Individualized growth assessment of fetal thigh circumference using three-dimensional ultrasonography

Abstract

Objectives: To develop individualized growth assessment (IGA) standards for upper (ThC(u)) and middle (ThC(m)) fetal thigh circumferences using three-dimensional ultrasonography. Methods: A prospective, longitudinal sonographic study of 30 fetuses was performed beginning at 18 weeks' menstrual age. Second-trimester sonographic parameters were measured from three-dimensional volume data to establish IGA standards. Normal infant growth outcomes were confirmed using modified Neonatal Growth Assessment Scores (m 3NGAS51). ThC(u) and ThC(m) were studied in more detail. Rossavik growth model specification procedures, based on the slopes of the second-trimester growth curves, were developed for both ThC(u) and ThC(m). Third-trimester growth trajectories and birth measurements were subsequently predicted for these parameters. Percentage deviations during the third trimester and percentage differences at actual birth age were used to compare observed and predicted measurements. The 95% ranges for Growth Potential Realization Index (GPRI) values for both types of thigh circumference were determined. Values for m3NGAS51 using GPRIThC(u), GPRI ThC(m) and GPRITh(o) (original method) were compared. Results: The 30 newborns had no postnatal evidence of abnormal growth. Two examiners demonstrated a satisfactory measurement bias of mean ± SD 2.1 ± 3.6 (95% limits of agreement, -4.9 to 9.1)% for ThC(m) and 3.3 ±4.1 (95% limits of agreement, -4.8 to 11.4)% for ThC(u). Rossavik functions fitted parameter trajectories well, with mean R2 values of 99.5 ± 0.4% for ThC(u) and 99.6 ± 0.3% for ThC(m). By fixing coefficients k at their mean values, their respective fits did not change, and the variabilities of coefficients c and s were significantly reduced. For ThC(u), coefficient c was significantly related to the second-trimester slope (R2 - 98.6%), as was s to c (R2 = 91.0%). For ThC(m), coefficient c was significantly related to the second-trimester slope (R 2 = 98.6%), as was s to c (R2 = 85.6%). Third-trimester growth trajectories, derived from second-trimester slopes for individual fetuses, had third-trimester deviations of 0.07 ±3.7% for ThC(u) and -0.04 ±3.7% for ThC(m). Percentage differences at birth age were 16.8 ± 10.2% for ThC(u) and 8.9 ± 9.5% for ThC(m). With correction for systematic overestimations, the mean GPRI values were 103.7 (95% range, 90-121)% for ThC(u) and 101.6 (95% range, 88-118)% for ThC(m). Corresponding mean ± SD m3NGAS51 values, using GPRIThC(u) GPRIThC(m) and GPRIThC(o), were 203 ±11%, 201 ±10% and 200 ± 9%, respectively. Conclusions: Fetal thigh circumference can be measured reliably and evaluated using standard IGA methods. Both ThC(u) and ThC(m) give similar results in the third trimester but neonatal thigh circumference predictions are improved by using ThC(m). Corresponding GPRIThC(m) values are closer to the ideal value of 100% and can be used in m3NGAS51 calculations for assessment of neonatal growth outcome. Copyright © 2008 ISUOG. Published by John Wiley & Sons, Ltd.