Quantification of Tumour Angiogenesis and Perfusion using Contrast-free Super-Resolution Ultrasound Imaging

Quantification of Tumour Angiogenesis and Perfusion using Contrast-free Super-Resolution Ultrasound Imaging | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Quantification of Tumour Angiogenesis and Perfusion using Contrast-free Super-Resolution Ultrasound Imaging Amy McDermott, Iman Taghavi, Anne Skovsbo Clausen, Lauge Naur Hansen, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6829687/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction. Angiogenesis and blood perfusion are key components of tumour development, which vary between tumour types and patients. Current clinical imaging methods have limited ability to qualitatively or quantitatively assess these parameters. This study describes a novel application of the ultrasound-based imaging modality, super-resolution ultrasound imaging using erythrocytes (SURE), and evaluates its performance against available in vivo and ex vivo imaging methods. Methods. An optimised SURE imaging technique was used to visualise the vasculature in a syngeneic murine model bearing subcutaneous B16-F10 melanoma tumours. For comparison, the tumour was also imaged in vivo with B-mode, colour Doppler, spectral Doppler ultrasound, and ex vivo with micro-CT angiography and immunohistochemistry. Results. SURE imaged the tumour microvasculature, visualising vessels down to ≈ 30 µm in diameter while enabling extensive characterisation of tumour haemodynamics. SURE significantly outperformed colour Doppler and spectral Doppler in visualising vasculature and quantitatively assessing flow. SURE enabled analysis of the variability in perfusion across the tumour section and can enable users to evaluate the health of individual vessels. Vascular density data obtained from SURE did not significantly differ from that obtained by immunohistochemistry, demonstrating comparability even to ex vivo methods. Conclusions . Using only 10 seconds of data acquisition, SURE imaging can provide vascular density in murine tumours at a resolution comparable to ex vivo techniques, along with haemodynamic data across the entire tumour section, distinct regions, or individual vessels. This technique has the potential to identify previously unknown biomarkers of cancer, elucidating characteristics of neoplastic angiogenesis which can be used for early diagnosis and treatment. Super-Resolution Ultrasound Contrast-free Ultrasound SURE Oncology Imaging Microvasculature Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. INTRODUCTION Angiogenesis plays a critical role in tumour development, enabling the formation of new blood vessels and the required blood perfusion rates to sustain tumour growth [ 1 ]. This process is widely recognised as a hallmark of cancer, allowing tumours to obtain the oxygen and nutrients needed for proliferation and development [ 2 ]. The significance of angiogenesis in tumour development has been extensively documented, highlighting the process as a potential therapeutic target in cancer treatment [ 3 ]. Despite this, current non-invasive in vivo imaging techniques lack the resolution to monitor angiogenesis accurately or to analyse tumour haemodynamics quantitatively. Clinical imaging methods, such as ultrasound, contrast-enhanced magnetic resonance imaging and computed tomography, can only detect large blood vessels [ 4 ]. Consequently, while these techniques can be utilised to aid in surgical planning and treatment monitoring, they cannot fully characterise the tumours’ vascular architecture. In research settings, ex vivo techniques such as histology and micro-computed tomography (µCT) have a significantly higher capability to quantitatively assess the microvasculature and provide data to evaluate treatment responses and differentiate between tumour types [ 5 ]. However, these methods are unsuitable for in vivo use as they cause significant tissue destruction. Therefore, a substantial gap exists between the capabilities of in vivo and ex vivo microvascular imaging methods. The development of super-resolution ultrasound scanning (SRUS) techniques, such as ultrasound localisation microscopy (ULM), has enabled high-resolution imaging of the vascular system in many tissue types, including tumours [ 6 – 8 ]. The majority of these techniques utilise intravascular contrast agents, usually microbubbles, to achieve an imaging resolution beyond the traditional ultrasound diffraction limit [ 8 , 9 ]. While these techniques are non-damaging to tissues, representing a significant advancement for in vivo vascular imaging capabilities, the need for intravascular infusion and a long data acquisition time of 1–10 minutes, has limited the technique’s implementation in clinical and preclinical settings [ 10 ]. Advancements by Jensen et al. (2022) have overcome this challenge, achieving non-invasive, high-resolution vascular imaging by tracking the backscattering of moving erythrocytes instead of microbubbles [ 11 ]. This technique, SU per- R esolution ultrasound using E rythrocytes (SURE), has been documented to achieve an imaging resolution of 28µm, with only a few seconds of acquisition time, enabling near-real-time microvascular imaging [ 12 ]. Tracking erythrocytes instead of microbubbles enables SURE to overcome many technical limitations of other SRUS techniques. Microbubble infusion is inconvenient for patients, requiring bolus or constant-rate infusion; conversely, SURE is non-invasive, with erythrocytes abundant in the vasculature of all tissues [ 13 , 14 ]. Erythrocytes are unaffected by therapeutic levels of ultrasound acoustic pressure, whilst microbubbles can be significantly affected, limiting the usable mechanical index to prevent microbubble rupture [ 15 , 16 ]. Using a higher mechanical index in SURE allows deeper imaging depth, enabling broader applications and greater clinical transferability compared to ULM [ 17 ]. In research settings, microbubble infusion is not always possible, particularly in small animals such as mice with a small circulating blood volume and high susceptibility to infusion fluid overload [ 18 ]. This is especially relevant in oncology research, where murine models are the most frequently used animal model [ 19 , 20 ]. Currently, SURE imaging is not achievable in real-time due to the need for manual post-processing, which allows for careful image optimisation and analysis, but also introduces additional steps between acquisition and final data output. Two SURE semi-automated post-processing methods have been described which are components of the imaging pipeline: 1. Vascular density maps – enabling qualitative assessment of the microvascular architecture of a tissue [ 12 ]. 2. Velocity maps – to identify blood flow direction, velocity and vessel luminal diameter measurement [ 21 ]. Both methods have been extensively validated, and the anatomical accuracy of the techniques has been validated through comparisons between microvascular imaging of select regions of the rat kidney with both contrast-enhanced µCT and ULM [ 11 , 12 , 22 ]. The techniques’ translatability to qualitatively image other species and other tissue types has been demonstrated in the rabbit kidney [ 23 ] and the human lymph node [ 24 ]. This study describes how the SURE technique can be optimised for qualitative vascular imaging while assessing its ability to analyse subcutaneous tumour haemodynamics. We will determine the technique’s susceptibility to variability between individuals and imaging planes. We hypothesise that the signal to noise ratio (SNR) of SURE scans may be impacted by biological variation in vasculature, such as vessel size (smaller vessels give weaker signals), blood velocity (too fast or too slow can degrade signal), or volumetric flow rates (which affects the number of erythrocyte scatterers contributing to the image). We will compare SURE's qualitative vascular imaging capabilities to the best available non-invasive in vivo techniques, colour Doppler and power Doppler. We also assess SURE's ability to analyse tumour angiogenesis quantitatively, comparing the data collected to gold-standard ex vivo microvascular imaging techniques: ex vivo immunohistochemistry and contrast-enhanced µCT. 2. METHODS 2.1 Animal Studies 2.1.1 Ethical Approval All animal experiments were conducted in accordance with Directive 2010/63/EU of the European Parliament and Council on the protection of animals used for scientific purposes and were approved by the Danish Animal Experiments Inspectorate under license no. 2021-15-0201-01041 (approved December 2021). As this study did not include human participants, Ethics, Consent to Participate, and Consent to Publish declarations are not applicable. 2.1.2 Study Population Six-week-old female C57Bl/6JRj mice (n=5) (Janiver labs, France) were used in this study. They were acclimated for one week upon arrival and housed in an individual, ventilated cage with a light-to-dark period of 12:12 hours under controlled environmental conditions. Access to fresh water and standard pellet diet was provided ad libitum. Mice were anaesthetised (3-4% sevoflurane in 65% N 2 and 35% O 2 ) and inoculated with the melanoma cell line B16-F10 (ATCC® CRL-6475™) via subcutaneous injection in the right flank with 1x10 6 cells in 100 µL of sterile phosphate-buffered saline per mouse. Tumours were allowed to grow for 20 days and were measured using a calliper. The tumour volume was estimated using the formula: volume = ½ (length × width 2 ). 2.1.3 In Vivo Experiments Mice were anaesthetised (2% isoflurane in 65% N 2 and 35% O 2 ) and placed in left lateral or ventral recumbency and immobilised using tape to minimise movement in the tumour tissue from respiration. A probe holder positioned a 168-channel GE L8-18iD Hockey stick linear array probe (GE HealthCare, USA) 1 cm away from the skin’s surface, with ultrasound gel providing the interface. The probe, with a wavelength of 154 µm, was operated at a transmission frequency of 10 MHz. Three imaging planes were randomly selected along the tumour’s long axis; no preview of the imaging region was performed to prevent operator bias. Each imaging plane was scanned first with SURE, then with B-mode and colour Doppler. Spectral Doppler was applied to areas highlighted by Colour Doppler to confirm pulsatile blood flow and obtain quantitative velocity measurements. All conventional ultrasound scans were collected using the same probe and a commercial GE LOGIQ E9 scanner at the exact location, ensured using a mechanical probe holder. For details of SURE imaging, see the Methods section 2.2. The animal was euthanised via cervical dislocation. 2.1.4 Post-Mortem Tissue Collection Immediately following euthanasia, the mice underwent intracardiac vascular filling with the µCT contrast agent Vascupaint (yellow colloidal bismuth suspension MDL-121, MediLumine, Montreal, Quebec, Canada). Vascupaint was reconstituted with the following proportions: 1 ml silicone, 1 ml dilutant, and 0.1 ml catalyst. Following post-mortem sternotomy, a blunted 20G needle was clamped within the left ventricle and the inferior vena cava was severed. The vascular system was flushed with heparin saline, 3 ml of 2% formalin, then 3 ml of reconstituted Vascupaint solution. All infusions were completed using a syringe pump at the rate of 1 ml/min unless it became blocked, in which case a manual injection was completed. The whole animal was stored at 5 °C for 24 hours post-mortem. The tumour tissue was then excised and fixed in formalin for 24 hours before being embedded in paraffin. 2.2. SURE Imaging 2.2.1 Data Collection SURE data was acquired using a Verasonics Vantage 256 scanner (Verasonics, Inc., Kirkland, WA, USA) and the aforementioned GE L8-18iD linear array probe. Data was collected using the SURE pulse inversion sequence, which employs a synthetic aperture scan technique to enhance spatial resolution by using multiple virtual sources to simulate a larger effective aperture [25]. This sequence utilised 24 emissions from 2×12 virtual sources, improving image reconstruction and target visualisation. The acquisition duration was 1 minute, with a transmission voltage of 95 V. Data were collected using both a manual and automated time gain compensation (TGC) settings to optimise signal quality—the automatic programme aimed to generate a smooth TGC curve, which prevented signal clipping within the data. Manual TGC generally yielded the best results; therefore, unless specified otherwise, the images presented in this paper were acquired using this setting. TGC adjustments, automated or manual, were made to optimise the image data by minimising signal clipping in the data. A single operator completed all manual adjustments to prevent operator-induced bias. 2.2.2 Data Processing - Vascular Density Maps The SURE processing method used was consistent with that described by Jensen et al. [12]. The key steps of the published SURE imaging pipeline are beamforming, motion estimation and correction with transverse oscillation, singular valve decomposition (SVD) for stationary echo cancelling, erythrocyte peak detection, density image formation and then velocity image formation. The following modifications were made to optimise the technique for oncology applications. To address clipping artefacts, clipped values in the raw radiofrequency data are identified and set to zero across all frames as a preprocessing step before SURE pipeline processing. A key refinement is selecting elements for removal based on beamformed clipped data rather than raw RF data alone. Elements contributing to clipping within the beamformed region are excluded, while those outside the region of interest (ROI) are ignored to avoid unnecessary data loss. However, excessive element exclusion was observed when clipping extended into the lower part of the image. To address this, we refine the selection criteria to remove elements only when their contribution degrades the beamformed region, preventing unnecessary signal loss. Following this, a ten-second segment of acquired data with the lowest motion was isolated from the full one-minute of data and processed. 2.2.3 Quantifying Angiogenesis SURE Vascular density maps were imported as RGB images (with varying dynamic ranges) into ImageJ (version 2.9.0/1.53t) (Fig 1a). To calculate the vascular density of the tumour region, the image was first converted to an 8-bit grayscale and binarised based on its dynamic range (Fig 1b). Vascular thresholding was performed manually based on the image’s dynamic range to isolate vasculature while excluding background signal (Fig 1c). When the background signal varies significantly across different image regions, it was divided into two sections to allow for separate vascular thresholding, accounting for the differences in signal intensity. Vascular thresholding was applied independently to each section before the binarised images were recombined into a single image. The inter-rater variability of this method has not been assessed, nor its degree of subjectivity to differences in image intensity. To minimise variability, the same individual completed thresholding for all images. The tumour region was selected manually using the ROI tool (Fig 1c). The ROI excluded tumour regions where imaging artefacts or background signal obscured the vasculature. The ROI also excluded the tumour margins as the tumour capsule is not clearly defined in SURE images; consequently, discerning the border between vasculature within the skin, subcutaneous tissue, and tumour is difficult. The pixels of thresholded vessels vs. background were measured within the specified ROI to quantify vascular density. 2.2.4 Data Processing - Velocity Maps One scan per individual mouse was selected for velocity map processing. The scan was selected based on the imaging plane demonstrating the highest detectable vascularity on both SURE and colour Doppler. From the total acquired data, a 1–1.5 second data segment corresponding to a low-motion period between respiratory cycles was utilised. This was then processed using a modified version of the method described by Naji et al. [21], adapted to include the use of a recursive nearest neighbour tracker. 2.2.5 Quantifying Haemodynamics Quantitative data were collected from multiple vessels in each velocity map to analyse the haemodynamic variation in different vessels across one tumour imaging plane and between tumours of other individuals. For each vascular region identified in a scan, the blood velocity and vessel lumen diameters were estimated using the published method described by Naji et al. [21]. Velocity magnitudes were extracted along a user-defined line oriented in both the lateral and axial directions. Approximately 50 cross-sectional slices were sampled across the vessel of interest from this. These individual profiles were then averaged to generate a raw data velocity curve (shown in blue). A parabolic function (red curve) was fitted to the averaged data to reflect the expected laminar flow distribution in cylindrical vessels, where velocity is maximal at the center and diminishes toward the periphery. Vessel diameter was determined using the Full Width at Half Maximum, calculated as the distance across the profile at which the velocity reaches 50% of its maximum value. Further characterisation of haemodynamics is performed by calculation of the estimated blood flow rate (volumetric flow rate) for each vessel using the following formula: Blood flow rate ( μL/s) = (blood velocity (mm/s) · vessel area (µm²) ) / 10 9 . Vessel area ( µm²) = (π · Diameter 2 (µm)) / 4. 2.3 Comparison to Other Vascular Imaging Techniques 2.3.1 Conventional Ultrasound One SURE image per mouse was qualitatively compared with colour Doppler. The scan was selected based on the imaging plane demonstrating the highest detectable vascularity on both SURE and conventional ultrasound imaging. 2.3.2 Contrast-Enhanced µCT Formalin-fixed paraffin-embedded (FFPE) samples were scanned with a Exciscope Polaris micro-CT scanner (Exciscope AB, Sweden), with an isotropic voxel size ranging from 4 µm to 5 µm. Scans were reconstructed using Exciscope’s own software and all analyses were performed using ITK-SNAP software (version 3.6.0) and MATLB (The MathWorks Inc., Natick, Massachusets) (Supplementary Fig 1A). A 3-dimensional ROI was manually created using the ITK-SNAP polygon function to isolate the tumour tissue from surrounding soft tissue (Supplementary Fig 1B). Data analysis was performed within this ROI. The contrast agent within the vasculature was isolated using a manually set intensity threshold (Supplementary Fig 1C). This enabled a proxy vascular density measurement for the tumour volume and for consecutive 20µm 2D image planes. 2.3.3 Histology FFPE samples were sectioned along the longitudinal axis. Three distinct regions along this axis were selected for analysis. From each region, two slides were prepared using adjacent tissue sections. One slide was stained with hematoxylin and eosin (H&E), while the other underwent fluorescent immunohistochemistry, utilising the same primary Anti-CD31 antibody [EPR17259] (ab182981, Abcam) along with an Alexa Fluor® 568-conjugated secondary antibody (Fig 1D-F). H&E-stained slides were used to guide the identification of the tumour border for fluorescent immunohistochemistry vascular density measurement. Vascular density analysis on fluorescent immunohistochemistry images was performed on ImageJ, using the same protocol listed in section 2.2.3. Another slide per tumour underwent chromogenic immunohistochemistry to detect CD31 expression using the Anti-CD31 antibody [EPR17259] and a secondary anti-rabbit HRP polymer (Nordic Biosite) (Supplementary Fig 2). Chromogenic staining enabled the qualitative assessment of contrast agent vascular filling. Whole-section slide imaging was performed using the AxioScan Z1 (ZEISS) at x10 magnification. 2.4. Statistical Analysis All data analysed within this study were examined for normality of distribution using the Shapiro-Wilk normality test. When analysing differences between tumours, where data is not normally distributed, the Kruskal-Wallis test with Dunn’s post-hoc test was used. When analysing differences between tumours where data is normally distributed, a one-way ANOVA with Tukey’s post-hoc test is used. The Wilcoxon signed-rank test was used to analyse the difference in data collected from the same tumour. Throughout the study, a p-value of < 0.05 was considered statistically significant. All statistical analyses were conducted using RStudio. Graphical representations of the data were displayed with significance indicated by asterisks: p < 0.05 was denoted by a single asterisk (*), p < 0.01 by two asterisks (**), and p < 0.001 by three asterisks (***). 3. RESULTS 3.1. Qualitative Vascular Imaging Fifteen SURE images were collected from five tumours (Fig 2). This naming system for the five individuals will be consistent throughout the paper. All SURE images showed detectable vasculature. Significant variations in image SNR were seen between scans collected from one individual and across the dataset. A high background signal in the deepest half of the tumour (beyond 10 mm) was observed in 60% of scans, limiting reliable assessment of this region. A high frequency of imaging artefacts was also seen, with 66% of total scans being impacted by a banding signal artefact, obscuring potential vascular imaging in the affected image regions. For all imaging regions compared, SURE images demonstrate superior vascular imaging compared to colour Doppler (Fig 3). Regions where blood flow was detected on colour Doppler vessels were all identified as containing vessels in SURE. SURE scans successfully imaged individual vessels and imaging of vascular in tumour regions not detected by Doppler. Of the compared scans, no ultrasound artefacts or variations in signal across the tumour tissue were seen in B-mode imaging. Three SURE scans were affected by banding artefacts (Fig 3B, 3C, and 3E), and high background signal in the deeper portion tissue impacted two tumour scans (Fig 3C and 3E). 3.2. Analysis of Individual Vessels The high resolution of SURE scans enabled detection of individual vessels, allowing the user to distinguish between neighbouring vessels in vascular tumour regions. This enabled vessels to be individually selected for secondary processing with SURE velocity mapping (Fig 4). A degree of vascular resolution was lost during secondary processing due to higher SNR thresholding (as illustrated by Fig 4A). Variation was seen in the number of measurements made between each scan, due to differences in vascularity across the imaging plane. For tumour A, 10 measurements were made; for tumour B, 18; tumour C, 18; tumour D, 9; and tumour E, 12. Information regarding the vessel lumen diameter, velocity, and flow direction was successfully obtained for all vessels retained during the velocity map processing step. Where velocity measurement across the vessel 3.3. Haemodynamic Characterisation of the Tumour As multiple vessels could be analysed per SURE image, we were able to provide an analysis of the variation in vessel characteristics and haemodynamics both within one tumour and between individuals (Fig 5). This could not be achieved using colour Doppler due to an inability to distinguish individual vessels or spectral Doppler, which could not measure blood flow rates in any scans. 3.3.1. Blood Velocity Across all measurements from all tumours,the mean blood velocity was 2.75 mm/s, with a median of 2.4 mm/s and a standard deviation of 1.17 mm/s. The range of velocities spans from 1.2 mm/s (minimum) to 7.5 mm/s (maximum), with an interquartile range of 1.1 mm/s.Velocity measurements varied significantly between tumours and were not normally distributed in all cases. The Shapiro-Wilk normality tests revealed that the blood velocity data for tumours A, B, D, and E follow a normal distribution. However, the velocity data for tumour C significantly deviates from normality ( p = 0.00196**) . A Kruskal–Wallis test revealed a statistically significant difference in velocity distributions across tumours (χ²(4) = 11.65, p = 0.020*) (Fig 5A). Post-hoc pairwise comparisons using Dunn’s test with Bonferroni correction identified a significant difference between tumours B and E ( p = 0.024*). No other pairwise comparisons reached statistical significance following correction. 3.3.2. Blood Vessel Luminal Diameter Across all measurements in this study, the mean lumen size was 121.31 µm. The median lumen size was 115.8 µm, and the standard deviation was 40.36 µm. The minimum lumen size recorded was 52.8 µm, while the maximum was 251.1 µm. The interquartile range (IQR) was 49.4 µm. The Shapiro-Wilks test showed that all tumours had normally distributed luminal diameter measurements. A one-way ANOVA indicates significant vessel size differences between the tumours (F(4, 72) = 4.118, p = 0.00463**) (Fig 5B). A post hoc Tukey's Honest Significant Difference (HSD) test was conducted, identifying significantly higher luminal diameter in Tumour E compared to tumour B ( p = 0.0068*), with a mean difference of 42.63 µm. No other significant pairwise differences were found. 3.3.3. Blood Flow Rate The blood flow rate measurements across the entire dataset had a mean of 3.63×10⁻⁵ µL/s, a median of 2.61×10⁻⁵ µL/s, and a standard deviation of 3.18×10⁻⁵ µL/s. The flow rates range from 3.97×10⁻⁶ µL/s (minimum) to 1.707×10⁻ 4 µL/s (maximum), with an interquartile range of 2.86×10⁻⁵ µL/s. The Shapiro-Wilk tests show that flow rate measurements from tumours A, B, and D are normally distributed, while data from tumours C ( p = 0.0429*) and E ( p = 0.00418**) are not normally distributed. Kruskal-Wallis test results indicate a significant difference in blood flow rates across tumour types (χ²(4) = 20.88, p = 0.000340***) (Fig 5C). Post-hoc pairwise comparisons using Dunn’s test with Bonferroni correction identified a significantly higher blood flow rate in tumour E compared to tumour A (0.0128*) and tumour B ( p = 0.000384***). No other pairwise comparisons reached statistical significance following correction. 3.4. Quantifying Angiogenesis 3.4.1. SURE Vascular Density Analysis We analysed tumour vascular density measurements using SURE across the five tumours with three regional measurements per tumour (Fig 6A). Tumour C had the highest mean vascular density at 7.73%, but the lowest variability (SD = 0.99%, IQR = 0.98%, coefficient of variation (CV) = 0.128), suggesting a consistent vascular density across regions. Followed by Tumour D with a mean vascular density of 5.54% (SD = 1.39%, IQR = 1.37%, CV = 0.251), indicating moderate variability. Tumours A, B and E had similar mean densities. Tumour A had a mean vascular density of 2.49% (SD = 0.51%, IQR = 0.51%, CV = 0.205), with moderate variability. Tumour B showed a mean of 3.30% (SD = 1.46% , IQR = 1.27% , CV = 0.443) , showing greater relative variability. Tumour E exhibited the greatest variability, with a mean vascular density of 3.03% (SD = 1.96% , IQR = 1.93% , CV = 0.644) , suggesting substantial heterogeneity across its regions. To statistically assess whether the variability in vascular density differed significantly between the tumors, a Fligner-Killeen test of homogeneity of variances was performed. The test yielded a chi-squared value of 1.4847 with 4 degrees of freedom, resulting in a non-significant p-value of 0.8293, demonstrating that there is no evidence of significant differences in variability among the tumors. 3.4.2. Comparison to Ex vivo methods Immunohistochemistry data from the whole dataset, 15 slides from 5 individuals, showed a mean vascular density of 3.00% and a standard deviation of 0.66% , indicating relatively low variability. The values ranged from a minimum of 2.01% to a maximum of 4.07%. The median was 2.84% , and the interquartile range (IQR) was 0.90% , showing that the central 50% of the data were tightly clustered. A Wilcoxon signed-rank test assessed whether vascular density measurements vary significantly between the SURE and histology methods within one tumour. There were no statistically significant differences between the results of the two measurement methods for any tumour (Fig 6A). Vascular density measured using µCT demonstrated substantial variability for different planes within a tumour (Fig 6B). There is also variation in average vascular densities between the five tumours. Tumours A, D, and E had vasculature present throughout the tumour. While the vascular density dropped close to 0.5% for large regions of Tumour B. Both Tumour A and Tumour B showed relatively low median values (1.17% and 0.644%) and interquartile ranges (IQR = 1.78% and 2.07%), with maximum vascular density measurements of 7.59% and 7.49%, respectively, indicating limited vascularisation. In contrast, Tumour D exhibited a markedly higher vascular density (median = 14.4%), with a wide distribution (IQR = 14.4%) and a maximum SURE value of 28.9%, reflecting substantial heterogeneity. Tumour E presented intermediate values (median = 4.18%, and IQR = 3.45%), with vascular density ranging from 0% to 19.8%. However we are limited in our ability to compare µCT density data with other methods due to incomplete and highly irregular contrast filling which was identified via qualitative assessment of chromogenic immunohistochemistry. Irregular Vascupaint filling was seen within all analysed tumours (Supplementary Fig 2). The contrast agent filling of tumour A could not be assessed due to damage during slide preparation. Tumour C was the only tumour with moderate erythrocyte extravasation throughout the tissue. Due to the irregular contrast agent filling and significant tissue distortion induced during the excision and fixation process, co-registration of individual vessels between SURE and µCT scans was impossible. Despite these limitations, vascular density measurements collected with SURE and fluorescent immunohistochemistry methods fell within the range of measured µCT densities. 4. DISCUSSION This study demonstrates that SURE, with its combination of high-resolution vascular imaging and quantitative haemodynamic analysis, significantly advances non-invasive in vivo vascular assessment capabilities in oncology research. SURE enables visualisation of individual vessels at a higher resolution than achieved by colour Doppler, more consistently than seen with contrast-enhanced µCT, and with vascular density data comparable to immunohistochemistry. SURE exceeds current capabilities of other super-resolution techniques, overcoming the need for intravascular contrast-agent use and significantly reducing data acquisition time to seconds. In this dataset, where tumours are newly formed, and vessels are small with low flow rates, conventional techniques have minimal usability, demonstrated by spectral Doppler’s inability to detect flow rates. Conversely, SURE successfully imaged individual vessels and enabled haemodynamic assessment in vessels as small as 50 µm. The greater vascular resolution achieved with SURE shows the technique to have considerable application possibilities in clinical settings, where the ability to assess a tumour’s vascular architecture quickly may have both diagnostic and prognostic value. The ability to reliably image the vasculature, while also performing haemodynamic analysis on the entire tissue or vessels of choice, gives SURE a unique advantage over other super-resolution techniques and varied applications in oncology. Variability in SURE Images As with other ultrasound-based techniques, each SURE scan must be collected with optimal data-collection parameters to overcome variations in signal intensity between different imaging planes. To address these variations, the operator tailors Time Gain Compensation (TGC) settings for each scan, adjusting for factors such as tissue depth and acoustic properties to ensure consistent image quality. SURE ultrasound sequences are otherwise consistent across participants, enabling the collection of comparable data. This study demonstrates that while SURE exceeds the vascular imaging capabilities of existing ultrasound techniques (Fig 3), it remains susceptible to many challenges associated with ultrasound-based technologies. In this dataset, there is a significant issue with high background signal in deeper portions of the image, which is unrelated to the image’s dynamic range (dB). This issue persisted with both manual and automated TGC settings. This is known to be related to ultrasound wave attenuation and divergence at depth, where signals weaken relative to system noise, high-frequency components are lost, and TGC cannot fully restore signal quality without amplifying noise, resulting in a drop-off in SNR [26]. However, in this study, we show that this impact is not uniform across a dataset of highly similar tumours or even consistent with different scan planes of the same tumour. In tumours C and E, the effect was present in all scans (Fig 2). Whilst it only affected some of the scans in others (Fig 2: tumours A and B). This study could not confirm or deny whether this effect is impacted by biological variation in the vasculature. There was limited variation in the biological parameters analysed in the study between the five tumours. Tumour E, which had significantly higher average velocities, vessel sizes and volumetric flow rates than tumour B, did not show better imaging quality, as expected. Manual vs automated TGC adjustment of images of the same tumour section did not impact the presence of this artefact, reducing the likelihood that TGC adjustment. Further research is needed to elucidate why these variations are seen and whether improvements in imaging strategies may address them. All tumours scanned in this study were affected by a significant imaging artefact: linear hyperechoic bands in the deep imaging section (Fig 2A, 2B, 2D, 2E, 2F, 2G, 2H, 2I, 2L, 2M, 2O). This artefact resembles the ‘incoherent clutter’ artefact, commonly seen in super-resolution imaging methods, which occurs for two reasons: off-axis scattering and complex tissue motion [27]. In this study, the primary cause for this artefact was suspected to be clipping within the data, with the horizontal bands occurring at depths corresponding to clipping occurrences. The effect is then distributed over the image during beamforming due to heterogeneous tissue motion from respiration or cardiac activity, introducing unpredictable signal fluctuations that standard SVD-based clutter filtering struggles to suppress. This artefact obscures weak blood flow signals and degrades the accuracy of vascular visualisation within tumours. These artefacts are known to be worse in the deeper imaging regions due to increased attenuation of the ultrasound signal and stronger off-axis scattering from surrounding tissues, which reduces signal-to-noise ratio and makes clutter harder to suppress [26]. As all scans were collected using methods designed to maximise signal without introducing data clipping, using either automated or manual TGC, the prevalence of this in the dataset represents a challenge. Highlighting that current signal amplification selection methods are insufficient for the selection of an appropriate TGC, and further improvements are needed. Haemodynamic Assessment using SURE Many ULM studies, utilising intravascular contrast agents, have successfully characterised vascular density in subcutaneous tumours within animal models [8, 28-30]. Some of these studies also gathered quantitative data on microbubble movement as a proxy for erythrocyte haemodynamics [31]. The appropriateness of using microbubbles as a direct proxy for vascular haemodynamics in newly formed tumour microvasculature is unknown and understudied. Consequently, SURE's ability to remove this variable and directly analyse the erythrocytes' haemodynamics is one significant advantage of the technique. Other groups have also achieved high-resolution imaging of the subcutaneous tumour vasculature without using a contrast agent [32, 33]. However, these techniques cannot provide quantitative analysis of the haemodynamics, limiting clinical applicability. Measuring blood velocity in tumours provides analysis of tissue vascular health and angiogenesis, with changes in vascular resistance or irregular flow patterns used to differentiate benign from malignant tumours [34]. Measuring volumetric flow enables assessment of tumour tissue perfusion. Insufficient flow is often seen in malignant tumours, where irregular, disorganised blood vessels result in hypoxia and necrosis, affecting fluid flow [35]. The ability to quantitatively analyse these haemodynamic parameters enables clinicians and researchers to characterise and assess the tumour tissue, allowing the monitoring of disease progression or response to treatments. Both colour and spectral Doppler are commonly used in clinical settings to assess tissue vascularity and perfusion. The techniques’ non-invasive and rapid qualitative assessment capabilities outweigh their limitations in microvascular analysis. While SURE’s current requirement for post-processing limits its real-time capabilities, further advancement in computing will likely result in comparable clinical usability. In this study, tumour blood velocity ranged from 1-5 mm/s (Fig 5), consistent with measurements achieved by other researchers using contrast-enhanced techniques [31]. This data confirms why conventional ultrasound techniques have limited function in these tissues. Colour Doppler and Spectral Doppler struggle to detect vessels smaller than 100 µm [36, 37]. They also require moderate flow rates (above 40 cm/s) to produce accurate images [38]. SURE measurements show that while a substantial proportion of the vasculature is larger than the required 100 µm to be detected by colour Doppler, all vessels had a velocity below that required for detection. Additionally, the spatial resolution of conventional Doppler (1–2 mm) is not sufficient for imaging the micron-sized vessels found in tumours [39]. SURE showed that haemodynamics can vary significantly between tumours of the same type, age, and clinical condition. We identified significantly higher blood velocities and vessel sizes in tumour E compared to tumour B, and significantly higher volumetric flow rates in tumour E compared to tumours A and B (Fig 5a). The ability to measure haemodynamics in different regions of one tumour scan also demonstrated that vascular architecture and haemodynamics can vary significantly between tumours of the same type. We identified that in two tumours, C and E, data from different regions of the tumour had very different blood velocities and flow rates, differing from the other tumours, where these parameters did not vary widely (Fig 5). The ability to characterise and sub-categorise tumour forms is a cornerstone of oncology prognostics. This finding is a prime example of how SURE can identify novel cancer biomarkers. Comparison to Ex Vivo Techniques Immunohistochemistry was utilised in this study to assess how SURE compares to an established ex vivo 2-dimensional imaging technique. We found no significant difference between vascular density datasets obtained with SURE or immunohistochemistry (Fig 6). Importantly, this comparison is inherently biased due to methodological differences. For example, immunohistochemistry typically reports lower vascular density than SURE, as CD31 staining labels only the vessel wall, whereas SURE visualises the vessel lumen. These differences contribute to discrepancies between datasets, which give a greater tolerance to SURE's lower vascular detection sensitivity. Despite this, the comparability of the datasets provides strong support for the use of SURE in research settings. To analyse tumour vascular heterogeneity more extensively, µCT was utilised. This method enabled the quantification of angiogenesis in slices of comparable thickness to the SURE imaging plane. Data showed an unexpectedly high degree of vascular heterogeneity, which did not align with immunohistochemistry results showing high vascular homogeneity in this tumour model (Fig 6). Qualitative assessment of chromogenic immunohistochemistry slides demonstrated that irregular contrast-agent filling of the vasculature impacted the µCT vascular density results (Supplementary Fig 2). While this means these data cannot be used to give a complete overview of the vascular density through the tumour, the data range represents a range of the minimal possible vascular density measurements. The vascular density measurements obtained using SURE and immunohistochemistry all fell within the range of possible vascular density values from µCT analysis. While all methods gave comparable vascular density data, which did not significantly differ, all were anatomically inaccurate for different reasons. Immunohistochemistry only quantifies the vessel's rim, measuring a lower density value than if the lumen were included. The µCT analysis quantifies the contrast agent as a proxy for the vasculature, subject to filling limitations. SURE images show significant variability as described above, which, combined with the subjective method of density analysis, reduces the accuracy of the data. Full co-registration of SURE scans to µCT volumes was impossible due to significant tissue distortion during excision and fixation. Consequently, the anatomical accuracy of SURE imaging of individual vessels in an image plane cannot be validated in this study. Study Strengths and Limitations As this is a novel method, no prior power analysis was performed due to the lack of available data on expected variability. A post-hoc power analysis showed the SURE vascular density comparison to have 72% power, demonstrating the study was moderately underpowered (G*Power; ANOVA approximation, f = 1.03, α = 0.05, N = 15, 5 groups). This may have limited the ability of this study to detect statistically significant differences between tumours. While limited to one tumour model and a small population size, this study provides critical pilot data which can inform future studies. The imaging model used in this work, involving small subcutaneous tumours, has many physical characteristics translatable to human pathology. Regarding measurement reliability, velocity and vessel size data have been validated in rat kidney models and ex vivo systems [12, 21]. Nevertheless, future studies should compare these measurements with in vivo data from larger tumour vessels and a wider range of tissues to confirm their accuracy and confirm their accuracy in clinical contexts. Future Directions Future research should prioritise methodological refinements to enhance data quality and imaging reliability. Key areas for improvement include reducing processing artefacts and increasing the signal-to-noise ratio (SNR), enabling clearer resolution of vascular structures in deeper regions of the tumour. Additionally, further investigation is needed to understand how experimental variables such as tumour size, tissue viability, and systemic factors like animal blood pressure affect image quality and consistency. Beyond technical advancements, future studies should explore the broader applications of this method for characterising tumour microvascular architecture. One promising direction is its potential use in distinguishing between benign and malignant tumour nodules, as well as differentiating between tumour classifications. The technique could also prove valuable for stratifying tumour subgroups, aiding in classification, and conducting longitudinal assessments to monitor disease progression or therapeutic response. Optimising this method for human applications is critical in furthering its translatability and expanding its impact in oncology. Conclusions This work has demonstrated that SURE can provide vascular density data at a resolution comparable to ex vivo techniques, together with haemodynamic quantitative data of significant value to clinical and research applications. While the method requires further processing optimisation to improve reliability and clinical usability, it can potentially identify previously unknown cancer biomarkers, elucidating characteristics of neoplastic angiogenesis which can be used for early diagnosis and treatment. Declarations Ethical Approval All animal experiments were conducted in accordance with Directive 2010/63/EU of the European Parliament and Council on the protection of animals used for scientific purposes and were approved by the Danish Animal Experiments Inspectorate under license no. 2021-15-0201-01041 (approved December 2021). As this study did not include human participants, Ethics, Consent to Participate, and Consent to Publish declarations are not applicable. Competing Interests The authors declare that they have no competing interests relevant to this research. Funding This work was funded by a Synergy grant from the European Research Council, project no. 854796. Author Contribution Author Contributions: A.M: Conceptualisation, Methodology (lead), Investigation (lead), Data Analysis, Writing - Original Draft, Data Visualisation (lead). I.T: Software - SURE Density Imaging, Investigation - In Vivo Data Collection, Data Curation, Writing - Review. A.S.C: Methodology, Investigation – provision of mouse model. L.N.H: Data Curation - micro-CT. Data Visualisation - Figure 6B. M.A.J: Software - SURE Velocity Imaging, Data Curation - velocity processing (joint with A.M.). Data Visualisation - Figure 4C. H.M.K: Software, Resources, Data Curation. A.K: Supervision. M.B.N: Supervision, Funding Acquisition. J.A.J: Supervision, Funding Acquisition, Software (lead). C.M.S: Supervision (lead), Methodology, Project Administration, Funding Acquisition, Writing - Review & Editing. All authors have read, reviewed and agreed to the published version of the manuscript. Data Availability Data is provided within the manuscript or supplementary information files. References De Palma, M., D. Biziato, and T.V. Petrova, Microenvironmental regulation of tumour angiogenesis. Nature Reviews Cancer, 2017. 17 (8): p. 457-474. Potente, M., H. Gerhardt, and P. Carmeliet, Basic and therapeutic aspects of angiogenesis. Cell, 2011. 146 (6): p. 873-87. Yadav, L., et al., Tumour Angiogenesis and Angiogenic Inhibitors: A Review. J Clin Diagn Res, 2015. 9 (6): p. Xe01-xe05. Jeswani, T. and A.R. Padhani, Imaging tumour angiogenesis. Cancer Imaging, 2005. 5 (1): p. 131-8. Pabst, A.M., et al., Imaging angiogenesis: Perspectives and opportunities in tumour research – A method display. Journal of Cranio-Maxillofacial Surgery, 2014. 42 (6): p. 915-923. Porte, C., et al., Ultrasound Localization Microscopy for Cancer Imaging. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2024. 71 (12: Breaking the Resolution Barrier in Ultrasound): p. 1785-1800. Couture, O., et al., Ultrasound Localization Microscopy and Super-Resolution: A State of the Art. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2018. 65 (8): p. 1304-1320. Dencks, S. and G. Schmitz, Ultrasound localization microscopy. Zeitschrift für Medizinische Physik, 2023. 33 (3): p. 292-308. Christensen-Jeffries, K., et al., Super-resolution Ultrasound Imaging. Ultrasound in Medicine & Biology, 2020. 46 (4): p. 865-891. Song, P., J.M. Rubin, and M.R. Lowerison, Super-resolution ultrasound microvascular imaging: Is it ready for clinical use? Zeitschrift für Medizinische Physik, 2023. 33 (3): p. 309-323. Jensen, J.A., et al. Fast super resolution ultrasound imaging using the erythrocytes . in Medical Imaging 2022: Ultrasonic Imaging and Tomography . 2022. SPIE. Arendt Jensen, J., et al., Super-Resolution Ultrasound Imaging Using the Erythrocytes-Part I: Density Images. IEEE Trans Ultrason Ferroelectr Freq Control, 2024. 71 (8): p. 925-944. Quaia, E., et al., Bolus versus continuous infusion of microbubble contrast agent for liver ultrasound by using an automatic power injector in humans: A pilot study. Journal of Clinical Ultrasound, 2016. 44 (3): p. 136-142. Helms, C.C., M.T. Gladwin, and D.B. Kim-Shapiro, Erythrocytes and Vascular Function: Oxygen and Nitric Oxide. Frontiers in Physiology, 2018. 9 . Wang, Q.X. and K. Manmi, Three dimensional microbubble dynamics near a wall subject to high intensity ultrasound. Physics of Fluids, 2014. 26 (3). Bezer, J.H., et al., Microbubble dynamics in brain microvessels. PLOS ONE, 2025. 20 (2): p. e0310425. Beaver, T.A., et al., Integrated Backscatter During Harmonic and Fundamental Frequency Imaging—Effect of Depth, Mechanical Index, and Tissue Anisotropy: Implications for Myocardial Tissue Characterization. Echocardiography, 2003. 20 (4): p. 337-343. Lichtenberger, M., Principles of shock and fluid therapy in special species. Seminars in Avian and Exotic Pet Medicine, 2004. 13 (3): p. 142-153. Gould, S.E., M.R. Junttila, and F.J. de Sauvage, Translational value of mouse models in oncology drug development. Nature Medicine, 2015. 21 (5): p. 431-439. Céspedes, M.V., et al., Mouse models in oncogenesis and cancer therapy. Clinical and Translational Oncology, 2006. 8 (5): p. 318-329. Naji, M.A., et al., Super-Resolution Ultrasound Imaging Using the Erythrocytes—Part II: Velocity Images. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2024. 71 (8): p. 945-959. Hansen, L.N., et al. Comparison of 2D SURE and 3D CT imaging of cortical vessels in a rat kidney . in 2023 IEEE International Ultrasonics Symposium (IUS) . 2023. Naji, M.A., et al., Transcutaneous Super-Resolution Ultrasound Imaging using Erythrocytes versus Microbubbles in a Rabbit Kidney . 2024: Proceedings of the 2024 IEEE Ultrasonics, Ferroelectrics, and Frequency Control Joint Symposium IEEE. 1-4. Naji, M.A., et al. Super-Resolution Ultrasound Imaging using Erythrocytes on an Axillary Human Lymph Node . in 2024 IEEE Ultrasonics, Ferroelectrics, and Frequency Control Joint Symposium (UFFC-JS) . 2024. Jensen, J.A., et al., Universal Synthetic Aperture Sequence for Anatomic, Functional, and Super Resolution Imaging. IEEE Trans Ultrason Ferroelectr Freq Control, 2023. 70 (7): p. 708-720. The British Medical Ultrasound Society. Guidelines for the safe use of diagnostic ultrasound equipment , in Diagnostic Ultrasound: Physics and Equipment , P.R. Hoskins, K. Martin, and A. Thrush, Editors. 2010, Cambridge University Press: Cambridge. p. 217-225. Huang, C., et al., Simultaneous Noise Suppression and Incoherent Artifact Reduction in Ultrafast Ultrasound Vascular Imaging. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2021. 68 (6): p. 2075-2085. Kasoji, S.K., et al., Early Assessment of Tumor Response to Radiation Therapy using High-Resolution Quantitative Microvascular Ultrasound Imaging. Theranostics, 2018. 8 (1): p. 156-168. Ghosh, D., et al. Monitoring early tumor response to vascular targeted therapy using super-resolution ultrasound imaging . in 2017 IEEE International Ultrasonics Symposium (IUS) . 2017. Hoyt, K., Super-Resolution Ultrasound Imaging for Monitoring the Therapeutic Efficacy of a Vascular Disrupting Agent in an Animal Model of Breast Cancer. Journal of Ultrasound in Medicine, 2024. 43 (6): p. 1099-1107. Yin, J., et al., Pattern recognition of microcirculation with super-resolution ultrasound imaging provides markers for early tumor response to anti-angiogenic therapy. Theranostics, 2024. 14 (3): p. 1312-1324. Huang, C., et al., Noninvasive Contrast-Free 3D Evaluation of Tumor Angiogenesis with Ultrasensitive Ultrasound Microvessel Imaging. Scientific Reports, 2019. 9 (1): p. 4907. Ternifi, R., et al., Quantitative Biomarkers for Cancer Detection Using Contrast-Free Ultrasound High-Definition Microvessel Imaging: Fractal Dimension, Murray’s Deviation, Bifurcation Angle & Spatial Vascularity Pattern. IEEE Transactions on Medical Imaging, 2021. 40 (12): p. 3891-3900. Tailor, A., et al., A comparison of intratumoural indices of blood flow velocity and impedance for the diagnosis of ovarian cancer. Ultrasound in Medicine & Biology, 1996. 22 (7): p. 837-843. Alamer, M. and X. Yun Xu, The influence of tumour vasculature on fluid flow in solid tumours: a mathematical modelling study. Biophys Rep, 2021. 7 (1): p. 35-54. Delorme, S., et al., Imaging the smallest tumor vessels using color Doppler ultrasound in an experiment. Radiologe, 2001. 41 (2): p. 168-72. Giavedoni, P., et al., Advanced Doppler Ultrasound Insights: A Multicenter Prospective Study on Healthy Skin. Diagnostics (Basel), 2025. 15 (5). von Bibra, H., et al., Limitations of flow detection by color Doppler: in vitro comparison to conventional Doppler. Echocardiography, 1991. 8 (6): p. 633-42. Fabiszewska, E., et al., Evaluation of Imaging Parameters of Ultrasound Scanners: Baseline for Future Testing. Pol J Radiol, 2017. 82 : p. 773-782. Additional Declarations No competing interests reported. Supplementary Files 0525SURETumourVascularImagingSupplementary.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6829687","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":476968689,"identity":"087e40a0-36d6-4430-91a3-4caf18a1e98a","order_by":0,"name":"Amy McDermott","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA70lEQVRIiWNgGAWjYPACZihdASISQDxmA2K0MDYcOEOyloNtRGjhn3bGgOlGjbWc+bTDxx9/nFeXuJ09gdmYh8HaGJcWids5Bsw5x9KNZW6nJTYc3HY4cWfPA+ZkHoZ0M5xOAmnJbTicOEM6xxCo5UDihhsJzId5GA7b4NIhD9VSD9Eyp46wFgOolgQJsJYGZrAWoMMO43SY4e20gsNAvxjOkE5LnHHm2GHjDWceNhvOMUjH6X2528kbH+fUWMtLSCcf+FBRUye74XjyYYk3FdaGDTj9z2FwAE2EEagYb0SyP8AnOwpGwSgYBaOAgQEATqxaoXqgMTkAAAAASUVORK5CYII=","orcid":"","institution":"University of Copenhagen","correspondingAuthor":true,"prefix":"","firstName":"Amy","middleName":"","lastName":"McDermott","suffix":""},{"id":476968691,"identity":"18765911-64b4-4437-82a9-937a08120fd2","order_by":1,"name":"Iman Taghavi","email":"","orcid":"","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Iman","middleName":"","lastName":"Taghavi","suffix":""},{"id":476968692,"identity":"98af8e3a-7d9f-45e9-a2cd-00b8c6c96c4f","order_by":2,"name":"Anne Skovsbo Clausen","email":"","orcid":"","institution":"University of Copenhagen","correspondingAuthor":false,"prefix":"","firstName":"Anne","middleName":"Skovsbo","lastName":"Clausen","suffix":""},{"id":476968693,"identity":"c31c5e6a-0442-4c2b-87a3-35ad76783cf8","order_by":3,"name":"Lauge Naur Hansen","email":"","orcid":"","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Lauge","middleName":"Naur","lastName":"Hansen","suffix":""},{"id":476968694,"identity":"7b3dea79-da4b-42dc-b741-759cf4771313","order_by":4,"name":"Mostafa Amin Naji","email":"","orcid":"","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Mostafa","middleName":"Amin","lastName":"Naji","suffix":""},{"id":476968695,"identity":"ef2867a2-f39a-43bc-86f1-d500ad1bec59","order_by":5,"name":"Hans Martin Kjer","email":"","orcid":"","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Hans","middleName":"Martin","lastName":"Kjer","suffix":""},{"id":476968696,"identity":"142e5122-03f2-4e81-baf9-5d43182459ca","order_by":6,"name":"Andreas Kjaer","email":"","orcid":"","institution":"University of Copenhagen","correspondingAuthor":false,"prefix":"","firstName":"Andreas","middleName":"","lastName":"Kjaer","suffix":""},{"id":476968697,"identity":"8cae2740-772f-4487-a9d0-cbd18a01a1b0","order_by":7,"name":"Michael Bachmann Nielsen","email":"","orcid":"","institution":"Rigshospitalet","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"Bachmann","lastName":"Nielsen","suffix":""},{"id":476968698,"identity":"521f0912-288e-4bab-adea-9cb0548def7f","order_by":8,"name":"Jørgen Arendt Jensen","email":"","orcid":"","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Jørgen","middleName":"Arendt","lastName":"Jensen","suffix":""},{"id":476968699,"identity":"bc9b71fa-c58e-43b4-bd9f-985184d00088","order_by":9,"name":"Charlotte Mehlin Sørensen","email":"","orcid":"","institution":"University of Copenhagen","correspondingAuthor":false,"prefix":"","firstName":"Charlotte","middleName":"Mehlin","lastName":"Sørensen","suffix":""}],"badges":[],"createdAt":"2025-06-05 13:38:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6829687/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6829687/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85649040,"identity":"ea67c611-466b-47d3-b73f-76809eaf644f","added_by":"auto","created_at":"2025-06-30 08:54:49","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2746439,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIllustrating the method of vascular density measurement for SURE (A-B) and immunohistochemistry (C-E) images. \u003c/strong\u003eThe SURE image (\u003cstrong\u003eA\u003c/strong\u003e), is converted from RGB to 8-bit grayscale and binarized based on signal intensity (\u003cstrong\u003eB\u003c/strong\u003e). The same process is completed on immunohistochemistry images (\u003cstrong\u003eC-D\u003c/strong\u003e). A magnified immunohistochemistry image section, outlined in orange, highlights the vascular resolution (\u003cstrong\u003eE\u003c/strong\u003e). The tumour region of interest (manually drawn in blue) excludes signal artefacts and high background signal regions (\u003cstrong\u003eB, D\u003c/strong\u003e). Images generated using ImageJ.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/bc269c1eb8ac5529ee5f83f8.png"},{"id":85649041,"identity":"97dd312a-c810-4961-a6f7-c93cc7711cf3","added_by":"auto","created_at":"2025-06-30 08:54:49","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2154297,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSURE scans in this dataset; three views of different imaging planes were per tumour. \u003c/strong\u003eEach row of panels displays the scans from each tumour: tumour A (\u003cstrong\u003eA-C\u003c/strong\u003e), tumour B (\u003cstrong\u003eD-F\u003c/strong\u003e), tumour C (\u003cstrong\u003eG-I\u003c/strong\u003e), and tumour D (\u003cstrong\u003eJ-L\u003c/strong\u003e). The x- and y-axes show distance from the probe (mm), and the scale bar on the right of each scan shows the image’s dynamic range (dB). Images generated using MATLAB.\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/ed0057b9dae6a857ce950f30.jpeg"},{"id":85649062,"identity":"78278f17-0a00-4117-b76e-164c354c8c92","added_by":"auto","created_at":"2025-06-30 08:54:54","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1149211,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA comparison between SURE (top image in each panel) and colour Doppler (bottom image in each panel) for matched tumour imaging planes. \u003c/strong\u003eOne view is presented per individual in this study (A-E).\u003cstrong\u003e \u003c/strong\u003eSURE images are displayed with a dual y-axis, showing the imaging depth as distance from the probe (mm) and the signal intensity as the image’s dynamic range (dB).\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/dbd99b67cf8b4e4300557070.jpeg"},{"id":85649071,"identity":"3e56eae9-90a9-477b-a236-3e5bcfeec520","added_by":"auto","created_at":"2025-06-30 08:54:57","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":338909,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAn overview of performing quantitative analysis on a vessel of interest with SURE. \u003c/strong\u003e\u0026nbsp;A region of interest is identified and processed into a SURE velocity map (\u003cstrong\u003e4A\u003c/strong\u003e) from a larger SURE density image (\u003cstrong\u003e4B\u003c/strong\u003e). The colour wheel (\u003cstrong\u003e4A\u003c/strong\u003e) displays the direction of flow. A vessel of interest is selected from the velocity image (white line perpendicular to blood flow direction (\u003cstrong\u003e4A\u003c/strong\u003e) for haemodynamic characterisation (\u003cstrong\u003e4C\u003c/strong\u003e). Blood flow velocity estimation is measured over the selected vessel profile (\u003cstrong\u003e4C\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe dual x-axis shows the different velocity measurements; the blue line shows the measurements per pixel, and the red line shows a normalised, fitted profile across the vessel (\u003cstrong\u003e4C\u003c/strong\u003e).\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/bf56c9253e1fb15c85f47d72.png"},{"id":85649046,"identity":"3e8db40e-6126-4258-a808-95dc4a6e1e76","added_by":"auto","created_at":"2025-06-30 08:54:50","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":182583,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eQuantitative, repeated measures assessment of tumour haemodynamics using SURE. \u003c/strong\u003eMeasurement of blood velocity (\u003cstrong\u003e5A\u003c/strong\u003e), vascular luminal diameter (\u003cstrong\u003e5B\u003c/strong\u003e) and blood volumetric flow rate (\u003cstrong\u003e5C\u003c/strong\u003e). \u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eAll measurements were collected from different regions of one SURE image per tumour. The SURE velocity map data processing method was used for data extraction, with manual selection of the vessel of interest using a MATLAB graphical user interface.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/1d00e7c1ab714fa5eceeb39f.jpeg"},{"id":85649043,"identity":"df0af62c-db53-4f71-90e7-ed8be80a9642","added_by":"auto","created_at":"2025-06-30 08:54:49","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":205747,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVascular density measurements using three imaging methods: SURE, histology and µCT. \u003c/strong\u003eAll measurements are made with non-matching imaging planes.\u003cstrong\u003e \u003c/strong\u003eThe first row (\u003cstrong\u003e6A\u003c/strong\u003e) compares vascular density analysis from 2-dimensional imaging methods, in vivo SURE (red) and ex vivo histology (grey). Each measurement was repeated three times per subject to assess variability. The second row (\u003cstrong\u003e6B\u003c/strong\u003e) illustrates the variability in vascular density along the longitudinal axis, throughout the 3-dimensional maximal intensity projection (MIP) for the analysed tumours A, B, D and E. No µCT data were obtained for tumour C. The blue curve displays the vessel density per consecutive µCT image slice along the longitudinal axis. The orange curve shows vascular density (%) per consecutive 2-dimensional MIP image of 0.5 mm width along the longitudinal axis of the tumour.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/7744b1da3c7ba0cc982d7e24.png"},{"id":86429169,"identity":"5c3e6deb-999a-46d1-aa96-56d593235374","added_by":"auto","created_at":"2025-07-10 14:08:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7470341,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/eb542011-d8c0-47e0-ae78-52306f39ebf5.pdf"},{"id":85649719,"identity":"7e1090ac-3260-4265-8058-eb0bc60dbf6a","added_by":"auto","created_at":"2025-06-30 09:02:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":2270681,"visible":true,"origin":"","legend":"","description":"","filename":"0525SURETumourVascularImagingSupplementary.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6829687/v1/09b433609d6273cd33f67a13.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Quantification of Tumour Angiogenesis and Perfusion using Contrast-free Super-Resolution Ultrasound Imaging","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eAngiogenesis plays a critical role in tumour development, enabling the formation of new blood vessels and the required blood perfusion rates to sustain tumour growth [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. This process is widely recognised as a hallmark of cancer, allowing tumours to obtain the oxygen and nutrients needed for proliferation and development [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. The significance of angiogenesis in tumour development has been extensively documented, highlighting the process as a potential therapeutic target in cancer treatment [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Despite this, current non-invasive in vivo imaging techniques lack the resolution to monitor angiogenesis accurately or to analyse tumour haemodynamics quantitatively. Clinical imaging methods, such as ultrasound, contrast-enhanced magnetic resonance imaging and computed tomography, can only detect large blood vessels [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Consequently, while these techniques can be utilised to aid in surgical planning and treatment monitoring, they cannot fully characterise the tumours\u0026rsquo; vascular architecture. In research settings, ex vivo techniques such as histology and micro-computed tomography (\u0026micro;CT) have a significantly higher capability to quantitatively assess the microvasculature and provide data to evaluate treatment responses and differentiate between tumour types [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. However, these methods are unsuitable for in vivo use as they cause significant tissue destruction. Therefore, a substantial gap exists between the capabilities of in vivo and ex vivo microvascular imaging methods.\u003c/p\u003e \u003cp\u003eThe development of super-resolution ultrasound scanning (SRUS) techniques, such as ultrasound localisation microscopy (ULM), has enabled high-resolution imaging of the vascular system in many tissue types, including tumours [\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. The majority of these techniques utilise intravascular contrast agents, usually microbubbles, to achieve an imaging resolution beyond the traditional ultrasound diffraction limit [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. While these techniques are non-damaging to tissues, representing a significant advancement for in vivo vascular imaging capabilities, the need for intravascular infusion and a long data acquisition time of 1\u0026ndash;10 minutes, has limited the technique\u0026rsquo;s implementation in clinical and preclinical settings [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Advancements by Jensen et al. (2022) have overcome this challenge, achieving non-invasive, high-resolution vascular imaging by tracking the backscattering of moving erythrocytes instead of microbubbles [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. This technique, \u003cb\u003eSU\u003c/b\u003eper-\u003cb\u003eR\u003c/b\u003eesolution ultrasound using \u003cb\u003eE\u003c/b\u003erythrocytes (SURE), has been documented to achieve an imaging resolution of 28\u0026micro;m, with only a few seconds of acquisition time, enabling near-real-time microvascular imaging [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTracking erythrocytes instead of microbubbles enables SURE to overcome many technical limitations of other SRUS techniques. Microbubble infusion is inconvenient for patients, requiring bolus or constant-rate infusion; conversely, SURE is non-invasive, with erythrocytes abundant in the vasculature of all tissues [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Erythrocytes are unaffected by therapeutic levels of ultrasound acoustic pressure, whilst microbubbles can be significantly affected, limiting the usable mechanical index to prevent microbubble rupture [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Using a higher mechanical index in SURE allows deeper imaging depth, enabling broader applications and greater clinical transferability compared to ULM [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. In research settings, microbubble infusion is not always possible, particularly in small animals such as mice with a small circulating blood volume and high susceptibility to infusion fluid overload [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. This is especially relevant in oncology research, where murine models are the most frequently used animal model [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCurrently, SURE imaging is not achievable in real-time due to the need for manual post-processing, which allows for careful image optimisation and analysis, but also introduces additional steps between acquisition and final data output. Two SURE semi-automated post-processing methods have been described which are components of the imaging pipeline: 1. Vascular density maps \u0026ndash; enabling qualitative assessment of the microvascular architecture of a tissue [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. 2. Velocity maps \u0026ndash; to identify blood flow direction, velocity and vessel luminal diameter measurement [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Both methods have been extensively validated, and the anatomical accuracy of the techniques has been validated through comparisons between microvascular imaging of select regions of the rat kidney with both contrast-enhanced \u0026micro;CT and ULM [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. The techniques\u0026rsquo; translatability to qualitatively image other species and other tissue types has been demonstrated in the rabbit kidney [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] and the human lymph node [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study describes how the SURE technique can be optimised for qualitative vascular imaging while assessing its ability to analyse subcutaneous tumour haemodynamics. We will determine the technique\u0026rsquo;s susceptibility to variability between individuals and imaging planes. We hypothesise that the signal to noise ratio (SNR) of SURE scans may be impacted by biological variation in vasculature, such as vessel size (smaller vessels give weaker signals), blood velocity (too fast or too slow can degrade signal), or volumetric flow rates (which affects the number of erythrocyte scatterers contributing to the image). We will compare SURE's qualitative vascular imaging capabilities to the best available non-invasive in vivo techniques, colour Doppler and power Doppler. We also assess SURE's ability to analyse tumour angiogenesis quantitatively, comparing the data collected to gold-standard ex vivo microvascular imaging techniques: ex vivo immunohistochemistry and contrast-enhanced \u0026micro;CT.\u003c/p\u003e"},{"header":"2. METHODS","content":"\u003cp\u003e\u003cstrong\u003e2.1\u0026nbsp;Animal Studies\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e2.1.1\u0026nbsp; \u0026nbsp;\u0026nbsp;Ethical Approval\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll animal experiments were conducted in accordance with Directive 2010/63/EU of the European Parliament and Council on the protection of animals used for scientific purposes and were approved by the Danish Animal Experiments Inspectorate under license no. 2021-15-0201-01041 (approved December 2021). As this study did not include human participants,\u0026nbsp;Ethics, Consent to Participate, and Consent to Publish declarations are not applicable.\u003c/p\u003e\n\u003cp\u003e2.1.2\u0026nbsp; \u0026nbsp;\u0026nbsp;Study Population\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSix-week-old female C57Bl/6JRj mice (n=5) (Janiver labs, France) were used in this study. They were acclimated for one week upon arrival and housed in an individual, ventilated cage with a light-to-dark period of 12:12 hours under controlled environmental conditions. Access to fresh water and standard pellet diet was provided ad libitum. Mice were anaesthetised (3-4% sevoflurane in 65% N\u003csub\u003e2\u003c/sub\u003e and 35% O\u003csub\u003e2\u003c/sub\u003e) and inoculated with the melanoma cell line B16-F10 (ATCC® CRL-6475™) via subcutaneous injection in the right flank with 1x10\u003csup\u003e6\u003c/sup\u003e cells in 100 µL of sterile phosphate-buffered saline per mouse. Tumours were allowed to grow for 20 days and were measured using a calliper. The tumour volume was estimated using the formula: volume = ½ (length × width\u003csup\u003e2\u003c/sup\u003e).\u003c/p\u003e\n\u003cp\u003e2.1.3\u0026nbsp; \u0026nbsp;\u0026nbsp;In Vivo Experiments\u003c/p\u003e\n\u003cp\u003eMice were anaesthetised (2% isoflurane in 65% N\u003csub\u003e2\u003c/sub\u003e and 35% O\u003csub\u003e2\u003c/sub\u003e) and placed in left lateral or ventral recumbency and immobilised using tape to minimise movement in the tumour tissue from respiration. A probe holder positioned a 168-channel GE L8-18iD Hockey stick linear array probe (GE HealthCare, USA) 1 cm away from the skin’s surface, with ultrasound gel providing the interface. The probe, with a wavelength of 154 µm, was operated at a transmission frequency of 10 MHz. Three imaging planes were randomly selected along the tumour’s long axis; no preview of the imaging region was performed to prevent operator bias. Each imaging plane was scanned first with SURE, then with B-mode and colour Doppler. Spectral Doppler was applied to areas highlighted by Colour Doppler to confirm pulsatile blood flow and obtain quantitative velocity measurements. All conventional ultrasound scans were collected using the same probe and a commercial GE LOGIQ E9 scanner at the exact location, ensured using a mechanical probe holder. For details of SURE imaging, see the Methods section 2.2. The animal was euthanised via cervical dislocation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.1.4\u0026nbsp; \u0026nbsp;\u0026nbsp;Post-Mortem Tissue Collection\u003c/p\u003e\n\u003cp\u003eImmediately following euthanasia, the mice underwent intracardiac vascular filling with the µCT contrast agent Vascupaint (yellow colloidal bismuth suspension MDL-121, MediLumine, Montreal, Quebec, Canada). Vascupaint was reconstituted with the following proportions: 1 ml silicone, 1 ml dilutant, and 0.1 ml catalyst. Following post-mortem sternotomy, a blunted 20G needle was clamped within the left ventricle and the inferior vena cava was severed. The vascular system was flushed with heparin saline, 3 ml of 2% formalin, then 3 ml of reconstituted Vascupaint solution. All infusions were completed using a syringe pump at the rate of 1 ml/min unless it became blocked, in which case a manual injection was completed. The whole animal was stored at 5 °C for 24 hours post-mortem. The tumour tissue was then excised and fixed in formalin for 24 hours before being embedded in paraffin.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. SURE Imaging\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e2.2.1\u0026nbsp; \u0026nbsp;\u0026nbsp;Data Collection\u003c/p\u003e\n\u003cp\u003eSURE data was acquired using a Verasonics Vantage 256 scanner (Verasonics, Inc., Kirkland, WA, USA) and the aforementioned GE L8-18iD linear array probe. Data was collected using the SURE pulse inversion sequence, which employs a synthetic aperture scan technique to enhance spatial resolution by using multiple virtual sources to simulate a larger effective aperture [25]. This sequence utilised 24 emissions from 2×12 virtual sources, improving image reconstruction and target visualisation. The acquisition duration was 1 minute, with a transmission voltage of 95 V. Data were collected using both a manual and automated time gain compensation (TGC) settings to optimise signal quality—the automatic programme aimed to generate a smooth TGC curve, which prevented signal clipping within the data. Manual TGC generally yielded the best results; therefore, unless specified otherwise, the images presented in this paper were acquired using this setting. TGC adjustments, automated or manual, were made to optimise the image data by minimising signal clipping in the data. A single operator completed all manual adjustments to prevent operator-induced bias.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.2.2\u0026nbsp; \u0026nbsp;\u0026nbsp;Data Processing - Vascular Density Maps\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe SURE processing method used was consistent with that described by Jensen et al. [12]. The key steps of the published SURE imaging pipeline are beamforming, motion estimation and correction with transverse oscillation, singular valve decomposition (SVD) for stationary echo cancelling, erythrocyte peak detection, density image formation and then velocity image formation. The following modifications were made to optimise the technique for oncology applications. To address clipping artefacts, clipped values in the raw radiofrequency data are identified and set to zero across all frames as a preprocessing step before SURE pipeline processing. A key refinement is selecting elements for removal based on beamformed clipped data rather than raw RF data alone. Elements contributing to clipping within the beamformed region are excluded, while those outside the region of interest (ROI) are ignored to avoid unnecessary data loss. However, excessive element exclusion was observed when clipping extended into the lower part of the image. To address this, we refine the selection criteria to remove elements only when their contribution degrades the beamformed region, preventing unnecessary signal loss. Following this, a ten-second segment of acquired data with the lowest motion was isolated from the full one-minute of data and processed.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.2.3\u0026nbsp; \u0026nbsp;\u0026nbsp;Quantifying Angiogenesis\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSURE Vascular density maps were imported as RGB images (with varying dynamic ranges) into ImageJ (version 2.9.0/1.53t) (Fig 1a). To calculate the vascular density of the tumour region, the image was first converted to an 8-bit grayscale and binarised based on its dynamic range (Fig 1b). Vascular thresholding was performed manually based on the image’s dynamic range to isolate vasculature while excluding background signal (Fig 1c). When the background signal varies significantly across different image regions, it was divided into two sections to allow for separate vascular thresholding, accounting for the differences in signal intensity. Vascular thresholding was applied independently to each section before the binarised images were recombined into a single image. The inter-rater variability of this method has not been assessed, nor its degree of subjectivity to differences in image intensity. To minimise variability, the same individual completed thresholding for all images. The tumour region was selected manually using the ROI tool (Fig 1c). The ROI excluded tumour regions where imaging artefacts or background signal obscured the vasculature. The ROI also excluded the tumour margins as the tumour capsule is not clearly defined in SURE images; consequently, discerning the border between vasculature within the skin, subcutaneous tissue, and tumour is difficult. The pixels of thresholded vessels vs. background were measured within the specified ROI to quantify vascular density.\u003c/p\u003e\n\u003cp\u003e2.2.4\u0026nbsp; \u0026nbsp;\u0026nbsp;Data Processing - Velocity Maps\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOne scan per individual mouse was selected for velocity map processing. The scan was selected based on the imaging plane demonstrating the highest detectable vascularity on both SURE and colour Doppler. From the total acquired data, a 1–1.5 second data segment corresponding to a low-motion period between respiratory cycles was utilised. This was then processed using a modified version of the method described by Naji et al. [21], adapted to include the use of a recursive nearest neighbour tracker.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.2.5\u0026nbsp; \u0026nbsp;\u0026nbsp;Quantifying Haemodynamics\u003c/p\u003e\n\u003cp\u003eQuantitative data were collected from multiple vessels in each velocity map to analyse the haemodynamic variation in different vessels across one tumour imaging plane and between tumours of other individuals. For each vascular region identified in a scan, the blood velocity and vessel lumen diameters were estimated using the published method described by Naji et al. [21]. Velocity magnitudes were extracted along a user-defined line oriented in both the lateral and axial directions. Approximately 50 cross-sectional slices were sampled across the vessel of interest from this. These individual profiles were then averaged to generate a raw data velocity curve (shown in blue). A parabolic function (red curve) was fitted to the averaged data to reflect the expected laminar flow distribution in cylindrical vessels, where velocity is maximal at the center and diminishes toward the periphery. Vessel diameter was determined using the Full Width at Half Maximum, calculated as the distance across the profile at which the velocity reaches 50% of its maximum value. Further characterisation of haemodynamics is performed by calculation of the estimated blood flow rate (volumetric flow rate) for each vessel using the following formula:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBlood flow rate\u003c/strong\u003e \u003cstrong\u003e(\u003c/strong\u003e\u003cstrong\u003eμL/s)\u0026nbsp;\u003c/strong\u003e= (blood velocity (mm/s)\u0026nbsp;·\u0026nbsp;vessel area (µm²) ) / 10\u003csup\u003e9\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVessel area (\u003c/strong\u003e\u003cstrong\u003eµm²)\u003c/strong\u003e = (π · Diameter\u003csup\u003e2\u003c/sup\u003e (µm)) / 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3\u0026nbsp; \u0026nbsp;Comparison to Other Vascular Imaging Techniques\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e2.3.1\u0026nbsp; \u0026nbsp;\u0026nbsp;Conventional Ultrasound\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOne SURE image per mouse was qualitatively compared with colour Doppler. The scan was selected based on the imaging plane demonstrating the highest detectable vascularity on both SURE and conventional ultrasound imaging.\u003c/p\u003e\n\u003cp\u003e2.3.2\u0026nbsp; \u0026nbsp;\u0026nbsp;Contrast-Enhanced µCT \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFormalin-fixed paraffin-embedded (FFPE) samples were scanned with a Exciscope Polaris micro-CT scanner (Exciscope AB, Sweden), with an isotropic voxel size ranging from 4 µm to 5 µm. Scans were reconstructed using Exciscope’s own software and all analyses were performed using ITK-SNAP software (version 3.6.0) and MATLB (The MathWorks Inc., Natick, Massachusets) (Supplementary Fig 1A). A 3-dimensional ROI was manually created using the ITK-SNAP polygon function to isolate the tumour tissue from surrounding soft tissue (Supplementary Fig 1B). Data analysis was performed within this ROI. The contrast agent within the vasculature was isolated using a manually set intensity threshold (Supplementary Fig 1C). This enabled a proxy vascular density measurement for the tumour volume and for consecutive 20µm 2D image planes. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.3.3\u0026nbsp; \u0026nbsp;\u0026nbsp;Histology\u003c/p\u003e\n\u003cp\u003eFFPE samples were sectioned along the longitudinal axis. Three distinct regions along this axis were selected for analysis. From each region, two slides were prepared using adjacent tissue sections. One slide was stained with hematoxylin and eosin (H\u0026amp;E), while the other underwent fluorescent immunohistochemistry, utilising the same primary Anti-CD31 antibody [EPR17259] (ab182981, Abcam) along with an Alexa Fluor® 568-conjugated secondary antibody (Fig 1D-F). H\u0026amp;E-stained slides were used to guide the identification of the tumour border for fluorescent immunohistochemistry vascular density measurement. Vascular density analysis on fluorescent immunohistochemistry images was performed on ImageJ, using the same protocol listed in section 2.2.3. Another slide per tumour underwent chromogenic immunohistochemistry to detect CD31 expression using the Anti-CD31 antibody [EPR17259] and a secondary anti-rabbit HRP polymer (Nordic Biosite) (Supplementary Fig 2). Chromogenic staining enabled the qualitative assessment of contrast agent vascular filling. Whole-section slide imaging was performed using the AxioScan Z1 (ZEISS) at x10 magnification.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4. Statistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data analysed within this study were examined for normality of distribution using the Shapiro-Wilk normality test. When analysing differences between tumours, where data is not normally distributed, the Kruskal-Wallis test with Dunn’s post-hoc test was used. When analysing differences between tumours where data is normally distributed, a one-way ANOVA with Tukey’s post-hoc test is used. The Wilcoxon signed-rank test was used to analyse the difference in data collected from the same tumour. Throughout the study, a p-value of \u0026lt; 0.05 was considered statistically significant. All statistical analyses were conducted using RStudio. Graphical representations of the data were displayed with significance indicated by asterisks: p \u0026lt; 0.05 was denoted by a single asterisk (*), p \u0026lt; 0.01 by two asterisks (**), and p \u0026lt; 0.001 by three asterisks (***).\u0026nbsp;\u003c/p\u003e"},{"header":"3.\tRESULTS ","content":"\u003cp\u003e\u003cstrong\u003e3.1. \u0026nbsp; \u0026nbsp; \u0026nbsp; Qualitative Vascular Imaging\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFifteen SURE images were collected from five tumours (Fig 2). This naming system for the five individuals will be consistent throughout the paper. All SURE images showed detectable vasculature. Significant variations in image SNR were seen between scans collected from one individual and across the dataset. A high background signal in the deepest half of the tumour (beyond 10 mm) was observed in 60% of scans, limiting reliable assessment of this region. A high frequency of imaging artefacts was also seen, with 66% of total scans being impacted by a banding signal artefact, obscuring potential vascular imaging in the affected image regions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor all imaging regions compared, SURE images demonstrate superior vascular imaging compared to colour Doppler (Fig 3). Regions where blood flow was detected on colour Doppler vessels were all identified as containing vessels in SURE. SURE scans successfully imaged individual vessels and imaging of vascular in tumour regions not detected by Doppler. Of the compared scans, no ultrasound artefacts or variations in signal across the tumour tissue were seen in B-mode imaging. Three SURE scans were affected by banding artefacts (Fig 3B, 3C, and 3E), and high background signal in the deeper portion tissue impacted two tumour scans (Fig 3C and 3E).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. \u0026nbsp; \u0026nbsp; \u0026nbsp; Analysis of Individual Vessels\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe high resolution of SURE scans enabled detection of individual vessels, allowing the user to distinguish between neighbouring vessels in vascular tumour regions. This enabled vessels to be individually selected for secondary processing with SURE velocity mapping (Fig 4). A degree of vascular resolution was lost during secondary processing due to higher SNR thresholding (as illustrated by Fig 4A). Variation was seen in the number of measurements made between each scan, due to differences in vascularity across the imaging plane. For tumour A, 10 measurements were made; for tumour B, 18; tumour C, 18; tumour D, 9; and tumour E, 12. Information regarding the vessel lumen diameter, velocity, and flow direction was successfully obtained for all vessels retained during the velocity map processing step. Where velocity measurement across the vessel\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. \u0026nbsp; \u0026nbsp; \u0026nbsp; Haemodynamic Characterisation of the Tumour\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs multiple vessels could be analysed per SURE image, we were able to provide an analysis of the variation in vessel characteristics and haemodynamics both within one tumour and between individuals (Fig 5). This could not be achieved using colour Doppler due to an inability to distinguish individual vessels or spectral Doppler, which could not measure blood flow rates in any scans.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e3.3.1. \u0026nbsp; Blood Velocity\u003c/p\u003e\n\u003cp\u003eAcross all measurements from all tumours,the mean blood velocity was 2.75 mm/s, with a median of 2.4 mm/s and a standard deviation of 1.17 mm/s. The range of velocities spans from 1.2 mm/s (minimum) to 7.5 mm/s (maximum), with an interquartile range of 1.1 mm/s.Velocity measurements varied significantly between tumours and were not normally distributed in all cases. The Shapiro-Wilk normality tests revealed that the blood velocity data for tumours A, B, D, and E follow a normal distribution. However, the velocity data for tumour C significantly deviates from normality (\u003cem\u003ep\u003c/em\u003e = 0.00196**) . A Kruskal–Wallis test revealed a statistically significant difference in velocity distributions across tumours (χ²(4) = 11.65, \u003cem\u003ep\u003c/em\u003e = 0.020*) (Fig 5A). Post-hoc pairwise comparisons using Dunn’s test with Bonferroni correction identified a significant difference between tumours B and E (\u003cem\u003ep\u003c/em\u003e = 0.024*). No other pairwise comparisons reached statistical significance following correction.\u003c/p\u003e\n\u003cp\u003e3.3.2. \u0026nbsp; Blood Vessel Luminal Diameter\u003c/p\u003e\n\u003cp\u003eAcross all measurements in this study, the mean lumen size was 121.31 µm. The median lumen size was 115.8 µm, and the standard deviation was 40.36 µm. The minimum lumen size recorded was 52.8 µm, while the maximum was 251.1 µm. The interquartile range (IQR) was 49.4 µm. The Shapiro-Wilks test showed that all tumours had normally distributed luminal diameter measurements. A one-way ANOVA indicates significant vessel size differences between the tumours (F(4, 72) = 4.118, \u003cem\u003ep\u003c/em\u003e = 0.00463**) (Fig 5B). A post hoc Tukey's Honest Significant Difference (HSD) test was conducted, identifying significantly higher luminal diameter in Tumour E compared to tumour B (\u003cem\u003ep\u003c/em\u003e = 0.0068*), with a mean difference of 42.63 µm. No other significant pairwise differences were found.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e3.3.3. \u0026nbsp; \u0026nbsp;Blood Flow Rate\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe blood flow rate measurements across the entire dataset had a mean of\u0026nbsp;3.63×10⁻⁵ µL/s, a median of\u0026nbsp;2.61×10⁻⁵ µL/s, and a standard deviation of\u0026nbsp;3.18×10⁻⁵ µL/s. The flow rates range from\u0026nbsp;3.97×10⁻⁶ µL/s\u0026nbsp;(minimum) to\u0026nbsp;1.707×10⁻\u003csup\u003e4\u003c/sup\u003e µL/s (maximum), with an interquartile range of 2.86×10⁻⁵ µL/s. The Shapiro-Wilk tests show that flow rate measurements from tumours A, B, and D are normally distributed, while data from tumours C (\u003cem\u003ep\u003c/em\u003e = 0.0429*) and E (\u003cem\u003ep\u003c/em\u003e = 0.00418**) are not normally distributed. Kruskal-Wallis test results indicate a significant difference in blood flow rates across tumour types (χ²(4) = 20.88, \u003cem\u003ep\u003c/em\u003e = 0.000340***) (Fig 5C). Post-hoc pairwise comparisons using Dunn’s test with Bonferroni correction identified a significantly higher blood flow rate in tumour E compared to tumour A (0.0128*) and tumour B (\u003cem\u003ep\u003c/em\u003e = 0.000384***). No other pairwise comparisons reached statistical significance following correction.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4. \u0026nbsp; \u0026nbsp; \u0026nbsp; Quantifying Angiogenesis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e3.4.1. \u0026nbsp; SURE Vascular Density Analysis\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe analysed tumour vascular density measurements using SURE across the five tumours with three regional measurements per tumour (Fig 6A). Tumour C had the highest mean vascular density at 7.73%, but the lowest variability (SD = 0.99%, IQR = 0.98%, coefficient of variation (CV) = 0.128), suggesting a consistent vascular density across regions. Followed by Tumour D with a mean vascular density of 5.54% (SD = 1.39%, IQR = 1.37%, CV = 0.251), indicating moderate variability. Tumours A, B and E had similar mean densities. Tumour A had a mean vascular density of 2.49% (SD = 0.51%, IQR = 0.51%, CV = 0.205), with moderate variability. \u003cstrong\u003eTumour B\u003c/strong\u003e showed a mean of \u003cstrong\u003e3.30%\u003c/strong\u003e (SD = \u003cstrong\u003e1.46%\u003c/strong\u003e, IQR = \u003cstrong\u003e1.27%\u003c/strong\u003e, CV = \u003cstrong\u003e0.443)\u003c/strong\u003e, showing greater relative variability. Tumour \u003cstrong\u003eE\u003c/strong\u003e exhibited the greatest variability, with a mean vascular density of \u003cstrong\u003e3.03%\u003c/strong\u003e (SD = \u003cstrong\u003e1.96%\u003c/strong\u003e, IQR = \u003cstrong\u003e1.93%\u003c/strong\u003e, CV = \u003cstrong\u003e0.644)\u003c/strong\u003e, suggesting substantial heterogeneity across its regions. To statistically assess whether the variability in vascular density differed significantly between the tumors, a Fligner-Killeen test of homogeneity of variances was performed. The test yielded a chi-squared value of 1.4847 with 4 degrees of freedom, resulting in a non-significant p-value of 0.8293, demonstrating that there is no evidence of significant differences in variability among the \u0026nbsp;tumors.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e3.4.2.\u0026nbsp; \u0026nbsp;Comparison to Ex vivo methods\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eImmunohistochemistry data from the whole dataset, 15 slides from 5 individuals, showed a \u003cstrong\u003emean\u003c/strong\u003e vascular density of \u003cstrong\u003e3.00%\u003c/strong\u003e and\u0026nbsp;a \u003cstrong\u003estandard deviation\u003c/strong\u003e of \u003cstrong\u003e0.66%\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e indicating relatively low variability. The values ranged from a \u003cstrong\u003eminimum\u003c/strong\u003e of \u003cstrong\u003e2.01%\u003c/strong\u003e to a\u003cstrong\u003emaximum\u003c/strong\u003eof\u003cstrong\u003e4.07%.\u0026nbsp;\u003c/strong\u003eThe \u003cstrong\u003emedian\u003c/strong\u003e was \u003cstrong\u003e2.84%\u003c/strong\u003e, and the\u003cstrong\u003einterquartile range (IQR)\u003c/strong\u003ewas \u003cstrong\u003e0.90%\u003c/strong\u003e, showing that the central 50% of the data were tightly clustered. A Wilcoxon signed-rank test assessed whether vascular density measurements vary significantly between the SURE and histology methods within one tumour. There were\u0026nbsp;no statistically significant differences\u0026nbsp;between the results of the two measurement methods for any tumour (Fig 6A).\u003c/p\u003e\n\u003cp\u003eVascular density measured using µCT demonstrated substantial variability for different planes within a tumour (Fig 6B). There is also variation in average vascular densities between the five tumours. Tumours A, D, and E had vasculature present throughout the tumour. While the vascular density dropped close to 0.5% for large regions of Tumour B. Both Tumour A and Tumour B showed relatively low median values (1.17% and 0.644%) and interquartile ranges (IQR = 1.78% and 2.07%), with maximum vascular density measurements of 7.59% and 7.49%, respectively, indicating limited vascularisation. In contrast, Tumour D exhibited a markedly higher vascular density (median = 14.4%), with a wide distribution (IQR = 14.4%) and a maximum SURE value of 28.9%, reflecting substantial heterogeneity. Tumour E presented intermediate values (median = 4.18%, and IQR = 3.45%), with vascular density ranging from 0% to 19.8%. \u0026nbsp;However we are limited in our ability to compare µCT density data with other methods due to incomplete and highly irregular contrast filling which was identified via qualitative assessment of chromogenic immunohistochemistry. Irregular Vascupaint filling was seen within all analysed tumours (Supplementary Fig 2). The contrast agent filling of tumour A could not be assessed due to damage during slide preparation. Tumour C was the only tumour with moderate erythrocyte extravasation throughout the tissue. Due to the irregular contrast agent filling and significant tissue distortion induced during the excision and fixation process, co-registration of individual vessels between SURE and µCT scans was impossible. Despite these limitations, vascular density measurements collected with SURE and fluorescent immunohistochemistry methods fell within the range of measured µCT densities.\u003c/p\u003e"},{"header":"4.\tDISCUSSION ","content":"\u003cp\u003eThis study demonstrates that SURE, with its combination of high-resolution vascular imaging and quantitative haemodynamic analysis, significantly advances non-invasive in vivo vascular assessment capabilities in oncology research. SURE enables visualisation of individual vessels at a higher resolution than achieved by colour Doppler, more consistently than seen with contrast-enhanced µCT, and with vascular density data comparable to immunohistochemistry. SURE exceeds current capabilities of other super-resolution techniques, overcoming the need for intravascular contrast-agent use and significantly reducing data acquisition time to seconds. In this dataset, where tumours are newly formed, and vessels are small with low flow rates, conventional techniques have minimal usability, demonstrated by spectral Doppler’s inability to detect flow rates. Conversely, SURE successfully imaged individual vessels and enabled haemodynamic assessment in vessels as small as 50 µm. The greater vascular resolution achieved with SURE shows the technique to have considerable application possibilities in clinical settings, where the ability to assess a tumour’s vascular architecture quickly may have both diagnostic and prognostic value. The ability to reliably image the vasculature, while also performing haemodynamic analysis on the entire tissue or vessels of choice, gives SURE a unique advantage over other super-resolution techniques and varied applications in oncology.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVariability in SURE Images\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAs with other ultrasound-based techniques, each SURE scan must be collected with optimal data-collection parameters to overcome variations in signal intensity between different imaging planes. To address these variations, the operator tailors Time Gain Compensation (TGC) settings for each scan, adjusting for factors such as tissue depth and acoustic properties to ensure consistent image quality. SURE ultrasound sequences are otherwise consistent across participants, enabling the collection of comparable data. This study demonstrates that while SURE exceeds the vascular imaging capabilities of existing ultrasound techniques (Fig 3), it remains susceptible to many challenges associated with ultrasound-based technologies.\u003c/p\u003e\n\u003cp\u003eIn this dataset, there is a significant issue with high background signal in deeper portions of the image, which is unrelated to the image’s \u003cstrong\u003edynamic range (dB). This issue persisted with both manual and automated TGC settings. This is known to be related to\u0026nbsp;\u003c/strong\u003eultrasound wave attenuation and divergence at depth, where signals weaken relative to system noise, high-frequency components are lost, and TGC cannot fully restore signal quality without amplifying noise, resulting in a drop-off in SNR\u0026nbsp;[26]. However, in this study, we show that this impact is not uniform across a dataset of highly similar tumours or even consistent with different scan planes of the same tumour. In tumours C and E, the effect was present in all scans (Fig 2). Whilst it only affected some of the scans in others (Fig 2: tumours A and B). This study could not confirm or deny whether this effect is impacted by biological variation in the vasculature. There was limited variation in the biological parameters analysed in the study between the five tumours. Tumour E, which had significantly higher average velocities, vessel sizes and volumetric flow rates than tumour B, did not show better imaging quality, as expected. Manual vs automated TGC adjustment of images of the same tumour section did not impact the presence of this artefact, reducing the likelihood that TGC adjustment. Further research is needed to elucidate why these variations are seen and whether improvements in\u0026nbsp;imaging strategies may address them.\u003c/p\u003e\n\u003cp\u003eAll tumours scanned in this study were affected by a significant imaging artefact: linear hyperechoic bands in the deep imaging section (Fig 2A, 2B, 2D, 2E, 2F, 2G, 2H, 2I, 2L, 2M, 2O). This artefact resembles the ‘incoherent clutter’ artefact, commonly seen in super-resolution imaging methods, which occurs for two reasons: off-axis scattering and complex tissue motion [27]. In this study, the primary cause for this artefact was suspected to be clipping within the data, with the horizontal bands occurring at depths corresponding to clipping occurrences. The effect is then distributed over the image during beamforming due to heterogeneous tissue motion from respiration or cardiac activity, introducing unpredictable signal fluctuations that standard SVD-based clutter filtering struggles to suppress. This artefact obscures weak blood flow signals and degrades the accuracy of vascular visualisation within tumours.\u0026nbsp;These artefacts are known to be worse in the deeper imaging regions due to increased attenuation of the ultrasound signal and stronger off-axis scattering from surrounding tissues, which reduces signal-to-noise ratio and makes clutter harder to suppress\u0026nbsp;[26]. As all scans were collected using methods designed to maximise signal without introducing data clipping, using either automated or manual TGC, the prevalence of this in the dataset represents a challenge. Highlighting that current signal amplification selection methods are insufficient for the selection of an appropriate TGC, and further\u0026nbsp;improvements are needed.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHaemodynamic Assessment using SURE\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eMany ULM studies, utilising intravascular contrast agents, have successfully characterised vascular density in subcutaneous tumours within animal models [8, 28-30]. Some of these studies also gathered quantitative data on microbubble movement as a proxy for erythrocyte haemodynamics [31]. The appropriateness of using microbubbles as a direct proxy for vascular haemodynamics in newly formed tumour microvasculature is unknown and understudied. Consequently, SURE's ability to remove this variable and directly analyse the erythrocytes' haemodynamics is one significant advantage of the technique. Other groups have also achieved high-resolution imaging of the subcutaneous tumour vasculature without using a contrast agent [32, 33]. However, these techniques cannot provide quantitative analysis of the haemodynamics,\u0026nbsp;limiting clinical applicability. Measuring blood velocity in tumours provides analysis of tissue vascular health and angiogenesis, with changes in vascular resistance or irregular flow patterns used to differentiate benign from malignant tumours\u0026nbsp;[34]. Measuring volumetric flow enables assessment of tumour tissue perfusion. Insufficient flow is often seen in malignant tumours, where irregular, disorganised blood vessels result in hypoxia and necrosis, affecting fluid flow\u0026nbsp;[35]. The ability to quantitatively analyse these haemodynamic parameters enables clinicians and researchers to characterise and assess the tumour tissue, allowing the monitoring of disease progression or response to treatments.\u003c/p\u003e\n\u003cp\u003eBoth colour and spectral Doppler are commonly used in clinical settings to assess tissue vascularity and perfusion. The techniques’ non-invasive and rapid qualitative assessment capabilities outweigh their limitations in microvascular analysis. While SURE’s current requirement for post-processing limits its real-time capabilities, further advancement in computing will likely result in comparable clinical usability. In this study, tumour blood velocity ranged from 1-5 mm/s (Fig 5), consistent with measurements achieved by other researchers using contrast-enhanced techniques [31]. This data confirms why conventional ultrasound techniques have limited function in these tissues. Colour Doppler and Spectral Doppler struggle to detect vessels smaller than 100\u0026nbsp;µm [36, 37]. They also require moderate flow rates (above 40 cm/s) to produce accurate images\u0026nbsp;[38]. SURE measurements show that while a substantial proportion of the vasculature is larger than the required 100\u0026nbsp;µm to be detected by colour Doppler, all vessels had a velocity below that required for detection. Additionally, the spatial resolution of conventional Doppler (1–2 mm) is not sufficient for imaging the micron-sized vessels found in tumours [39].\u003c/p\u003e\n\u003cp\u003eSURE showed that haemodynamics can vary significantly between tumours of the same type, age, and clinical condition. We identified significantly higher blood velocities and vessel sizes in tumour E compared to tumour B, and significantly higher volumetric flow rates in tumour E compared to tumours A and B (Fig 5a). The ability to measure haemodynamics in different regions of one tumour scan also demonstrated that vascular architecture and haemodynamics can vary significantly between tumours of the same type. We identified that in two tumours, C and E, data from different regions of the tumour had very different blood velocities and flow rates, differing from the other tumours, where these parameters did not vary widely (Fig 5). The ability to characterise and sub-categorise tumour forms is a cornerstone of oncology prognostics. This finding is a prime example of how SURE can identify novel cancer biomarkers.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eComparison to Ex Vivo Techniques\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eImmunohistochemistry was utilised in this study to assess how SURE compares to an established ex vivo 2-dimensional imaging technique. We found no significant difference between vascular density datasets obtained with SURE or immunohistochemistry (Fig 6). Importantly, this comparison is inherently biased due to methodological differences. For example, immunohistochemistry typically reports lower vascular density than SURE, as CD31 staining labels only the vessel wall, whereas SURE visualises the vessel lumen. These differences contribute to discrepancies between datasets, which give a greater tolerance to SURE's lower vascular detection sensitivity. Despite this, the comparability of the datasets provides strong support for the use of SURE in research settings.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo analyse tumour vascular heterogeneity more extensively, µCT was utilised. This method enabled the quantification of angiogenesis in slices of comparable thickness to the SURE imaging plane. Data showed an unexpectedly high degree of vascular heterogeneity, which did not align with immunohistochemistry results showing high vascular homogeneity in this tumour model (Fig 6). Qualitative assessment of chromogenic immunohistochemistry slides demonstrated that irregular contrast-agent filling of the vasculature impacted the µCT vascular density results (Supplementary Fig 2). While this means these data cannot be used to give a complete overview of the vascular density through the tumour, the data range represents a range of the minimal possible vascular density measurements. The vascular density measurements obtained using SURE and immunohistochemistry all fell within the range of possible vascular density values from µCT analysis. While all methods gave comparable vascular density data, which did not significantly differ, all were anatomically inaccurate for different reasons. Immunohistochemistry only quantifies the vessel's rim, measuring a lower density value than if the lumen were included. The µCT analysis quantifies the contrast agent as a proxy for the vasculature, subject to filling limitations. SURE images show significant variability as described above, which, combined with the subjective method of density analysis, reduces the accuracy of the data. Full co-registration of SURE scans to µCT volumes was impossible due to significant tissue distortion during excision and fixation. Consequently, the anatomical accuracy of SURE imaging of individual vessels in an image plane cannot be validated in this study.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStudy Strengths and Limitations\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAs this is a novel method, no prior power analysis was performed due to the lack of available data on expected variability. A post-hoc power analysis showed the SURE vascular density comparison to have 72% power, demonstrating the study was moderately underpowered (G*Power; ANOVA approximation, f = 1.03, α = 0.05, N = 15, 5 groups). This may have limited the ability of this study to detect statistically significant differences between tumours. While limited to one tumour model and a small population size, this study provides critical pilot data which can inform future studies. The imaging model used in this work, involving small subcutaneous tumours, has many physical characteristics translatable to human pathology. Regarding measurement reliability, velocity and vessel size data have been validated in rat kidney models and ex vivo systems [12, 21]. Nevertheless, future studies should compare these measurements with in vivo data from larger tumour vessels and a wider range of tissues to confirm their accuracy and confirm their accuracy in clinical contexts.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFuture Directions\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFuture research should prioritise methodological refinements to enhance data quality and imaging reliability. Key areas for improvement include reducing processing artefacts and increasing the signal-to-noise ratio (SNR), enabling clearer resolution of vascular structures in deeper regions of the tumour. Additionally, further investigation is needed to understand how experimental variables such as tumour size, tissue viability, and systemic factors like animal blood pressure affect image quality and consistency. Beyond technical advancements, future studies should explore the broader applications of this method for characterising tumour microvascular architecture. One promising direction is its potential use in distinguishing between benign and malignant tumour nodules, as well as differentiating between tumour classifications. The technique could also prove valuable for stratifying tumour subgroups, aiding in classification, and conducting longitudinal assessments to monitor disease progression or therapeutic response. Optimising this method for human applications is critical in furthering its translatability and expanding its impact in oncology.\u003c/p\u003e"},{"header":"Conclusions ","content":"\u003cp\u003eThis work has demonstrated that SURE can provide vascular density data at a resolution comparable to ex vivo techniques, together with haemodynamic quantitative data of significant value to clinical and research applications. While the method requires further processing optimisation to improve reliability and clinical usability, it can potentially identify previously unknown cancer biomarkers, elucidating characteristics of neoplastic angiogenesis which can be used for early diagnosis and treatment.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eEthical Approval\u003c/h2\u003e \u003cp\u003e All animal experiments were conducted in accordance with Directive 2010/63/EU of the European Parliament and Council on the protection of animals used for scientific purposes and were approved by the Danish Animal Experiments Inspectorate under license no. 2021-15-0201-01041 (approved December 2021). As this study did not include human participants, Ethics, Consent to Participate, and Consent to Publish declarations are not applicable.\u003c/p\u003e \u003ch2\u003eCompeting Interests\u003c/h2\u003e \u003cp\u003eThe authors declare that they have no competing interests relevant to this research.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was funded by a Synergy grant from the European Research Council, project no. 854796.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthor Contributions: A.M: Conceptualisation, Methodology (lead), Investigation (lead), Data Analysis, Writing - Original Draft, Data Visualisation (lead). I.T: Software - SURE Density Imaging, Investigation - In Vivo Data Collection, Data Curation, Writing - Review. A.S.C: Methodology, Investigation \u0026ndash; provision of mouse model. L.N.H: Data Curation - micro-CT. Data Visualisation - Figure 6B. M.A.J: Software - SURE Velocity Imaging, Data Curation - velocity processing (joint with A.M.). Data Visualisation - Figure 4C. H.M.K: Software, Resources, Data Curation. A.K: Supervision. M.B.N: Supervision, Funding Acquisition. J.A.J: Supervision, Funding Acquisition, Software (lead). C.M.S: Supervision (lead), Methodology, Project Administration, Funding Acquisition, Writing - Review \u0026amp; Editing. All authors have read, reviewed and agreed to the published version of the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eData is provided within the manuscript or supplementary information files.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eDe Palma, M., D. Biziato, and T.V. Petrova, \u003cem\u003eMicroenvironmental regulation of tumour angiogenesis.\u003c/em\u003e Nature Reviews Cancer, 2017. \u003cstrong\u003e17\u003c/strong\u003e(8): p. 457-474.\u003c/li\u003e\n\u003cli\u003ePotente, M., H. Gerhardt, and P. Carmeliet, \u003cem\u003eBasic and therapeutic aspects of angiogenesis.\u003c/em\u003e Cell, 2011. \u003cstrong\u003e146\u003c/strong\u003e(6): p. 873-87.\u003c/li\u003e\n\u003cli\u003eYadav, L., et al., \u003cem\u003eTumour Angiogenesis and Angiogenic Inhibitors: A Review.\u003c/em\u003e J Clin Diagn Res, 2015. \u003cstrong\u003e9\u003c/strong\u003e(6): p. Xe01-xe05.\u003c/li\u003e\n\u003cli\u003eJeswani, T. and A.R. Padhani, \u003cem\u003eImaging tumour angiogenesis.\u003c/em\u003e Cancer Imaging, 2005. \u003cstrong\u003e5\u003c/strong\u003e(1): p. 131-8.\u003c/li\u003e\n\u003cli\u003ePabst, A.M., et al., \u003cem\u003eImaging angiogenesis: Perspectives and opportunities in tumour research \u0026ndash; A method display.\u003c/em\u003e Journal of Cranio-Maxillofacial Surgery, 2014. \u003cstrong\u003e42\u003c/strong\u003e(6): p. 915-923.\u003c/li\u003e\n\u003cli\u003ePorte, C., et al., \u003cem\u003eUltrasound Localization Microscopy for Cancer Imaging.\u003c/em\u003e IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2024. \u003cstrong\u003e71\u003c/strong\u003e(12: Breaking the Resolution Barrier in Ultrasound): p. 1785-1800.\u003c/li\u003e\n\u003cli\u003eCouture, O., et al., \u003cem\u003eUltrasound Localization Microscopy and Super-Resolution: A State of the Art.\u003c/em\u003e IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2018. \u003cstrong\u003e65\u003c/strong\u003e(8): p. 1304-1320.\u003c/li\u003e\n\u003cli\u003eDencks, S. and G. Schmitz, \u003cem\u003eUltrasound localization microscopy.\u003c/em\u003e Zeitschrift f\u0026uuml;r Medizinische Physik, 2023. \u003cstrong\u003e33\u003c/strong\u003e(3): p. 292-308.\u003c/li\u003e\n\u003cli\u003eChristensen-Jeffries, K., et al., \u003cem\u003eSuper-resolution Ultrasound Imaging.\u003c/em\u003e Ultrasound in Medicine \u0026amp; Biology, 2020. \u003cstrong\u003e46\u003c/strong\u003e(4): p. 865-891.\u003c/li\u003e\n\u003cli\u003eSong, P., J.M. Rubin, and M.R. Lowerison, \u003cem\u003eSuper-resolution ultrasound microvascular imaging: Is it ready for clinical use?\u003c/em\u003e Zeitschrift f\u0026uuml;r Medizinische Physik, 2023. \u003cstrong\u003e33\u003c/strong\u003e(3): p. 309-323.\u003c/li\u003e\n\u003cli\u003eJensen, J.A., et al. \u003cem\u003eFast super resolution ultrasound imaging using the erythrocytes\u003c/em\u003e. in \u003cem\u003eMedical Imaging 2022: Ultrasonic Imaging and Tomography\u003c/em\u003e. 2022. SPIE.\u003c/li\u003e\n\u003cli\u003eArendt Jensen, J., et al., \u003cem\u003eSuper-Resolution Ultrasound Imaging Using the Erythrocytes-Part I: Density Images.\u003c/em\u003e IEEE Trans Ultrason Ferroelectr Freq Control, 2024. \u003cstrong\u003e71\u003c/strong\u003e(8): p. 925-944.\u003c/li\u003e\n\u003cli\u003eQuaia, E., et al., \u003cem\u003eBolus versus continuous infusion of microbubble contrast agent for liver ultrasound by using an automatic power injector in humans: A pilot study.\u003c/em\u003e Journal of Clinical Ultrasound, 2016. \u003cstrong\u003e44\u003c/strong\u003e(3): p. 136-142.\u003c/li\u003e\n\u003cli\u003eHelms, C.C., M.T. Gladwin, and D.B. Kim-Shapiro, \u003cem\u003eErythrocytes and Vascular Function: Oxygen and Nitric Oxide.\u003c/em\u003e Frontiers in Physiology, 2018. \u003cstrong\u003e9\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eWang, Q.X. and K. Manmi, \u003cem\u003eThree dimensional microbubble dynamics near a wall subject to high intensity ultrasound.\u003c/em\u003e Physics of Fluids, 2014. \u003cstrong\u003e26\u003c/strong\u003e(3).\u003c/li\u003e\n\u003cli\u003eBezer, J.H., et al., \u003cem\u003eMicrobubble dynamics in brain microvessels.\u003c/em\u003e PLOS ONE, 2025. \u003cstrong\u003e20\u003c/strong\u003e(2): p. e0310425.\u003c/li\u003e\n\u003cli\u003eBeaver, T.A., et al., \u003cem\u003eIntegrated Backscatter During Harmonic and Fundamental Frequency Imaging\u0026mdash;Effect of Depth, Mechanical Index, and Tissue Anisotropy: Implications for Myocardial Tissue Characterization.\u003c/em\u003e Echocardiography, 2003. \u003cstrong\u003e20\u003c/strong\u003e(4): p. 337-343.\u003c/li\u003e\n\u003cli\u003eLichtenberger, M., \u003cem\u003ePrinciples of shock and fluid therapy in special species.\u003c/em\u003e Seminars in Avian and Exotic Pet Medicine, 2004. \u003cstrong\u003e13\u003c/strong\u003e(3): p. 142-153.\u003c/li\u003e\n\u003cli\u003eGould, S.E., M.R. Junttila, and F.J. de Sauvage, \u003cem\u003eTranslational value of mouse models in oncology drug development.\u003c/em\u003e Nature Medicine, 2015. \u003cstrong\u003e21\u003c/strong\u003e(5): p. 431-439.\u003c/li\u003e\n\u003cli\u003eC\u0026eacute;spedes, M.V., et al., \u003cem\u003eMouse models in oncogenesis and cancer therapy.\u003c/em\u003e Clinical and Translational Oncology, 2006. \u003cstrong\u003e8\u003c/strong\u003e(5): p. 318-329.\u003c/li\u003e\n\u003cli\u003eNaji, M.A., et al., \u003cem\u003eSuper-Resolution Ultrasound Imaging Using the Erythrocytes\u0026mdash;Part II: Velocity Images.\u003c/em\u003e IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2024. \u003cstrong\u003e71\u003c/strong\u003e(8): p. 945-959.\u003c/li\u003e\n\u003cli\u003eHansen, L.N., et al. \u003cem\u003eComparison of 2D SURE and 3D CT imaging of cortical vessels in a rat kidney\u003c/em\u003e. in \u003cem\u003e2023 IEEE International Ultrasonics Symposium (IUS)\u003c/em\u003e. 2023.\u003c/li\u003e\n\u003cli\u003eNaji, M.A., et al., \u003cem\u003eTranscutaneous Super-Resolution Ultrasound Imaging using Erythrocytes versus Microbubbles in a Rabbit Kidney\u003c/em\u003e. 2024: Proceedings of the 2024 IEEE Ultrasonics, Ferroelectrics, and Frequency Control Joint Symposium IEEE. 1-4.\u003c/li\u003e\n\u003cli\u003eNaji, M.A., et al. \u003cem\u003eSuper-Resolution Ultrasound Imaging using Erythrocytes on an Axillary Human Lymph Node\u003c/em\u003e. in \u003cem\u003e2024 IEEE Ultrasonics, Ferroelectrics, and Frequency Control Joint Symposium (UFFC-JS)\u003c/em\u003e. 2024.\u003c/li\u003e\n\u003cli\u003eJensen, J.A., et al., \u003cem\u003eUniversal Synthetic Aperture Sequence for Anatomic, Functional, and Super Resolution Imaging.\u003c/em\u003e IEEE Trans Ultrason Ferroelectr Freq Control, 2023. \u003cstrong\u003e70\u003c/strong\u003e(7): p. 708-720.\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eThe British Medical Ultrasound Society. Guidelines for the safe use of diagnostic ultrasound equipment\u003c/em\u003e, in \u003cem\u003eDiagnostic Ultrasound: Physics and Equipment\u003c/em\u003e, P.R. Hoskins, K. Martin, and A. Thrush, Editors. 2010, Cambridge University Press: Cambridge. p. 217-225.\u003c/li\u003e\n\u003cli\u003eHuang, C., et al., \u003cem\u003eSimultaneous Noise Suppression and Incoherent Artifact Reduction in Ultrafast Ultrasound Vascular Imaging.\u003c/em\u003e IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2021. \u003cstrong\u003e68\u003c/strong\u003e(6): p. 2075-2085.\u003c/li\u003e\n\u003cli\u003eKasoji, S.K., et al., \u003cem\u003eEarly Assessment of Tumor Response to Radiation Therapy using High-Resolution Quantitative Microvascular Ultrasound Imaging.\u003c/em\u003e Theranostics, 2018. \u003cstrong\u003e8\u003c/strong\u003e(1): p. 156-168.\u003c/li\u003e\n\u003cli\u003eGhosh, D., et al. \u003cem\u003eMonitoring early tumor response to vascular targeted therapy using super-resolution ultrasound imaging\u003c/em\u003e. in \u003cem\u003e2017 IEEE International Ultrasonics Symposium (IUS)\u003c/em\u003e. 2017.\u003c/li\u003e\n\u003cli\u003eHoyt, K., \u003cem\u003eSuper-Resolution Ultrasound Imaging for Monitoring the Therapeutic Efficacy of a Vascular Disrupting Agent in an Animal Model of Breast Cancer.\u003c/em\u003e Journal of Ultrasound in Medicine, 2024. \u003cstrong\u003e43\u003c/strong\u003e(6): p. 1099-1107.\u003c/li\u003e\n\u003cli\u003eYin, J., et al., \u003cem\u003ePattern recognition of microcirculation with super-resolution ultrasound imaging provides markers for early tumor response to anti-angiogenic therapy.\u003c/em\u003e Theranostics, 2024. \u003cstrong\u003e14\u003c/strong\u003e(3): p. 1312-1324.\u003c/li\u003e\n\u003cli\u003eHuang, C., et al., \u003cem\u003eNoninvasive Contrast-Free 3D Evaluation of Tumor Angiogenesis with Ultrasensitive Ultrasound Microvessel Imaging.\u003c/em\u003e Scientific Reports, 2019. \u003cstrong\u003e9\u003c/strong\u003e(1): p. 4907.\u003c/li\u003e\n\u003cli\u003eTernifi, R., et al., \u003cem\u003eQuantitative Biomarkers for Cancer Detection Using Contrast-Free Ultrasound High-Definition Microvessel Imaging: Fractal Dimension, Murray\u0026rsquo;s Deviation, Bifurcation Angle \u0026amp; Spatial Vascularity Pattern.\u003c/em\u003e IEEE Transactions on Medical Imaging, 2021. \u003cstrong\u003e40\u003c/strong\u003e(12): p. 3891-3900.\u003c/li\u003e\n\u003cli\u003eTailor, A., et al., \u003cem\u003eA comparison of intratumoural indices of blood flow velocity and impedance for the diagnosis of ovarian cancer.\u003c/em\u003e Ultrasound in Medicine \u0026amp; Biology, 1996. \u003cstrong\u003e22\u003c/strong\u003e(7): p. 837-843.\u003c/li\u003e\n\u003cli\u003eAlamer, M. and X. Yun Xu, \u003cem\u003eThe influence of tumour vasculature on fluid flow in solid tumours: a mathematical modelling study.\u003c/em\u003e Biophys Rep, 2021. \u003cstrong\u003e7\u003c/strong\u003e(1): p. 35-54.\u003c/li\u003e\n\u003cli\u003eDelorme, S., et al., \u003cem\u003eImaging the smallest tumor vessels using color Doppler ultrasound in an experiment.\u003c/em\u003e Radiologe, 2001. \u003cstrong\u003e41\u003c/strong\u003e(2): p. 168-72.\u003c/li\u003e\n\u003cli\u003eGiavedoni, P., et al., \u003cem\u003eAdvanced Doppler Ultrasound Insights: A Multicenter Prospective Study on Healthy Skin.\u003c/em\u003e Diagnostics (Basel), 2025. \u003cstrong\u003e15\u003c/strong\u003e(5).\u003c/li\u003e\n\u003cli\u003evon Bibra, H., et al., \u003cem\u003eLimitations of flow detection by color Doppler: in vitro comparison to conventional Doppler.\u003c/em\u003e Echocardiography, 1991. \u003cstrong\u003e8\u003c/strong\u003e(6): p. 633-42.\u003c/li\u003e\n\u003cli\u003eFabiszewska, E., et al., \u003cem\u003eEvaluation of Imaging Parameters of Ultrasound Scanners: Baseline for Future Testing.\u003c/em\u003e Pol J Radiol, 2017. \u003cstrong\u003e82\u003c/strong\u003e: p. 773-782.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Super-Resolution Ultrasound, Contrast-free Ultrasound, SURE, Oncology, Imaging, Microvasculature","lastPublishedDoi":"10.21203/rs.3.rs-6829687/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6829687/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction.\u003c/strong\u003e Angiogenesis and blood perfusion are key components of tumour development, which vary between tumour types and patients. Current clinical imaging methods have limited ability to qualitatively or quantitatively assess these parameters. This study describes a novel application of the ultrasound-based imaging modality, super-resolution ultrasound imaging using erythrocytes (SURE), and evaluates its performance against available in vivo and ex vivo imaging methods.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods.\u003c/strong\u003e An optimised SURE imaging technique was used to visualise the vasculature in a syngeneic murine model bearing subcutaneous B16-F10 melanoma tumours. For comparison, the tumour was also imaged in vivo with B-mode, colour Doppler, spectral Doppler ultrasound, and ex vivo with micro-CT angiography and immunohistochemistry.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults.\u003c/strong\u003e SURE imaged the tumour microvasculature, visualising vessels down to ≈ 30 µm in diameter while enabling extensive characterisation of tumour haemodynamics. SURE significantly outperformed colour Doppler and spectral Doppler in visualising vasculature and quantitatively assessing flow. SURE enabled analysis of the variability in perfusion across the tumour section and can enable users to evaluate the health of individual vessels. Vascular density data obtained from SURE did not significantly differ from that obtained by immunohistochemistry, demonstrating comparability even to ex vivo methods.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e. Using only 10 seconds of data acquisition, SURE imaging can provide vascular density in murine tumours at a resolution comparable to ex vivo techniques, along with haemodynamic data across the entire tumour section, distinct regions, or individual vessels. This technique has the potential to identify previously unknown biomarkers of cancer, elucidating characteristics of neoplastic angiogenesis which can be used for early diagnosis and treatment.\u003c/p\u003e","manuscriptTitle":"Quantification of Tumour Angiogenesis and Perfusion using Contrast-free Super-Resolution Ultrasound Imaging","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-30 08:54:10","doi":"10.21203/rs.3.rs-6829687/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9439d2d1-09a5-48fc-9ab3-cfb7d9e08ee1","owner":[],"postedDate":"June 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-07-10T14:08:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-30 08:54:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6829687","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6829687","identity":"rs-6829687","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}