20–24 October 2018, Singapore Oral communication abstracts Results: Mean BW was 3432 ± 557 g. All weight prediction models were accurate. Random errors for the Hadlock model (8.2%) were similar to the INTERGROWTH model (8.6%); both random errors were greater when compared to limb volume models (5.4 to 5.9%). The BW prediction error (TVol) was 1.7 ± 2.5% for BW > 4,000 g (n=6). Conclusions: Automated fractional limb volume provides a practi- cal means for including a nutritional soft tissue component as part of the fetal weight estimation process. In this diabetic cohort, these third trimester weight estimations were accurate and more precise than results from using 2D measurements alone. OC17.03: Table 1. Performance of 2D and 3D fetal weight estimation models in diabetic pregnancies (n=40) Weight prediction model Systematic error (%) Random error (%) Hadlock (BPD, AC, FDL) +2.3 8.2 INTERGROWTH (HC, AC) +0.4 8.6 Lee (BPD, AC, AVol) -0.6 5.4 Lee (BPD, AC, TVol) -1.1 5.9 Accuracy (systematic error) was expressed as signed mean percent difference from actual birth weight. Precision (random error) was expressed as the standard deviation of these percent differences. OC17.04 Adding fetal growth velocity parameters to maternal biochemical biomarkers improves the detection of small-for-gestational-age neonates M. Hendrix 1 , J. Bons 3 , R. Snellings 1 , S. van Kuijk 2 , O. Bekers 3 , M. Spaanderman 1 , S. Al Nasiry 1 1 Department of Obstetrics and Gynecology, Maastricht Uni- versity Medical Centre, Maastricht, Netherlands; 2 Department of Clinical Epidemiology and Medical Technology Assessment (KEMTA), Maastricht University Medical Centre, Maastricht, Netherlands; 3 Central Diagnostic Laboratory, Maastricht Uni- versity Medical Centre, Maastricht, Netherlands Objectives: To evaluate the value of adding fetal growth velocity parameters to maternal biochemical biomarkers for the detection of SGA neonates. Methods: A retrospective cohort study of 1094 singleton pregnan- cies, in the Maastricht University Medical Centre (MUMC) between 2011 and 2016. All women had ultrasound data of fetal growth from two periods: 18-22 and 30-34 weeks. The PAPP-A, β-hCG, PlGF and sFlt-1 values were measured at 11-13 weeks. Differences in maternal biochemical biomarkers and fetal growth velocities (mm/week) of the abdominal circumference (AC), biparietal diam- eter (BPD), head circumference (HC) and femur length (FL) were compared between the AGA (birth weight percentiles 10-90) and SGA (birth weight percentiles <10), using one-way ANOVA and post-hoc. Detection of SGA was calculated using ROC-curve and chi-square test. Results: Compared to AGA (n=1049) as reference group, SGA (n=45) had significant lower growth velocities ACv (mean±SD) (10.08±0.98 vs 11.26±1.00, p<.0001), BPDv (2.78±0.29 vs 3.08±0.27, p<.0001), HCv (9.98±0.83 vs 10.67±0.82, p<.0001) and FLv (2.34±0.252 vs 2.52±0.20, p<.0001). SGA compared with AGA had lower PAPP-a MoM (0.87±0.48 vs 1.13±0.70, p=0.015), higher sFlt-1 (1283.96±699.36 vs 1088.21±480.80, p=0.048), and a higher sFlt-1/PlGF ratio (50.16±41.56 vs 35.67±19.68, p=0.002). Combining all maternal biomarkers resulted in an AUC of 0.762 (0.655-0.869), with a sensitivity 84.4% (95%CI 69.9-93.0%) and negative predictive value (NPV) 93.8% (95%CI, 87.1-97.2%) for the prediction of SGA. However, addition of fetal growth velocities to the maternal biomarkers, improves the prediction of SGA with an AUC of 0.823 (0.741-0.906) sensitivity 92.3% (95%CI 79.4-97.7%); specificity 50.0% (95%CI 43.7-56.3%); and NPV 97.3% (95%CI 92.3-99.2%). Conclusions: Detection of small-for-gestational-neonates can be improved using the combination of fetal growth velocity parameters and maternal biochemical markers. A larger prospective study including serial maternal biochemical markers is needed. OC17. 05 Increasing fetal growth velocity increases the risk of shoulder dystocia among non-macrosomic fetuses T.M. MacDonald 1,2 , L. Hui 1,2 , A. Robinson 1 , K. Dane 1 , A. Middleton 1 , S. Tong 1,2 , S.P. Walker 1,2 1 Mercy Perinatal, Mercy Hospital for Women, Melbourne, VIC, Australia; 2 Obstetrics and Gynecology, University of Melbourne, Melbourne, VIC, Australia Objectives: Shoulder dystocia can lead to severe morbidity, but prediction for non-macrosomic fetuses is poor. We investigated whether fetuses who demonstrate increasing estimated fetal weight (EFW) or abdominal circumference (AC) centiles across the third trimester are at increased risk of shoulder dystocia. Methods: We performed a prospective, longitudinal study of 347 nulliparous women. EFW and AC centile were measured at 28 and 36 weeks, and the change in EFW or AC centile over exactly 8 weeks was calculated. Only live births delivered vaginally were included. We excluded cases of EFW >95 th centile at 36 week ultrasound, as these fetuses are already known to be at increased shoulder dystocia risk. We calculated the relative risk (RR) of shoulder dystocia for fetuses who demonstrated an increase in EFW or AC of >30 centiles over 8 weeks, irrespective of birthweight centile, compared to the rest of the cohort. We also correlated AC and EFW change in centile with neonatal body fat percentage and Ponderal Index. Results: 36 week EFW >95 th centile demonstrated 81.8% sensitivity and 91.0% specificity for birthweight >95 th centile. Of the 347 participants, 39(11.2%) had EFW >95 th centile at 36 week ultrasound and were excluded. Of the 308 participants remaining, 226(73.4%) delivered their baby vaginally and were included in the analysis, with 6(2.7%) cases of shoulder dystocia. Increasing EFW and AC centile were both significantly associated with increased shoulder dystocia risk. An increase in EFW by >30 centiles over 8 weeks was associated with a RR of 8.9 (p=0.03) for shoulder dystocia, and an increase in AC by >30 centiles over 8 weeks was associated with a RR of 7.7 (p=0.02). Change in EFW and AC centile were both significantly correlated with neonatal fat measures. Conclusions: Increasing EFW or AC centile across the third trimester is significantly associated with increased risk of shoulder dystocia in normal weight fetuses. This may assist the prediction of shoulder dystocia for the non-macrosomic infant. OC17.06 Prediction of Caesarean section due to intrapartum dystocia using antenatal customised fetal growth charts A. Dall’Asta 1 , T. Ghi 1 , G. Rizzo 2 , A. Kiener 1 , N. Volpe 1 , A. Bacigalupi 2 , A. Quarto 2 , G. Maruotti 3 , G. Saccone 3 , L. Sarno 3 , A. Fichera 4 , F. Prefumo 4 , T. Frusca 1 1 Department of Medicine and Surgery, Obstetrics and Gynecology Unit, University of Parma, Parma, Italy; 2 Division of Maternal and Fetal Medicine, Ospedale Cristo Re, University of Rome Tor Vergata, Rome, Italy; 3 Department of Neuroscience, Reproductive Sciences and Dentistry, School  The Authors 2018  Ultrasound in Obstetrics & Gynecology 2018; 52 (Suppl. 1): 1–65. 41