Super-resolution imaging of urethral vasculature in healthy pre- and post-menopausal females

Summary

Previous studies have postulated that the urethral vasculature (UV) might play an important role in urinary continence for women. The goal of this research was to compare the UV in pre- and post-menopausal women using a super-resolution ultrasound imaging method called Super Ultrasound for Greater Accuracy and Resolution (SUGAR). We found that post-menopausal women exhibited decreased UV parameters such as fractal dimension, vessel proportion, and mean blood vessel diameter than pre-menopausal women. We also discriminated the vascular pattern in several layers of the urethra and its surrounding in vivo, including the urethral mucosa and submucosa, urethral muscle, and anterior vaginal wall. Besides, the statistical analysis of the vasculature pattern showed that most of the UV parameters peaked at mid-urethra. Ultimately, the UV parameters exhibited a tendency of first increasing, then reducing, and finally decreasing with age.

Graphical abstract

Highlights

• •Super-resolution sonography reveals female UV alterations before and after menopause
• •Middle segment of the urethra plays a pivotal role in urinary control
• •UV parameters exhibited a tendency first increasing, then reducing, and finally decreasing with age

Artificial intelligence

Introduction

Urinary incontinence (UI) is a significant global health issue that affects approximately 30% of reproductive and post-menopausal women.¹Current hypotheses suggest that the urethral vasculature (UV) may play a critical role in female urinary control. To maintain continence, the closing pressure of the urethra must exceed the pressure induced from the bladder. There are several beliefs that the urethral submucosal vascular plexus accounts for around 31% of the urethral closure pressure.^2,3,4 It may contribute to continence by forming a watertight seal via coaptation of the mucosal surfaces. However, the submucosal vascular plexus of the urethra has not been visualized comprehensively in vivo.

The remaining 69% of the urethral closure pressure is contributed by the urethral muscles, nerves, and paraurethral supportive structures. The perfusion status of these tissues may reflect their metabolic state.⁵ Furthermore, the density of muscle fibers^6,7 and nerve density⁸ in the female urethra were also shown, by histological methods, to decrease as women age, using sections of female urethral specimens.^6,7,8 However, the histological slice approach used to study these tissues can only uncover the static characteristics of the urethral architecture and, hence, cannot reveal the dynamic properties of blood flow or its changes in these tissues with age. The pattern of blood flow distribution in these tissues with age is unclear.

There are also some theories that the middle segment of the urethra, as opposed to the proximal and distal segments of the urethra, plays a major role in female urinary control. The anatomical rationale for this conclusion is that the muscle fiber density reaches its highest value in the middle segment of the female urethra. This static feature can be directly demonstrated by medical imaging methods. However, the distribution characteristics of the urethral vessels in different segments cannot be shown directly, and whether the distribution of blood flow in the mid-urethra is the most abundant needs to be proved. In brief, these scientific hypotheses cannot be confirmed or falsified because we have not yet provided an accurate map of the female UV at the microscopic level. Therefore, it is essential to investigate the features of the distribution of urethral microvasculature in women, as well as their changes with age.

To address this issue, an increasing number of studies have been using medical imaging methods to observe UV features, such as power Doppler (PD) and color Doppler flow imaging (CDFI).^{3,9,10,11,12,13,14,15} The Doppler imaging modality relies on red blood cells inside blood vessels as the scattering medium. However, because the number of red blood cells in small vessels is extremely limited, a large number of vessels in the urethra are not visible for Doppler imaging. To overcome this drawback, Siracusano et al.¹⁶ used contrast-enhanced ultrasonography (CEUS) as a promising alternative. This involves the infusion of microbubble agents into veins to highlight blood vessel sites for evaluating UV in healthy pre- and post-menopausal females.¹⁶ Although having deep penetration ability, the CEUS method could not display small vessels due to the acoustic diffraction limit of half wavelength. The study thereby only provided a relative assessment of UV characteristics at a macroscopic level of several millimeters. Traditional imaging modalities for UV still struggle to balance precision and depth.

To address the trade-off between resolution and penetration in imaging methods, we and other researchers have developed a new super resolution approach called Super Ultrasound for Greater Accuracy and Resolution (SUGAR).^17,18 SUGAR can precisely locate the center of each sparsely dispersed microbubble with a precision of microns, surpassing the ultrasonic diffraction limit.^19,20 Due to its exceptional resolution capabilities, SUGAR has garnered significant interest in recent years, with applications primarily in the field of fundamental or preclinical research on cancer,²¹ cardiology,²² and brain functions.^17,23 These investigations have demonstrated the immense potential of SUGAR in the field of UV characterization. However, current super resolution imaging research is mainly conducted in laboratories. The special ultrasound acquisition requirements adopted for super-resolution imaging in the laboratory setting, such as ultrafast ultrasound scanning speed, low microbubble concentration, and smaller probe vibration, may be difficult to achieve under routine clinical conditions. Therefore, further exploration is needed to determine the precise distribution of female UV via SUGAR.

This study aimed to reconstruct UV in pre- and post-menopausal women at the microscale using the super-resolution SUGAR method under routine clinical conditions, in order to check the validity of previous scientific theories about the importance of female UV in urinary control.

Results

The SUGAR approach was used to observe UV in women in the pre-menopausal (PreM) and post-menopausal (PostM) groups (Figure 1, left panel). Detailed clinical data of the participants in this study can be found in Table S3.

Figure 1.

Flowchart depicting the participation of volunteers in the study of urethral vasculature

(UV) imaging, stratified by pre- and post-menopausal status (left panel); the method of Super Ultrasound for Greater Accuracy and Resolution (SUGAR) was employed to generate super-resolution imaging results of UV from conventional contrast-enhanced ultrasound data (right panel).

(A) Acquisition of raw CEUS data using transrectal ultrasonography (TRUS).

(B) Extraction of microbubble signals.

(C) Localization of individual microbubbles with subpixel resolution.

(D) Generation of super-resolution images using SUGAR, with a resolution of ∼60 μm. PreM, pre-menopausal; PostM, post-menopausal; E, estrogen; P, progestin; BMI, body mass index.

The dimensional features of urethra for the pre- and post-menopausal volunteers were analyzed on the basis of B-mode imaging. All three parameters, i.e., urethral length, urethral area, and urethral circumference, were smaller for the PostM group than for the PreM group (see Table S4).

Next, the SUGAR imaging approach (Figure 1, right panel) obtained a microscopic distribution of urinary vasculature (UV) in the urethra for both pre- (Figures 2A and 2B) and post-menopausal women (Figures 2C and 2D). Our imaging results showed that SUGAR achieved an order of 60 μm for super-resolution imaging of the female urethra, which is much better than the millimeter resolution level of conventional ultrasound imaging and magnetic resonance imaging (MRI) (Figure S3). The spatial distribution of fine vessels in different areas of the urethra were clearly demonstrated in the acquired SUGAR super-resolution images (Figures 2A2, 2B2, 2C2, and 2D2). In addition to the topographical map, the UV velocity vectorial distribution (Figures 2A1, 2B1, 2C1, and 2D1) was obtained.

Figure 2.

Super-resolution imaging results of urinary vasculature (UV) generated by Super Ultrasound for Greater Accuracy and Resolution (SUGAR), accompanied by statistical analysis of UV parameters

(A and B) The super-resolution imaging result of a 32-year-old and a 44-year-old woman from Group PreM, respectively.

(C and D) The super-resolution imaging result of a 58-year-old woman and a 71-year-old woman from Group PostM, respectively. A1, B1, C1, and D1 show the UV velocity vectorial map; the red color represents blood flow moving toward the ultrasound probe, and the blue color represents blood flow moving away from the ultrasound probe; AVW, anterior vaginal wall; MUS, muscle layer; ML, mucosal layer; subML, submucosal layer. A2, B2, C2, and D2 show the static structure of urethral blood vessels; the color represents the probability of a microbubble occurring at a certain location. A3, B3, C3, and D3 show the maximum intensity projection (MIP). A4, B4, C4, and D4 show B-mode imaging; U, urethra; V, vagina; PS, pubic symphysis. A5, B5, C5, and D5 show conventional color Doppler flow imaging (CDFI) of UV. A6, B6, C6, and D6 show super-resolution images of UV velocity vectorial map superimposed on B-mode image.

(E) The statistical analysis results of various UV parameters; the horizontal lines represent the mean values.

The PreM group had higher average values for fractal dimension, vessel proportion, maximum blood velocity, mean blood velocity, vascular tortuosity index, mean vessel diameter, diameter of large vessels, and diameter of small vessels compared with the PostM group (Table S5). Statistically significant differences were observed in fractal dimension, vessel proportion, and mean diameter.

It is noteworthy that the SUGAR method has clearly demonstrated the details inside different layers of the female urethra and its surroundings. The blood flow status of the mucosal layer (ML), submucosal layer (subML), muscles (MUS), and anterior vaginal wall (AVW) has been successfully achieved via SUGAR, whereas conventional CDFI struggles to provide similar information. For all the cases herein, subML possesses a richer blood flow morphology than MUS (Figures 2A1, 2A2, 2B1, 2B2, 2C1, 2C2, 2D1, and 2D2). The blood flow in AVW decreased after menopause (Figures 2A1 and 2B1 versus 2C1 and 2D1).

To further analyze the variation of the statistical parameters above in different regions of the urethra, we also plotted the variation of the SUGAR outcome parameters in different locations of the urethra with regard to the urethral midline (Figure 3). In terms of UV parameters, the PreM and PostM groups have certain similarities. The distance 5–35 mm, which roughly corresponds to the mid-urethra, exhibited the highest values of UV characteristics. The UV parameters increased with increasing distance in the range of 0∼5 mm, which roughly corresponds to the proximal urethra. These parameters were also shown to decline with distance in the range of 35∼40 mm, i.e., the distal urethra. On the other hand, there are differences between the two groups about the parameters. There is no discernible change in UV measurements between the PreM and PostM groups in the proximal and distal parts. The UV parameters in the mid urethra, however, are greater in the PreM group than in the PostM group.

Figure 3.

Variation in urinary vasculature (UV) parameters across the internal to external urethral orifice

Each parameter is defined as a function of the form y = f(x), and a coordinate system is established on the B-mode image. The zero point of the x and y axes is defined as the location of the internal urethral orifice. The x axis represents the distance along the urethral lumen, with dS denoting the infinitesimal region perpendicular to the x axis at the distance of x. The y axis represents the value of each UV parameter, including (A), fractal dimension, (B), vessel proportion, (C), maximum blood velocity, (D), mean blood velocity, (E), vascular tortuosity index, (F), mean vessel diameter, (G), diameter of large vessels, and (H), diameter of tiny vessels.

For the PreM and PostM groups of participants, our SUGAR imaging approach produced super-resolution blood flow data while achieving macroscopic ultrasonic contrast time-intensity curves (TIC) (Figure 4). The PreM group had lower average values of AT and TTP than the PostM group. The PreM group showed higher mean values for PI, AS, DS, AUC, and MTT than the PostM group, with statistically significant differences observed in AS and AUC.

Figure 4.

Characterization of urethral vasculature (UV) in pre- and post-menopausal women using time intensity curves (TICs)

(A) TICs for the PreM group, and (B) for the PostM group. A1–A4 and B1–B4 show images of contrast enhanced ultrasonography (CEUS) that represent the intensity of microbubbles at 20 s, 30 s, 40 s, and 50 s, respectively. A5 and B5 show TIC curves.

(C) The statistical analysis results of TIC for each participant in the groups PreM and PostM. AT, arrival time; TTP, time to peak; PI, peak intensity; AS, ascending slope; DS, descending slope; AUC, area under curve; and MTT, mean transit time. The horizontal lines represent the mean values.

Finally, all the super-resolution UV parameters were regrouped according to different age ranges (Figure 5). The fractal dimension peaked in the third decade. Other parameters including the vessel proportion, mean velocity, maximum velocity, and mean diameter reached their maximum in the fourth decade. The vessel tortuosity showed its largest value in the fifth decade. Then, all these parameters declined with age until the age of about 64 years. During the age range of 64–74 years, all these UV parameters started to increase.

Figure 5.

Parameters of super resolution urethral vasculature (UV) stratified by age

All of the 22 participants, from both the PreM and PostM groups, were divided into incremental age ranges of 20∼29, 30–39, 40–49, 50–59, 60–69, and 70–79 years. For each age range, the parameters of UV and the ages of the participants, inside this age range, underwent arithmetic average.

(A) Fractal dimension; the inset in (A) represents the prevalence rate of urinary incontinence (UI) for females, using epidemiological data compiled from ref.²⁴

(B) Vessel proportion.

(C) Maximum blood velocity.

(D) Mean blood velocity.

(E) Vascular tortuosity index.

(F) Mean vessel diameter.

Discussion

Our results have three implications. Firstly, the pattern of UV alterations was revealed in women after menopause, which may explain why post-menopausal women are prone to UI. Secondly, the super-resolution imaging of female urethra verified an earlier scientific hypothesis that the middle segment of the urethra plays a pivotal role in urinary control. Finally, the UV parameters exhibited a tendency of first increasing, then reducing, and finally decreasing with age.

The study demonstrated significant differences in statistical parameters between the PreM and PostM groups (Table S5). The absence of estrogen in the PostM group restricts the division and maturation of vascular endothelial cells,²⁵ leading to a decrease in the submucosal vessels in the urethra. Additionally, there were more vaginal deliveries in the PostM group than in the PreM group, which can cause damage to the urethral tissue and the blood vessels embedded in the tissue.^26,27 These factors contribute to the weaker angiogenesis ability of the PostM group, with existing vessels prone to atrophy. Consequently, the values of fractal dimension, vessel proportion, and vessel diameter were reduced in the PostM group, as observed in our super-resolution imaging experiments. Notably, VTI values do not decrease after menopause (Figure 2E5). This could be due to the compensatory mechanism triggered in women after menopause. The body improves the degree of urethral perfusion by increasing the degree of vascular tortuosity, which compensates for the reduced vascular regenerative capacity and vessel proportion after menopause.

Macroscopic epidemiological surveys suggest that UI is more prevalent among post-menopausal women, for which our super-resolution results may provide an explanation through transient analysis. By utilizing the super-resolution blood flow information as input to a dynamic model that considers fluid capacitance, fluid resistance, and fluid inductance (as shown in Figure 6), we can gain insight into why post-menopausal women are at a higher risk of developing UI. The model can be viewed as a transfer function that aims to study the buffering and absorption capacity of the urethral vascular system against external pressures under various urethral perfusion conditions. We utilized the model to solve for the output u_o. The u_o, buffered by the urethral vascular system, is against urethral closure pressure. To investigate the most extreme conditions, we have set u_i as a periodic pulse that mimics the scenario of coughing that causes urinary leaks in women. Further information on the model and parameter values can be found in the supplementary material (S1) for this article. According to the findings of super-resolution imaging (Figure 2), post-menopausal women have decreased fluid capacitance due to a decrease in vascular percentage and blood flow velocity. As illustrated in Figure 6, the result of this model computation is that u_o(Post) > u_o(Pre). This suggests that the entire urethral dynamic system has a decreased capacity to withstand spike pulses of external pressure when the fluid capacitance of the urethral blood flow diminishes. Such result indicates that post-menopausal women, whose UV is comparatively thin, have less tolerance to the shock pressure from the urethra and are thus more susceptible to urinary incontinence.

Figure 6.

A urethral vasculature (UV) dynamic model was established, utilizing super-resolution data yielded by ultrasound localization microscopy (SUGAR) as input, to explain why post-menopausal women are more susceptible to urinary incontinence (UI) compared with pre-menopausal women

Detailed explanation on modeling and parameter choice can be found in Figure S1.

The acquired UV parameters carry their own clinical significance, with details shown in the S2 section of supplementary materials.

Our super-resolution imaging results from SUGAR not only allow for direct analysis of UV parameters but also provide information on the perfusion status within the different layers of the urethra and its surroundings. Using the SUGAR method, we were able to clearly demonstrate the distribution of blood flow in the urethral ML and subML, MUS, and AVW regions.⁴ We observed a significant reduction in blood perfusion in these layers in the PostM group compared with the PreM group (Figures 2A1 and 2B1 versus 2C1 and 2D1). Additionally, we identified a rich submucosal vascular plexus (Figures 2A1 and 2B1), allowing for visualization of the complete cushion-like structure that closes the urethral lumen. Although previous studies have mentioned this structure, it could not be observed using regular medical imaging methods. When the surrounding muscles slightly compress this vascular plexus, it produces an evident closure process of the urethral lumen. Regarding the distribution of blood flow in the MUS, the blood flow near the subML region belongs to the urethral smooth muscle, whereas the blood flow near the anterior and posterior borders of the urethra belongs to the urethral striated muscle. However, due to the lack of a clear demarcation between the two types of muscles, we did not further delineate them in this study. Meanwhile, the periurethral supportive tissue, which includes the AVW and endopelvic fascia, plays a crucial role in urinary continence by forming the "hammock" structure.²⁴ This structure compresses the urethra when abdominal pressure increases, preventing urinary leakage. Our SUGAR method provides a clear map of blood flow in the AVW, thereby providing robust evidence for the assessment of the periurethral supportive tissue. In the future, we anticipate observing blood flow in other surrounding supportive structures of the urethra, such as the endopelvic fascia and pubic nerves. This development can enhance the clinical assessment of UI by identifying which tissue has insufficient blood perfusion. Furthermore, it can refine the typology of stressed UI (SUI), urgency UI (UUI), and mixed UI (MUI), allowing for pinpointing the specific point of ischemia. By doing so, targeted treatments can be effectively implemented. The SUGAR method is also expected to evaluate the efficacy of various treatments including pelvic floor muscle training, stem cell therapy, electrical stimulation, and acupuncture. As such, our SUGAR method has the potential to improve our understanding of the pathophysiology of UI and enhance the clinical management of this condition.

Besides analyzing different layers of the urethra, our super-resolution images of microvasculature have provided a comprehensive pattern of the variation of UV parameters across the internal to external urethral orifice (Figure 3). First, the parameter curves of the PreM group, except for VTI (Figure 3E), were generally higher than those of the PostM group. Second, we found that all UV parameters, both in the PreM group and PostM groups, peaked in the midline of the urethra at a distance range of 5∼35 mm (Figure 3), corresponding to the mid-urethra. This pattern helps validate a previously proposed hypothesis in the mechanism of urinary continence that the mid-urethra plays a vital role in urethral closure compared with other segments.^28,29,30

In addition to its super-resolution ability, SUGAR allows the output of conventional ultrasound time intensity curves (TICs). The PreM group exhibited a greater ascending slope (AS), larger peak intensity (PI), and area under curve (AUC). An increase in AS indicates a faster perfusion of microbubbles in the UV for the PreM group. The larger value of PI for the PreM group indicates that the microbubble concentration is denser than that of the PostM group. The increased value of AUC suggests that the total amount of UV influx is greater in the PreM group than in the PostM group. These AUC results are consistent with the classic study conducted by Siracusano et al.¹⁶

The aforementioned seminal study has successfully proven that CEUS is a safe and effective tool for studying UV. However, the reperfusion curve only offered a relative assessment of flow characteristics and could not directly evaluate blood velocity flow and signal intensity. In contrast, our SUGAR approach uses super-resolution reconstruction of UV and can achieve resolutions in microns beyond the acoustic diffraction limit of half wavelength. Based on classic studies, our research introduced more refined super-resolution parameters for microcirculation and realized direct parameter analysis. Our findings not only confirmed the classic study but also revealed that blood flow velocity changes subtly after menopause. Such a refined result is difficult to observe using traditional CEUS.

Macroscopic epidemiological studies showed that the occurrence of female UI did not rise monotonically with age, leaving an unanticipated drop between the ages of 64 and 74 years. Our super-resolution findings suggest that there must be a microscopic mechanism behind this phenomenon. Interestingly, some of the UV parameters do not change as expected with age, too. Generally, we would propose that these UV parameters should decrease monotonically with age, especially after menopause. However, the actual result was that VTI begins to decline after approximately 55 years old, and the remaining UV parameters begin to decrease between 40 and 60 years (Figure 5). These trends are in line with conventional assumptions. The reason behind this phenomenon may be related to (1) the beginning of estrogen decline and (2) the decreased density of urethral muscle fibers and nerves between the ages of 40 and 60 years. However, between the ages of 64 and 74 years, there is an increasing trend in all parameters of UV. Coincidentally, the epidemiological results of UI also show a negative correlation with UV parameters when the incidence of UI continues to increase until the age of 64 years but decreases between the ages of 64 and 74 years (see the inset in Figure 5A). Thus, the question arises: how should we interpret the correlation between microscopic super-resolution results and macroscopic epidemiological findings? It is now accepted that the density of urethral muscle and nerves decreases with age in women.^31,32 Accordingly, we can propose an educated hypothesis that the decrease in UI incidence between the ages of 64 and 74 years may be associated with a compensatory increase in UV. Of course, more evidence is needed to confirm this assumption. Nonetheless, our super-resolution approach has observed subtle trends in UV parameters varying with age, which helps propose new scientific hypotheses about the mechanism of UI.

To sum up, we reported a large difference in female UV before and after menopause, at the microscopic level, using super-resolution SUGAR. Our work uncovered, for the first time, the complete anatomical structure of the submucosal vascular plexus. We also discovered that UV parameters are highest in the mid-urethra, supporting the scientific theory in the urological community that the mid-urethra plays a key role in female urinary control. Finally, we reported that the UV parameters exhibited a tendency of first increasing, then reducing, and finally decreasing with age. A critical future study is that we may create more customized therapy for female patients with UI using super-resolution SUGAR methods by highlighting the precise location of UV ischemic zones.

Limitations of the study

Our study has two limitations. First, the SUGAR method requires an offline post-data processing procedure after acquiring raw ultrasonic data. In response to this limitation, we have tried to decrease the post-processing time of generating super-resolution data in most clinical examination settings to less than 10 min. The second limitation is that our study is limited to a single section, albeit the most representative one. The examination of the whole urethral vascularity requires the use of 3D imaging methods. In future investigations, we will employ our 3D ULM approach for urethral imaging to improve the outcomes.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Ultrasonic contrast agent	Bracco Imaging B.V, Plan-Les-Ouates Geneva, Switzerland	CAS: 2551-62-4
MRI contrast agent	Shanghai Xudong Haipu Pharmaceutical Co., LTD, Shanghai, China	N/A
Critical commercial assays
Ultrasonic scanner	Shenzhen Mindray Bio-Medical Electronics Co., Ltd., Shenzhen, China	https://www.mindray.com/en/products/ultrasound
MRI scanner	Siemens AG, Germany	https://www.siemens-healthineers.com/magnetic-resonance-imaging
Software and algorithms
SUGAR	This paper	https://orcid.org/0000-0002-5532-7604
ImageJ	National Institute of Health, USA	https://imagej.nih.gov/ij/
GraphPad Prism	GraphPad software, San Diego, California, USA	https://www.graphpad.com/

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Chen Hua (Ch.Hua@sjtu.edu.cn).

Materials availability

• This study did not generate new unique reagents.

Data and code availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •SUGAR source code is available from the lead contact upon request.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Our team recruited healthy female volunteers from the community who met specific criteria, including having no pelvic floor disease, urinary tract infection, history of urinary incontinence or sexual disorders, and no exposure to outside interventions such as birth control pills or other hormonal therapies. We included a total of 43 women in the study, but two were excluded due to their use of hormone medication and a history of incontinence (Figure 1, left panel). The majority of the participants were Han ethnic people who lived in China. The remaining 41 women were divided into two groups based on their menstrual status: pre-menopausal (PreM) and post-menopausal (PostM) volunteers. Initially, we had 30 PreM candidates, but ultimately, we classified 11 females in the follicular period as PreM group participants to avoid the influence of hormones. We assigned the remaining 11 participants who had no menstruation to the PostM group. We defined post-menopausal status as no menstruation for at least one year. Before conducting ultrasonography, we collected clinical information from all participants, such as age, body mass index (BMI), estrogen (E), and progestin (P) levels, number of pregnancies and number of vaginal deliveries. Prior to participating in our SUGAR trials, all participants had been informed of the study’s objective, and The Ethics Committee of The Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine authorized our investigations.

Method details

Imaging steps of Super Ultrasound for Greater Accuracy and Resolution (SUGAR)

We then performed ultrasonic examination for both PreM and PostM groups using Mindray ultrasonography equipment (Resona R9, Shenzhen Mindray Bio-Medical Electronics Co., Ltd., Shenzhen, PRC). We conducted transrectal ultrasonography (TRUS) with a biplane multifrequency transducer (3.2–12.8 MHz linear array and 3.5–9.5 MHz convex array) on all patients while they lay on their left side (Figure 1A). We captured B-mode and CDFI images on both sagittal and cross-sections.

After transrectal ultrasonography (TRUS), we administered an intravenous bolus of ultrasound contrast agents (Bracco Imaging B.V, Plan-Les-Ouates Geneva, Switzerland) to each participant. The participants received 1.6 mL of diluted SonoVue agents via cubital vein injection after diluting the agents to a volume of 5 mL. Using a frame rate of 30–75 fps, we conducted conventional contrast imaging on the urethral region of interest (ROI) and recorded the movement of the urethral tissues during B-mode acquisition. To ensure that we accurately captured the complete microbubble movement process, including appearance, rise, and disappearance, we recorded the imaging for 200 s. The imaging parameters were consistent for all participants. We mainly used sagittal section footage for inspection because the acoustic beam’s path was parallel to the urethral axis, and the frequency was relatively high, ensuring resolution ability. After raw ultrasonic data acquisition, the B-mode images of all the participants were analyzed by ImageJ 1.53t (National Institute of Health, USA) to yield macroscopic size information of the urethra.

Then we started super resolution imaging reconstruction. One investigator, who was blinded to the patient categorization displayed in Figure 1 (left panel), performed offline processing of the raw ultrasonic video data, including ultrasound contrast and B-mode signals. Our custom-built SUGAR method for generating super resolution images completed the post-processing operation, which initially cropped the images on a temporal scale to ensure that microbubbles were visible to the naked eye. Then, the data underwent motion correction to account for motion errors caused by transducer movement or human breathing, which could impact the image quality. We then extracted the microbubble signal from the captured video using a bandpass filter that could identify the microbubble signal source from other sources across all frames of the video (Figure 1B). We determined the precise location of each microbubble within each frame of the collected video (Figure 1C) and obtained a precise topological distribution of large and tiny blood vessels inside the ROI after microbubble localization (Figure 1D).

Quantification and statistical analysis

After obtaining structural and velocity information of UV through SUGAR, we retrieved various parameters for further statistical analysis. These parameters include fractal dimension, vessel proportion, maximum blood velocity, mean blood velocity, vascular tortuosity index, mean vessel diameter, diameter of large vessels, and diameter of tiny vessels. We will provide a detailed explanation of how to define or calculate each parameter. The value of fractal dimension is calculated using the box counting method to reflect the branching tendency of blood vessels. The region of interest (ROI) of the super resolution images is meshed with small squares with an edge length of η. The number of square boxes with blood vessels inside is counted as N(η). This process is repeated for different values of η, resulting in a series of N(η) that changes with η. The fractal dimension value is obtained by using the following expression once η is small enough:

$$Fractaldimension=\underset{\eta \to 0}{\lim }\frac{\log N\left(\eta \right)}{\log \left(\frac{1}{\eta }\right)}$$ (Equation 1)

The vessel proportion is defined as the percentage of the area that blood vessels occupy in a given ROI. Blood velocity, including both maximum and mean velocities, is calculated by taking the first derivative of the displacement vector of each microbubble with respect to time t. The vascular tortuosity index is calculated as the ratio of two lengths, namely $=\frac{\hat{C}}{L}$ . Here, $\hat{C}$ is the length of one curved blood vessel, and L is the length of the straight line that points from the head to the tail of the target blood vessel, as illustrated in Figure S2. The mean vessel diameter is the average of the diameter of each blood vessel within a given ROI. We then use a threshold of 2 mm to separate large vessels from small ones, which is a typical value of conventional ultrasound resolution used in clinical settings. Vessels with a diameter larger than 2 mm are categorized as large vessels, and those less than 2 mm are categorized as tiny vessels. The diameter of large vessels is defined as the average diameter of large vessels, and the diameter of tiny vessels is defined as the average diameter of tiny vessels.

We have plotted the variations of these parameters across different urethral anatomical segments to illustrate UV’s location-dependent characteristics. We used the urethral luminal line as the x axis, with the internal urethral orifice as the origin. The displacement vector $\overrightarrow{x}$ along the urethral luminal line was set as the independent variable, while the parameters of UV were set as the dependent variable $y\left(\overrightarrow{x}\right)$. In other words, $y\left(\overrightarrow{x}\right)$ represents the values of the UV parameters within the infinitesimal area of dS at the site of $\overrightarrow{x}$ (see the central part of Figure 3).

Furthermore, we analyzed the time-intensity curve (TIC) of the urethra, using parameters such as arrival time (AT), time to peak (TTP), peak intensity (PI), ascending slope (AS), descending slope (DS), area under the curve (AUC), and mean transit time (MTT) to analyze urethral blood perfusion characteristics.

Finally, we regrouped all the super resolution UV parameters according to different age ranges. We conducted statistical analysis using the GraphPad Prism 9.5.0 software package (GraphPad Software, San Diego, California, USA). The T-test was utilized to compare the parameters of the PreM and PostM groups, and the data were presented as mean standard deviation. Cases with p values lower than 0.05 were considered statistically significant.

Additional resources

Ethics approval was provided by The Ethics Committee of The Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine [2022-KY-170(K) and 2021-138]. All of the participants involved in this study signed written informed consent. ChiCTR registration: ChiCTR2100048361.

Published: February 22, 2024