AI-Powered, Point-of-Care Testing for Preeclampsia Prediction

Background & Unmet Need

• Preeclampsia is a serious hypertensive disorder of pregnancy that typically occurs ≥20 weeks of gestation and affects 2-8% of pregnancies worldwide
• Preeclampsia is the third leading cause of maternal mortality in the U.S., and ~10% of perinatal death is due to pregnancy complicated by preeclampsia
• Early interventions, such as low-dose aspirin, may be useful for preventing pre-term preeclampsia in at-risk women, but screening is challenging as many patients are asymptomatic
• Existing prediction methods largely depend on non-specific laboratory testing, clinical judgment and specialized knowledge, which may restrict their accessibility across different healthcare settings
• Moreover, the best predictive biomarker, the SFIt/PIGF ratio, has limited positive predictive value and is primarily used to rule out the condition
• Unmet Need: Early, accurate, and universally accessible methods to predict preeclampsia

Technology Overview

• The Technology: Integrated system combining blood testing and predictive modeling using AI and machine learning algorithms to accurately predict the risk of preeclampsia in pregnant women
• The model integrates data on maternal characteristics (e.g., age, race, gestational age, medical history) and routine lab results (e.g., CBC, basic metabolic panel, urine protein analysis)
• The model generates a rolling risk estimates at different gestational age windows between 28-40 weeks, forecasting preeclampsia onset within 1, 2, and 4 weeks
• This provides dynamic, short-term prediction of preeclampsia in late gestation using routine data
• PoC Data: The model was trained on ~36k patient records across three demographically distinct hospitals, achieving rolling AUCs between 0.75-0.88 across gestational weeks from 28 to 40

Technology Applications

• Routine prediction of preeclampsia in both acute clinical and routine prenatal care settings to facilitate early intervention
• Integration with EHR or telehealth platforms for continuous monitoring of high-risk patients
• Use as a research tool to analyze patterns and outcomes for preeclampsia

Technology Advantages

• Exclusively uses standard clinical lab results and patient data, streamlining integration with existing EHR systems
• Output does not require specialized interpretation by obstetrics providers, increasing utility in low resource or rural settings
• Provides real time processing and results to inform care