Keywords: Positronemissiontomography ; Pleurisy ; Pleural mesothelioma ; Pleural metastasis ; Lung cancer
Pleural disorders occur frequently in pulmonology and include a variety of inflammatory, infectious and industrial diseases and also malignancies. One of the most common presentations is pleural thickening or pleurisy. Researching the aetiology of pleural disease involves a very broad range of differential diagnoses, sometimes complicated by the fact that there are no criteria that are truly specific to malignancies on the conventional imaging exploration techniques such as plain radiographs, CT-scans and MRI [1,2]. It is true that the chemistry, bacteriology and cytology of the pleural fluid can make a significant contribution to the diagnosis [3,4]. However, although thoracocentesis is an essential examination, its sensitivity in the diagnosis of pleurisy caused by tuberculosis or malignancies is only 28 and 62% respectively . In both these aetiologies the sensitivity of pleural needle aspiration biopsy, performed blind, is only 51 and 44% [5-7]. Thoracoscopy is therefore often essential and in some cases unavoidable to research an accurate aetiological diagnosis and remains the gold standard for exploring pleural disorders. Its diagnostic yield is above 95% , although it is an invasive procedure and can sometimes be difficult to perform, especially in elderly patients whose
3 Universit´e de Rennes 1, LTSI, Rennes, F-35000, France
4 Centre de Recherche en Information Biomedicale Sino-francais (CRIBs), Rennes, France
This paper proposes a new sequential weighted least squares (SWLS) method for positronemissiontomography (PET) reconstruction. The SWLS algorithm is noniterative and can be considered as equivalent to the penalized WLS method under certain initial conditions. However, a full implementation of SWLS is computationally intensive. To overcome this problem, we propose a simplified SWLS as a reasonable alternative to the SWLS. The performance of this SWLS method is evaluated in experiments using both simulated and clinical data. The results show that the method can be advantageously compared with the original SWLS both in computation time and reconstruction quality.
as an important pillar of evidence-based medicine for diagnosis, follow-up of disease evolution and assessment of treatment effi- cacy 1 . In contrast with most bioanalytical techniques, in vivo imag- ing is non-invasive and non-destructive, and therefore repeatable. It can also be directly translated clinically, but comes with increased cost and complexity and typically yields just one or a few biological parameters at a time. This low parametric output is far from the multi-parametric data derived from omics technologies and greatly limits the capacity of in vivo imaging to decipher complex diseases with multiple hallmarks, such as cancer 2 . A preferred approach to overcoming this limitation is to combine different imaging tech- niques that superimpose co-registered information from the same subject. Accordingly, modern imaging calls increasingly on bimodal instruments, such as positronemissiontomography (PET)–com- puted tomography (CT) 3 , single-photon emission computed tomography–CT (SPECT–CT) 4 or the more recent PET–magnetic resonance imaging (MRI) 5,6 .
The TV-EM algorithm is less robust to the Gaussian noise than to the Poisson’s one, but its overall performance is not very high compared to the other methods.
The ML-EM method is the most general statistical approach and becomes the standard in positronemissiontomography. However, the reconstructed images become increasingly noisy as the number of iterations increases. To limit the noise accumulation, we have presented an iterative algorithm for PET reconstruction based on fuzzy anisotropic diffusion penalty. The proposed method is capable of taking advantages of fuzzy theory and anisotropic diffusion regularization. It appears more accurate compared to the ML-EM, OS-EM, Gaussian-MAP, MRP, and TV-EM algorithms. The fuzzy anisotropic diffusion penalty has an effect of edge-preserving and denoising. It allows suppression of noise and Gibbs phenomena without significantly affecting the edges. Simulation results showed that incorporation of the fuzzy anisotropic diffusion penalty improves the reconstruction quality for both noise-free and noisy
Xiao Tong 1,2 , Anikitos Garofalakis 1,2 , Albertine Dubois 1,2 , Raphaël Boisgard 1,2 , Frédéric Ducongé 1,2 , Régine Trébossen 1 and Bertrand Tavitian 1,2*
Background: Bimodal molecular imaging with fluorescence diffuse optical tomography (fDOT) and positronemissiontomography (PET) has the capacity to provide multiple molecular information of mouse tumors. The objective of the present study is to co-register fDOT and PET molecular images of tumors in mice automatically. Methods: The coordinates of bimodal fiducial markers (FM) in regions of detection were automatically detected in planar optical images (x, y positions) in laser pattern optical surface images (z position) and in 3-D PET images. A transformation matrix was calculated from the coordinates of the FM in fDOT and in PET and applied in order to co-register images of mice bearing neuroendocrine tumors.
A total body positronemissiontomography (PET) scanner (Panel A) showed foci of uptake in the right inferior man- dibula (arrowhead, an asymptomatic infected supplemental root), the thyroid (arrow, a papil- lary carcinoma), and the rectum (star). A colonoscopy reveals a non-uniform adenoma-
Molecular- and nuclear-based imaging techniques offer a large scale opportunity for detection of diseases. These techniques are based on the use of tracers labeled with radioactive isotopes and allow non-invasive in vivo detec- tion of different physiologic and pathologic phenomena with high sensitivity. PositronEmissionTomography (PET) is one of those that are increasingly used to diagnose and characterize disease activity in the setting of inflammatory disorders such as RA [7-10]. PET is even more sensitive than Single Photon Emission Computed Tomography (SPECT) and can also provide quantitative measurements. The low spatial resolution can make assigning the signal to specific anatomical structures difficult and can be partially compensated by combining PET with CT (X-ray Computed Tomography). [ 18 F]Fluorodeoxyglucose ([ 18 F]FDG), a radio- fluorinated analogue of glucose and the most widely used radiopharmaceutical worldwide today, was proposed for imaging RA [11,12]. Using this radiotracer, inflamed joints could be detected even though [ 18 F]FDG is not a specific marker of inflammation. Macrophage infiltration has been identified in early stages of RA , and therefore a spe- cific tracer of such a process would be more specific and possibly also enable an earlier detection of inflammation. Recently, expression of the folate receptor has been inves- tigated in a rat model of RA using [ 18 F]fluoro-PEG-folate  illustrating the interest for molecular imaging in this type of pathology.
PositronEmissionTomography (PET) is a functional, nuclear imaging modality that has become an integral part of patient management in a clinical setting. Nuclear imaging permits the observation of a physiological process as opposed to anatomical imaging which show the structures inside the body. As we explain in Chapter 1, PET enables the imaging of the spatiotemporal distribution of a radiotracer injected in the patient’s body. As the attached radionuclide undergoes a radioactive decay, a positron is emitted and eventually encounters an electron leading to the production of two annihilations 511 keV photons that are detected by the PET camera. Reconstruction algorithms, incorporating data correction factors, produce a quantitative image reflecting the distribution of the radiotracer which has been designed to target a specific process in the body. The multitude of radiotracers available explains the broad use of PET imaging in the clinic and research settings. In this work, we focus on one recently developed radiotracer: [ 18 F]-AV-1451 which permits the observation of the distribution of Paired Helical Filament (PHF) tau protein in the brain.
during sunitinib treatment. The identification and early detection of such events will be highly-beneficial for the clinical management of anti-cancer therapies with sunitinib.
1.3. PositronEmissionTomography (PET)
PET is a nuclear medical imaging modality (Divgi, 2009) whose principle consists in the simultaneous detection of two pairs of collinear gamma rays of 511 keV each. Those rays are emitted at the same time, under a relative angle of 180˚ to each other and are registered by opposing detectors in a ring scanner. The process of emission is called β + decay, in which photons are liberated after annihilation of a positron emitted from an unstable nucleus rich in protons (the radionuclide) and an electron of a nearby atom (Cherry, 2001; Ghosh et al., 2010). The detectors define a line of response with each true event of coincidence, in which the annihilation occurs somewhere along the line between the detectors (Anagnostopoulos et al., 2013). The line response is used to build multiple and sequential tomographic images which are then reconstructed three-dimensionally using a
as the best way to provide relevant information on a significant number of various cancer hallmarks (Culver et al., 2008; Hanahan and Weinberg, 2011). The complexity of tumors interactions with their environment calls for imaging methods capable of detecting a diversity of tumor hallmarks (Egeblad et al., 2010; Hanahan and Weinberg, 2011). Positronemissiontomography (PET) with [18F]2-deoxy-2-fluoro-D-glucose (FDG), the most efficient imaging method to detect cancer, is an indicator of tumor energy metabolism dominated by aerobic glycolysis in both cancer and tumor-associated inflammatory cells (Gillies et al., 2008). However, FDG-PET carries no information about other cancer hallmarks such as angiogenesis, replicative immortality, evasion of growth suppressors, capacity to metastasize, and yields at best indirect information on resistance to apoptosis and proliferation (Hanahan and Weinberg, 2011). PET imaging with other radiotracers can complement FDG, although radiolabeling of multiple tracers add in the complexity of an experiment. As far as experimental molecular imaging is concerned, multiple PET sessions are difficult to envision on a large scale because of high cost and low practicability.
paper on resting state brain metabolism measured by positronemissiontomography (PET) was included and discussed. 2 We were most surprised to see that the authors
of the review seem to have misunderstood the findings of our study, which concerned patients with psychogenic non-epileptic seizures (PNES). The authors state that the 16 patients included in our study “were later found to have PNES with comorbid epilepsy”. This is incorrect, since our study included only patients with PNES in whom comorbid epilepsy was excluded. This crucial point is indeed detailed in the Methods section of our article and clearly stated in the abstract: “in all patients, the diagnosis was subsequently confirmed to be PNES with no coexisting epilepsy.” It is thus on the basis of incorrect understanding of our results that Drs Ejareh dar and Kanaan discuss the possible significance of hypometabolism in the anterior cingulate region described in our paper, and erroneously suggest that interpretation of PET findings is complicated by coexistent epilepsy, which was not in fact the case.
This report is preliminary and will be completed by investigations of the metabolic activity of the aneurysm wall. A comparison of PET imaging with morphological and biochemical analyses of specimens of excised aneurysm wall should provide more insight in the pathogenesis of aneurysmal disease. Positronemissiontomography could also help the clinician to proceed to operation electively However, this specialised investigational procedure is not routinely available, and has not yet acquired a definitive place in the diagnosis or treatment of aortic aneurysms.
C. Dosimetry analysis
Full organ segmentation was performed by a single observer (FB) on CT images for clearly visible organs (i.e. brain, gallbladder, heart, kidneys, lung, spleen, liver and testes) using PMOD (Version 3.306, PMOD Technologies, Zurich, Switzerland). CT-based volumes of interest (VOI) were used on co-registered PET images to obtain average organ activity per volume in kBq/cc for each frame. Although the bladder was visible in the CT, the volume increase of the organ due to filling with urine throughout the PET emission scan created a misalignment between the location and size of the VOI in CT and PET emission scan. Average activity per volume was applied to organ volumes of the standard 70 kg hermaphroditic phantom  implemented in OLINDA/EXM (Version 1.1)  and normalized for injected activity. Heart wall and heart blood pool could not be separated in the images due to insufficient contrast and were summarized as heart wall. Time-activity curves (TAC) were linearly interpolated between time points (0.1 min, 5 min, 12 min, 26 min, 66 min, 101 min) and physical decay only was assumed after the last time point up to 10 h post injection. Biological clearance of activity from organs after 2h was neglected assuming the activity was trapped inside the organ. Time-integrated activity coefficients, which are mathematically equal to the number of disintegrations occurring within the source organ, were calculated from these TACs by trapezoidal numerical integration as proposed in literature [6, 13] using MATLAB software version 7.12.0 (Mathworks, Natick, USA). Since CT- based segmentation was not feasible for the urinary bladder and only a limited number of time points were available, urinary excretion scenarios with voiding every 4 h and a biological half-life