Early pregnancy volume measurements: validation of ultrasound techniques and new perspectives


M Rousian, Department of Obstetrics and Gynaecology, Division of Obstetrics and Prenatal Medicine, Erasmus MC, University Medical Centre Rotterdam, PO Box 2040, 3000 CA Rotterdam, the Netherlands. Email m.rousian@erasmusmc.nl


Objective  To investigate accuracy and reliability of four different ultrasound-related volume-measuring methods.

Design  Observational study.

Setting  Both in vitro and in vivo.

Population or Sample  Ten phantoms for in vitro measurements and 28 pregnancies with gestational ages ranging from 6 to 11 weeks for in vivo measurements were included.

Methods  Three-dimensional (3D) ultrasound images of phantoms (with known variable contents) and yolk sacs were used to calculate volumes using four different methods: Virtual Organ Computed-Aided AnaLysis (VOCAL), inversion mode, Sono Automatic Volume Calculation (SonoAVC) and V-Scope. V-Scope is a newly developed 3D volume visualisation application using a Barco I-Space virtual reality system. Intra- and interobserver agreement was established by calculating intraclass correlation coefficients (ICC).

Main outcome measure  Evaluation of accuracy and reliability by comparing the different techniques with true volumes (in vitro) and with each other (in vitro and in vivo).

Results  In the in vitro study, volume measurements by VOCAL, inversion mode and V-Scope proved to be accurate. SonoAVC measurements resulted in a substantial systematic underestimation. Correlation coefficients of measured versus true volumes were excellent in all four techniques. For all techniques, an intra- and interobserver agreement of at least 0.91 was found. Yolk sac measurements by the different techniques proved to be highly correlated (ICCs > 0.91).

Conclusions  We demonstrated that VOCAL, inversion mode and V-Scope can all be used to measure volumes of hypoechoic structures. The newly introduced V-Scope application proved to be accurate and reliable.