The use of maternal plasma cell-free DNA (cfDNA) profiles can decrease performance in preeclampsia prediction models, according to a recent study in Frontiers in Medicine.
Takeaways
- Maternal plasma cell-free DNA (cfDNA) profiles offer a promising avenue for enhancing the accuracy of preeclampsia (PE) prediction models, potentially revolutionizing screening methods for this serious pregnancy complication.
- Traditional PE prediction models, relying on clinical risk factors such as PE history, chronic hypertension, and age, often yield high rates of false positives, highlighting the need for more accurate screening approaches.
- The study involved pregnant women aged 24 to 45 years who underwent non-invasive prenatal testing (NIPT), revealing associations between cfDNA profiles and early- or late-onset PE risks.
- Two classifiers, CEPE and CLPE, were developed using cfDNA profiles and clinical risk factors, demonstrating impressive accuracies of 0.87 and 0.90, respectively, with area under the curve scores indicating high predictive performance.
- Integration of cfDNA profiles into PE prediction models showed reduced variation and improved accuracy compared to traditional methods, offering a promising avenue for broader application of NIPT in PE screening.
PE, defined by new onset hypertension and proteinuria, presents in approximately 2% to 8% of patients. Categories of PE include early- and late-onset PE, and premature delivery is associated with multiple adverse fetal consequences. Therefore, accurate PE prediction is crucial, especially for early-onset PE.
Clinical risk factors of PE include PE history, chronic hypertension and diabetes mellitus, increased body mass index, older age, and in vitro fertilization (IVF). However, models utilizing these risk factors often have reduced accuracy, leading to high rates of false positives.
Approximately 10% to 15% of cfDNA in pregnant women is derived from placental trophoblasts. Gene expression in the plasma may be reflected by abundant cfDNA in maternal plasma, indicating potential of cfDNA as a biomarker for diseases caused by placental dysfunction during pregnancy.
Investigators conducted a study to determine the predictive performance of cfDNA biomarkers for PE. Participants included pregnant women aged 24 to 45 years who received non-invasive prenatal testing (NIPT) between 2019 and 2021 at 12+0~22+6 weeks’ gestation.
PE was determined by maternal systolic blood pressure of 140 mmHG or more or diastolic blood pressure of 90 mmHG or more after 20 weeks’ gestation. Early-onset PE was defined as PE diagnosis before 34 weeks’ gestation.
Exclusion criteria included malignant neoplastic diseases and fetal chromosomal abnormalities. Controls included patients without pregnancy complications, severe cardiac, hepatic, and renal insufficiency, malignant neoplastic diseases, or fetal chromosomal abnormalities.
Streck Cell Free DNA BCT blood collection tubes (Streck, La Vista, NE, United States) were used to store maternal peripheral blood. CfDNA extraction was completed using 200 microliters of maternal plasma.
Maternal plasma cfDNA coverages were evaluated at primary transcription start sites (pTSSs) to determine cfDNA profiles that could be used for predicting early- or late-onset PE risk among pregnant women. Pathway enrichment analysis was performed using g: Profiler.
NIPT data was collected from 2727 pregnant individuals. Early- or late-onset PE was more common in patients with increased body mass index (BMI), maternal age, past medical histories, or IVF. Late-onset PE was less common in parous women.
There were 117 pTSS coverages of promoters significantly higher among early-onset PE samples vs healthy controls, and 137 that were significantly lower. There were 266 and 344 promoters, respectively, for late-onset PE vs healthy controls. Lower pTSS coverages in both early- and late-onset PE samples were found for CACNB2 and NRF1 gene promoters.
During pathway enrichment analyses, there were genes enriched in multiple PE-associated pathways. For early-onset PE, these included the Hedgehog signaling pathway, Hippo signaling pathway, AMPK signaling pathway, and more. For late-onset PE, these included the HIF-1 signaling pathway, Ras signaling pathway, PI3K-Akt signaling pathway, and MAPK signaling pathway.
The early-onset PE classifier (CEPE) was developed using the cfDNA profile and clinical risk factors. Genes with pTSS coverage in the cfDNA profile included FOSL2, CAMKK2, CCND1, ITPR1, PRKACB and WNT7B. These genes were associated with at least 3 PE-associated pathways. Risk factors included maternal age, BMI, medical histories, conception method, and parity.
The late-onset PE classifier (CLPE) was developed using a cfDNA profile with pTSS coverage of the FLT3LG and EGF genes, as well as the same clinical risk factors used for the CEPE. When based on the logistic regression or random forest model, the CEPE and CLPE had accuracies of 0.87 and 0.90, respectively.
The CEPE had an area under the curve (AUC) of 87%, vs 96% for CLPE. When using 2 internal and 1 external validation datasets, the CEPE had AUCs ranging from 80% to 90%, while the CLPE had AUCs of 76%, 74%, and 72%, respectively.
These results indicated reduced variation in PE prediction models when utilizing cfDNA profiles. Investigators concluded this method may broaden future application of NIPT in PE screening.
Reference
Yu Y, Xu W, Zhang S, et al. Non-invasive prediction of preeclampsia using the maternal plasma cell-free DNA profile and clinical risk factors. Front Med (Lausanne). 2024;11:1254467. doi:10.3389/fmed.2024.1254467.
