| Title |
|---|
| Early Prediction of Preeclampsia in Pregnancy Using Noninvasive Retinal Vascular Features |
| Authors |
|---|
| Yuxuan Wu, Xiaohang Wu, Haotian Lin |
| Presenting |
|---|
| Yuxuan Wu |
| PURPOSE: |
|---|
| Preeclampsia (PE), a severe hypertensive disorder during pregnancy, significantly contributes to maternal and neonatal mortality. Current prediction biomarkers are often invasive and expensive, hindering widespread application. This study introduces PROMPT (Preeclampsia Risk factor + Ophthalmic data + Mean arterial pressure Prediction Test), an AI-based model using retinal photography for PE prediction. |
| METHODS: |
|---|
| Analyzing 1812 pregnancies before 14 gestational weeks, retinal parameters were extracted via a deep learning system. PE prediction models were constructed using different machine learning algorithms based on meta data, novel biomarker, and retinal vessel parameters. Detection rate of adverse outcomes was analyzed across different screening strategies. A decision tree model was built for the cost-effectiveness analysis of different prediction models. |
| RESULTS: |
|---|
| The PROMPT achieved an AUC of 0.87 (0.83-0.90) for PE prediction and 0.91 (0.85-0.97) for preterm PE prediction using machine learning, significantly outperforming baseline model (P<0.001). It improved detection of severe adverse pregnancy outcomes from 35% to 41%. Economically, PROMPT was estimated to avert 1,809 PE cases and saved over $50 million per 100,000 screenings. |
| CONCLUSIONS: |
|---|
| These findings highlight PROMPT as a cost-effective, non-invasive tool for prenatal care, especially in low- and middle-income countries. |