We developed a new ultrasound-based technique to noninvasively assess urethral tissue microstructure using a multiparametric approach. By combining angle-dependent backscatter analysis with radiomic texture features, we created a novel "urethra score" that tracks age-related tissue atrophy with high sensitivity. This method provides a promising, biopsy-free tool to quantify urethral tissue quality and could inform future diagnostic and therapeutic strategies for urinary incontinence.

This figure illustrates the step-by-step workflow for analyzing transperineal ultrasound (US) data to assess urethral microstructure.

A) The region of interest (ROI) of the urethra (red dotted line) is manually delineated in each B-mode frame, and representative anatomical landmarks like the pubic bone (black arrow) and bladder neck (white arrow) are annotated for orientation.

B) The central axis of the urethra is automatically extracted using morphological skeletonization and B-spline fitting (red curve), with perpendicular lines (blue) calculated at each point along the axis to define local urethral angles.

C) The US beam direction is superimposed on the image (blue lines), enabling calculation of the angle between the beam and each point on the urethra, shown using a heatmap (red hues indicate larger angles).

D) Representative frames at two angles (30° in red and 60° in green) are used to illustrate how angle-dependent backscatter behavior is quantified. Image intensity values from urethral regions at matching angles are averaged across frames to generate a function that reflects the angular sensitivity of backscattered US signals, which serves as a proxy for tissue microstructural organization.

This figure shows how multiparametric texture features are derived from B-mode ultrasound images to evaluate urethral microstructure. The top row presents maps generated from first-order statistical features (mean, variance, skewness, and kurtosis), which describe pixel intensity distributions within the urethral region. The bottom row includes higher-order features extracted from advanced radiomic matrices—namely the Gray Level Co-occurrence Matrix (GLCM, orange boxes), Gray Level Dependence Matrix (GLDM, purple), Gray Level Run Length Matrix (GLRLM, red), and Gray Level Size Zone Matrix (GLSZM, green). These features quantify spatial relationships between pixel intensities, providing insight into structural organization and heterogeneity.  Notably, the variance and GLCM contrast maps exhibit similar spatial patterns, highlighting redundancy between certain features. Each texture map reveals distinct spatial variations within the urethra, which are sensitive to changes in angle and age-related microstructural alterations. Together, these features form the basis for data-driven biomarker development in the proposed urethra score.

[1] H. Tai, K. Kalayeh, J. A. Ashton-Miller, J. O. L. DeLancey and J. B. Fowlkes, “Urethral tissue characterization using multiparametric ultrasound imaging,” Ultrasonics, vol. 145, pp. 107481, 2025. https://www.ncbi.nlm.nih.gov/pubmed/39348748