Quantitative Multimodality PET/CT and PET/MR Imaging
Although diagnosis is currently the predominant application of multimodality imaging, therapy response monitoring, assessment and dosimetry are increasingly gaining ground. A prerequisite for these applications is image quantitation, which concerns all of the current clinically available multimodality imaging devices, including PET/CT, SPECT/CT and PET/MRI. Different data corrections, including those for attenuation, scatter and respiratory motion effects are needed in order to ensure the quantitative accuracy of the reconstructed images. The objective of this presentation is to provide an up to date view of recent advances in quantitative multimodality imaging and the applications mostly impacted by these developments.
The Promise of Radiomics and Multi-Parametric Modeling in Nuclear Medical Imaging
Prediction and follow up of therapy is of increasing interest in the field of multi-modality clinical imaging for oncology applications. In positron emission tomography (PET) imaging the image index predominantly used for the assessment of metabolic response is the maximum standardized uptake value corresponding to the normalized highest activity pixel value SUVmax and/or the normalized mean tumor activity concentration known as the mean Standardized Uptake Value SUVmean, within a region of interest around the tumor. Within the same context, it has been recently proposed that tumor functional volumes should be part of the classification criteria for response to therapy studies incorporating PET imaging. On the other hand, FDG tumor uptake has been shown to be associated not only with increased metabolism, but also with several other physiological parameters such as perfusion, cell proliferation, tumor viability, aggressiveness or hypoxia all of which may in turn be responsible for tumor uptake heterogeneity. Therefore, the hypothesis can be made that characterizing tumor FDG distribution, through its relationship to underlying tumor biological characteristics, may be useful in predicting therapy response and overall survival. The development of multi-parametric prognostic and predictive models based on combining multimodality imaging and biological indices for oncology applications, a field known as radiogenomics, will be clearly part of the future developments in this domain. The objective of this presentation is to provide an overview of the radiomics and radiogenomics field in nuclear medical multimodality imaging.
Motion Modeling/Correction in Multimodality PET/CT/MR Imaging
Respiratory motion is a major source of reduced quality in multi-modality imaging. In order to minimize its effects, it has been suggested to use respiratory synchronized acquisitions, leading to respiratory gated frames. Although such a solution may be appropriate for CT only imaging, in the case of nuclear imaging such gated frames are of low signal-to-noise ratio as they contain reduced statistics leading to a significant loss of contrast. Several existing compensation techniques are based on registering all the respiratory gated frames against a reference frame. Such a deformable registration can be performed based on the 4D CT or 4D PET images. The displacement vectors derived from such a registration can be subsequently applied either during, or after the reconstruction process allowing the combination of all respiratory gated frames making use of all available statistics acquired throughout a respiratory average image acquisition. Variations to this approach concern its application on restricted image regions of interest leading to local motion correction algorithms. The development of patient specific and generic respiration motion models represents a field of increasing interest and future research and development. PET respiratory motion correction approaches previously proposed in the field of PET/CT are directly applicable to PET/MR, although different MR sequences can be used to provide the necessary respiratory motion related dynamic information. This presentation will provide a comprehensive review of motion correction approaches in multimodality imaging.
Artificial Intelligence in the Nuclear Medical Imaging Context
The evolution of machine learning approaches with the introduction of deep learning techniques, has led to an exponential increase in the use of Artificial Intelligence (AI) in a variety of application areas, including the healthcare field. Within this context, AI has been introduced at different steps of medical imaging, covering the simulation of imaging devices, image formation including tomographic reconstruction and image related corrections to image processing for the subsequent exploitation of reconstructed images. The latest advances in the field of medical image formation and processing will be covered during this presentation including an insight on the potential interest of AI from a clinical perspective.
Dimitris Visvikis is a director of research with the National Institute of Health and Medical Research (INSERM) in France and co-director of the Medical Image Processing Lab in Brest (LaTIM, UMR1101). His current research interests focus on improvement in PET/CT image quantitation for specific oncology applications, such as response to therapy and radiotherapy treatment planning, through the development of methodologies for detection and correction of respiratory motion, 4D PET image reconstruction, tumor radiomics multiparametric and multimodality modelling, as well as the development of computer assisted interventional radiotherapy and Monte Carlo based dosimetry applications. He is a member of numerous professional societies such as IPEM (Fellow, Past Vice-President International), IEEE (Senior Member, Past NPSS NMISC chair), AAPM, EANM (Physics committee chair). He has won numerous awards including the SNMMI EJ Hoffman award 2020 for outstanding contributions in quantitative PET imaging, and the IEEE NPSS Shea Distinguished Member Award in 2019. He is member of numerous editorial boards and the first Editor in Chief of the IEEE Transactions in Radiation and Plasma Medical Sciences.