OC104: Soft tissue parameters improve the precision of fetal weight estimation



Current weight estimation models do not consider the contribution of fetal soft tissue to fetal weight. We have investigated new weight estimation models that use a combination of conventional biometry and soft tissue parameters.


A cross-sectional study of 248 fetuses was performed using 3DUS to acquire standard biometry (BPD, AC, FDL), fractional arm volume (Avol), and fractional thigh volume (TVol) within 4 days of delivery. Multiple linear regression was used to develop a ‘modified Hadlock’ model (BPD, AC, FDL) and similar models that also included soft tissue parameters for predicting birth weight (BW). New models were prospectively applied in 55 pregnancies. Predicted and observed weights were compared and signed percent differences calculated. Mean and SD of % differences were taken as measures of systematic and random prediction errors, respectively. Proportion of subjects with predicted BW within 5 or 10% (inclusive) were compared using McNemar's test.


For the validation group, the mean maternal age was 27 ± 7.0 (SD) years. Ethnicities included: Caucasian (19), African American (30), Asian (5), Hispanic (1). BW ranged from 390 to 5143 grams (28 males, 27 females). The ‘modified Hadlock’ model was Log BW = 0.0026(BPD2) + 0.0467(AC) + 0.1865(FDL)− 0.0042(AC × FDL) + 1.346. The best new model was 11.1372 (BPD2 − 67.2281(BPD) + 1.2175(AC2)− 17.3004(AC)− 0.0490(TVol2) + 25.3052(TVol) + 285.429.

Model ParametersR2Mean % DiffPredicted BW Within 5%Predicted BW Within 10%
  • *

    p = 0.002

  • **

    p = 0.008.

BPD, AC, FDL0.980.28 ± 7.6%28.9% (15/52)73.1% (38/52)
BPD, AC, TVol0.990.35 ± 6.2%63.5% (33/52)*92.3% (48/52)**


Inclusion of fractional thigh volume improves the precision of weight estimates when compared to conventional methods that consider only 2D measurements of the head, trunk, and limb. Soft tissue assessment may offer a novel paradigm for the intrauterine detection of fetal size and growth abnormalities.