OBJECTIVE: The use of the micro positron emission tomography (microPET) technique provides a powerful means for molecular imaging on small animals, while its inferior spatial resolution offers insufficient anatomical information which impedes the interpretations of the scans. To improve this limitation, it often relies on a clinical magnetic resonance imaging (MRI) for providing anatomical details. In this study, we designed and developed a new image co-registration platform which contains a stereotactic frame and external fiducial markers for microPET and MRI studies. The image co-registration accuracies were also validated by this new platform using various imaging protocols for microPET and MRI. METHODS: The microPET images were reconstructed by filtered back-projection (FBP) and ordered subset expectation maximization (OSEM) methods. Two MRI pulse sequences, two-dimensional T1-weighted fast spin-echo (FSE) and three-dimensional spoiled gradient recalled (SPGR), were employed in the studies. Two MRI scanning protocols were proposed for small animal imaging: the whole-body high-speed mode and the partial high-resolution mode. RESULTS: Reconstructed images from two different modalities were integrated by point-to-point registration via the external fiducials. Four inter-modality matched co-registration pairs (FBP-FSE, FBP-SPGR, OSEM-FSE, OSEM-SPGR) were obtained for both the high speed and high resolution modes. Co-registration accuracy was given as the average fiducial registration error (FRE) between the centroids of six markers from the registered images. The overall systemic FREs were about 0.50 mm. CONCLUSIONS: From the inter-modality FRE comparison, MRI imaging with FSE performed better than that with SPGR sequence, due to its higher signal-to-noise ratio and less magnetic susceptibility effects. In the microPET perspective, the OSEM was superior to the FBP, as a result of fewer image artifacts.
OBJECTIVE: The use of the micro positron emission tomography (microPET) technique provides a powerful means for molecular imaging on small animals, while its inferior spatial resolution offers insufficient anatomical information which impedes the interpretations of the scans. To improve this limitation, it often relies on a clinical magnetic resonance imaging (MRI) for providing anatomical details. In this study, we designed and developed a new image co-registration platform which contains a stereotactic frame and external fiducial markers for microPET and MRI studies. The image co-registration accuracies were also validated by this new platform using various imaging protocols for microPET and MRI. METHODS: The microPET images were reconstructed by filtered back-projection (FBP) and ordered subset expectation maximization (OSEM) methods. Two MRI pulse sequences, two-dimensional T1-weighted fast spin-echo (FSE) and three-dimensional spoiled gradient recalled (SPGR), were employed in the studies. Two MRI scanning protocols were proposed for small animal imaging: the whole-body high-speed mode and the partial high-resolution mode. RESULTS: Reconstructed images from two different modalities were integrated by point-to-point registration via the external fiducials. Four inter-modality matched co-registration pairs (FBP-FSE, FBP-SPGR, OSEM-FSE, OSEM-SPGR) were obtained for both the high speed and high resolution modes. Co-registration accuracy was given as the average fiducial registration error (FRE) between the centroids of six markers from the registered images. The overall systemic FREs were about 0.50 mm. CONCLUSIONS: From the inter-modality FRE comparison, MRI imaging with FSE performed better than that with SPGR sequence, due to its higher signal-to-noise ratio and less magnetic susceptibility effects. In the microPET perspective, the OSEM was superior to the FBP, as a result of fewer image artifacts.
Authors: Thomas Rodt; Matthias Luepke; Claudia Boehm; Katja Hueper; Roman Halter; Silke Glage; Ludwig Hoy; Frank Wacker; Juergen Borlak; Christian von Falck Journal: PLoS One Date: 2012-09-21 Impact factor: 3.240