摘要：

Positron emission tomography (PET) has the potential to play a role in early cancer detection. However, a major technical challenge is the spatial resolution of PET images. Besides the system resolution of 4~7mm for most clinical PET scanners nowadays, studies have shown that patients' Respiratory motion also contribute, sometimes significantly, to the degradation of image spatial resolution for the organs of lungs, liver, etc, in clinical F18-FDG oncological PET imaging. To improve the PET image resolution, we developed a new 4D image reconstruction method with respiratory motion for F18-FDG PET imaging. Our major contribution to this area is that we estimate the respiratory motion and reconstruct a “motion-free” PET image all based on an innovative 4D Poisson likelihood model for the respiratory-gated PET sinograms. In this talk, we first explain why we chose this approach to solve the respiratory motion problem. Then we discuss the details of our method and some preliminary results of evaluation. At the end, we conclude the work of the past 4 years and propose some future plans.

 

PET/MR has been one of the hottest topics in multi-modality medical imaging field during the past 6 or 7 years. Last year at RSNA, Siemens announced the first and so far the only clinical simultaneous whole body PET/MR imaging system in the world. In the pre-clinical world, simultaneous SPECT/MR could be as useful as PET/MR, if not more, for certain applications. There are two reasons: more variability of radiotracers and better spatial resolution for SPECT than PET when imaging small animals. Since 2006, we have been working on MR-compatible small animal SPECT systems based on the CZT detector technology. In this talk, we first discuss our work on characterizing the performances of the early generations of our MR-compatible SPECT systems as well as our initial studies of simultaneous SPECT/MR imaging for small animals. Then we introduce the current generation of our MR-compatible SPECT system, which has the potential to achieve sub-millimeter spatial resolution and relatively high sensitivity in simultaneous SPECT/MR imaging.