How deep learning transforms bladder neck motion analysis in SUI

News
Article

A new study demonstrates the efficacy of a deep learning-based transperineal ultrasound system for evaluating bladder neck motion in women with stress urinary incontinence.

How deep learning transforms bladder neck motion analysis in SUI | Image Credit: © natali_mis - © natali_mis - stock.adobe.com.

How deep learning transforms bladder neck motion analysis in SUI | Image Credit: © natali_mis - © natali_mis - stock.adobe.com.

Bladder neck (BN) motion in women with stress urinary incontinence (SUI) may be quantified using a deep learning (DL) system, according to a recent study published in the American Journal of Obstetrics & Gynecology.1

BN mobility has been determined as vital for SUI, which comprises approximately 50% of urinary incontinence cases in women.2 According to investigators, the complex interplay of nerves, hormones, pelvic floor muscles, and other factors behind SUI makes it difficult to observe the process of BN motion.1

“Transperineal ultrasound (TPUS) is highly recommended for evaluating SUI owing to its advantages of visualizing pelvic morphology, ease of access, noninvasiveness, and cost-effectiveness,” wrote investigators, who conducted the study to evaluate the efficacy of a DL-based TPUS toward evaluating BN motion.

Participants included women referred to Zhejiang Provincial People's Hospital between December 2022 and September 2023 for postpartum care or pelvic floor dysfunction. The electronic medical system was assessed for data about these patients’ height, weight, age, menopausal status, parity, and gynecologic or pelvic surgical procedure history.

The SUI group consisted of patients with an SUI diagnosis, while the control group consisted of healthy women. Exclusion criteria included history of pelvic surgery or SUI, being unable to perform Valsalva, and having pelvic organ prolapse beyond the hymen.

During routine clinical practice, clinicians often recorded two-dimensional (2D) TPUS videos for the DL-based system. A Voluson E8 device (GE Healthcare, Chicago, IL) with a 4- to 8-MHz 4-dimensional volume transducer was used by 2 radiologists with over 3 years of TPUS experience to conduct TPUS.

Bladder neck descent (BND), β angle, and urethral rotation angle (URA) were available for each frame using the DL-based AutoPelvic system (RayShape Medical Technology, Shenzhen, China). The maximum and average speed of these characteristics were determined based on the data provided by the AutoPelvic system.

There were 173 women included in the final analysis, 47.4% of whom were in the SUI group and 52.6% were in the control group. Older age, increased body mass index, and lower parity were observed in the SUI group vs controls.

No significant differences between groups were reported for maximum and average speed of BND (BNDm, BNDa) nor the speed variance of BND. However, the maximum and average speed of β angle (β anglem, β anglea), and URA (URAm, URAa) did significantly differ between groups.

In SUI patients β anglem, β anglea, URAm, and URAa values were 151.2, 6, 105.5, and 101, respectively. In controls, these values were 109, 3.1, 69.6, and 7.9, respectively. Differences in speed variation for β angle and URA between SUI patients vs controls were also significant, at 844.8 vs 336.4 and 347.6 vs 131.1, respectively.

The β angle, URA, and BND during the Valsalva maneuver were visualized using fitted curves. In SUI patients, the BND, β angle, and URA were usually larger vs control patients, with the curves revealing increased speeds in this population.

Significant diagnostic parameters after adjustment included β anglem, β anglea, URAm, and URAa. These parameters had odds ratios for SUI of 1.01, 1.40, 1.02, and 1.08, respectively. Area under the receiver operating characteristic curves were 0.67, 0.74, 0.72, and 0.60, respectively.

According to investigators, this application of DL in TPUS highlighted promising SUI diagnostic parameters. “This approach provides more valuable information, helps simplify and improve clinical work, and enhances efficiency,” wrote investigators.

References

  1. Wang J, Yang X, Wu Y, et al. Deep learning–assisted two-dimensional transperineal ultrasound for analyzing bladder neck motion in women with stress urinary incontinence. Am J Obstet Gynecol. 2025;232:112.e1-12. doi:10.1016/j.ajog.2024.07.021
  2. Hunskaar S, Lose G, Sykes D, Voss S. The prevalence of urinary incontinence in women in four European countries. BJU Int. 2004;93(3):324-30. doi:10.1111/j.1464-410x.2003.04609.x
Recent Videos
Expert on OAB
Expert on OAB
Experts on OAB
Expert on OAB
Expert on OAB
Expert on OAB
Experts on OAB
Expert on OAB
Expert on OAB
Expert on OAB
Related Content
© 2025 MJH Life Sciences

All rights reserved.