Machine learning in quantitative PET: A review of attenuation correction and low-count image reconstruction methods

The rapid expansion of machine learning is offering a new wave of opportunities for nuclear medicine. This paper reviews applications of machine learning for the study of attenuation correction (AC) and low-count image reconstruction in quantitative positron emission tomography (PET). Specifically, we present the developments of machine learning methodology, ranging from random forest and dictionary learning to the latest convolutional neural network-based architectures. For application in PET attenuation correction, two general strategies are reviewed: 1) generating synthetic CT from MR or non-AC PET for the purposes of PET AC, and 2) direct conversion from non-AC PET to AC PET. For low-count PET reconstruction, recent deep learning-based studies and the potential advantages over conventional machine learning-based methods are presented and discussed. In each application, the proposed methods, study designs and performance of published studies are listed and compared with a brief discussion. Finally, the overall contributions and remaining challenges are summarized.     

Micro-CT of rodents: State-of-the-art and future perspectives

Micron-scale computed tomography (micro-CT) is an essential tool for phenotyping and for elucidating diseases and their therapies. This work is focused on preclinical micro-CT imaging, reviewing relevant principles, technologies, and applications. Commonly, micro-CT provides high-resolution anatomic information, either on its own or in conjunction with lower-resolution functional imaging modalities such as positron emission tomography (PET) and single-photon emission computed tomography (SPECT). More recently, however, advanced applications of micro-CT produce functional information by translating clinical applications to model systems (e.g. measuring cardiac functional metrics) and by pioneering new ones (e.g. measuring tumor vascular permeability with nanoparticle contrast agents). The primary limitations of micro-CT imaging are the associated radiation dose and relatively poor soft tissue contrast. We review several image reconstruction strategies based on iterative, statistical, and gradient sparsity regularization, demonstrating that high image quality is achievable with low radiation dose given ever more powerful computational resources. We also review two contrast mechanisms under intense development. The first is spectral contrast for quantitative material discrimination in combination with passive or actively targeted nanoparticle contrast agents. The second is phase contrast which measures refraction in biological tissues for improved contrast and potentially reduced radiation dose relative to standard absorption imaging. These technological advancements promise to develop micro-CT into a commonplace, functional and even molecular imaging modality.     

La TEP et la TEMP pour l’étude des épilepsies

Positron emission tomography (PET) and single photon emission computed tomography (SPECT) imaging are very useful for the management of patients with medically refractory partial epilepsy. Presurgical evaluation of patients with medically refractory partial epilepsy often included PET imaging using FDG. The use of SPECT in these patients adds some more information and gives the clinicians the possibility of having ictal imaging. Furthermore, PET and SPECT imaging are performed to better understand the pathophysiology of epilepsy.     

Targeted molecular imaging of angiogenesis in PET and SPECT: a review

Over the past few decades, there have been significant advancements in the imaging techniques of positron emission tomography (PET) and single photon emission tomography (SPECT). These changes have allowed for the targeted imaging of cellular processes and the development of hybrid imaging systems (e.g., SPECT/CT and PET/CT), which provide both functional and structural images of biological systems. One area that has garnered particular attention is angiogenesis as it relates to ischemic heart disease and limb ischemia. Though the aforementioned techniques have benefits and consequences, they enable scientists and clinicians to identify regions that are vulnerable to or have been exposed to ischemic injury via non-invasive means. This literature review highlights the advancements in molecular imaging techniques and specific probes as they pertain to the process of angiogenesis in cardiovascular disease.     

Imaging lung perfusion

From the first measurements of the distribution of pulmonary blood flow using radioactive tracers by West and colleagues (J Clin Invest 40: 1-12, 1961) allowing gravitational differences in pulmonary blood flow to be described, the imaging of pulmonary blood flow has made considerable progress. The researcher employing modern imaging techniques now has the choice of several techniques, including magnetic resonance imaging (MRI), computerized tomography (CT), positron emission tomography (PET), and single photon emission computed tomography (SPECT). These techniques differ in several important ways: the resolution of the measurement, the type of contrast or tag used to image flow, and the amount of ionizing radiation associated with each measurement. In addition, the techniques vary in what is actually measured, whether it is capillary perfusion such as with PET and SPECT, or larger vessel information in addition to capillary perfusion such as with MRI and CT. Combined, these issues affect quantification and interpretation of data as well as the type of experiments possible using different techniques. The goal of this review is to give an overview of the techniques most commonly in use for physiological experiments along with the issues unique to each technique.     

Current and future anatomical and functional imaging approaches to pheochromocytoma and paraganglioma

After establishing a biochemical diagnosis, pheochromocytomas and extra-adrenal paragangliomas (PPGLs) can be localized using different anatomical and functional imaging modalities. These include computed tomography, magnetic resonance imaging, single-photon emission computed tomography (SPECT) using 123I-metaiodobenzylguanidine or 111In-DTPA-pentetreotide, and positron emission tomography (PET) using 6-[18F]-fluorodopamine (18F-FDA), 6-[18F]-fluoro-l-3,4-dihydroxyphenylalanine (18F-DOPA), and 2-[18F]-fluoro-2-deoxy-d-glucose. We review the currently available data on the performance of anatomical imaging, SPECT, and PET for the detection of (metastatic) PPGL as well as parasympathetic head and neck paragangliomas. We show that there appears to be no 'gold-standard' imaging technique for all patients with (suspected) PPGL. A tailor-made approach is warranted, guided by clinical, biochemical, and genetic characteristics. In the current era of a growing number of PET tracers, PPGL imaging has moved beyond tumor localization towards functional characterization of tumors.     

State of the art and challenges of time-of-flight PET

After a brief review of the history of time-of-flight (TOF) positron emission tomography (PET) instrumentation from the 1980s to present, the principles of TOF PET are introduced, the concept of time resolution and its effect on TOF gain in signal-to-noise ratio (SNR) are discussed. The factors influencing the time resolution of a TOF PET scanner are presented, with focus on the intrinsic properties of scintillators of particular interest for TOF PET. Finally, some open issues, challenges and achievements of today's TOF PET reconstruction are reviewed: the structure of the data organization, the choice of analytical or iterative method, the recent experimental assessment of TOF image quality, and the most promising applications of TOF PET.     

A Review of the Deep Learning Methods for Medical Images Super Resolution Problems

Super resolution problems are widely discussed in medical imaging. Spatial resolution of medical images are not sufficient due to the constraints such as image acquisition time, low irradiation dose or hardware limits. To address these problems, different super resolution methods have been proposed, such as optimization or learning-based approaches. Recently, deep learning methods become a thriving technology and are developing at an exponential speed. We think it is necessary to write a review to present the current situation of deep learning in medical imaging super resolution. In this paper, we first briefly introduce deep learning methods, then present a number of important deep learning approaches to solve super resolution problems, different architectures as well as up-sampling operations will be introduced. Afterwards, we focus on the applications of deep learning methods in medical imaging super resolution problems, the challenges to overcome will be presented as well.     

A realistic utilization of nanotechnology in molecular imaging and targeted radiotherapy of solid tumors

Precise dose delivery to malignant tissue in radiotherapy is of paramount importance for treatment efficacy while minimizing morbidity of surrounding normal tissues. Current conventional imaging techniques, such as magnetic resonance imaging (MRI) and computerized tomography (CT), are used to define the three-dimensional shape and volume of the tumor for radiation therapy. In many cases, these radiographic imaging (RI) techniques are ambiguous or provide limited information with regard to tumor margins and histopathology. Molecular imaging (MI) modalities, such as positron emission tomography (PET) and single photon-emission computed-tomography (SPECT) that can characterize tumor tissue, are rapidly becoming routine in radiation therapy. However, their inherent low spatial resolution impedes tumor delineation for the purposes of radiation treatment planning. This review will focus on applications of nanotechnology to synergize imaging modalities in order to accurately highlight, as well as subsequently target, tumor cells. Furthermore, using such nano-agents for imaging, simultaneous coupling of novel therapeutics including radiosensitizers can be delivered specifically to the tumor to maximize tumor cell killing while sparing normal tissue.     

Imagerie moléculaire dans les démences

Nuclear neuroimaging, with single photon emission computed tomography (SPECT) or with positron emission tomography (PET), allows study of functional and neurochemical aspects of human brain. It provides in vivo informations about pathophysiology in dementia. The routinely available tracers are perfusional tracers (^99mTc-HMPAO/GE Healthcare and ^99mTc-ECD/BMS). In June 2006, a new indication for the DaTSCAN^® (GE Healthcare) has been adopted. DaTSCAN^® could now be used to help differentiate probable dementia with Lewy bodies from Alzheimer's disease. Since a few years, there have been tremendous efforts expended to develop specific radioligands for several neurochemical systems and for imaging of beta-amyloid plaques. The aim of this paper is to review the most common radiotracers for SPECT and PET brain imaging in dementia and to focus on the new tracers.     

Advances in micro-CT imaging of small animals

PurposeMicron-scale computed tomography (micro-CT) imaging is a ubiquitous, cost-effective, and non-invasive three-dimensional imaging modality. We review recent developments and applications of micro-CT for preclinical research.MethodsBased on a comprehensive review of recent micro-CT literature, we summarize features of state-of-the-art hardware and ongoing challenges and promising research directions in the field.ResultsRepresentative features of commercially available micro-CT scanners and some new applications for both in vivo and ex vivo imaging are described. New advancements include spectral scanning using dual-energy micro-CT based on energy-integrating detectors or a new generation of photon-counting x-ray detectors (PCDs). Beyond two-material discrimination, PCDs enable quantitative differentiation of intrinsic tissues from one or more extrinsic contrast agents. When these extrinsic contrast agents are incorporated into a nanoparticle platform (e.g. liposomes), novel micro-CT imaging applications are possible such as combined therapy and diagnostic imaging in the field of cancer theranostics. Another major area of research in micro-CT is in x-ray phase contrast (XPC) imaging. XPC imaging opens CT to many new imaging applications because phase changes are more sensitive to density variations in soft tissues than standard absorption imaging. We further review the impact of deep learning on micro-CT. We feature several recent works which have successfully applied deep learning to micro-CT data, and we outline several challenges specific to micro-CT.ConclusionsAll of these advancements establish micro-CT imaging at the forefront of preclinical research, able to provide anatomical, functional, and even molecular information while serving as a testbench for translational research.     

Optimization of SPECT-CT Hybrid Imaging Using Iterative Image Reconstruction for Low-Dose CT: A Phantom Study

Background

Hybrid imaging combines nuclear medicine imaging such as single photon emission computed tomography (SPECT) or positron emission tomography (PET) with computed tomography (CT). Through this hybrid design, scanned patients accumulate radiation exposure from both applications. Imaging modalities have been the subject of long-term optimization efforts, focusing on diagnostic applications. It was the aim of this study to investigate the influence of an iterative CT image reconstruction algorithm (ASIR) on the image quality of the low-dose CT images.

Methodology/Principal Findings

Examinations were performed with a SPECT-CT scanner with standardized CT and SPECT-phantom geometries and CT protocols with systematically reduced X-ray tube currents. Analyses included image quality with respect to photon flux. Results were compared to the standard FBP reconstructed images. The general impact of the CT-based attenuation maps used during SPECT reconstruction was examined for two SPECT phantoms. Using ASIR for image reconstructions, image noise was reduced compared to FBP reconstructions for the same X-ray tube current. The Hounsfield unit (HU) values reconstructed by ASIR were correlated to the FBP HU values(R² ≥ 0.88) and the contrast-to-noise ratio (CNR) was improved by ASIR. However, for a phantom with increased attenuation, the HU values shifted for low X-ray tube currents I ≤ 60 mA (p ≤ 0.04). In addition, the shift of the HU values was observed within the attenuation corrected SPECT images for very low X-ray tube currents (I ≤ 20 mA, p ≤ 0.001).

Conclusion/Significance

In general, the decrease in X-ray tube current up to 30 mA in combination with ASIR led to a reduction of CT-related radiation exposure without a significant decrease in image quality.     

Apport de la tomographie par émission de positons dans les tumeurs endocrines digestives : choix du traceur

Digestive endocrine tumors represent a heterogeneous group of neoplasm sharing common characteristics such as their high density of peptide receptors, their ability to take up amino acids and decarboxylate them into biogenic amines and their low glycolytic activity. These features are used for nuclear imaging targeting. To date, somatostatin receptor scintigraphy is considered the “gold standard” imaging procedure of well-differentiated tumors. Despite the significant contribution of SPECT/CT, the use of positron emission tomography imaging (PET) is growing rapidly. Three PET imaging modalities are currently available: 68Ga-labeled somatostatin analogs PET, 18F-dihydroxyphenylalanine PET (¹⁸F-DOPA) and 18F-deoxyglucose PET (¹⁸F-FDG). This article focuses on the current targets of molecular imaging and highlights the potential clinical applications of new targets.     

Diagnostic imaging tests and microbial infections

Despite significant advances in the understanding of its pathogenesis, infection remains a major cause of patient morbidity and mortality. While the presence of infection may be suggested by signs and symptoms, imaging tests are often used to localize or confirm its presence. There are two principal imaging test types: morphological and functional. Morphological tests include radiographs, computed tomography (CT), magnetic resonance imaging, and sonongraphy. These procedures detect anatomic, or structural, alterations produced by microbial invasion and host response. Functional imaging tests reflect the physiological changes that are part of this process. Prototypical functional tests are radionuclide procedures such as bone, gallium, labelled leukocyte and fluorodeoxyglucose (FDG)-positron emission tomography (PET) imaging. In-line functional/morphological tomographic imaging systems, PET/CT and single photon emission tomography (SPECT)/CT, have revolutionized diagnostic imaging. These devices consist of a functional imaging device (PET or SPECT) joined together with a CT scanner. The patient undergoes both tests sequentially without leaving the examination table. Images from each study can be viewed separately and as fused images, providing precisely localized anatomic and functional information. It must be noted, however, that none of the current morphological or functional tests, either alone or in combination, are specific for infection and the goal of finding such an imaging test remains elusive.     

PET as a potential tool for imaging molecular mechanisms of oncology in man

During the past ten years, positron emission tomography (PET) has been increasingly developed for imaging and quantifying molecular mechanisms in oncology. The technique uses radionuclides to label molecules, which can then be imaged in man. The inherent sensitivity and specificity of PET is unrivalled because it can image molecular interactions and pathways, providing quantitative kinetic information down to the subpicomolar level. This technology has the potential to answer a large number of important clinical questions in translational research in oncology. However, the challenges in the methodology are substantial. Molecular imaging has the potential to assist in the optimization of molecular-based targeted therapies in cancer and to investigate the function of the genome.     

Imaging neurodegeneration in Parkinson's disease

Neuroimaging techniques have evolved over the past several years giving us unprecedented information about the degenerative process in Parkinson's disease (PD) and other movement disorders. Functional imaging approaches such as positron emission tomography (PET) and single photon emission computerised tomography (SPECT) have been successfully employed to detect dopaminergic dysfunction in PD, even while at a preclinical stage, and to demonstrate the effects of therapies on function of intact dopaminergic neurons within the affected striatum. PET and SPECT can also monitor PD progression as reflected by changes in brain levodopa and glucose metabolism and dopamine transporter binding. Structural imaging approaches include magnetic resonance imaging (MRI) and transcranial sonography (TCS). Recent advances in voxel-based morphometry and diffusion-weighted MRI have provided exciting potential applications for the differential diagnosis of parkinsonian syndromes. Substantia nigra hyperechogenicity, detected with TCS, may provide a marker of susceptibility to PD, probably reflecting disturbances of iron metabolism, but does not appear to correlate well with disease severity or change with disease progression. In the future novel radiotracers may help us assess the involvement of non-dopaminergic brain pathways in the pathology of both motor and non-motor complications in PD.     

Artificial intelligence and machine learning for medical imaging: A technology review

Artificial intelligence (AI) has recently become a very popular buzzword, as a consequence of disruptive technical advances and impressive experimental results, notably in the field of image analysis and processing. In medicine, specialties where images are central, like radiology, pathology or oncology, have seized the opportunity and considerable efforts in research and development have been deployed to transfer the potential of AI to clinical applications. With AI becoming a more mainstream tool for typical medical imaging analysis tasks, such as diagnosis, segmentation, or classification, the key for a safe and efficient use of clinical AI applications relies, in part, on informed practitioners. The aim of this review is to present the basic technological pillars of AI, together with the state-of-the-art machine learning methods and their application to medical imaging. In addition, we discuss the new trends and future research directions. This will help the reader to understand how AI methods are now becoming an ubiquitous tool in any medical image analysis workflow and pave the way for the clinical implementation of AI-based solutions.     

Nuclear cardiac imaging for the assessment of myocardial viability

An important aspect of the diagnostic and prognostic work-up of patients with ischaemic cardiomyopathy is the assessment of myocardial viability. Patients with left ventricular dysfunction who have viable myocardium are the patients at highest risk because of the potential for ischaemia but at the same time benefit most from revascularisation. It is important to identify viable myocardium in these patients, and radionuclide myocardial scintigraphy is an excellent tool for this. Single-photon emission computed tomography perfusion scintigraphy (SPECT), whether using ²⁰¹thallium, ^99mTc-sestamibi, or ^99mTc- tetrofosmin, in stress and/or rest protocols, has consistently been shown to be an effective modality for identifying myocardial viability and guiding appropriate management.Metabolic and perfusion imaging with positron emission tomography radiotracers frequently adds additional information and is a powerful tool for predicting which patients will have an improved outcome from revascularisation. New techniques in the nuclear cardiology field, such as attenuation corrected SPECT, dual isotope simultaneous acquisition (DISA) SPECT and gated FDG PET are promising and will further improve the detection of myocardial viability. Also the combination of multislice computed tomography scanners with PET opens possibilities of adding coronary calcium scoring and noninvasive coronary angiography to myocardial perfusion imaging and quantification.