Clinical Report: Automated Analysis of Transperineal Ultrasound for PFD

Overview

This study introduces an automated deep learning approach for analyzing transperineal ultrasound (TPUS) images to assess pelvic floor dysfunction (PFD) post-total hysterectomy. The method demonstrates high accuracy in segmentation and key point localization, providing a reliable tool for postoperative monitoring and evaluation.

Background

Pelvic floor dysfunction is a common issue following total hysterectomy, significantly affecting patients' quality of life. Traditional assessment methods are often inefficient and subjective, highlighting the need for a more objective and quantitative approach. This study addresses these limitations by utilizing deep learning techniques for automated TPUS analysis.

Data Highlights

Key Findings

• The automated model achieved an average Dice coefficient of 88.67% for segmentation accuracy.
• Key point localization errors were maintained within 2 mm.
• There was a strong correlation (Pearson coefficient up to 0.92) between automated measurements and manual annotations.
• The method effectively distinguished functional differences among patients with varying surgical approaches.
• This approach is applicable to both benign and gynecologic oncology patients, enhancing postoperative functional monitoring.

Clinical Implications

The automated TPUS analysis offers a structured and objective method for evaluating pelvic floor function post-hysterectomy, which can improve patient management and rehabilitation strategies. This tool may facilitate timely interventions and better long-term outcomes for patients at risk of PFD.

Conclusion

The proposed deep learning method for TPUS analysis represents a significant advancement in the assessment of pelvic floor dysfunction, with potential applications in both benign and malignant gynecological conditions. Its implementation could enhance clinical practice by providing reliable and efficient evaluations.

Related Resources & Content

1. Frontiers in Oncology, 2026 -- Automated Transperineal Ultrasound Analysis Using Deep Learning for Pelvic Floor Dysfunction Assessment after Total Hysterectomy
2. Frontiers in Oncology — Ultrasound-derived pelvic floor parameters and their association with functional impairment in gynecologic cancer survivors: a retrospective cohort study
3. Techniques in Coloproctology — Automated Identification of Male Pelvic Floor Soft Tissue Anatomy for Simulation and Morphological Evaluation in Lower Rectal Cancer Procedures
4. Surgical Endoscopy — Utilizing Deep Learning for Pelvic MRI Analysis to Evaluate Surgical Complexity in Total Mesorectal Excision
5. Frontiers in Surgery — Robot-assisted anterior abdomino-vaginal mesh suspension for stress urinary incontinence associated with anterior compartment pelvic organ prolapse: technique, imaging workflow, and 12-month pilot outcomes
6. U-POP and LUTD care refined in international consensus report | AUA
7. The Uncorrected Defect: A Risk Stratification Model for Persistent Levator Ballooning Following Pelvic Organ Prolapse Repair | International Urogynecology Journal | Springer Nature Link
8. Urogenital Hiatus Closure System: A Framework for Understanding how Muscle, Motor Control, and Fascial Connections Interact in Normal and Failed Closure | International Urogynecology Journal | Springer Nature Link
9. Rehabilitation of pelvic floor muscle for women with urinary incontinence post hysterectomy: A randomized controlled trial - ScienceDirect
10. Telerehabilitation reduced urinary incontinence, pelvic pain and dyspareunia in women after treatment for gynaecological cancer: a randomised trial - ScienceDirect
11. Frontiers | Automated Transperineal Ultrasound Analysis Using Deep Learning for Pelvic Floor Dysfunction Assessment after Total Hysterectomy
12. Deep Learning for Medical Ultrasound Image Segmentation: A Systematic Review of the Current Research | Journal of Imaging Informatics in Medicine | Springer Nature Link
13. PFUS1: Premier pelvic floor ultrasound segmentation dataset. A resource for advancing research - ScienceDirect
14. The copyright holder for this preprint
15. Ultrasound assessment of the perineal body: A scoping review - ScienceDirect