Numerical Analysis in the Detection of Abnormal Uptake of Radionuclide by the Brain

Abstract

Radionuclide brain imaging is a well-established diagnostic investigation, and electronic data processing is often used to improve the information yielded. Two methods of data analysis have been researched. The first used statistical techniques to analyse images of the distribution of a radionuclide in the brain to detect areas where the uptake pattern differed from the expected value, indicating disease. The normal uptake pattern, and the normal variability of this pattern in a population, was established by combining proven normal patterns into a composite. Novel methods of normalizing individual patient data were developed, and these methods revealed previously unreported features of the radionuclide scan in certain pathological conditions, which could be useful as diagnostic indicators. It was also found that the normalized patterns could be used in the automatic screening of large series of patient data to search for specific features within scans. This latter technique was used in the evaluation of the efficiencies of two choroid plexus blocking agents. The results of the statistical analysis showed that the computer assessment was closely comparable in diagnostic accuracy to an experienced radiologist.

The second method of analysis used the computer in conjunction with a commercially available scanner to produce transverse axial images of a radionuclide distribution in the brain. Substantial modifications to both the scanner and data processor were made, and new electronic circuitry was developed. A pilot study using programs developed elsewhere was conducted, and this showed the potential value of the method. New programs were then developed, suitable for routine use and intended to maximize the speed of the reconstruction and to minimize the size of the computer required. These programs were used in a clinical trial involving approximately 100 cases. The results obtained are discussed and compared with the results obtained with a similar technique using an X-ray scanner.

Date of Award	Mar 1977
Original language	English
Awarding Institution	Aston University