OC18.03: SonoPelvic™: artificial intelligence for the measurement of hiatal dimensions

The levator hiatus is the largest potential hernial portal in the human body and an important determinant of pelvic organ prolapse and prolapse recurrence. Manufacturers of ultrasound equipment have started to offer automatic determination of hiatal dimensions. We aimed to validate such data against manual measurement. This was an observational study using routinely obtained data sets of 62 women who attended a urogynecological unit in 2023. A Voluson E 22 system with SonoPelvic software was used for automated determination of hiatal dimensions. The latter were re-measured manually on de-identified stored data sets at a later date, using the software 4D View, blinded against all other data. The results show a comparison of manual (left) and automated (right) measurements of the levator hiatus on Valsalva. Four data sets were unavailable, leaving 58 for a comparison of manual and automatic measurements. The former provided a mean of 27.9 (range, 7- 53.8, SD 10.8)cm2 while the SonoPelvic automated measurements yielded a mean of 23.7 (5-44.8, SD 8.32)cm2, giving a Pearson's correlation of 0.663 (P < 0.0001). The difference between automated and manual measurements was statistically significant (P = 0.019). In 18 out of 58 cases, the discrepancy between automatic and manual measurement was 20% or higher in either direction, and these cases were reviewed. In some instances the discrepancy was due to incorrect placement of the reference plane as a result of misidentification of symphysis pubis or anorectal angle. In 11 others the hiatal circumference was mis- identified, most commonly due to a large cystocele, the contour of which was mis-identified as the hiatus (n = 9). Bland Altman analysis demonstrated a lack of systematic bias, but a substantial number of outliers in the sense of under-estimation of hiatal area by the AI. At this point in time, inexperienced users should not consider AI-based hiatal measurements to be authoritative. This is particularly true in women with large cystocele. Improvement is to be expected as development is ongoing.

Authors

Institutions

Publication Details

Journal
Ultrasound in Obstetrics and Gynecology
Published
2024-09-01
DOI
https://doi.org/10.1002/uog.27847
Citations
3
Primary Topic
Gastroesophageal reflux and treatments
Type
article
Field-Weighted Citation Impact
1.39
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

OC18.03: SonoPelvic™: artificial intelligence for the measurement of hiatal dimensions

Ka Lai Shek, H. Dietz
3 citations
Ultrasound in Obstetrics and Gynecology
Gastroesophageal reflux and treatments
1.39
article

OC18.03: SonoPelvic™: artificial intelligence for the measurement of hiatal dimensions

Ka Lai Shek, H. Dietz
article en
3 citations

Abstract

The levator hiatus is the largest potential hernial portal in the human body and an important determinant of pelvic organ prolapse and prolapse recurrence. Manufacturers of ultrasound equipment have started to offer automatic determination of hiatal dimensions. We aimed to validate such data against manual measurement. This was an observational study using routinely obtained data sets of 62 women who attended a urogynecological unit in 2023. A Voluson E 22 system with SonoPelvic software was used for automated determination of hiatal dimensions. The latter were re-measured manually on de-identified stored data sets at a later date, using the software 4D View, blinded against all other data. The results show a comparison of manual (left) and automated (right) measurements of the levator hiatus on Valsalva. Four data sets were unavailable, leaving 58 for a comparison of manual and automatic measurements. The former provided a mean of 27.9 (range, 7- 53.8, SD 10.8)cm2 while the SonoPelvic automated measurements yielded a mean of 23.7 (5-44.8, SD 8.32)cm2, giving a Pearson's correlation of 0.663 (P < 0.0001). The difference between automated and manual measurements was statistically significant (P = 0.019). In 18 out of 58 cases, the discrepancy between automatic and manual measurement was 20% or higher in either direction, and these cases were reviewed. In some instances the discrepancy was due to incorrect placement of the reference plane as a result of misidentification of symphysis pubis or anorectal angle. In 11 others the hiatal circumference was mis- identified, most commonly due to a large cystocele, the contour of which was mis-identified as the hiatus (n = 9). Bland Altman analysis demonstrated a lack of systematic bias, but a substantial number of outliers in the sense of under-estimation of hiatal area by the AI. At this point in time, inexperienced users should not consider AI-based hiatal measurements to be authoritative. This is particularly true in women with large cystocele. Improvement is to be expected as development is ongoing.

Ultrasound in Obstetrics and GynecologyVol. 64(S1)
Western Sydney University (AU)
Gender equality
Openalex Percentile: Top 19%
Gastroesophageal reflux and treatments
1.39
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

OC18.03: SonoPelvic™: artificial intelligence for the measurement of hiatal dimensions — Ka Lai Shek, H. Dietz · Ultrasound in Obstetrics and Gynecology (2024) | TGRS Research Map | TGRS