Integrated Ultrasound Device for Precision Bladder Volume Monitoring via Acoustic Focusing and Machine Learning

Bladder volume monitoring is critical for managing lower urinary tract dysfunctions, yet existing methods remain invasive or operator-dependent and are unsuitable for continuous use. Here, we present a conformable wearable ultrasound system that combines lens-assisted acoustic focusing with machine-learning regression to enable non-invasive bladder volume estimation, while providing a clear path toward future real-time implementation. A flexible PZT array integrated with a concave acoustic lens enhances lateral energy concentration and depth selectivity, while a Random Forest model was used to map echo-derived features to bladder volume estimates. In a pilot study, bladder-volume estimates generated offline after data collection showed good agreement with a benchtop electrical impedance-based measurement system, supporting the feasibility of non-invasive bladder volume estimation. The device was operated using conservative low-voltage, low-duty-cycle excitation settings designed to minimize acoustic exposure and be consistent with diagnostic-ultrasound safety guidance, and biocompatible, flexible encapsulation is designed to support extended wear. Together with compact packaging and low-power wireless transmission, these attributes support ambulatory, longitudinal bladder monitoring and offer design insights for future wearable ultrasound systems targeting precise and ultimately continuous physiological monitoring.