Masters Theses

Date of Award

8-1994

Degree Type

Thesis

Degree Name

Master of Science

Major

Nuclear Engineering

Major Professor

Laurence F. Miller

Committee Members

Gary Smith, Hedley Bond

Abstract

Positron emission tomography may be used to predict quantitative metabolic values useful in clinical diagnosis. Compartmental models requiring arterial input functions are necessary to make determinations of the values. By use of nuclear instrumentation, arterial input functions may be determined by invasive blood collection from a patient's arteries. This method has the drawbacks associated with arterial cannulation. In theory, PET may be used to determine arterial input functions provided an artery is located within the scanner field of view.

The primary goals of this work have been to (1) characterize and test a BGO scintillation detector that was used to measure the activity concentrations of positron- emitting isotopes in solution, (2) develop and use an automated arterial sampling system for determining arterial input functions, and (3) develop the necessary tools to determine arterial input functions from PET scans being performed for clinical diagnosis of ovarian cancer.

The Fluid Monitor was characterized in terms of its range of linear counting sensitivity, dead time, and operable constraints. A nonparametric function for converting count rates to specific activity values was determined over the near linear counting activity range of 0.01 µCi/ml to 140 µCi/ml. An average deadtime of 5.66 ± 1.04 µsec was determined for the detection system over this range of activity values. The activity range of 0.01 µCi/ml to 140 µCi/ml was established as the clinically useful and reliable activity counting range. Outside of this range, one may encounter difficulties establishing specific activity values that accurately reflect the true source strength.

An automated arterial sampling system has been developed, validated and used to collect arterial samples from five patients. The system's determined input function correlated very well (r = 0.997 and slope = 1.003 ± 0.005) to a theoretical function predicted in a model study. Furthermore, the arterial sampling system operated consistently and without failure during the patient studies performed. The software developed for the system provides a quick and simple method for determining high temporal resolution arterial input functions from the count rates collected.

A protocol was developed and used to measure arterial input functions from PET scans of the abdominal aorta or iliac arteries. The protocol calls for creating a region of interest within an artery to determine the arterial input functions. The abdominal aorta is generally larger than the resolution (≈ 10.65 mm) of the scanner when a Hann backprojection reconstruction filter with cutoff frequency of 0.4 cycles/cm is used. The iliac arteries, however, are typically smaller than the resolution. The abdominal aorta was captured in an image in only one of the five patient studies one or both of the iliac arteries was captured in the other four patient studies. A correlation of r = 0.949 and slope of 1.01 ± 0.05 was calculated between the integrated time-activity curves of the PET scan measurements and arterial sampling measurements. The method is a viable method for determining arterial input functions noninvasively, provided that clinical judgment is exercised.

A practical method for calculating recovery coefficients (RCs) was devised and validated in phantom studies. The calculated RCs compared to measured values with a correlation coefficient of r = 0.997 and a slope of 1.017± 0.013. A true test of the RC calculational method for human studies is as yet incomplete. The patient studies performed did not fully meet the criteria required for testing the RC calculational method. The method requires that the cross-sectional area of the blood vessel be measurable, which is straightforward with the abdominal aorta, but more difficult with the iliac arteries.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS