Mark Corson

By P. Jerek. Eastern Oregon University.

Relatively pragmatic methods have been developed to correct for scattered coincidences registered in brain studies using coin­ cidence events recorded outside the head [6] order 2.5 mg lozol with visa. However lozol 2.5mg low cost, direct monitoring of scatter has been achieved by recording coincidences that occur within the Compton region of the energy spectrum [7] purchase lozol 1.5mg with amex. This approach promises to be the most effective for cor­ recting data recorded over the chest and abdominal areas. Care is required to ensure that the subtraction of random rates, as recorded by delayed window coincidences, imposes the minimum statistical degradation of the data set. There is growing awareness that the first step in this process should not necessarily be that of fitting the data to a predetermined model. This follows, since the data may not be sufficiently accurate to produce a fully identifiable fit and in many cases the kinetic fate of the tracer is not fully defined or understood. A good example of this is the case of tumours within which there is known to be a wide heterogeneity of tissue types, including necrosis. As a result, a data lead approach is preferable whereby the data is first inspected to define the principal rate constants actually present in the data at the voxel level. Cluster, factor and spectral analysis [9] methods are being employed on a voxel by voxel basis to provide a first pass means for extracting the kinetic components that are actually present in the data set and not what the investigator thinks is there! Which of these kinetic survey techniques should be used remains an area for ongoing study and depends on the nature of the data set and confounding variables, such as the presence of labelled metabolites, etc. Within these strategies, the goal must remain to try and maintain the spatial resolution inherent in the reconstructed images. Once voxel by voxel parameters are derived, this provides powerful data sets which may be inter­ rogated for change between subjects in time using statistical techniques [10]. In recent years, powerful statistical techniques have been developed which allow the data to be interrogated globally and not just on a regions of interest basis [11]. Defining the statistical variations within the data set as a whole provides a baseline whereby focal changes are statistically identified. It is projected that medical imaging, as it becomes more available in digital form, will be subjected more and more to such statistical analysis and interpretation techniques. One conclusion derived from this meeting was that it is possible to separate imaging science into two major components: the quality of the data collected and the processing of the data. Scope exists for improving the specificity of tracers with further investment in radiolabelling with 99Tcm and 123I. The presence of the lead collimator severely impedes the full use of the flux photons emitted, and hence the radiation dose received by the subject studied. To overcome this, alternative detection principles are needed with the concept of the Compton camera [12] providing a lead contender for development. Bartholomew’s Hospital and Medical College and Imperial Cancer Research Fund Nuclear Medicine Group and St. The paper reviews some of the processes leading to the diagnosis of cancer using nuclear medicine. On the one hand, there are the general ‘catch all’ techniques starting with 67Ga and currently l8F-deoxyglucose. These form a type of nuclear radiology where it is identification of the cancer that is important, with high sensitivity but usually with low speci­ ficity. On the other hand, there are the developments of increasingly specific cancer identify­ ing techniques using tissue characterization based on those properties of the cancer cell which differ from the normal. These include surface antigens and receptors, but in the future perhaps direct imaging of the oncogene abnormality that is the basis of cancer will be possible. This combined approach to cancer distinguishes the discipline of nuclear medicinefrom diagnosticradiology andradiotherapy. Thisapproachdepended inthe paston exploiting the crude anatomical and pathophysiological differences between the cancer mass and the normal tissues interms of size, site, vascularity and some functional differences, such as the abilityofdifferentiatedthyroid cancer totake up l3lI when all normal competing tissue has been removed. These differencesexistintheir surfaceattributes: therange, quantitativeand qualitativedifferences insurfaceanti­ gens and receptors exposed to blood. On theone hand, there is the search for more and more cancer specific (and sensitive) radiolabelledtargeting[1](suchasradiolabelledantimelanomaantibodiesspecificto melanoma [2]);on theotherhand, therearemore sensitive,but rathernon-specific, ‘catch all disease’ agents, such as 6? A second featureisthe increasedpermeability and lackofnormal control factorsof tumour blood supply due to neovascularization. The thirdfeatureisexploitationofactivetransport, ofwhich thebestexample isthe iodine trap for 131I. Itisthentrapped afterphosphoryla­ tionby hexokinaseand neverreaches thepentoseorcitricacidcyclemetabolicpath­ ways. Another istheuptake by thewhite cellsand theirattractiontotheinflamma­ tory response that many tumours cause. All theseagentsaregenerallynotspecifictotumour typeand have thedisadvantagethat only a percentage ofalltumours ofa particulartype, such asbreastor lung cancer, willtakeup these agents. It also means thatloss ofuptake with treatmentdoes not mean loss of living tumour. A sickcell may not eat for some time, but itdoes not mean thatitwill not recover itsappetite in the future. Secondly, there are a few oncogenes whose presence causes prolifera­ tion directly. Since an oncogene isa set of altered D N A on a chromosome in the nucleus ofthe cell, one has to ask how such alterationsarose. The answer appears to be that a series of somatic mutations must take place to move from the normal D N A tothe D N A ofan oncogene causing cancer. The change from normal mucosa todysplasiatoa small adenoma toa larger adenoma to a cancer has a defined setofoncogenes on various chromo­ somes (ch). The primary alterationmay itselfhave ahereditary basis, as in familial adenomatous polyposis. Lack of oncoproteins action throughthesynthesisofnon-functioninganaloguesalterstheinternalmetabolism and theexternalcellsurfaceofthecancer: notonly initsdegreeand qualityofantigenic expression, butalsointhedegree, qualityand quantityofvariouscellsurfacerecep­ tors.

Accurate pre-operative localization of seizure origin is essential to obtain complete or nearly complete control of seizures generic 2.5mg lozol. The patients had been seizure free for at least 24 h order lozol 2.5 mg fast delivery, except in seven cases where the seizures were too frequent (seizure free for 4-20 h buy lozol 2.5mg with visa, with a mean time of 10 h). The patients were all closely observed for 10 min after tracer administration with no evidence of seizure during this time. The mean delay from seizure end to injection of the tracer was 86 ± 19 s (range 0-400 s). Sixty-four images were acquired with an acquisition time of 30 s per frame and an average of 70 000 counts per frame on a 64 x 64 matrix. Transaxial images were primarily obtained parallel to the long axis of the temporal lobe. The analysis was qualitative and an area was reported as abnormal if the increase or decrease of the tracer uptake was observed in at least two adjacent slices in comparison with the opposite temporal lobe. In two of them, lateralization differed from that obtained by the other methods and eventually the side of surgery. In two patients the hypoperfusion extended to the adjacent cortical areas, and in one case there was a contralateral cerebellar hypoper­ fusion. Another two cases demonstrated hypoperfusion of the temporal lobe extending to the adjacent cortical area, which corresponded to the eventual side of surgery (delay from seizure to injection was 70 and 400 s). Interictal hypoperfusion and ictal hyperperfusion in the temporal lobe corresponding to the eventual site of surgery was observed in six patients. In one of them, the interictal temporal hypoperfusion was associated with decreased ipsilateral parietal perfusion. The positive predictive value of combined ictal hyperperfusion and interictal hypoperfusion was 6/6 (100%). Ictal study reveals a markedly increased perfusion of the entire left temporal lobe. Ten patients were seizure free, four had nearly complete control of the seizure frequency (greater than 90% reduction in seizure frequency) and four had occasional simple partial seizures. For the successful excision of the epileptogenic focus, it is essential to obtain accurate pre-operative localization. It is therefore important to develop accurate, non-invasive methods of seizure focus localization. The reasons for these wide discrepancies may include the different popula­ tions of patients studied, the radiopharmaceutical used and the spatial resolution of the acquisition system. The degree of hypoperfusion seems to correlate inversely with the age of seizure onset [9]. When the tracer is administered in the post-ictal period, the sensitivity is lower, but seems to be higher than for interictal scans [9]. Consequently, better localization is obtained with ictal than with post-ictal studies and its sensitivity decreases with the time lag between seizure end and tracer injection [15]. During the seizure, it shows hyper- perfusion of the whole temporal lobe, up to 2 min post-ictally mesial hyperperfusion associated with ipsilateral lateral temporal lobe hypoperfusion, and hypoperfusion of the whole temporal lobe for the first few minutes after seizure onset [15]. Interictal hypoperfusion represents an area of dysfunction which may or may not be epileptogenic. However, when there is ictal hyperperfusion in the same focal area, it is highly suggestive of a seizure focus [15, 16]. It consists of two stages: initially it performs rotational and centring corrections in the coronal and transverse planes by maximizing two indices of bilateral hemispheric sym­ metry, and later it performs rotational correction in the sagittal plane from the automatic determination of the inclination angle of images in the sagittal plane. This is computed performing a line fit of the set of points of a curve that was defined by the inferior border of the encephalous in the image of the total sum of the sagittal planes. Using this value, and determining the orbito-meatal (O-M) position by a previous calibration study, the rotational correction in the sagittal plane is performed. In order to evaluate the algorithm, 20 patients were studied and a set of simulated symmetrical brain images with single and multiple lesions (software phantom) were examined. The relationship between the inclination angle of the volume and the O-M plane, computed in the calibration study, had a value of -8. The rotational and centring corrections showed satisfactory results, the rotational correction errors being less than 1. Este consta de dos etapas: en la primera se realiza la corrección de la orientación y centrado de los cortes en los planos coronal y transversal a partir de la maximización de dos índices de simetría inter­ hemisférica, y en la segunda se corrige la orientación en el plano sagital a partir de la deter­ minación automática del ángulo de inclinación de las imágenes en el plano sagital, calculado por el ajuste lineal de los puntos de una curva, que es definida por el borde inferior del encéfalo en la imagen formada por la superposición de los cortes en el plano sagital. Con este valor, y conociendo la posición del plano órbito meatal (O-M) a partir de un estudio de calibra­ ción realizado previamente, se hace la corrección por rotación en el plano sagital. Para la valoración de este método se estudió a 20 pacientes, y se examinó un juego de imágenes simuladas con lesiones únicas y múltiples (software phantom). La reorientación y centrado de las imágenes en los tres planos arrojó resultados satisfactorios, presentando un error menor que 1,4° en la reorientación, y menor que 0,5 pixeles en el centrado. La orientación y centrado de estos estudios tienen una importancia preponderante ya que permiten obtener niveles de cortes predeterminados que son comparados con patrones conocidos de perfusión, los cuales son utilizados como referencia para evaluar los resultados. La exactitud en el cálculo de índices relativos de perfusión interhemisférica a través de regiones de interés, también requiere de un adecuado centrado y alineamiento de las imágenes. Existen varios métodos empleados para garantizar una adecuada orientación y centrado de estas imágenes. Algunos de estos procedimientos consisten en posicionar la cabeza del paciente de tal forma que el plano orbito-meatal sea perpendicular al plano del detector y, otros, en reorientar las imágenes después de procesadas por “ software” de forma interactiva. Estos métodos son muy inexactos y tienen una gran dependencia de las habilidades y experiencias del operador, aunque en la actualidad se emplean sistemas con posicionamiento por láser con los que se logra una gran exactitud. Se adquireron 128 proyecciones de 15 s cada una, en formato de 64 X 64 y zoom de 1,14.