A cross-shaped combined ultrasound array patch for continuous bladder volume monitoring Authors: Ziqi Li, Cong Pu, Ting Zhang, Chang Peng Published: 2026 ABSTRACT AbstractContinuous monitoring of bladder volume is important for the assessment and management of lower urinary tract dysfunction. The study aims to develop a wearable non-invasive ultrasound patch for continuous bladder volume monitoring, addressing operator dependence, limited field of view (FOV), and manual probe rotation required in conventional handheld ultrasound. A cross-shaped combined ultrasound array patch (CCUA-Patch) was proposed using three 64-element linear arrays, including two vertically aligned arrays for extended sagittal imaging and one horizontal array for transverse imaging. Time-division scanning enabled multi-plane acquisition, and a feature-based stitching algorithm reconstructed a complete sagittal image. Bladder contours were automatically extracted, and volume was estimated using the prolate ellipsoid method. A 2 mm self-adhesive hydrogel layer was used to ensure stable acoustic coupling for hands-free continuous monitoring. The fabricated transducers exhibited a center frequency of 2.60 ± 0.02 MHz and a −6 dB bandwidth of 86.67 ± 2.73%. In phantom experiments, the estimated bladder volume was 126.14 mL versus a calibrated volume of 137 mL, with only 0.89 mL difference from a commercial ultrasound probe. In a pilot study of 10 healthy volunteers, the CCUA-Patch showed good agreement with the commercial device, with a mean Bland-Altman bias of 17.48 mL and most measurements falling within the 95% limits of agreement. The proposed CCUA-Patch expands the effective sagittal FOV, eliminates manual probe reorientation, and provides a practical solution for automated monitoring of the bladder volume. However, the present device used a benchtop Vantage 256 platform and only ten healthy volunteers; it therefore demonstrates the wearability and feasibility of the patch structure, not a fully wearable clinical system. Larger studies in target patient populations and integration with miniaturized, low-power readout electronics are required before clinical or home use.Keywords: Bladder volume monitoring Wearable ultrasound patch Cross-shaped ultrasound array Field-of-view expansion Continuous monitoring Self-adhesive hydrogel Introduction Bladder dysfunction refers to various functional abnormalities occurring during the processes of urine storage and voiding. Its clinical manifestations encompass a wide range of symptoms, including urinary frequency, urgency, incontinence, dysuria, and urinary retention. The etiologies of chronic bladder dysfunction are diverse, spanning neurological disorders such as spinal cord injury and multiple sclerosis, urinary tract infections, prostatic diseases, and physiological decline associated with aging [1,2]. Without timely intervention, it may trigger a series of severe complications [[3], [4], [5]], such as recurrent urinary tract infections, bladder wall thickening, detrusor underactivity, and even chronic renal failure [6,7]. Accurate and continuous monitoring of bladder capacity and lower urinary tract function plays a vital role in clinical diagnosis, treatment, complication prevention, and postoperative recovery [[8], [9], [10]]. Compared to voiding diaries that rely on subjective descriptions [11], continuous capacity monitoring captures transient pathological features through real-time recording, providing an objective diagnostic basis for overactive bladder (OAB) and lower urinary tract dysfunction [11,12]. Based on dynamic capacity curves, clinicians can quantify the extent to which medications, such as M-receptor antagonists, improve bladder compliance and adjust dosages accordingly [13]. For patients requiring intermittent catheterization, continuous monitoring guides voiding at the optimal capacity point, thereby enhancing the precision and efficacy of treatment. Particularly for patients with impaired sensory function, continuous monitoring provides timely voiding reminders to prevent high intravesical pressure from causing vesicoureteral reflux, reducing the risk of severe complications such as hydronephrosis, renal failure, and recurrent urinary tract infections [[14], [15], [16], [17]]. Furthermore, dynamic monitoring of post-void residual (PVR) volume during the perioperative period can scientifically guide the timing of catheter removal, reducing the risk of urinary retention from premature removal or infection caused by delayed removal [[18], [19], [20]]. During the rehabilitation phase, real-time capacity feedback assists patients in bladder training to regain sensory awareness, significantly shortening the recovery cycle of urinary function [21]. In current clinical practice, the gold standard for measuring bladder volume remains invasive catheterization. Although accurate, the procedure is associated with significant pain and a high risk of secondary urinary tract infections [22,23]. Furthermore, it demands high professional proficiency from operators [24], making it difficult to popularize for daily home use or long-term monitoring scenarios. As a non-invasive alternative, ultrasound imaging has become the most common tool for assessing bladder status due to its lack of radiation, low cost, and ease of use. Its physical principle lies in the distinct acoustic impedance mismatch between urine and the bladder wall; as ultrasound waves penetrate human tissue, they generate clear reflection echoes that allow for accurate identification of bladder boundaries. To achieve sufficient precision, bladder volume calculation typically requires scanning different bladder planes, followed by volume estimation using signal processing or mathematical models, such as the prolate ellipsoid method [25], double-area method [[26], [27], [28]], and the super-ellipsoid method [29]. Among these, the prolate ellipsoid method is the most classic and widely used algorithm in clinical and commercial portable ultrasound devices. It requires obtaining two mutually perpendicular sections in the sagittal and transverse planes and measuring the maximum bladder diameters along three orthogonal axes. Traditional ultrasound examinations rely heavily on bulky equipment and handheld probes, requiring professional physicians to manually locate sections and adjust angles, which precludes accurate, long-term continuous monitoring. In recent years, advancements in flexible electronics and piezoelectric materials have led to the development of various portable and wearable bladder monitoring devices. Based on their operating modes, these devices are categorized into one-dimensional echo (A-mode), two-dimensional imaging (B-mode), and three-dimensional reconstruction (3D-mode). In 2010, Tanaka et al. [30] improved the single-probe A-mode format by using 12 fixed-array probes and a geometric segmentation model. By summing segmented bladder slices, they significantly enhanced measurement accuracy; however, the resulting device was bulky, lacked portability, and had poor skin conformability. In 2023, Zhang et al. [31] developed a patch composed of five 64-element phased-array transducers. Through a specific array layout, they achieved B-mode bladder imaging and measurement without manual rotation. While the connections between probes were made of flexible materials for better conformability and validated on four volunteers, the multi-probe design served primarily to facilitate bladder localization. Multi-image feature fusion remained in the phantom validation stage and was not realized in human bladder imaging; the two sectional images presented in the study were limited to a single probe’s field of view (FOV), showing incomplete bladder structures. Moreover, the use of five transducers increased the complexity of the patch, hindering efficient clinical application. In 2025, Park et al. [32] proposed an automated bladder volume measurement system using a T-shaped cross-array ultrasound probe combined with manual shifting for 3D reconstruction. The system utilized transverse images for motion compensation and reconstructed continuous sagittal images into a 3D volume to reduce probe positioning errors. However, 3D-mode is relatively time-consuming, unsuitable for real-time continuous monitoring, and involves high overall costs [33]. Overall, while B-mode bladder ultrasound offers a balance between accuracy and cost, several challenges remain. The narrow FOV of a single probe often results in incomplete imaging of large-volume bladders. Combined with the need for scanning mutually perpendicular cross-sections in volume calculations, this usually necessitates manual probe rotation, making long-term continuous monitoring difficult. To enable complete, dual-plane imaging of large bladders without manual adjustment, this work developed a cross-shaped combined ultrasound array patch (CCUA-Patch) for continuous bladder volume monitoring. By utilizing three 64-element ultrasound arrays placed at a specific interval and employing feature-based image stitching and fusion, the system expands the sagittal imaging FOV. Its cross-shaped design allows for the automatic and rapid acquisition of bladder cross-sections in two directions, while a self-adhesive hydrogel at the base eliminates the need for coupling agents and handheld operation, facilitating continuous monitoring. Validation with ten volunteers demonstrates that this design effectively overcomes the FOV limitations of traditional single-probe imaging in continuous volume monitoring. Methods Fabrication of the CCUA-Patch The bladder is a muscular, sac-like organ for urine storage; its shape, size, and wall thickness vary according to the degree of fullness. In healthy adults, the average bladder capacity ranges from 350 to 500 mL, with a maximum capacity reaching 800 mL [34]. When the volume exceeds 500 mL, excessive tension in the bladder wall triggers pain. For patients with mechanical obstruction, neurogenic dysfunction, or chronic detrusor damage, bladder capacity often far exceeds normal limits. To avoid the "truncation effect", large-volume bladders require a wider FOV to ensure complete image acquisition, which is essential for accurate volume monitoring. Furthermore, constrained by pelvic anatomy and surrounding ligaments, the bladder primarily expands toward the cranial side and the anterior abdominal wall during filling, appearing as an approximate ovoid shape that extends above the superior border of the pubic symphysis [[35], [36], [37]]. Therefore, to fully capture both the superior and inferior walls of a high-capacity bladder without moving the probe, a longer longitudinal FOV is required in the sagittal plane, necessitating an imaging system with sufficient longitudinal span. To this end, we proposed a cross-shaped combined ultrasound array layout. Fig. 1 presents a schematic illustration of bladder volume monitoring using the CCUA-Patch. This design symmetrically arranges two longitudinal arrays (L1 and L2) along the vertical axis and embeds a third horizontal array (H1) transversely along their central vertical axis. The core advantage of this design is that the sectional images generated by the two longitudinal arrays exist in the same physical plane. Utilizing feature point extraction and image fusion algorithms, these two images can be seamlessly stitched together, achieving a longitudinal FOV sufficient to cover a large-capacity bladder without repositioning the patch. Schematic illustration of bladder volume monitoring using the CCUA-Patch. The system utilizes three 1D linear arrays with a center frequency of 2.5 MHz as the core sensing units. Each array integrates 64 elements, with active area dimensions of 19.15 mm × 12.00 mm, a pitch of 0.30 mm. During transducer design, the multilayer structure was evaluated using a one-dimensional Krimholtz–Leedom–Matthaei (KLM) model. The simulated frequency response predicted a center frequency of 2.51 MHz and a −6 dB fractional bandwidth of 83.7%. Based on the KLM-derived structural design, fabrication process began by adhering conductive epoxy strip (Esolder 3022, Von Roll USA, Inc.) onto one side of a 0.9-mm-thick 1–3 PZT-5H composite. The PZT-5H ceramic volume fraction of the composite was 52.5%. The kt of commercially procured 1–3 PZT-5H composite was 0.65 ± 0.0012. A 1 mm wide high-temperature polyimide tape (7413D, 3 M, USA) was applied to the top edge of the piezoelectric material on the same side. After sputtering a Cr/Au electrode onto the upper surface, the tape was removed and dicing was performed by an automatic dicing saw (DISCO, Model DAD 323, Tokyo, Japan) using a 50 µm-thick blade to create 65 dicing lines with a depth of 800 µm. These kerfs were filled with epoxy (EPO-TEK 301, Epoxy Technology Inc. USA) and the material was ground down to a target thickness of 720 µm after curing. A Cr/Au electrode was then sputtered onto the bottom surface. The acoustic matching structure was constructed using 280 µm of Al2O3/epoxy and 270 µm of EPO-TEK 301 as the first and second matching layers respectively. A customized FPC was attached to the top of the material along with a 6 mm thick tungsten/epoxy mixture serving as the backing layer. Finally, the entire assembly was encapsulated in epoxy to produce the lateral array H1 with final dimensions of 23 mm × 15.6 mm. Furthermore, the packaging for the longitudinal arrays (L1 and L2) was redesigned to reduce dimensions for clinical portability. Using epoxy for potting and curing, the external dimensions were reduced to 20.2 mm × 13.9 mm. Regarding structural support, a 4.5 mm thick cross-shaped reinforcement frame was precisely modeled using SolidWorks software (2025, Dassault Systèmes, France). To maintain structural integrity and prevent fatigue fracture, the frame struts were widened. Based on physical stress testing, the strut width was finalized at 4 mm to balance structural strength with skin conformability. To address the limited scanning range found in female with bladder capacities exceeding 450 mL, this design introduces a 6 mm structural buffer gap at the intersection of the transverse and longitudinal transducers. This geometric spacing was determined based on the array’s deflection scan angle and the average anatomical depth of the bladder. By “lengthening” the baseline, the imaging FOV is effectively expanded while keeping the near-field dead zone within a reasonable tolerance range. The final long-cross frame was high-precision 3D printed (ProJet MJP 2500 Plus, 3D Systems, USA) using VisiJet M2R-WT with a layer thickness resolution of 32 µm. The overall dimensions of the CCUA-Patch are shown in Fig. 2(a). Optical photograph of the CCUA-Patch. (a) The dimensions of the CCUA-Patch excluding the hydrogel portion. (b) The CCUA-Patch can adhere to the human body via the hydrogel in the standing position. Building on this structure, self-adhesive hydrogel serves as the acoustic coupling medium to replace traditional medical ultrasound gels. The synthesis of the hydrogel followed a specific procedure. Initially, 2.0 g of AMPS powder was dissolved in 2.0 g of deionized water under magnetic stirring until the solution reached complete transparent. Subsequently, 20 mg of MBAA powder was added to the mixture and stirred until fully dissolved. To improve the flexibility of the final hydrogel, 2.0 g of glycerol was incorporated and mixed thoroughly. Following this, 12 μL of TEMED neat solution and 50 μL of 10% APS solution were added as the accelerator and initiator, respectively. After 30 s of rapid stirring, the precursor solution was immediately cast into a mold and sealed. The polymerization was carried out in an oven at 65 ℃ for 2 h. This process produced a self-adhesive hydrogel with a thickness of 2 mm. This 2 mm hydrogel layer is attached to the bottom of the patch. Its measured acoustic impedance was 1.93 ± 0.07 MRayl. Unlike traditional ultrasound probes that require handheld operation or mechanical fixtures, this patch is hands-free; the self-adhesive properties of the hydrogel ensure a seamless fit with the skin surface, significantly reducing clinical operational complexity and nursing workload. The self-adhesive properties of the hydrogel on the bottom of the CCUA-Patch even allow the arrays to remain securely attached to the lower abdomen while in an upright position, as shown in Fig. 2(b). Moreover, compared to traditional liquid coupling agents—which are prone to evaporation and drying, leading to acoustic impedance mismatch—solid hydrogels offer superior stability. This effectively eliminates the need for repeated gel application before each measurement, facilitating continuous, long-term monitoring. Beamforming and bladder image processing To achieve rapid acquisition of multi-dimensional bladder image information, this study designed a time-division scanning imaging sequence based on a three-linear-array configuration. The experimental system comprises three independently controlled 64-element linear arrays: two arrays (L1 and L2) are utilized to capture the sagittal features of the target area, while a third array (H1) is positioned horizontally to obtain morphological information in the transverse plane. During program execution, the system first activates the horizontal array H1 for a full-array scan, recording the raw data of the cross-section. Subsequently, upon completing data storage, the system seamlessly switches to the vertical arrays L1 and L2. Through sequential excitation, it acquires and saves the raw radiofrequency (RF) data of two spatially correlated longitudinal sections. This sequence ensures that anatomical information from different spatial dimensions is collected within a short time window, thereby reducing spatial registration errors caused by tissue motion. Within each subsequence, each array independently executes a linear beamforming algorithm, forming continuous imaging scanlines within a preset depth range through precise pulse-launch delay control. Following the acquisition of raw echo data from the three arrays, the signals undergo normalization and enhancement via a processing routine. The wavelength λ is precisely calculated based on the speed of sound and the center frequency of the transducer, serving as the baseline for constructing the axial and lateral coordinate matrices in physical space. I/Q data are processed using modulus operations for envelope detection, followed by logarithmic compression to adjust the signal’s dynamic range. To improve imaging quality and ensure clear bladder boundaries, the raw RF data is further processed. Furthermore, to overcome the limited FOV of a single linear array and obtain a complete image of a large-volume bladder, an image stitching and fusion method based on geometric features is employed to merge the two longitudinal sectional images captured in the same plane. The method first utilizes an edge detection algorithm to automatically extract the bladder contours from the two vertically acquired B-mode images. For the extracted discrete contour points, an Arc-length Parameterization interpolation algorithm is introduced for smoothing. Through precise resampling of non-closed contours, this effectively fills detection gaps and eliminates coordinate jitter. Subsequently, a global coordinate system is established based on the relative spatial positions of the three arrays on the substrate, mapping local images to a unified physical space via translation and rotation transformation matrices. During the image fusion stage, the algorithm uses the identified bladder edges as fiducial markers. By calculating the coherence of overlapping regions for weighted pixel superposition, a feature-weighted linear fusion is executed at the image boundaries. This achieves seamless stitching across arrays and completely reconstructs the morphological characteristics of the bladder in the sagittal and coronal planes. To quantitatively evaluate the gain in spatial coverage provided by the multi-array configuration and to define the effective imaging region for subsequent morphometric analysis, the FOV of the stitched images was further characterized geometrically. Each array was operated with a 90° sector FOV, corresponding to a steering range of −45° to + 45° relative to the array normal. Considering the longitudinal active-aperture length of 19.15 mm, the geometric imaging width of a single array at depth z is Wsinglez=19.15+2ztan45∘. The two longitudinal arrays are coplanar, with an approximately 28.6 mm edge-to-edge separation between their active apertures. After spatial registration and image stitching, the total outer-to-outer longitudinal coverage is therefore Wstitchedz=219.15+28.6+2ztan45∘=66.9+2z mm, corresponding to approximately 146.9, 226.9, 306.9, and 386.9 mm at depths of 40, 80, 120, and 160 mm, respectively. The near-field gap between the two sectors decreases with depth according to Gz=max28.6−2ztan45∘,0 and closes at a depth of approximately 14.3 mm, consistent with the approximately 15 mm blind-zone depth observed experimentally. These results demonstrate that, beyond the shallow near-field blind zone, the stitched images provide continuous and sufficiently wide spatial coverage for complete bladder visualization, thereby establishing a reliable geometric basis for subsequent dimensional measurements. Building upon this, the system performs automated measurement and labeling of bladder diameters in the three directions. The process begins with Contrast Limited Adaptive Histogram Equalization (CLAHE) and Gaussian filtering for pre-processing, combined with Otsu’s method to accurately extract the bladder region mask. After obtaining the closed boundary, Savitzky-Golay smoothing filters and Piecewise Cubic Hermite Interpolating Polynomials (PCHIP) are applied to fit the curves, eliminating the sawtooth effect caused by segmentation. The measurement algorithm performs equidistant sampling along the depth axis (Z-axis) and calculates the lateral width at each depth using polygon intersections, designating the maximum value as the transverse diameter. Similarly, by sampling contour intersections along the horizontal axis (X-axis), the maximum longitudinal span is locked. Finally, the system overlays the smoothed contours in solid yellow lines on the original images, while measuring positions and values for transverse and longitudinal diameters are labeled with dashed lines, achieving automated quantitative analysis and visual presentation of bladder geometric characteristics. CCUA-Patch performance characterization Following the fabrication of the three 1D linear arrays, several fundamental electrical and acoustic characterizations were performed. The electrical properties of each transducer element, such as resonant frequency and electrical impedance, were evaluated using an impedance analyzer (E4990A, Keysight, USA). To assess acoustic characteristics, pulse-echo experiments were conducted by immersing the transducer array in degassed water at room temperature. A steel block was positioned 20 mm in front of the array to serve as a stable target for ultrasonic reflection. A pulse/receiver (DPR 500, JSR Ultrasonics, NY, USA) was used to generate 150 V negative pulses with a pulse energy of 2.5 µJ. The resulting echo signals were filtered and amplified by the same device and subsequently recorded using a digital oscilloscope (DSOX3054G, Keysight, USA). Interelement crosstalk was experimentally characterized for each of the L1, L2, and H1 arrays after application of copper-foil shielding. Once the CCUA-Patch assembly was complete, the imaging performance of the integrated patch was characterized. A multi-purpose phantom (Sono403, Sun Nuclear, USA) was utilized to evaluate the imaging performance of the ultrasound arrays as shown in Fig. 3(a). Subsequently, an ultrasound tissue-mimicking phantom (KS215BP, Institute of Acoustics, Chinese Academy of Sciences, China) was employed for simulated bladder imaging, as illustrated in Fig. 3(b). Optical images of the CCUA-Patch on the top surface of (a) a multi-purpose phantom, and (b) a tissue-mimicking phantom. Human clinical study The prolate ellipsoid method is the standard clinical approach and the most classical algorithm used in commercial portable ultrasound devices. It involves acquiring two mutually perpendicular sections—the sagittal plane and the transverse plane—to measure the maximum diameters along three orthogonal axes: height (H), length (L), and width (W). The calculation formula is as follows [38,39]: A total of ten healthy participants (6 males, 4 females) participated in the imaging trials. Inclusion criteria included an age range of 23–26 years and a body mass index (BMI) between 18.67 and 23.74 kg/m2. All participants provided written informed consent before enrollment. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of ShanghaiTech University (approval no. Q2023–020). For quantitative analysis, bladder volume was calculated via the prolate ellipsoid algorithm, utilizing the maximum diameters measured across three orthogonal planes. Data acquisition was conducted for two physiological states: a full-bladder state characterized by a strong urge to urinate, and a non-full state approximately 10 min post-voiding. The imaging equipment included a Stork handheld ultrasound system and the CCUA-Patch connected to an ultrasound imaging platform (Vantage 256, Verasonics, USA). During the procedure, volunteers remained in a supine position with the lower abdomen exposed. Initially, a handheld convex array probe with acoustic coupling gel was used to determine the bladder’s status and location. By manipulating the probe orientation, the transverse and sagittal planes were localized. The maximum bladder dimensions—specifically the transverse diameter, anteroposterior diameter, longitudinal diameter—were manually measured and recorded for volume estimation. The Stork device was used as a clinical reference comparator rather than an absolute gold standard, because its result remains dependent on the operator-selected section and the same prolate ellipsoid approximation. After removal of the handheld-probe coupling gel, the skin-projected bladder location identified during the reference scan was used as the placement center. The CCUA-Patch was centered over this suprapubic location, with its longitudinal arm aligned parallel to the body's sagittal midline and its transverse arm aligned with the anatomical transverse plane. The self-adhesive hydrogel was brought into full contact with the skin, and images from H1, L1, and L2 were checked to confirm that the bladder was visible in the intended planes before quantitative acquisition. This procedure avoided probe rotation during subsequent measurements, although the initial localization step in the present protocol was operator assisted. Consequently, the system enabled the simultaneous acquisition of both transverse and sagittal bladder images without the need for manual probe repositioning. The specific placement and orientation of the device during the experimental trials are illustrated in Fig. 4. It should be noted that the CCUA-Patch consists of a reusable transducer assembly and a detachable hydrogel coupling layer. The hydrogel was independently attached to the outer surface of the ultrasound transducers and could be removed or replaced without altering the transducer assembly. It retained sufficient adhesiveness after repeated attachment and was therefore reusable for serial measurements in the same participant. Before the CCUA-Patch was transferred to another participant, the hydrogel layer was removed and discarded. All accessible external surfaces of the reusable transducer assembly and frame were then thoroughly wiped with 75% ethanol, and allowed to air-dry completely. A new hydrogel layer was attached before use on the next participant. Thus, the direct hydrogel–skin contact interface was not shared between participants, reducing the risk of cross-contamination. The CCUA-Patch on the human’s lower abdomen for bladder volume monitoring. Results Structural design of the CCUA-Patch Given that the anatomical length of a large-volume bladder in the longitudinal plane is significantly greater than that in the transverse plane, conventional single-probe scanning often fails to capture the complete bladder contour. To address this, the present study utilizes a cross-shaped combined array structure, characterized by equal-length vertical arms and shorter horizontal arms, with overall dimensions of 76.42 mm × 30.20 mm and a weight of 23.07 g. Under this configuration, for large-volume bladders, wide FOV imaging is achieved by stitching images from the two vertically aligned arrays, thereby obtaining a complete ultrasound image of the bladder. The corresponding imaging results from the three arrays are shown in Fig. 5(a), (b), and (c). Imaging results of the three ultrasound arrays on the CCUA-Patch over a distended and a small-volume bladder. (a), (b), and (c) show the imaging results of L1, L2, and H1, respectively, under the distended bladder condition. (d), (e), and (f) show the imaging results for the small-volume bladder. Notably, no bladder ultrasound image is visible in the field of view of (d) L1. Conversely, when bladder volume is low, the upper vertical array’s FOV may not encompass the bladder (as illustrated in Fig. 5(d)), while the lower array alone can capture the full longitudinal section (as shown in Fig. 5(e)). In this state, the operational portion of the array assembly effectively functions as a T-shape. CCUA-Patch performance evaluation Fig. 6 illustrates the experimentally measured electrical and acoustic characteristics of a representative 64-element transducer array. Specifically, Fig. 6(a) displays the resonant frequency, anti-resonant frequency, electrical impedance, and phase angle of the 64 elements, while Fig. 6(b) shows the pulse-echo response of a representative transducer element. For the 64 elements, center frequency was measured at 2.60 ± 0.02 MHz, with a −6 dB bandwidth of 86.67 ± 2.73%. The corresponding KLM simulation predicted a center frequency of 2.51 MHz and a −6 dB fractional bandwidth of 83.7%. The measured center frequency was therefore 0.09 MHz, or approximately 3.6%, higher than the simulated value, while the measured bandwidth was 2.97 %age points higher. These results indicate reasonable agreement between the KLM simulation and experimental characterization, with the remaining differences potentially arising from fabrication tolerances and deviations between the nominal model parameters and the properties of the fabricated transducer. The measured −20 dB pulse duration was 1.21 ± 0.17 µs, and the overall Vp-p sensitivity was 2.22 dB. Finite interelement crosstalk was detected in all three arrays. After application of copper-foil shielding, the L1 array exhibited a mean interelement crosstalk of −27.30 ± 1.89 dB. The measured results provide a quantitative assessment of the residual coupling between the array elements. Table 1 is included to report array-level performance. Performance of a representative 64-element transducer array. (a) Electrical impedance and phase angle spectrum of all elements. (b) Pulse-echo response and frequency spectrum of a representative element (element 002). (c) Center frequency, (d) overall Vp-p sensitivity, (e) −6 dB bandwidth, and (f) −20 dB pulse duration for all elements. Table 1Array-level electrical and acoustic consistency of L1, L2, and H1.ArrayCenter frequency (MHz)-6 dB bandwidthVp-p sensitivity (dB)Nearest-neighbor crosstalk (dB)L12.60 ± 0.0286.67 ± 2.73%2.22−27.30 ± 1.89L22.54 ± 0.0682.52 ± 1.53%2.54−28.82 ± 1.60H12.59 ± 0.1479.08 ± 3.32%3.60−29.14 ± 0.94 Array-level electrical and acoustic consistency of L1, L2, and H1. The encapsulated array was attached to a Sono403 Multi-Purpose Phantom for imaging, with the focal depth set at 60 mm. Notably, no ultrasonic coupling agent or manual compression was applied during the procedure. Fig. 7(a) illustrates the ultrasound Point-Spread Functions (PSFs) obtained from wire targets at various depths. The corresponding lateral and axial resolutions are plotted in Fig. 7(b); it is observed that the lateral resolution first decreases and then increases, with the minimum value occurring at the 60 mm focal point, while the axial resolution remains essentially constant. Imaging performance of the CCUA-Patch on the phantom. (a) Image of the L1 array on the top surface of the phantom with self-adhesive hydrogel. (b) Imaging resolution of a representative array at different depths. (c) Reconstructed image from L1 and L2 arrays and (d) image from H1 array on the tissue-mimicking phantom with hydrogel. (e) and (f) The automatic annotation results of the bladder contour and dimensions. Subsequently, imaging was performed on an ultrasound tissue-mimicking phantom (KS215BP, Institute of Acoustics, Chinese Academy of Sciences, China), again without the use of couplant or manual pressure. Following feature-based image stitching, the cross-sectional images are shown in Fig. 7(c) and (d), while Fig. 7(e) and (f) present the results of edge extraction and the three-diameter calculation. The measured diameters were 54.57 mm, 81.18 mm, and 54.76 mm, respectively, resulting in a calculated bladder volume of 126.14 mL. Compared to the phantom’s calibrated ovoid target volume of 137 mL, the absolute error was 10.86 mL, with a relative error of approximately 7.93%. For comparison, the three-diameter measurements obtained using a commercial handheld convex probe were 54.41 mm, 54.59 mm, and 81.09 mm, respectively, corresponding to a volume of 125.25 mL. The volume discrepancy between the CCUA-Patch and the commercial probe was only 0.89 mL. In vivo bladder volume monitoring To evaluate the reliability of the CCUA-Patch in clinical applications, a volunteer study was conducted focusing on bladder volume monitoring under different filling states. Ten healthy volunteers (6 males and 4 females) aged 23–26, with a BMI ranging from 18.67 to 23.74 kg/m2, were recruited. The experiment was divided into two groups: Case I targeted the distended bladder, involving 3 males and 2 females who provided bladder images while experiencing a strong urge to urinate; Case II targeted the small bladder, involving the remaining 3 males and 2 females, with bladder images captured approximately 10 min post-urination when no urge was felt even under light pressure. After image processing and bladder contour identification, the spatial measurement accuracy of the system was validated by comparing the length, height, width, and estimated volume results of the CCUA-Patch against those obtained from a commercial Stork handheld ultrasound probe. Fig. 8 and Fig. 9 display the bladder images in these two states. For the distended bladders in Case I, two longitudinal images were stitched and annotated with bladder diameters, and the bladder volumes were calculated using the formula (1). Results showed the male bladder volumes of 410.20 mL, 559.34 mL, and 327.04 mL, respectively, and the female volumes of 590.60 mL and 330.20 mL, respectively. For the small bladders in Case II, the complete contour could be captured using only the central and lower transducers without image stitching. The calculated male bladder volumes were 88.05 mL, 7.67 mL, and 19.89 mL, respectively, while the female volumes were 22.13 mL and 126.16 mL, respectively. Imaging results on the volunteers in Case I. (a) and (b) The ultrasound images of the distended bladders for Case I. (c) Comparison of the three-axis measurement results between the CCUA-Patch and a commercial probe. Imaging results on the volunteers in Case II. (a) and (b) The ultrasound images of the small bladders for Case II. (c) Comparison of the three-axis measurement results between the CCUA-Patch and a commercial probe. Fig. 8c and Fig. 9(c) compare the three-axis measurements between the CCUA-Patch and the Stork probe across genders and filling states. In Case I, two devices showed good consistency, though slight deviations existed between genders and individuals. Specifically, the male volunteers exhibited slightly higher measurement deviations than the females. The maximum deviation occurred in the width measurement of the male volunteer #2, where the CCUA-Patch measured 119.11 mm and the commercial probe measured 109.91 mm. However, the absolute error of 0.92 cm remained within the 1 cm threshold. In contrast, the female volunteers demonstrated higher consistency, with all measurement errors maintained below 0.5 cm. In Case II, despite the significantly smaller target size, the correlation between the two devices improved, with absolute errors for all of the five volunteers dropping below 0.2 cm across all dimensions. Fig. 10 provides a quantitative comparison of bladder volume measurement results for all of the 10 volunteers. The data indicate high consistency in volume quantification, especially in small-volume scenarios. In Case II (male volunteers #4–6 and female volunteers #3–4), the volume difference was controlled within 8 mL. Notably, the absolute deviation for the female volunteer #3 was only 0.63 mL (22.13 mL vs. 22.76 mL). In Case I, while absolute errors increased with target size, relative errors remained low. For instance, the female volunteer #2 showed a difference of only 3.67 mL (1.1% relative error), and the male volunteer #2 showed 13.46 mL (2.4% relative error). The maximum deviation occurred in the male volunteer #1 (approximately 21.8%), though his three-axis dimension differences remained within 9 mm. Table 2 reports each volunteer's characteristics, CCUA-Patch and Stork measurements, volumes, and error fields. Comparison of the bladder volume measurement results between the CCUA-Patch and Stork handheld probe under Case I and Case II. Table 2Individual volunteer characteristics and measurement results.VolunteerStateBMICCUA L/H/W (mm)Stork L/H/W (mm)CCUA V* (mL)Stork V* (mL)Error** (mL)Male #1Distended22.0283.99 / 98.41 / 95.4480.10 / 89.89 / 89.92410.20336.67+73.53Male #2Distended23.7491.00 / 99.24 / 119.1197.89 / 97.57 / 109.91559.34545.88+13.46Male #3Distended23.0493.63 / 85.72 / 78.3677.77 / 85.45 / 79.43327.04274.48+52.56Female #1Distended19.60113.36 / 85.70 / 116.91113.75 / 83.69 / 112.22590.60555.52+35.08Female #2Distended18.6779.62 / 91.43 / 87.2382.06 / 88.35 / 88.56330.20333.87−3.67Male #4Small22.2373.15 / 47.56 / 48.6773.12 / 48.20 / 48.9388.0589.67−1.62Male #5Small23.5339.59 / 20.48 / 18.1838.30 / 21.34 / 15.697.676.67+1.00Male #6Small24.1040.26 / 36.71 / 25.8844.50 / 36.15 / 26.9519.8922.54−2.65Female #3Small20.7661.23 / 32.55 / 21.3562.04 / 31.07 / 22.7122.1322.76−0.63Female #4Small19.4986.82 / 68.46 / 40.8286.57 / 67.37 / 39.03126.16118.37+7.79*CCUA V and Stork V denote the bladder volumes.** Error was calculated as CCUA V − Stork V. Individual volunteer characteristics and measurement results. *CCUA V and Stork V denote the bladder volumes. ** Error was calculated as CCUA V − Stork V. To further assess the agreement between the CCUA-Patch and the Stork probe, a Bland-Altman analysis was performed, as shown in Fig. 11. The horizontal axis represents the mean volume of the two devices, while the vertical axis represents the difference calculated as minus. Statistical results show a mean bias of 17.48 mL with a standard deviation of 27.1 mL, indicating that CCUA-Patch measurements are slightly higher than the commercial probe, though the overall systemic error is minimal. Most data points are distributed within the 95% Limits of Agreement (LoA), ranging from −35.63 mL to 70.6 mL. The points in the small-volume range are highly concentrated near the zero-bias line, demonstrating exceptional precision. In the large-volume range, although absolute deviations increased for certain samples due to anatomical complexity and manual operation variance, the overall trend remained stable. The scatter plot shows no significant systematic bias or proportional error resulting from increasing measurement values. Bland-Altman analysis of the bladder volume measurement results for the 10 volunteers. Discussion The development and validation of the CCUA-Patch represent an advancement in non-invasive continuous bladder volume monitoring by addressing the inherent limitations of traditional single-probe ultrasound systems such as restricted fields of view and the requirement for manual rotation. Utilizing an innovative cross-shaped tiling layout that leverages multi-element synthetic fields of view and self-adhesive hydrogel coupling, this system effectively eliminates the truncation effects commonly encountered during the imaging of large bladder volumes. This approach ensures the complete capture of bladder morphology without the need for professional manipulation and enables the continuous measurement of bladder volume across various filling states. Table 3 compares the proposed method with those reported in the literature and indicates that the main contribution of this work lies in combining fixed dual-plane acquisition with longitudinal image stitching in a three-array patch. Table 3Comparison of representative ultrasound approaches for bladder-volume monitoring.StudyModeArray architectureFOV / plane-acquisition strategyPrincipal limitationTanaka et al., 2010 [30]A-mode12 fixed-array probesGeometric segmentation and summed bladder slicesBulky device; poor skin conformabilityZhang et al., 2023 [31]B-modeFive 64-element phased arraysMultiple probes for localization; multi-image fusion remained at phantom stageHuman images remained within a single-probe FOV; higher patch complexityPark et al., 2025 [32]3D-modeT-shaped cross-array probeManual shifting with motion-based 3D reconstructionManual motion; higher time and system costThis workB-modeThree 64-element linear arraysTwo fixed longitudinal arrays stitched for sagittal coverage plus one transverse arrayBenchtop Vantage 256 Comparison of representative ultrasound approaches for bladder-volume monitoring. In phantom imaging validation, the CCUA-Patch achieved an imaging depth exceeding 14 cm, meeting the requirements for most bladder imaging scenarios. Furthermore, sagittal images synthesized through feature-matching showed a blind zone depth between the two arrays of approximately 1.5 cm, which is significantly shallower than typical bladder depths, thereby ensuring complete image acquisition. For tissue-mimicking phantoms, although the volume error between the CCUA-Patch and the calibrated volume reached 10.86 mL, the discrepancy compared to commercial ultrasound results was only 0.89 mL, with three-axis dimensions matching almost perfectly. This suggests that the deviation from the calibrated volume likely stems from the geometric simplification of the prolate ellipsoid formula relative to the ovoid phantom shape, rather than measurement error from the CCUA-Patch itself. Clinical validation with 10 volunteers confirmed that the CCUA-Patch maintains high measurement consistency with commercial handheld probes across different filling states. In Case I (distended bladders), the system exhibited a significant FOV advantage. Traditional single-probe linear arrays are often limited by physical aperture, frequently resulting in "truncation effects" when imaging bladders with long longitudinal diameters, which prevents capturing the complete apex or base contours. The cross layout utilizes two longitudinally arranged 64-element arrays (L1 and L2) for time-division scanning. Combined with feature-based image registration and stitching, it successfully achieves wide-field sagittal imaging. For instance, in male volunteer #2, whose longitudinal bladder span approached 10 cm, the CCUA-Patch measurements for length (91.00 mm) and height (99.24 mm) were highly consistent with the commercial convex probe (97.89 mm and 97.57 mm, respectively). This design ensures full coverage of the bladder contour, eliminating operational errors caused by clinicians frequently rotating or moving a probe to locate the maximum cross-section. For Case II (small bladders), the CCUA-Patch demonstrated exceptional precision and sensitivity. Data showed that absolute errors for the three-axis measurements in all five volunteers dropped below 0.2 cm, with the volume error for female volunteer #3 being only 0.63 mL. In contrast, handheld convex probes often struggle to maintain contact in small-volume scenarios due to their bulk and interference with the iliac bone or pelvic structures. While some distended bladder samples showed larger volume deviations (approximately 21.8%), the absolute three-axis deviations remained within 9 mm. This amplification of error is primarily due to the multiplicative effect of the prolate ellipsoid formula on individual dimensional errors, while also reflecting the limitations of manual operation. In comparative experiments, handheld probes easily shift when switching between transverse and longitudinal planes; in high-volume states, the visual differences between these planes are subtle, making it difficult to guarantee a strictly orthogonal maximum cross-section. The fixed structure of the CCUA-Patch eliminates this human uncertainty, providing a more reproducible measurement baseline. Bland-Altman analysis quantified agreement between the CCUA-Patch and the Stork clinical reference device, with a mean bias of 17.48 mL and 95% limits of agreement from −35.63–70.60 mL. Because both devices used operator-selected images and the prolate ellipsoid approximation, and neither was compared with catheterization or 3D planimetry, the observed differences cannot be attributed exclusively to either device. During large-volume measurements, manual rotation and angle adjustment can cause the operator to inadvertently select a "maximum cross-section" that occupies a different spatial position than the original plane. The CCUA-Patch, with its fixed spatial coordinate system, guarantees the orthogonality and positional stability of the transverse and longitudinal sections. Additionally, the 2 mm self-adhesive hydrogel solves the problem of traditional coupling agents drying out and requiring reapplication. Its self-adhesive nature enables hands-free monitoring, and its excellent acoustic impedance matching ensures a high signal-to-noise ratio (SNR) over hours of monitoring, significantly reducing the clinical nursing burden. This is of great value for automated bladder volume monitoring and early warning. In addition to supporting stable, hands-free coupling during the reported imaging sessions, during a preliminary device-development observation, the hydrogel remained adhered to human skin for 6 h, and no visible skin redness was observed after the attachment period. Nevertheless, this qualitative observation was not a formal dermatological safety assessment, and the maximum wear duration, delayed skin reactions, and effects of repeated exposure were not evaluated. Therefore, the current evidence supports short-term attachment under the tested conditions but does not establish long-term skin compatibility. Synthesizing the experimental results, it is evident that although the CCUA-Patch is currently constrained by the inherent errors of geometric simplification models such as the prolate ellipsoid method and exhibits individual amplified errors in large volume measurements due to anatomical variations or differences in maximum plane selection, the findings demonstrate its reliability as a substitute for traditional handheld convex array probes in long-term and automated bladder volume monitoring. Its imaging precision is sufficient to support real-time clinical assessment of bladder filling status. Overall, the CCUA-Patch exhibits high measurement accuracy in both large and small volume states, allowing the system to stably obtain the three-diameter parameters of anatomical structures and achieve automated urine volume assessment. The core advantages of the system lie in its spatial structural stability and monitoring continuity. Compared to handheld probes and certain flexible wearable patches, the structural stability of the cruciform array configuration offers distinct benefits. While traditional handheld scanning relies on the ability of the operator to locate and maintain orthogonal planes, the CCUA-Patch ensures the positional stability and orthogonality of transverse and longitudinal sections. This fixed orthogonal scanning mechanism reduces measurement random errors introduced by human operation and enhances the robustness of volume estimation. Furthermore, the ability to obtain full field-of-view and dual-section images without manual adjustment provides the fundamental basis for long-term continuous monitoring. Coupled with a self-adhesive solid hydrogel coupling layer that enables hands-free operation over extended periods, the device significantly reduces the clinical nursing burden and eliminates the need for reusable auxiliary agents, granting the system the capability for long-term monitoring without human intervention. The clinical significance of the CCUA-Patch is extensive, offering substantial practical value for the early warning of urinary retention, the evaluation of postoperative voiding function recovery, and the care management of patients with incontinence. Integrating the device host computer into mobile terminals can provide timely alerts for patients with neurogenic bladder dysfunction or impaired sensory function, thereby preventing serious complications such as vesicoureteral reflux and chronic renal failure. Additionally, the system can scientifically guide the timing of catheter removal during postoperative recovery and assist in bladder training programs to help patients regain sensory awareness, showing broad application prospects in home care and telemedicine. Despite these advantages, the current study has certain limitations that necessitate further exploration. The current in vivo experimental validation involved ten healthy volunteers with a narrow range of age and body mass index. Future research should encompass larger and more diverse populations to evaluate the performance of the system under complex anatomical and pathological conditions. Moreover, while the prolate ellipsoid formula remains the clinical standard, its geometric simplification leads to inherent errors in volume conversion, particularly as target dimensions increase. Subsequent research will focus on introducing advanced deep learning algorithms to improve the automation and precision of bladder wall segmentation. Furthermore, the patch itself is lightweight (23.07 g), compact, and hands-free after attachment, but the complete experimental system is not yet portable or wearable because image acquisition used the Vantage 256 platform. Translation toward ambulatory or home use will require a miniaturized, low-power multichannel front end, integrated beamforming or multiplexing, wireless data transfer, and a battery-powered architecture. In conclusion, the CCUA-Patch provides a robust and accurate solution for automated and non-invasive bladder volume monitoring. By balancing wearable comfort with imaging precision, it overcomes the conflict between field-of-view expansion, wearing comfort, and operational complexity inherent in portable bladder monitoring devices. This work establishes a feasible clinical alternative to traditional handheld probes and provides a new technical pathway for the automated management of lower urinary tract functions in both clinical and domestic environments. At the same time, the current prototype therefore supports feasibility only. Larger and more diverse cohorts, including target patient populations, together with an independent volume reference, and a miniaturized imaging system architecture are required before clinical or home deployment. The structural stability and unique cruciform geometry of the CCUA-Patch provide an exceptional platform for extending the system's capabilities into Photoacoustic Imaging (PAI) through a decoupled excitation-detection scheme. In this potential configuration, the CCUA-Patch serves as a dedicated, high-performance wearable receiver array, while optical excitation is provided by an external laser source. This decoupled approach offers several distinct advantages: it mitigates the thermal management challenges associated with on-patch light sources and allows for the delivery of higher-energy pulses, which are essential for achieving the penetration depths required for pelvic organ monitoring. The orthogonal layout of the CCUA-Patch is particularly well-suited for capturing photoacoustic signals generated by diffuse external illumination. By providing dual-plane acoustic reception, the system can effectively address the "limited-view" problem, enabling more accurate 3D localization and reconstruction of deep-tissue optical absorbers compared to traditional single-element or linear-array wearable sensors. Clinically, this synergy would allow the CCUA-Patch to perform continuous, non-invasive monitoring of hemodynamic parameters, such as oxygen saturation and total hemoglobin concentration within the bladder wall and surrounding pelvic tissues, during external laser therapy or diagnostic scanning. This functional insight is of paramount importance for the early detection of bladder wall ischemia, chronic inflammatory responses, and the assessment of tumor angiogenesis in bladder malignancies. Furthermore, the self-adhesive solid hydrogel layer developed in this study provides excellent optical transparency and acoustic coupling, ensuring that externally delivered light can penetrate to target depths while photoacoustic waves are efficiently transmitted back to the cruciform array. This roadmap positions the CCUA-Patch not just as a bladder volume monitor, but as a core component of a future multi-modal pelvic health surveillance system capable of identifying physiological anomalies like tissue hypoxia or abnormal angiogenesis. CRediT authorship contribution statement Ziqi Li: Writing – original draft, Visualization, Validation, Investigation, Formal analysis, Data curation. Cong Pu: Visualization, Validation, Methodology, Investigation. Ting Zhang: Writing – review & editing, Validation, Methodology. Chang Peng: Writing – review & editing, Supervision, Software, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. Ethics declaration Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed. This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the ShanghaiTech University (Approval No. Q2023-020). Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This work was supported in part by the National Natural Science Foundation of China (Grant No. 12204307). Biographies Ziqi Li received the B.S. degree in biomedical engineering from Capital Medical University, Beijing, China, in 2024. She is currently pursuing a master’s degree with the Department of Biomedical Engineering, ShanghaiTech University, Shanghai, China. Her research interests mainly focus on the ultrasound imaging and flexible and wearable ultrasound devices. Cong Pu received the B.E. degree in biomedical engineering from the University of Shanghai for Science and Technology, Shanghai, China, in 2022. He is currently pursuing the Ph.D. degree in electronic science and technology from the School of Biomedical Engineering, ShanghaiTech University, Shanghai. His current research interests include wearable ultrasound devices and ultrasound neuromodulation. Ting Zhang is a doctoral candidate in the Department of Integrative Medicine, Huashan Hospital, Fudan University, under the supervision of Professor Yanqing Wang. Her research focuses on acupuncture for refractory chronic diseases and integrative Chinese–Western medicine in the diagnosis, treatment, and mechanistic study of malignant tumors. She has led one National Natural Science Foundation of China Youth Program project and one Huashan Hospital Basic Research Youth Cultivation Program, and has participated in several national and provincial research projects. She has published seven papers in recent years and contributed to textbooks and guidelines including Schools of Chinese Acupuncture and Moxibustion. She currently serves as a member of several academic committees, including relevant CACA professional committees and the Shanghai TCM Oncology Branch. Chang Peng, Ph.D., is an Assistant Professor and Ph.D. advisor at the School of Biomedical Engineering, ShanghaiTech University. He received his Ph.D. in Mechanical Engineering from the University of Florida in 2018 and conducted postdoctoral research at North Carolina State University from 2018 to 2021. His research mainly focuses on miniaturized medical ultrasound transducers and wearable ultrasound devices. He has published over 40 journal papers, filed 17 U.S. and Chinese invention patents, and led or participated in seven national, provincial, and industry-funded projects. His work has been featured by IEEE Spectrum, Forbes, and other media. He also serves on the IEEE IUS Technical Program Committee and other professional committees. Data availability Data will be made available on request. References Associated Data Data Availability Statement Data will be made available on request.