Biomedical subjects
Publications and source records attributed to P F Sharp.
The measurement of image quality is an essential stage in the evaluation of imaging techniques. Yet there is no accepted way of quantitatively assessing image quality. Current theories suggest that the quality of the raw data acquired by the device can be assessed independently from that of the displayed image. In this paper the various experimental techniques suitable for making psychometric measurements of displayed image quality are discussed.
Vascular leakage causing macular oedema is an important cause of visual loss in retinal vascular disorders such as diabetic retinopathy and retinal vein occlusion. By digitising selected frames from fundus fluorescein angiographic sequences we have developed a technique which uses the observed rate of change of fluorescence to detect and quantify areas of leakage. Reproducibility studies yielded a coefficient of variation of less than 6% for large areas of leakage, and 27% for very small areas of leakage. We believe this technique is sufficiently sensitive and robust for clinical use, and should find a significant role in the management of retinal vascular disorders.
Patients with a clinical diagnosis of Alzheimer's disease were studied using MRI, SPECT, and psychometric tests. Significant correlations between focal perfusion deficits and focal cognitive deficits were found. Significant correlations between regional relaxation time of white matter and psychometric tests of diffuse and focal categories were also found. Pathological examination confirmed Alzheimer's disease as the only diagnosis.
The aim of this study was to compare technetium-99m-hexamethylpropyleneamineoxime (99mTc-HMPAO) single-photon emission computed tomography (SPECT) with regional cerebral blood flow (rCBF) imaging using positron emission tomography (PET). As investigation of dementia is likely to be one of the main uses of routine rCBF imaging, 18 demented patients were imaged with both techniques. The PET data were compared quantitatively with three versions of the SPECT data. These were, first, data normalized to the SPECT cerebellar uptake, second, data linearly corrected using the PET cerebellar value and, finally, data Lassen corrected for washout from the high flow areas. Both the linearly-corrected (r = 0.81) and the Lassen-corrected (r = 0.79) HMPAO SPECT data showed good correlation with the PET rCBF data. The relationship between the normalized HMPAO SPECT data and the PET data was nonlinear. It is not yet possible to obtain rCBF values in absolute units from HMPAO SPECT without knowledge of the true rCBF in one reference region for each patient.
Explore the source record for details and available documents.
The interpretation of NMR images at Aberdeen is most often performed by the individual study of the proton density, difference, inversion recovery and T1 image. This report investigates the usefulness of presenting the information from a pair of images, in this instance the T1 and proton density images, in a single composite image. Information from the two images is combined such that the values from one image are represented by a change in hue (colour), and the values from the other by a change in luminance (intensity). To test the advantages of such combined images, a trial was run using a selection of prediagnosed abnormal brain scans. The information perceived as hue and luminance in the combined images was compared with that from separate conventional monochrome proton density and T1 images. Medical and nonmedical users were told the final diagnosis and were asked whether it was possible to see more, less, or the same information with regard to clinically relevant details in the structure of the abnormality and the image as whole. The results have revealed that for the majority of cases the combined image format can effectively represent the information normally contained in both the monochrome proton density and T1 images--thus speeding up the image viewing process.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
An earlier paper by the authors on the linearisation of computer-interfaced display devices has described a radiometric procedure for quantifying device non-linearity and a software technique for using the data to linearise screen luminance with respect to display grey level. As the display non-linearity varies with the particular settings of the brightness and contrast controls, any change in these settings requires that the data used with the software also be modified. While it is possible to store data for a number of selected control settings, it would be more convenient if linearisation could be performed using hardware for all relevant settings of these controls. The electronic circuit described in this paper makes use of the same non-linearity data to achieve device linearity with the added advantage of increased flexibility of use. The circuit is suitable for connection immediately following the digital-to-analogue converter or video processor and provides two useful new controls. One of these determines the amount of non-linearity correction to be applied to the video signal, while the second enables screen contrast to be varied without affecting the linearity achieved. The standard brightness and contrast controls of the display device are left as initially preset and require no subsequent adjustment.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
SPECT images of the brain can be obtained using either 123I labelled amines or 99mTc-hexamethylpropyleneamineoxime (HM-PAO). Both materials produce images which are blood flow dominated and so appear similar in normal subjects, although the respective mechanisms of uptake are not yet finally established. It seems likely, however, that the different mechanisms of uptake are responsible for recent reports of some differences seen in images obtained with the two types of agent in patients with cerebral pathology, mainly cerebrovascular disease. In this study, 12 demented patients, 6 with dementia of the Alzheimer type (DAT) and 6 with multi infarct dementia (MID), were imaged with 123I-isopropylamphetamine (IMP) and 99mTc-HM-PAO and the images compared. Significantly more lesions were seen with IMP than HM-PAO (P less than 0.02); out of a possible 120 sites, 41 lesions were seen with IMP compared to 28 with HM-PAO, 23 being seen with both agents. However, it is concluded that either agent can be used for the differential diagnosis of dementia, a task for which the new cerebral blood flow agents seem well suited.
Explore the source record for details and available documents.
One hundred fourteen patients suffering from neuropsychiatric conditions have been studied using 99mTc-labeled hexamethylpropyleneamine oxime (HM-PAO) and single photon emission computed tomography (SPECT). Ninety-one patients had a firm clinical diagnosis while 23 were examined without knowledge of the clinical diagnosis. Of the 91 patients, 51 were suffering from dementia, 25 multi-infarct type and 26 Alzheimer's disease. In 19 of the Alzheimer's patients, a characteristic pattern of decreased perfusion in the parieto-occipital regions was demonstrated while those with multi-infarct type showed varying degrees of irregular uptake in the cerebral cortex. These appearances are similar to those shown with positron emission tomography (PET) and we believe that HM-PAO will provide a widely available method for identifying patients with Alzheimer's disease. Twenty-nine patients were suffering from diseases involving the basal ganglia. Fifteen patients with Parkinson's disease showed no significant abnormality in basal ganglia uptake, while 7 or 8 patients with Huntington's disease who had full examinations showed decreased uptake in the caudate nuclei. Similarly, four of six patients with other basal ganglia diseases showed impaired uptake by basal ganglia, and it is concluded that HM-PAO may be useful for the diagnosis and management of this type of patient. Twenty-three patients received HM-PAO imaging as part of their diagnostic work-up; in 19 of them, detailed follow-up was obtained, which indicated that in 7 cases the result of the HM-PAO scan altered the clinical diagnosis and in 9 cases resulted in a change in management. In the remaining 13 cases, the study was found to be helpful in confirming the diagnosis.
Procedures for the adjustment of computer-interfaced display devices have been based on a subjective approach, generally relying on the visual examination of digital test patterns. The lack of an objective adjustment procedure means that the best image quality may not always be achieved. This paper describes how a simple home-made radiometer can be used to make screen luminance as well as density measurements on transparency film. The measurements are first used to determine the brightness and contrast settings for the display which enable the film to be used in the linear range of its optical densities. The display device characteristics, showing how luminance varies with grey level, are then determined. Using this information a simple display-linearising mapping can be produced to ensure that screen luminance (and hence film density) is a linear function of display grey-level value. Film-density measurements are also used to investigate the luminance and hence phosphor uniformity characteristics of the screen. The linearisation of the display device is an essential step in optimising the quality of recorded images. The measurements described are reproducible and sufficiently simple to provide the basis for a reliable quality-control procedure for the periodic assessment of the performance and settings of the display device.
Single photon emission computed tomography (SPECT) imaging of six Huntington's disease patients revealed a striking reduction in regional uptake of cerebral blood flow markers in vivo. Similar changes were found in one patient with "early stage" disease. The findings are compared with parallel magnetic resonance imaging (MRI) studies, and in one case, results of postmortem examination.
A patient with spasms of the neck, occurring when he turned his head to the left, responded to treatment with benzhexol. Cerebral blood flow imaging demonstrated reduced uptake in the right corpus striatum compared with the left. The study demonstrates the presence of an abnormality in the basal ganglia; it also illustrates response to drug treatment. Cerebral blood flow imaging may be useful in the detection of basal ganglia abnormalities in spasmodic torticollis and assist in the selection of cases which should be targeted for treatment with drugs.
Seventeen patients who had suffered a stroke in the previous 72 h have been studied using technetium 99m hexamethyl propylene amine oxime (HMPAO) to assess cerebral blood flow. Comparison of the scan appearances and clinical signs were made in all cases. Twelve of the patients had a repeat HMPAO scan 14 days later, together with further clinical assessment. Good correlation between the size and site of perfusion deficit and the clinical signs was found in 14 of the patients. When the results of the 12 patients who had repeat scans were examined at 14 days the correlation between the scan and the clinical signs was less accurate and it seems unlikely that the assessment of the size of cerebral infarction as demonstrated by an HMPAO scan will provide an accurate prognostic sign. Three of the patients, all of whom had suffered right hemiplegia and were dysphasic, had regions of increased uptake adjacent to the area of ischaemia and non deteriorated clinically; it is thought that this sign may represent hyperperfusion around infarction and indicates a fair prognosis. In patients suffering from transient neurological symptoms, the use of HMPAO may be useful by excluding the presence of other cerebral disease.