Fractional thigh volume by three-dimensional ultrasonography for birth weight prediction

Objective: To determine the correlation between actual birthweight (BW) and fetal weight calculated from fractional thigh volume (ThiV). Material and Method: The authors have conducted a prospective, cross-sectional study of normal Thai fetal thigh volume. There were 176 eligible pregnant women who met the criteria of singleton with no fetal anomaly were recruited into the present study. Prior to the present study, 3 operators had been trained and standardized for fractional ThiV measurement by an expert for the first 20 cases. To generate the fetal weight calculating formula, fetal ThiV data from the first 100 cases were employed. Then, the authors' new prediction formula was compared and validated with the Hadlock's and the Tongsong's formula in 56 normal late-third-trimester fetuses. All patients were assessed for 2D fetal biometry and 3D fractional ThiV within one week before delivery. Results: A total of 176 fetuses underwent ultrasound at the gestational age of 38.5 ± 2.1 weeks. The reproducibility of fractional ThiV measurement technique showned very good correlation in both interand intra-observer reliability as observed by the high intraclass correlation (0.971-0.994). By using the regression model, fractional ThiV presents a superior correlation to actual BW (r = 0.965). The fitting formula is characterized by predicted fetal BW (g) = 774.744 + 32.658 x fractional ThiV (ml). The presented new formula shows the smallest absolute percentage error (APE) for BW estimation when compared to that of Hadlock's and Tongsong's. Conclusion: Fetal 3D-fractional ThiV is consistent with actual BW. The measurement of fractional ThiV can improve the accuracy of fetal weight prediction especially in some eventful conditions.