Date of Graduation

5-2019

Document Type

Thesis

Degree Name

Master of Science in Biomedical Engineering (MSBME)

Degree Level

Graduate

Department

Biomedical Engineering

Advisor/Mentor

Morten Jensen

Committee Member

Hanna Jensen

Second Committee Member

Jingxian Wu

Keywords

Machine Learning Predition Models, Medical Triage, Minimum Alveolar Concentration, Peripheral Venous Pressure, Propofol, Volume Change Assessment

Abstract

Analysis of peripheral venous pressure (PVP) waveforms is a novel method of monitoring intravascular volume. Two cohorts were used to study the hemodynamics change of the body state and its influence on the PVP using (1) dehydration setting with infants suffering from pyloric stenosis and (2) hemorrhage setting during a craniosynostosis elective surgery. The goal of this research is to develop a minimally invasive method of analyzing the PVP waveforms and find correlations with volume loss.

Twenty-three pyloric stenosis patients PVP were acquired at five stages and were divided into euvolemic, normal fluid volume, and hypovolemic, significant fluid loss. Seven craniosynostosis patients were enrolled and the PVP was acquired at the intervention to explore if the isoflurane dosage influences the PVP. A multivariate analysis of variances (MANOVA) was used to test if the PVP was influenced by the volume change and the anesthetic drugs effect. Prediction algorithms based on Fast Fourier Transform were utilized at the two cohort patients analyses to classify an arbitrary PVP into its correct classification.

Our research found that PVP signal is influenced by the different hemodynamics states of the body. Based on MANOVA outcomes, we built prediction systems and they were able to categorize an arbitrary PVP signal into its correct classification. The k-nearest neighbor (k-NN) model correctly predicted 77% of the data in the euvolemic and hypovolemic groups. The k-NN models of the anesthetic drugs were able to correctly predict correctly at least 85% of the preoperative and intraoperative signals of the pyloric stenosis patients and the different isoflurane dosages of the craniosynostosis patients.

Analyzing the PVP signal is a promising tool for measuring the dehydration level in acute settings. Our results imply that the subsequent changes in vascular resistance due to inhaled and infused anesthetics are reflected in the peripheral veins. A technology that would accurately assess the volume status of a patient to guide triage and treatment would be a significant improvement in various care settings. This minimally invasive technology utilizes a standard peripheral intravenous line and a commercial pressure-monitoring transducer, which exist today and requires no new clinical skills.

Share

COinS