DOI RECORD
MV‑Flow imaging for predicting adverse pregnancy outcomes in gestational diabetes mellitus
Abstract
Abstract Objectives To assess longitudinal placental vascular index (VI) changes using MV-Flow ultrasound in gestational diabetes mellitus (GDM) and evaluate its predictive value for adverse perinatal outcomes. Methods We enrolled 112 women with GDM from a prospective cohort of 547 singleton pregnancies screened by 75g OGTT at 24–28 weeks. Diagnosis followed IADPSG criteria; exclusions left 60 with favorable outcomes and 29 with adverse outcomes. VIUp, VIMed, and VIDown were measured in upper, mid, and lower placental regions at 22–27 + 6 and 28–33 + 6 weeks using standardized MV-Flow. We calculated VIMean and ΔVI (third minus second trimester). Conventional Doppler parameters were collected concurrently. Group differences were tested parametrically or nonparametrically. Independent predictors were identified via multivariable logistic regression; model discrimination was assessed by ROC analysis and internally validated with 1,000 bootstrap resamples. Results Pre-pregnancy BMI and fasting OGTT glucose were higher in the adverse-outcome group (both P < 0.05). Third-trimester VIMean (36.6 ± 7.8 vs. 41.7 ± 9.3; P = 0.009) and VIMed (27.3 [27.3–48.7] vs. 34.7 [34.7–54.0]; P = 0.023) were lower in the adverse-outcome group. ΔVI and all Doppler parameters showed no group differences (all P > 0.05). Independent predictors were pre-pregnancy BMI (OR = 1.18; 95% CI: 1.03–1.35; P = 0.016), fasting OGTT glucose (OR = 1.12; 95% CI: 0.99–1.26; P = 0.053), and VIMed.2 (OR = 0.94; 95% CI: 0.90–0.98; P = 0.005). The combined model (BMI + fasting OGTT glucose + VIMed.2) had AUC = 0.770 (95% CI: 0.671–0.869), versus 0.718 for BMI alone (ΔAUC = 0.052; DeLong test, P = 0.262). Bootstrap-corrected AUC was 0.743. Conclusions Third-trimester VIMed is significantly lower in GDM pregnancies with adverse outcomes. The model combining pre-pregnancy BMI, fasting OGTT glucose, and VIMed.2 shows moderate predictive ability and stable performance—supporting its use as an adjunctive risk-stratification tool in GDM.