Author

Ronald Soule

Abstract

Doppler blood velocity waveform analysis is conducted to affect clinical diagnosis. Current analysis codes developed at RIT posses the capability to assess gross hemodynamic parameters such as heart rate, mean pulse velocity, peak systolic velocity and also the beat to beat variability of these parameters. These computer algorithms have, however, lacked the ability to determine hemodynamic indices such as the pulsatility and resistance index as well as the AB ratio. This latter deficiency stems from an algorithmic need to accurately determine end diastolic velocity in every cardiac cycle. The current thesis specifically augments current algorithms to accurately compute end diastolic velocity. The end diastolic velocity, peak systolic velocity and mean pulse velocity determined in each cardiac cycle are then used to compute the various pulsevelocity waveform indices noted above. In addition, the use of end diastolic velocity in conjunction with peak systolic velocity allows the velocity waveform to be dissected into diastolic subsections, which resemble decaying exponential curves. These exponential decay curves will be characterized via curve fitting. The goal of this thesis is to assess whether traditional pulsatility indices and/or the decay curve parameters are adequate to assess fetal developmental age between 10-13 weeks gestation. Discrimination assessment is conducted using neural network analysis techniques. Whether entire pulsevelocity waveforms extracted between successive end-diastolic velocities provides a more robust data set for gestational age discrimination is also explored. The results suggest that hemodynamic indices computed for fetuses between 10 to 13 weeks gestation provide insufficient data for effective neural network classification. Use of the entire pulse-velocity waveform data in neural network analysis showed better fetal gestational age classification than use of waveform indices. However, similarity of waveforms between 10-13 weeks gestation prevented robust classification using either hemodynamic indices or entire pulse-velocity waveforms based on the fetal data records used for this study.

Library of Congress Subject Headings

Fetal blood; Doppler ultrasonography; Blood flow--Measurement; Placenta--Blood-vessels; Fetus--Anatomy; Coronary circulation--Measurement

Publication Date

8-1-1999

Document Type

Thesis

Department, Program, or Center

Mechanical Engineering (KGCOE)

Advisor

Kempski, Mark

Advisor/Committee Member

Robinson, Risa

Advisor/Committee Member

Nye, Alan

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: RG618 .S68 1999

Campus

RIT – Main Campus

Share

COinS