Frequently, testing of Single Photon Emission Computed Tomography (SPECT) motion correction algorithms is done either by using simplistic deformations that do not accurately simulate true patient motion or by applying the algorithms directly to data acquired from a real patient, where the true internal motion is unknown. In this work, we describe a way to combine these two approaches by using imaging data acquired from real volunteers to simulate the data that the motion correction algorithms would normally observe. The goal is to provide an assessment framework which can both: simulate realistic SPECT acquisitions that incorporate realistic body deformations and provide a ground truth volume to compare against. Every part of the motion correction algorithm needs to be exercised: from parameter estimation of the motion model, to the final reconstruction results. In order to build the ground truth anthropomorphic numerical phantoms, we acquire high resolution MRI scans and motion observation data of a volunteer in multiple different configurations. We then extract the organ boundaries using thresholding, active contours, and morphology. Phantoms of radioactivity uptake and density inside the body can be generated from these boundaries to be used to simulate SPECT acquisitions. We present results on extraction of the ribs, lungs, heart, spine, and the rest of the soft tissue in the thorax using our segmentation approach. In general, extracting the lungs, heart, and ribs in images that do not contain the spine works well, but the spine could be better extracted using other methods that we discuss. We also go in depth into the software development component of this work, describing the C++ coding framework we used and the High Level Interactive GUI Language (HLING). HLING solved a lot of problems but introduced a fair bit of its own. We include a set of requirements to provide a foundation for the next attempt at developing a declarative and minimally restrictive methodology for writing interactive image processing applications in C++ based on lessons learned during the development of HLING.

Hokkaido University (北海道大学) / 博士 / 医学

Positron Emission Tomography (PET) plays a very important role in the field of drugdevelopment already in the preclinical phase. This is done by using positron labeled molecules for different approaches/methodologies such as cell analysis, frozen section autoradiography, homogenate binding, organ distribution and at the end in vivo small-animal PET imaging. The technique is also used for integrated animal studies in which both functional information from PET or Single Photon Emission Tomography (SPECT) and structural information from Computed Tomography (CT) are integrated. However, significant improvements in technical aspects of the animal scanner such as resolution (under 1 mm), high sensitivity and ease of the operational procedures have affected the usage of these types of imaging. This study aims to test the technical and operational performance of the FLEX Triumph preclinical PET/SPECT/CT imaging system. Spatial resolution, sensitivity and partial volume effects have been the parameters in focus but a performance comparison between different isotopes and an in vivo mouse study has also been performed. The results show that the Triumph is capable of producing high quality images for all modalities and also high quality PET/CT fusions.

No description available.

The objective of this work is to estimate and simulate organ uptake variability and correlations using measured data from the FastSPECT II Single Photon Emission Computed Tomography (SPECT) imaging system. We will investigate various methods that attempt to determine organ-uptake within a set of organs in a digital phantom; these methods include Region-of-Interest, Gauss-Markov, Wiener, and Reconstruction Estimation algorithms. In addition to the estimators, we will also test whether moving the phantom with respect to the imaging system and gathering multiple images from different positions will improve the performance of our estimators. The variations and correlations in the object will be modeled using Gaussian distributions with first and second-order statistics known exactly. We will present a mathematical formulation of this model, in a texture-free context, as well as some results on image-quality assessment. The object model will be the MOBY digital mouse phantom; the 4-D MOBY Mouse Model is a digital phantom developed by Paul Segars\citep{moby}, which provides a useful digital model for nuclear-medicine and CT imaging.

Inom nuklearmedicinsk diagnostik används radioaktiva läkemedel för att utvärdera olika organs metabolism och fysiologiska beteende. Genom att använda en scintillationskamera kan strålningen som erhålls när det administrerade läkemedlet sönderfaller i kroppen detekteras och en funktionell bild över aktivitetsfördelningen erhålls. En tredimensionell bildvolym kan erhållas om gammakameran får rotera runt patienten vilket kallas SPECT (Single Photon Emission Computed Tomography). Bildernas kvalitet är av stor betydelse för att kunna göra en noggrann bedömning av olika patologiska tillstånd. Kvaliteten begränsas av en mängd faktorer och en av dem är Comptonspridda fotoner. I denna studie optimerades bildkvaliteten för SPECT-undersökningar med 111In-Octreoscan för ett Infinia Hawkeye 4 (GE Healthcare, Wisconsin, USA) SPECT-system vid Norrlands universitetssjukhus (NUS). Optimeringen gjordes med avseende på detektion av små tumörer vid visuell inspektion av 111In-Octreoscan-bilder. Monte Carlo simuleringar användes för att utvärdera tre olika parallellhålskollimatorer med fyra olika fönsterinställningar. Rekonstruktion av bilder gjordes med den iterativa tekniken OSEM (Ordered Subset Expectation Maximization) med olika antal iterationer. Bilderna postfiltrerades med två olika filter med tre kritiska frekvenser vardera. Den lämpligaste inställningen för NUS visade sig vara MEGP-kollimatorn (Medium Energy General Purpose) tillsammans med en fönsterinställning med två huvudfönster centrerade kring 171 respektive 245 keV utan spridningskorrektion. Mest optimal rekonstruktion visade sig vara med två OSEM-iterationer (10 subsets) och postfiltrering med ett Butterworth-filter med kritisk frekvens 0.40 cm-1 och powerfaktor 8. I övrigt visade sig ELEGP-kollimatorn (Extended Low Energy General Purpose) vara den kollimator som optimerar bildkvaliteten mest med avseende på detektion av små tumörer, men den finns ännu inte på NUS. / Optimization of image quality for SPECT imaging with 111In-Octreoscan at the University Hospital of Umeå – a Monte Carlo study In nuclear medicine diagnostics, radiopharmaceuticals are used for evaluating metabolism and physiological behavior of various organs. By using a scintillation camera, radiation can be detected when the administered drug decays in the body, and the result is a functional image of the activity distribution within the patient. A three-dimensional image volume can be obtained by letting the gamma camera rotate around the patient. This method is called SPECT (Single Photon Emission Computed Tomography). Image quality is very important to make an accurate assessment of various pathological conditions. The quality is limited by many factors and one of them is the Compton scattered photons. In this study image quality of SPECT-examinations with 111In-Octreoscan were optimized for an Infinia Hawkeye 4 (GE Healthcare, Wisconsin, USA) SPECT-system at the University Hospital of Umeå (NUS). The optimization was made with respect to detecting small tumors for visual inspection of 111In-Octreoscan images. Monte Carlo simulations were used to evaluate three different parallel hole collimators with four different window settings. Reconstruction of images was performed with the iterative technique OSEM (Ordered Subset Expectation Maximization) with different numbers of iterations. The images were post-filtered with two different filters with three critical frequencies each. The most appropriate setting for the SPECT-system at NUS is the MEGP-collimator (Medium Energy General Purpose) with a window setting of two main windows centered around 171 and 245 keV, without scatter correction. The most optimal reconstruction is obtained by using two OSEM-iterations (10 subsets) and post-filtering with a Butterworth-filter with critical frequency 0.40 cm-1 and power factor 8. The ELEGP-collimator (Extended Low Energy General Purpose) proved however to be the most optimal collimator for detecting small tumors, but this collimator is currently not available at NUS.

No description available.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第19962号 / 医博第4152号 / 新制||医||1017(附属図書館) / 33058 / 京都大学大学院医学研究科医学専攻 / (主査)教授 髙橋 良輔, 教授 伊佐 正, 教授 村井 俊哉 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

O objetivo deste trabalho foi avaliar a influência das correções de atenuação e espalhamento na reconstrução quantitativa em SPECT. O estudo foi baseado em diversas simulações de Monte Carlo, com ênfase especial no modelo torso-cardíaco matemático (MCAT). Para a reconstrução, foi utilizado o algoritmo iterativo ML-EM com projetor-retroprojetor modificado pelo mapa de atenuação. Para avaliar a correção de espalhamento, foram simulados os espectros energéticos, com múltiplas ordens de espalhamento Compton. O método da dupla janela de energia (Jaszczak) foi aplicado, devido a sua simplicidade, e as imagens corrigidas foram comparadas com as de fótons primários. Foram analisadas as escolhas das janelas do fotopico e espalhamento, além da dependência do fator de espalhamento k com a distribuição de atividades do objeto. Duas abordagens foram adotadas para a obtenção dos mapas de atenuação: a estimativa do mapa uniforme diretamente dos dados de emissão, sem o uso de imagens de transmissão; e o borramento de mapas não-uniformes, reconstruídos a partir das projeções por transmissão. A estimativa do mapa de atenuação diretamente dos sinogramas de emissão baseou-se nas condições de consistência da transformada de Radon atenuada. Neste caso, foram estudados os efeitos de diferentes contagens e vários coeficientes de atenuação iniciais sobre as imagens corrigidas. Os mapas de atenuação não-uniformes foram borrados com um \"kernei\" gaussiano, aplicados nas correções e os efeitos na quantificação foram analisados. Os espectros energéticos emitidos pelo modelo MCAT mostraram que os fótons espalhados não poderiam ser excluídos a contento, mesmo que fossem utilizadas janelas de aquisição estreitas sobre o fotopico. Em relação a correção de Jaszczak, verificou-se que a escolha das janelas de fotopico e espalhamento é crucial e confirmou-se que o valor de k é altamente dependente do objeto examinado. Dada uma estimativa inicial do mapa de atenuação, o uso das condições de consistência para estimar o mapa de atenuação uniforme, consistente com os dados de emissão do modelo MCAT simulado, resultou sempre em uma mesma forma, para quaisquer valores iniciais do conjunto de parâmetros. Apesar do erro diminuir com o aumento da contagem, o melhor coeficiente de atenuação não pôde ser obtido, mesmo em altas contagens. Isto se deve a presença dos fótons espalhados, que alteraram a solução das condições de consistência, reduzindo as dimensões do mapa. Os resultados indicaram que a correção de espalhamento é o fator mais importante na reconstrução quantitativa em SPECT. Com referência aos efeitos quantitativos da correção de atenuação, não foram observadas diferenças significativas com a utilização dos mapas borrados, enquanto que a correção com mapas uniformes mostrou-se menos eficaz. / The goal of this work is to evaluate the influence of scatter and attenuation correction methods in quantitative SPECT reconstruction. The study was based on several Monte Carlo simulations, with special emphasis on the mathematical cardiac-torso phantom (MCAT). Iterative ML-EM reconstruction with modified projector-backprojector was used. To evaluate the scatter correction, energy spectra were simulated for SPECT imaging including multiple order Compton scattered photons. The dual energy window method proposed by Jaszczak was applied and scatter corrected images were compared with primary photons images. The choice of the scattering and photopeak windows and the dependence of the scatter factor k with the activity distribution were also analysed. Two approaches were adopted for obtaining the maps for attenuation correction: the estimation of the attenuation maps directly from the emission data, without transmission imaging, and the blurring of non-uniform attenuation maps, reconstructed from transmission data. The estimation of attenuation maps directly from the emission sinograms was based on the consistency conditions of attenuated Radon transform. In this case, the effects of different counting rates and various initial attenuation coefficients on the corrected images were studied. The non-uniform attenuation maps were blurred with a gaussian kernel with different variances, applied in further corrections and their effects on quantitation were examined. Analysis of energy spectra emitted from the MCAT phantom showed that scattered photons cannot be totally excluded, even when narrow acquisition windows were used. As far as the Jaszczak correction is concerned, results showed that the choice of photopeak and secondary windows is crucial and that the value of k is highly dependent on the imaged object. Given an initial estimation of the attenuation map with a constant coefficient, the use of consistency conditions to estimate the uniform map, consistent with the emission data of simulated MCAT phantom, resulted in the same shape for any set of initial parameters. In spite of the fact that the error falls with increasing counting rate, higher counts are not able to determine the best attenuation coefficient. This is due to scattered photons, which alter the solution of consistency conditions, reducing the size of estimated maps. Results indicated that the scatter correction is the most important factor inquantitative SPECT reconstruction. Furthermore, no significant differences were observed in the quantitation, when using the blurred non- uniform attenuation maps in attenuation correction, while corrections with uniform maps proved to be less efficient

Este foi um estudo retrospectivo de 18 pacientes com atrofia hipocampal (AH) e 21 pacientes com AH associada ao blurring do polo temporal (BPT), nos quais realizamos a investigação das alterações perfusionais ao SPECT (Single Photon Emission Computed Tomography) crítico, ao SPECT intercrítico e ao SISCOM (Subtraction Ictal SPECT Coregistered to MRI) e sua comparação com os dados de avaliação pré-cirúrgica. Os pacientes com BPT apresentaram início mais precoce da epilepsia, uma duração de epilepsia maior e um período maior de seguimento pós-operatório (PO). O padrão pefusional ipsilateral ao SPECT crítico, o qual denominamos de padrão -típico?, foi o padrão perfusional mais frequentemente encontrado em ambos os grupos de pacientes com e sem BPT. Ao SPECT intercrítico, a maioria dos pacientes com BPT apresentaram hipoperfusão ipsilateral no lobo temporal (LT) epileptogênico, enquanto que no grupo sem BPT esta hipoperfusão foi observada em metade dos pacientes. Ao SISCOM, o padrão perfusional -típico? foi novamente o mais encontrado em ambos os grupos com e sem BPT. Entretanto, os padrões considerados -atípicos? foram encontrados mais frequentemente nos pacientes com BPT, o que nos sugere um padrão de propagação das crises epilépticas discretamente diferente neste grupo com BPT, o qual envolve outras áreas dos LT bilateralmente. Contrariamente à nossa hipótese inicial, não encontramos associação entre a presença de BPT e a presença de hiperperfusão no polo temporal (PoT) seja ao SPECT crítico ou ao SISCOM. Por este achado, nossos dados não suportam a idéia de que o BPT seja resultado de alterações teciduais intercríticas secundárias a um maior envolvimento ou participação deste polo na geração ou propagação de crises epilépticas na epilepsia do lobo temporal (ELT) mesial. Por outro lado, nossos resultados sugerem que o PoT é uma região frequentemente envolvida pelas crises do LT. Particularmente, a hiperperfusão no PoT ipsilateral à AH ao SISCOM foi mais encontrada nos pacientes com casos clínicos típicos de ELT mesial, sugerindo que o PoT é mais frequentemente envolvido (um padrão típico) em crises epilépticas em casos clínicos com informações mais concordantes (clear cut) e sugestivas de ELT mesial unilateral / This was a retrospective study of 18 patients with hippocampal atrophy (HA) and 21 patients with HA associated to temporopolar blurring (TPB). We performed an investigation of perfusion changes of ictal single photon emission computed tomography (SPECT), interictal SPECT and subtraction ictal SPECT coregistered to magnetic resonance imaging (SISCOM) and its comparison with results of presurgical evaluation. Patients with TPB showed earlier epilepsy onset, longer duration of epilepsy and a longer period of post surgical follow-up. The ipsilateral perfusion pattern in the ictal SPECT, which we denominated -typical? pattern, was the most frequent in both groups of patients, with and wihout TPB. In the interictal SPECT, most of patients with TPB showed ipsilateral hypoperfusion to the epileptogenic temporal pole (TP), while in the group without TPB this hypoperfusion was observed in half of the patients. Through SISCOM, the typical perfusion pattern was the most frequent in both groups. However, the patterns considered -atypical? were found more frequently in patients with TPB, which suggests a pattern of propagation of epileptic seizures slightly different in the group with TPB, which involves other areas of TL bilaterally. Adversely to our hypothesis, we did not found association between the occurrence of TPB and the hyperperfusion in the TP neither in the ictal SPECT nor SISCOM. Based on this finding, our data does not support the idea that the BTP is a result of interictal secondary tissue changes due to a higher involvement or role of this pole in the generation or propagation of epileptic seizures in the mesial temporal lobe epilepsy (TLE). On the other hand, our results suggest that the TP is a region frequently involved by seizures of TL. Particularly, the hyperperfusion in the TP which is ipsilateral to the HA through SISCOM was the most found in patients with typical clinical cases of mesial TLE, suggesting that the TP is frequently involved (a typical pattern) in epileptic seizures of clinical cases with more concordant information (clearcut) and suggestive of unilateral mesial TLE