Artificial Intelligence for the Automated Measurement of Hiatal Dimensions.

INTRODUCTION AND HYPOTHESIS: The levator hiatus is an important determinant of pelvic organ prolapse and prolapse recurrence after surgery. Manufacturers are now offering automatic determination of hiatal dimensions.

METHODS

This was an observational study using data sets of 437 women attending a urogynaecological unit in 2024 and 2025. All had 4D pelvic floor ultrasound using a Voluson E 22 system with SonoPelvic™ software. Data sets obtained on Valsalva were used for automated measurements. Manual measurements were performed later on stored data sets blinded against all other data. For statistical analysis we employed ICC and Bland Altman plots. Patients in whom measurements differed by ≥20% were reviewed.

RESULTS

The 437 women seen during the study period presented with stress urinary incontinence (n = 304), urgency urinary incontinence (n = 334) and symptoms of pelvic organ prolapse (n = 229). Automated measurements yielded a mean of 25.45 (4.08-64.76) cm2, whereas manual hiatal area measurements provided a mean of 26.8 (6.97-65.7) cm2 (Pearson's correlation 0.59, p < 0.0001). The difference between individual measurements was statistically nonsignificant on t test. A direct comparison was possible in 425, with 80/425 (19%) showing a discrepancy of ≥20%. These were reviewed separately by the authors and found to be due to loss of symphysis pubis (n = 20) or anorectal angle (n = 35) or both (n = 10) in higher stages of prolapse.

CONCLUSION

Automated analysis by the current version of SonoPelvic™ shows a strong correlation with manual measurements, except in women with highly abnormal pelvic organ support. The software is under continuous development for which several recommendations are made.