Electrical and Computer Engineering ETDs

Publication Date

5-6-1977

Abstract

The evaluation of kidney function is a necessary procedure in assessing the extent of renal disease in today's population. In particular, this assessment should include measurements of the patient's plasma flow into and urine flow from each kidney. Changes in either measurement from accepted normal ranges indicate the presence of a disease process.

The combination of radioactive tracer methods with data analysis using models describing biological kinetics has demonstrated the potential of large computers in the assessment of renal function. However, the requirement for sophi􀂵ticated computer resources has limited the widespread application these techniques might otherwise receive. This work investigates the application of minicomputers for performing this data analysis using digital filtering (deconvolution) methods.

A multiparameter compartmental model describing the distribution of 131I-orthoiodohippurate in the renal-vascular system is presented. Differential equations are developed which, when solved, describe the time behavior of this tracer material following its intravenous injection into the body. Analysis of this modeled system shows that only a few parameter values need be identified to characterize the functional vistate of thP renal system in terms of plasma flow and urine flow. The values for these parameter values can be obtained from the renal impulse responses, if these responses are available. Since the renogram can be modeled by a convolution of the renal impulse response with an appropriate renal input function, it follows that isolation of the renal impulse response can be accomplished by deconvolution. The development of this digital filter for performing these deconvolutions is developed in detail.

Following the deconvolution of the renogram, the resulting sequence can be analyzed and values assigned to the parameters requiring estimation. These parameter values, in turn, are used to provide measures of plasma flow to and urine flow from each kidney. Computer simulation demonstrated the proper behavior of the deconvolution filter. Clinical application demonstrated the ability to estimate unilateral plasma flow but urine flow could not be computed confidently.

In conclusion, it is felt the model does represent an adequate description of hippurate kinetics but does warrant further investigation. Although deconvolution techniques present an economical and efficient means for renal data analysis, the clinical utility cannot be fully realized until the signal-to-noise ratio of the data is improved.

Document Type

Thesis

Language

English

Degree Name

Electrical Engineering

Level of Degree

Masters

Department Name

Electrical and Computer Engineering

First Committee Member (Chair)

Harold Kund Knudsen

Second Committee Member

Joseph Thomas Coradro Jr.

Third Committee Member

John Marvin Brayer

Third Advisor

Daniel P. Peterson

Download

Share

COinS