Fetal weight estimation by three-dimensional ultrasound

Objective To assess the value of three-dimensional volume scanning in predi cting fetal weight at birth. Study design Prospective cross-sectional study within 7 days of delivery A total of 190 patients were considered for final analysis (formula-finding g roup: n = 125, formula evaluation group: n = 65). Inclusion criteria were a singleton pregnancy and absence of chromosomal or significant structural a nomalies. Three-dimensional (3D) volumetric measurements of the fetal thigh , upper arm and abdomen were performed together with conventional two-dimen sional (2D) biometry. Results All measurements were completed successfully in each patient. Polyn omial regression analysis with standard biometric parameters and volumes of the upper arm, the thigh and the abdomen was employed to yield the best-fi t formula for prediction of fetal weight at birth. The new 3D formula (esti mated fetal weight (EFW) = -1478.557 + 7.242 x thigh volume + 13.309 x uppe r arm volume +852.998 x log(10) abdominal volume + 0.526 x BPD3) proved to be superior to established 2D equations with the lowest mean error (25.8 +/ - 194.4 g), the lowest mean absolute error (155.2 +/- 118.2 g) and the lowe st mean absolute percentage error (6.1 +/- 5.0%) when studied prospectively in the evaluation group. Conclusion 3D sonography allows superior fetal weight estimation by includi ng soft tissue volume. Further studies at the extremes of fetal weight are needed to confirm the value of our formula in these subsets.