Prenatal screening for trisomy 21 in the first trimester: A comparison of a Thai non-biochemical predictive model and the Fetal Medicine Foundation algorithm.

Abstract

Objective: To evaluate the performance of a newly developed Thai non-biochemical predictive model compared with the established Fetal Medicine Foundation (FMF) algorithm for prenatal screening of trisomy 21 in Southeast Asian pregnant women at 11-13⁺⁶ weeks of pregnancy.

Methods: A secondary analysis was conducted on data from pregnant women attending Maharaj Nakorn Chiang Mai Hospital, Thailand, between 2011 and 2023. Trisomy 21 risk estimates were calculated using maternal characteristics and ultrasound parameters-crown-rump length (CRL), nuchal translucency (NT), and fetal heart rate (FHR)-via the FMF algorithm. The Thai model was developed using generalized linear regression incorporating maternal age, NT, and FHR. Risk classification thresholds were set at <1:250 (low-risk) and >1:250 (high-risk), and the models' diagnostic performances were compared.

Results: Among 8473 participants, 28 cases of trisomy 21 were identified (0.33%). The Thai model demonstrated comparable diagnostic accuracy to the FMF algorithm, with an area under the receiver operating characteristic curve (AUC) of 0.890 (95% confidence interval [CI] 0.819-0.961) versus 0.882 (95% CI 0.797-0.969) (P = 0.736). At a 1:250 cut-off, the FMF algorithm achieved 60.7% sensitivity and a 2.4% false-positive rate, whereas the Thai model performed optimally at a 1:100 cut-off, yielding 67.9% sensitivity with a 3.5% false-positive rate.

Conclusions: The Thai non-biochemical predictive model demonstrated similar performance to the FMF algorithm for trisomy 21 screening in Southeast Asian populations. The FMF algorithm remains a practical screening tool in low-resource settings, facilitating early and accurate risk stratification where advanced biochemical screening is unavailable.