AI Summary & Key Takeaways
AI InsightCompared with the biochemical model, the multimarker model showed higher sensitivity (81.0% vs 69.0%), specificity (95.7% vs 92.6%), positive predictive value (39.5% vs 24.6%), and overall accuracy (95.2% vs 91.8%).
📌 3 Key Takeaways
Abstract
Aims: To compare the diagnostic performance of biochemical risk assessment alone with that of a multimarker first-trimester risk assessment combining biochemical and sonographic parameters for screening of fetal chromosomal anomalies. Methods: This retrospective observational study included singleton pregnancies evaluated at a tertiary perinatology clinic between January 2022 and January 2026. Pregnancies undergoing first-trimester screening at 11+0 to 13+6 weeks of gestation were included. Biochemical risk was calculated using maternal serum free beta-human chorionic gonadotropin and pregnancy-associated plasma protein-A, whereas multimarker risk was derived from maternal characteristics, biochemical markers, and sonographic findings, including nuchal translucency and selected additional markers when available. The primary outcome was confirmed fetal chromosomal anomaly. Diagnostic indices and receiver operating characteristic (ROC) analyses were performed. Results: A total of 1,248 singleton pregnancies were analyzed, including 42 fetuses (3.4%) with chromosomal anomalies. Compared with the biochemical model, the multimarker model showed higher sensitivity (81.0% vs 69.0%), specificity (95.7% vs 92.6%), positive predictive value (39.5% vs 24.6%), and overall accuracy (95.2% vs 91.8%). The false-positive rate was lower with the multimarker approach (4.3% vs 7.4%). ROC analysis demonstrated superior discriminative performance for the multimarker model (AUC: 0.903 vs 0.812, p<0.001). Conclusion: Multimarker first-trimester risk assessment provided better overall diagnostic performance than biochemical risk assessment alone, particularly in terms of specificity, accuracy, and reduction of false-positive results. These estimates should be interpreted in light of the retrospective design, partial verification bias, and the limited number of affected pregnancies.