Perinatal outcome of monochorionic twin pregnancy complicated by selective fetal growth restriction: ISUOG vs Delphi diagnostic criteria

OBJECTIVE: To evaluate the diagnostic performance of the Delphi consensus definition for selective fetal growth restriction (sFGR), compared with the traditional definition recommended by the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG), in predicting adverse perinatal outcome in monochorionic diamniotic (MCDA) twin pregnancy. METHODS: This was a retrospective cohort study of MCDA twin pregnancies followed at a tertiary fetal medicine unit between January 2000 and January 2024. Cases diagnosed with twin-to-twin transfusion syndrome or twin anemia-polycythemia sequence before or at the time of sFGR diagnosis and those with fetal structural or genetic anomaly were excluded. Fetal growth was assessed using chorionicity-specific twin reference charts and sFGR was diagnosed using the ISUOG or Delphi definition. Logistic regression analysis was used to evaluate the performance of each constituent criterion of the Delphi definition in identifying cases at risk of adverse outcome. The diagnostic performance of the ISUOG and Delphi criteria was assessed using receiver-operating-characteristics (ROC)-curve analysis. RESULTS: The final analysis included 363 MCDA twin pregnancies, of which 110 (30.3%) were diagnosed with sFGR using the Delphi consensus definition. The ISUOG criteria identified only 53/363 (14.6%) cases as sFGR. The rate of intact survival of both twins was significantly lower among the 53 cases diagnosed using ISUOG criteria compared with the 57 cases diagnosed solely using Delphi criteria (26.4% vs 63.2%), with significantly lower neonatal morbidity in the latter group. Logistic regression analysis showed that each constituent criterion of the Delphi definition was associated independently with significantly reduced intact survival of both twins. All combinations of Delphi criteria showed low-to-moderate discriminative ability in predicting the demise of the smaller and/or larger twin (all areas under the ROC curve > 0.6). The Delphi criteria had slightly higher sensitivity (0.840 vs 0.789) but lower specificity (0.743 vs 0.877) compared with the ISUOG criteria for predicting the demise of the smaller twin. Similar results were obtained for the prediction of larger twin demise and double fetal demise. CONCLUSIONS: While the detection rate of sFGR was higher using the Delphi criteria compared with the ISUOG criteria, the additional cases identified solely using the Delphi definition had significantly lower perinatal morbidity and mortality compared with those meeting the ISUOG definition for sFGR. Nonetheless, each constituent criterion within the Delphi definition was independently associated with adverse outcome in sFGR twin pregnancy. Further research is needed to elucidate the most appropriate tools for diagnosing and classifying MCDA twin pregnancies complicated by sFGR. © 2026 International Society of Ultrasound in Obstetrics and Gynecology.