Classifying healthy women and preeclamptic patients from cardiovascular data using recurrence and complex network...

Area of Science:

• Cardiovascular physiology
• Prenatal medicine
• Complex systems analysis

Background:

• Preeclampsia is a life-threatening pregnancy complication requiring early detection.
• Current diagnostic methods often identify preeclampsia after clinical manifestation.

Purpose of the Study:

• To develop and evaluate recurrence-based methods for early preeclampsia detection.
• To distinguish between healthy and preeclamptic pregnancies in the second trimester.

Main Methods:

• Utilized cardiovascular time series: heart rate variability, systolic and diastolic blood pressures.
• Applied Recurrence Quantification Analysis (RQA) and novel ε-recurrence networks.
• Analyzed phase space coupling structures and derived various RQA and network measures.
• Employed quadratic discriminant analysis for classification.

Main Results:

• Recurrence-based methods successfully classified pregnancies.
• ε-recurrence networks achieved higher specificity (68%) compared to RQA (45.8%) at 91.7% sensitivity.
• Both methods demonstrated potential for early preeclampsia identification.

Conclusions:

• Recurrence quantification analysis and ε-recurrence networks are effective tools for early preeclampsia detection.
• These methods offer a non-invasive approach to identify at-risk pregnancies in the second trimester.
• Further validation is warranted for clinical implementation.