes

Figs. 1 and 2 show the results for two different axillary lymph nodes. Both figures present B-mode, power Doppler, and SURE density and velocity images generated over two seconds of data, acquired using a Verasonics research scanner and a GE L8-18i-D linear array Hockey stick probe with a wavelength of 166 \(\mu\)m. SURE density and velocity profiles are shown for selected vessels. Power Doppler and color Doppler images from a commercial GE LOGIQ E9 scanner are also included for comparison. An analysis of the microvascular structure of the lymph nodes, as visualized in the SURE density images and power Doppler images from both the Verasonics and GE scanners in Figs. 1 and 2, reveals that the SURE density image significantly enhances the visualization of the microvasculature. It depicts a greater number of vascular branches with higher clarity and contrast compared to the power Doppler images. Similarly, the SURE velocity image provides superior visualization of the microvasculature compared to the GE scanner’s color Doppler image. Additionally, it provides a depiction of flow direction across a full 360-degree range, as indicated by the color wheel, and enables blood flow velocity estimation. For the first axillary lymph node shown in Fig.1, the estimated vessel diameters derived from the SURE density and velocity profiles are 66 \(\mu\)m and 67.8 \(\mu\)m, respectively, which are less than half the wavelength. The estimated peak velocity of the selected vessel determined from the velocity profile is 15.3 mm/s. For the second axillary lymph node in Fig. 2, the estimated vessel diameters from the SURE density and velocity profiles are 55.8 \(\mu\)m and 57.5 \(\mu\)m, respectively, also less than half the wavelength. The estimated peak velocity of the selected vessel is 2.2 mm/s, as determined from the velocity profile. This vessel is located more distally from the hilus compared to the first axillary lymph node, resulting in a slower peak velocity.

Fig. 1
figure 1

Results for the first normal axillary lymph node. (a and b) B-mode and power Doppler images acquired using the Verasonics research scanner, with the color bar indicating the dB range. (c and d) Power Doppler and color Doppler images from a commercial GE LOGIQ E9 ultrasound scanner. (e and f) The SURE density and velocity images, where the color wheel indicates the direction of blood flow. (g and h) The SURE density and velocity profiles of the same vessel, highlighted in the SURE density and velocity images with a white line.

Full size image
Fig. 2
figure 2

Results for the second normal axillary lymph node. (a and b) B-mode and power Doppler images acquired using the Verasonics research scanner, with the color bar indicating the dB range. (c and d) Power Doppler and color Doppler images from a commercial GE LOGIQ E9 ultrasound scanner. (e and f) The SURE density and velocity images, where the color wheel indicates the direction of blood flow. (g and h) The SURE density and velocity profiles of the same vessel, highlighted in the SURE density and velocity images with a white line.

Full size image

Fig. 3 provides a detailed comparison of the Verasonics power Doppler and SURE density images after two seconds of data acquisition, where enlarged sections from Fig. 1 are shown. The density profiles along the blue lines in both the power Doppler and SURE density images are also depicted. The microvascular structures and vessel sizes are not clearly discernible in the power Doppler image and its corresponding profile. In contrast, they are more distinctly visualized in the SURE density image and profile. The estimated vessel diameters derived from the SURE density profile are 72, 66, 75, and 65 \(\mu\)m, all of which are less than half of the imaging wavelength.

Fig. 3
figure 3

Density profile results for the first normal axillary lymph node in Fig. 1. (a and b) The magnified area in the Verasonics power Doppler image and SURE density image of the axillary lymph node, with blue lines indicating the locations where the density profiles were measured. (c) The density profiles, with the power Doppler profile shown in red and the SURE density profile in blue. The estimated vessel diameters are indicated in the SURE density profile.

Full size image

Evolution of SURE

Fig. 4 illustrates the time evolution of a SURE density image over 10 seconds for the same axillary lymph node shown in Fig. 1. The SURE density images in Fig. 4, are generated after 0.1, 0.25, 0.5, 1, 2, 3, 5, and 10 seconds. A complete video of the 10-second evolution can be found in the Supplementary Information. Within 0.5 seconds, the microvasculature becomes clearly visible, and after 2 seconds, further improvements to the image are minimal. Therefore, a two-second acquisition time was selected for all lymph node experiments. This demonstrates the capability of the SURE pipeline to rapidly create microvascular density and velocity images.

Fig. 4
figure 4

Evolution of SURE images for the first normal axillary lymph node from Fig. 1 over time. The SURE images were generated after 0.1, 0.25, 0.5, 1, 2, 3, 5, 7, and 10 seconds of data acquisition using the same processing method. A video showing the time evolution can be found in the Supplementary Information.

Full size image

Consistency of SURE

The reliability and consistency of SURE images for the human lymph node can be demonstrated by generating images from independent datasets of the same object. This is achieved by splitting the long 10-second data into 2-second segments and creating SURE density images for each segment. Fig. 5 illustrates this approach using starting times of 0, 3, 6, and 8 seconds to ensure the independence of the images. Additionally, a video has been produced with the starting time incrementing in 0.1-second steps, resulting in 84 frames as the starting time progresses from 0 to 8 seconds. The video is available in the Supplementary Information.

Fig. 5
figure 5

Consistency of SURE imaging for the first normal axillary lymph node shown in Fig. 1. Each image is generated from 2 seconds of data, with starting times of 0, 3, 6, and 8 seconds within a total recording duration of 10 seconds. A video in the Supplementary Information is provided, showing images created for all start times from 0 to 8 seconds in 0.1-second increments.

Full size image

Inguinal lymph nodes

Similar to the axillary lymph nodes experiments, Figs. 6 and 7 show the results for two different inguinal lymph nodes. The figures include B-mode, power Doppler and SURE density and velocity images acquired using the Verasonics research scanner, along with SURE density and velocity profiles. Power Doppler and color Doppler images from a commercial GE LOGIQ E9 scanner are shown for comparison. Again, the SURE images provide significantly enhanced visualization, revealing greater detail in the microvasculature compared to power Doppler and color Doppler images, while also enabling blood flow velocity estimation. For the first inguinal lymph node in Fig. 6, the estimated vessel diameters from the SURE density and velocity profiles are 70.7 and 66.2 \(\mu\)m, respectively. The estimated peak blood flow velocity of the selected vessel is 5.1 mm/s, as determined from the velocity profile. For the second inguinal lymph node in Fig. 7, the estimated vessel diameters are 41.2 \(\mu\)m in the density profile and 33.8 \(\mu\)m in the velocity profile. The estimated peak velocity of the selected vessel is 8.7 mm/s, as determined from the velocity profile. All vessel diameters are less than half the wavelength.

Fig. 6
figure 6

Results for the first normal inguinal lymph node. (a and b) B-mode and power Doppler images acquired using the Verasonics research scanner, with the color bar indicating the dB range. (c and d) Power Doppler and color Doppler images from a commercial GE LOGIQ E9 ultrasound scanner. (e and f) The SURE density and velocity images, where the color wheel indicates the direction of blood flow. (g and h) The SURE density and velocity profiles of the same vessel, highlighted in the SURE density and velocity images with a white line.

Full size image
Fig. 7
figure 7

Results for the second normal inguinal lymph node. (a and b) B-mode and power Doppler images acquired using the Verasonics research scanner, with the color bar indicating the dB range. (c and d) Power Doppler and color Doppler images from a commercial GE LOGIQ E9 ultrasound scanner. (e and f) The SURE density and velocity images, where the color wheel indicates the direction of blood flow. (g and h) The SURE density and velocity profiles of the same vessel, highlighted in the SURE density and velocity images with a white line.

Full size image

Comparison of SURE with ULM and micro-CT

In this study, microbubble injection in human lymph nodes was not permitted due to permission constraints, which precluded a direct comparison of SURE and ULM in human subjects. Therefore, the comparison of SURE with ULM and micro-CT was performed using a rat kidney. Fig. 8 compares the rat kidney microvasculature obtained using SURE from 2 seconds of data, ULM from 10 minutes of data, and ex-vivo micro-CT image from 2 hours and 11 minutes of scanning. Visual inspection shows that all three modalities capture highly similar vascular structures, with well-matched patterns of vessel branching and fine structure details. This indicates that SURE reliably reproduces the microvasculature, comparable to the established ULM and micro-CT techniques. Magnified images of selected regions from ULM, SURE, and micro-CT are shown in Fig. 8 (e, f, and g), further emphasizing the similarity among the modalities. The estimated vessel sizes along the green line are 105, 96, 54, and 70 \(\mu\)m for ULM; 92, 95, 61, and 75 \(\mu\)m for SURE; and 79, 65, 63, and 68 \(\mu\)m for micro-CT. The signal-to-noise ratio (SNR) plot in Fig. 8 (d) indicates that SURE consistently achieves higher SNR values, typically ranging from 25 to 35 dB, while ULM shows slightly lower SNR values in the range of 15 to 25 dB. The reduced SNR in ULM is attributed to the lower transmit voltage required to prevent microbubble destruction. For the SNR calculation, the mean of 20 motion corrected frames was considered the signal, while the noise was defined as the difference between a single frame and the mean of all frames. In addition, Fig. 9 presents the B-mode image and the corresponding SNR profile of the first axillary lymph node. In the SNR plot, a gradual decrease from approximately 35 dB near the surface to about 20 dB at deeper regions is observed.

Fig. 8
figure 8

Results of SURE, ULM, micro-CT of a rat kidney. (a) The ULM track density image acquired after 10 minutes of data collection. (b) The SURE density image obtained after 2 seconds. (c) Micro-CT image with a voxel size of 5 \(\mu\)m from 2 hours and 11 minutes of scanning. (d) SNR comparison between SURE and ULM. (e-g) Magnified views of the ULM, SURE, and micro-CT images. The estimated vessel sizes estimated along the green line are 105, 96, 54, and 70 \(\mu\)m for ULM; 92, 95, 61, and 75 \(\mu\)m for SURE; and 79, 65, 63, and 68 \(\mu\)m for micro-CT.

Full size image
Fig. 9
figure 9

B-mode image and signal-to-noise ratio (SNR) of the first normal axillary lymph node in Fig. 1. (a) The B-mode image of an axillary human lymph node from the Verasonics research scanner. (b) The SNR of the B-mode image to the depth of the image.

Full size image

Discussion

This study demonstrates the feasibility of visualizing the microvasculature of human lymph nodes using a fast contrast-free SRUS approach based on erythrocytes (SURE) and quantifying vessel diameters and blood flow direction and velocities. Vessels as small as 40\(\mu\)m in diameter and blood flow velocities ranging from 2 to 16 mm/s were successfully identified. The robustness of SURE was validated by repeatedly visualizing the same lymph node at different time points, demonstrating and confirming its consistency and reliability. This highlights the potential of this non-invasive image modality to characterize the lymph node microvasculature with just 2 seconds of data acquisition.

In comparison with the ULM studies of human lymph nodes13,20, ULM benefits from the high detectability of microbubbles, which provide a strong contrast between blood vessels and surrounding tissues, enabling precise delineation of vessel structures, and noise reduction. In contrast, SURE offers several advantages. Unlike ULM, it does not require the injection of contrast agents or microbubble density monitoring. Additionally, SURE requires only two seconds of data acquisition, allowing for fast follow-up scans that improve workflow and patient comfort. Compared to the contrast-free HDMI method21, SURE not only visualizes vessels with diameters smaller than half the wavelength but also provides images of microvascular flow direction and estimates of blood flow velocity.

Since ULM on human lymph nodes was not applicable in the current study due to permission constraints, a direct comparison of SURE with ULM in human lymph nodes was not possible. However, a comprehensive resolution comparison between SURE, ULM, and micro-CT in rat and rabbit kidneys was previously demonstrated25,26,29. Here, the SURE and ULM comparisons were conducted using the same ultrasound system, transducer, imaging position, and identical settings. The results showed that although ULM achieved slightly higher resolution, SURE attained a resolution below the ultrasound diffraction limit and produced vessel sizes and structures that closely matched those obtained via micro-CT and ULM. Additionally, for this paper, a separate study involving SURE, ULM, and micro-CT was performed on a different rat kidney, as illustrated in Fig. 8.

In the SNR profile in Fig. 9 for an axillary human lymph node, SNR decreases from \(\sim\)35 dB near the surface to \(\sim\)20 dB at deeper regions due to attenuation. The decline with increasing axial distance is due to acoustic attenuation and other propagation effects, which reduce the signal strength and thereby lower the SNR at greater depths. At the time of this study, permission to perform ULM on human lymph nodes was not granted due to clinical constraints regarding microbubble injection. Therefore, the SNR comparison between SURE and ULM was performed using a rat kidney. However, a direct comparison of SNR between SURE and ULM can be misleading because the two methods operate under different transmit conditions and sequences. In ULM, the MI must remain low (typically < 0.3) to avoid destroying the microbubbles, which limits the transmit voltage. In contrast, SURE does not rely on microbubbles and can therefore use a higher MI (e.g., up to 1.9, which is within the FDA limit) and a higher transmit voltage. In the rat kindey, the SURE and ULM data were acquired using the same transducer, ultrasound system, and experimental conditions. However, the transmit voltage for SURE was 96 V, while for ULM it was 12 V, because using a higher voltage in ULM causes microbubble destruction and cavitation. This fundamental difference in acoustic power constraints should be taken into account when comparing the SNR of SURE and ULM. Consequently, ULM exhibited lower SNR values at all depths compared to SURE in the rat kidney, as shown in Fig. 8, despite microbubbles scattering more strongly. But, both SURE and ULM visualized the same vasculature to the same depth. Although erythrocytes have a lower scattering cross-section than microbubbles, resulting in lower raw signal amplitude, SURE compensates by using higher transmit voltage and MI, along with advanced signal processing (e.g., SVD filtering and accurate motion compensation). Animal studies in rat and rabbit kidneys25,26,29 show that SURE visualizes the same micovasculature to ULM and micro-CT at depths up to 20–25 mm. While our 10 MHz probe offers high resolution, deeper imaging might require a lower-frequency probe (e.g., 5–7 MHz), trading off some spatial resolution. Ongoing work aims to develop SURE across various frequency ranges and transducer designs to further enhance clinical applicability. Also, comparative studies in animal studies demonstrate that, although both SURE and ULM encounter challenges in accurately detecting very low velocities (e.g., below 2 mm/s), SURE achieves the same performance as ULM for detecting microvascular flow ranges26.

For translating SURE into a clinical setting, the implementation of 3D imaging and real-time processing is crucial30,31,32. Capturing the entire microvasculature of a lymph node in 3D is essential for a full assessment and for detecting micrometastases. In addition, real-time imaging further enhances SURE’s practicality and clinical utility31. Furthermore, developing automated analysis algorithms is key to effectively quantifying microvascular parameters, as these algorithms would enable immediate characterization of the microvasculature, thereby increasing clinical applicability. Currently, a MATLAB-based graphical user interface (GUI) has been developed for data acquisition and processing. This GUI automates critical steps–including data acquisition, beamforming, motion estimation, clutter filtering, detection thresholding, and tracking–with default parameter settings (e.g., SVD cutoff values, detection threshold, and tracking parameters) optimized based on extensive animal and human pilot studies, while also allowing user adjustments as needed. These tools significantly reduce the complexity of scanning and data processing, enabling clinicians and sonographers to perform SURE imaging without requiring in-depth programming knowledge.

There are some other limitations to this study. The feasibility study included a small sample size. Larger-scale studies in diverse populations are necessary to characterize the microvasculature in normal, reactive, and various types of malignant lymph nodes. This could contribute to developing quantitative tools to differentiate between benign and malignant lymph nodes, potentially improving diagnostic accuracy and optimizing lymph node excision. Additionally, there is a need for further incorporation of quantitative microvascular parameters beyond vessel diameters and blood flow velocities. These parameters should assess the microvascular structure of lymph nodes and could include vessel density, tortuosity, degree of vascular disorganization, variations in flow velocities and presence of necrotic areas. These parameters have shown promising results for SRUS16,20,21,33.

This study scanned cervical, inguinal, and axillary lymph nodes, observing better imaging results in the axillary and inguinal regions than in the cervical region. The cervical region is more heterogeneous and contains more complex anatomy, including muscles, glands, and large blood vessels. Scanning in heterogeneous areas with significant acoustic impedance variations can affect image quality. The pulsation of the carotid artery contributes to motion artifacts, affecting image acquisition and analysis. The axillary and inguinal regions are less complex, allowing for better acoustic penetration. Further research is needed to improve image quality in the cervical region, especially to improve motion correction. This study highlights the feasibility of imaging the lymph node microvasculature with SURE. Visualizing and quantifying the microvasculature in normal lymph nodes is a foundation for further research into malignant lymph nodes, intending to develop quantitative tools to characterize the microvasculature and differentiate between benign and malignant lymph nodes.

Methods

This feasibility study was registered on clinicaltrial.com (Identifier: NCT05754814, March 3, 2023). Ethical approval was granted by the Danish Committee on Health Research Ethics (Identifier: 2204911). This study was conducted in accordance with the Declaration of Helsinki and the Council for International Organizations of Medical Sciences (CIOMS) international ethical guidelines and the applicable national guidelines, regulations, and legislation.

Participants

Healthy participants were recruited from public institutions. Inclusion criteria were age between 18 and 70 years, ability to cooperate during the scanning, capacity to give informed consent, being overtly healthy, and having a superficial normal lymph node. Exclusion criteria were pregnancy, any ongoing or recent (within the last four weeks) infectious disease (bacterial, viral, fungal, or protozoal) that may cause reactive lymph nodes or any disease or medication that induces lymphadenopathy. Written and verbal informed consent was obtained before scanning.

Ultrasound Scanning

All participants were scanned at the Center for Fast Ultrasound Imaging, Department of Health Technology, Technical University of Denmark. Initially, participants were scanned using a commercial GE Logic E9 ultrasound scanner with a GE L8-18i-D linear array Hockey stick transducer (GE HealthCare, Chicago, IL, USA). The axillary and inguinal region were scanned to locate suitable lymph nodes with a normal ultrasonic appearance. The specific region to be scanned was determined before the scanning session. B-mode, Color Doppler, and Power Doppler images of the selected lymph node were acquired in both the transverse and longitudinal planes. SURE scanning of the same lymph nodes was performed using a Verasonics Vantage 256 ultrasound system (Verasonics, Inc., Kirkland, WA, USA) and the same 168-channel GE L8-18i-D linear array Hockey stick transducer (GE HealthCare, Chicago, IL, USA), which was operated at 10 MHz. In the Verasonics research scanner, the pulse-width modulation transmit stage produced a center frequency of 9.29 MHz in tissue with a corresponding wavelength of 166 \(\mu\)m, as estimated from the mean frequency of the received signal’s power density spectrum25,26. An interleaved synthetic aperture (SA) sequence with 12 virtual sources with a frame rate of 208.3 Hz was used34. During scanning, participants were instructed to lie still and hold their breath. Scans with excessive motion were excluded from the study. The data acquisition using the SURE sequence captured 10 seconds of imaging data, but only two seconds were used to generate the SURE images. The remaining data were used to investigate the evolution and consistency of the SURE images.

SURE pipeline

The SURE pipeline for lymph nodes is depicted in Fig. 10. The beamforming of stored data was achieved using GPU-accelerated code implemented in Matlab 2021b (MathWorks, Massachusetts, USA)35, utilizing recursive SA processing26,36,37,38. Tissue motion was estimated and corrected using transverse oscillation25. Stationary tissue signals were removed using singular value decomposition, with cutoff values of 18% for low singular values (representing tissue) and 75% for high singular values (representing noise) at an erythrocyte detection threshold of 35 dB. The positions of local peaks corresponding to erythrocytes in the remaining signal were extracted to generate the SURE density image. The SURE density image was then constructed by accumulating the detected erythrocyte peaks. A recursive nearest-neighbor (NN) tracking method was employed to track the erythrocyte peaks, yielding microvascular flow images and velocity profile estimations26. The NN tracker was applied 12 times with a 12-frame interval in recursive imaging, following the recursive NN tracking method strategy. To generate the velocity map and profile, all erythrocyte peak trajectories were accumulated into an image. Velocity maps were visualized by scaling and mapping velocity vectors to an RGB color wheel, where hue represents flow direction and brightness indicates velocity magnitude. Velocity profiles were extracted from the velocity images by measuring the velocity along a defined line. The estimated velocity was adjusted by a factor of 1.33, as suggested by the theoretical framework in39. According to this model, the peak velocity of a three-dimensional parabolic velocity profile is underestimated by up to 33% in two-dimensional SRUS imaging when the vessel diameter is smaller than the elevation thickness.

Fig. 10
figure 10

Stages of the SURE imaging pipeline for human lymph node processing. The scanner image is used with permission from Verasonics, and the lymph node illustration was created using BioRender.com.

Full size image

ULM and micro-CT

In this study, due to permission constraints, ULM imaging could not be performed on human lymph nodes. Therefore, ULM and micro-CT imaging of a Sprague-Dawley rat kidney were conducted for a multi-modality comparison to validate the SURE images25,26,27,28,29,40. The animal experiment was conducted at the Department of Biomedical Sciences, University of Copenhagen, under experimental license number 2020-15-0201–00547, in accordance with the Danish Animal Experiments Inspectorate25. ULM imaging was performed for 10 minutes using a 1:5 microbubble dilution and an infusion rate of 55 \(\mu\)l/min, with data acquired at a frame rate of 150 Hz. An amplitude modulation synthetic aperture sequence with 3\(\times\)3 virtual sources was employed of for each virtual source, three emissions were performed using all elements, even elements, and odd elements, to obtain full and half amplitude emissions. Summing signals from these three emissions gives the non-linear bubble response. SURE and ULM imaging were conducted using the same GE L8-18iD transducer at the same position on the rat kidney. Additionally, two seconds of SURE imaging data, using the same sequence as that employed for the human lymph node, were acquired from the same kidney. The transmit voltage for SURE was 96 V, while for ULM it was 12 V. The ULM pipeline for generating track density images used a hierarchical Kalman tracker10. The kidney was prepared using silicone-based intravascular contrast injection to clearly delineate blood vessels40,41. Imaging was performed with a high-flux, laboratory-based phase-contrast micro-CT scanner (Exciscope Polaris), allowing voxel resolutions on the order of 5 \(\mu\)m. The rat kidney was scanned for 2 hours and 11 minutes to produce the micro-CT image.