Development of a predictive model for preeclampsia utilizing time-of-flight mass spectrometry based on novel nanomaterials integrated with machine learning.

BACKGROUND

Preeclampsia (PE) is a serious complication of pregnancy, causing irreversible damage to multiple systems and organs of both mother and baby, and can even be life-threatening. Early diagnosis and intervention are key to improving maternal and fetal outcomes. Traditional diagnostic methods primarily rely on clinical symptoms and relatively single laboratory indicators, suffering from issues such as low sensitivity and insufficient specificity. Treatment is limited to symptomatic drug therapy for hypertension, and apart from terminating the pregnancy, there is a lack of effective treatment options. In recent years, Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) has shown great potential in disease biomarker screening due to its advantages of high sensitivity, high throughput, and rapid analysis.

METHODS

This study utilized MALDI-TOF MS technology based on novel inorganic nanosilica material, combined with high-throughput multi-omics(proteomics, peptidomics and metabolomics) analysis, to perform machine learning modeling and optimization analysis on 159 samples from the First People's Hospital of Yunnan Province.

RESULTS

Testing on the validation cohort showed an AUC value as high as 0.93 for preeclampsia detection, with the model's efficiency and accuracy surpassing traditional diagnostic methods. Furthermore, the machine learning model analysis identified 20 potential biomarkers associated with preeclampsia.

CONCLUSIONS

By constructing a diagnostic model for preeclampsia onset, this study can provide a basis for exploring the pathological mechanisms of preeclampsia. Simultaneously, it offers a novel technical approach for the screening and prediction of preeclampsia, holding significant clinical application value.