2,281 posters, 63 videos, 56 audios, 43 topics, 18 sessions, 6,174 authors, 2,323 institutions
ePostersLive by SciGen Technologies S.A. All rights reserved.
RCOG World Congress 2025
June 23 - 25, 2025 | ExCeL London, UK
PP.0002
AI-Assistance Improves External Anal Sphincter Defect Detection and Workflow Efficiency
Poster Presenter
Authors
Affiliations
Part of Topic
AI & digital health
Objective
Can an in-house artificial intelligence (AI) toolhelp novice clinicians identify significant external anal sphincter (EAS) defects on transperineal ultrasound (TPUS)?
We compared AI-assisted novice assessments to bothnovice and expert manual analyses in terms of: accuracy, perceived workload and interpretation time.
Methods:
102 TPUS volumes were manually analysed with tomographic ultrasound imaging (TUI) sequences (Fig. 1) [1] by:
- A novice (1 year of TPUS experience)
- An expert (> 5 years).
Metrics: novice disagreement rate (% of slices where a novice marked a defect/intact while the expert marked intact/defect). Time and NASA Task Load Index (NASA-TLX).
Results:
Across the entire dataset, novice manual analysis disagreement rate was 12.9%. In the subset (n=17) (Table 1), the novice’s manual disagreement rate doubled to 23.5%, and they identified 3 of 6 severe EAS defect cases. With AI-assistance: novice disagreement rate reduced to 15.7%, and 5 out of 6 patients were identified with significant EAS defects.
Conclusion: AI-assistance improved novice detection of EAS defects, enabling identification of two additional patients with significant EAS defects. AI-assistance also halved workload and significantly reduced analysis time.