Rigid-body motion correction in hybrid PET/MRI using spherical navigator echoes

Swin Transformer and the Unet Architecture to Correct Motion Artifacts in Magnetic Resonance Image Reconstruction

We present a deep learning-based method that corrects motion artifacts and thus accelerates data acquisition and reconstruction of magnetic resonance images. The novel model, the Motion Artifact Correction by Swin Network (MACS-Net), uses a Swin transformer layer as the fundamental block and the Unet architecture as the neural network backbone. We employ a hierarchical transformer with shifted windows to extract multiscale contextual features during encoding. A new dual upsampling technique is employed to enhance the spatial resolutions of feature maps in the Swin transformer-based decoder layer. A raw magnetic resonance imaging dataset is used for network training and testing; the data contain various motion artifacts with ground truth images of the same subjects. The results were compared to six state-of-the-art MRI image motion correction methods using two types of motions. When motions were brief (within 5 s), the method reduced the average normalized root mean square error (NRMSE) from 45.25% to 17.51%, increased the mean structural similarity index measure (SSIM) from 79.43% to 91.72%, and increased the peak signal-to-noise ratio (PSNR) from 18.24 to 26.57 dB. Similarly, when motions were extended from 5 to 10 s, our approach decreased the average NRMSE from 60.30% to 21.04%, improved the mean SSIM from 33.86% to 90.33%, and increased the PSNR from 15.64 to 24.99 dB. The anatomical structures of the corrected images and the motion-free brain data were similar.

Prospective motion correction for brain MRI using spherical navigators

PURPOSE

To demonstrate for the first time the feasibility of performing prospective motion correction using spherical navigators (SNAVs).

METHODS

SNAVs were interleaved in a 3D FLASH sequence with an additional short baseline scan (6.8 s) for fast rotation estimation. Assessment of SNAV-based prospective motion correction was performed in six volunteers. Participant motion was guided using randomly generated stepwise prompts as well as prompts derived from real motion cases. Experiments were performed on a 3 T MRI scanner using a 32-channel head coil.

RESULTS

When optimized for real-time application, SNAV-based motion estimates were computed in 25.8 ± 1.3 ms. Phantom-based quantification of rotation and translation accuracy indicated mean absolute errors of 0.10 ± 0.09° and 0.25 ± 0.14 mm, respectively. Implementing SNAV-based motion estimates for prospective motion correction led to a clear improvement in image quality with minimal increase in scan time (<5%).

CONCLUSION

Optimization of SNAV processing for real-time application enables prospective motion correction with low latency and minimal scan time requirements.

Evaluation of a Dedicated Radiofrequency Carotid PET/MRI Coil

Carotid radiofrequency coils inside a PET/MRI system can result in PET quantification errors. We compared the performance of a dedicated PET/MRI carotid coil against a coil for MRI-only use. An 18F-fluorodeoxyglucose (18F-FDG) phantom was scanned without and with an MRI-only coil and with the PET/MRI coil. The decay-corrected normalized activity was compared for the different coil configurations. Eighteen patients were scanned with the three coil configurations. The maximal standardized uptake values (SUVmax) and signal-to-noise ratios (SNR) were calculated. Repeated measures ANOVA was performed to assess the differences in SUVmax and SNR between the coil configurations. In the phantom study, the PET/MRI coil demonstrated a slight decrease (<5%), while the MRI-only coil showed a substantial decrease (up to 10%) in normalized activity at the position of coil elements compared to no dedicated coil configuration. In the patient study, the SUVmax values for both no surface coil (3.59 ± 0.15) and PET/MRI coil (3.54 ± 0.15) were significantly higher (p = 0.03 and p = 0.04, respectively) as compared to the MRI-only coil (3.28 ± 0.16). No significant difference was observed between PET/MRI and no surface coil (p = 1.0). The SNR values for both PET/MRI (7.31 ± 0.44) and MRI-only (7.62 ± 0.42) configurations demonstrated significantly higher (p < 0.001) SNR values as compared to the no surface coil (3.78 ± 0.22), while no significant difference was observed in SNR between the PET/MRI and MRI-only coil (p = 1.0). This study demonstrated that the PET/MRI coil can be used for PET imaging without requiring attenuation correction while acquiring high-resolution MR images.

caliPER: A software for blood-free parametric Patlak mapping using PET/MRI input function

Routine clinical use of absolute PET quantification techniques is limited by the need for serial arterial blood sampling for input function and more importantly by the lack of automated pharmacokinetic analysis tools that can be readily implemented in clinic with minimal effort. PET/MRI provides the ability for absolute quantification of PET probes without the need for serial arterial blood sampling using image-derived input functions (IDIFs). Here we introduce caliPER, a modular and scalable software for simplified pharmacokinetic modelling of PET probes with irreversible uptake or binding based on PET/MR IDIFs and Patlak Plot analysis. caliPER generates regional values or parametric maps of net influx rate (K_i) using reconstructed dynamic PET images and anatomical MRI aligned to PET for IDIF vessel delineation. We evaluated the performance of caliPER for blood-free region-based and pixel-wise Patlak analyses of [¹⁸F] FDG by comparing caliPER IDIF to serial arterial blood input functions and its application in imaging brain glucose hypometabolism in Frontotemporal dementia. IDIFs corrected for partial volume errors including spill-out and spill-in effects were similar to arterial blood input functions with a general bias of around 6-8%, even for arteries <5 mm. The K_i and cerebral metabolic rate of glucose estimated using caliPER IDIF were similar to estimates using arterial blood sampling (<2%) and within limits of whole brain values reported in literature. Overall, caliPER is a promising tool for irreversible PET tracer quantification and can simplify the ability to perform parametric analysis in clinical settings without the need for blood sampling.

An iterative image-based inter-frame motion compensation method for dynamic brain PET imaging

As a non-invasive imaging tool, positron emission tomography (PET) plays an important role in brain science and disease research. Dynamic acquisition is one way of brain PET imaging. Its wide application in clinical research has often been hindered by practical challenges, such as patient involuntary movement, which could degrade both image quality and the accuracy of the quantification. This is even more obvious in scans of patients with neurodegeneration or mental disorders. Conventional motion compensation methods were either based on images or raw measured data, were shown to be able to reduce the effect of motion on the image quality. As for a dynamic PET scan, motion compensation can be challenging as tracer kinetics and relatively high noise can be present in dynamic frames. In this work, we propose an image-based inter-frame motion compensation approach specifically designed for dynamic brain PET imaging. Our method has an iterative implementation that only requires reconstructed images, based on which the inter-frame subject movement can be estimated and compensated. The method utilized tracer-specific kinetic modelling and can deal with simple and complex movement patterns. The synthesized phantom study showed that the proposed method can compensate for the simulated motion in scans with¹⁸F-FDG,¹⁸F-Fallypride and¹⁸F-AV45. Fifteen dynamic¹⁸F-FDG patient scans with motion artifacts were also processed. The quality of the recovered image was superior to the one of the non-corrected images and the corrected images with other image-based methods. The proposed method enables retrospective image quality control for dynamic brain PET imaging, hence facilitating the applications of dynamic PET in clinics and research.

Synergistic motion compensation strategies for positron emission tomography when acquired simultaneously with magnetic resonance imaging

Subject motion in positron emission tomography (PET) is a key factor that degrades image resolution and quality, limiting its potential capabilities. Correcting for it is complicated due to the lack of sufficient measured PET data from each position. This poses a significant barrier in calculating the amount of motion occurring during a scan. Motion correction can be implemented at different stages of data processing either during or after image reconstruction, and once applied accurately can substantially improve image quality and information accuracy. With the development of integrated PET-MRI (magnetic resonance imaging) scanners, internal organ motion can be measured concurrently with both PET and MRI. In this review paper, we explore the synergistic use of PET and MRI data to correct for any motion that affects the PET images. Different types of motion that can occur during PET-MRI acquisitions are presented and the associated motion detection, estimation and correction methods are reviewed. Finally, some highlights from recent literature in selected human and animal imaging applications are presented and the importance of motion correction for accurate kinetic modelling in dynamic PET-MRI is emphasized. This article is part of the theme issue 'Synergistic tomographic image reconstruction: part 2'.

Abdominal and pelvic 18F-FDG PET/MR: a review of current and emerging oncologic applications

Positron emission tomography (PET) using fluorodeoxyglucose (¹⁸F-FDG) combined with magnetic resonance imaging (MR) is an emerging hybrid modality that has shown utility in evaluating abdominal and pelvic disease entities. Together, the high soft tissue contrast and metabolic/functional imaging capabilities make this modality ideal for oncologic imaging in many organ systems. Its clinical utility continues to evolve and future research will help solidify its role in oncologic imaging. In this manuscript, we aim to (1) provide an overview of the various PET/MR systems, describing the strengths and weaknesses of each system, and (2) review the oncologic applications for ¹⁸F-FDG PET/MR in the abdomen and pelvis.

An evaluation of the diagnostic equivalence of 18F-FDG-PET between hybrid PET/MRI and PET/CT in drug-resistant epilepsy: A pilot study

OBJECTIVE

Hybrid PET/MRI may improve detection of seizure-onset zone (SOZ) in drug-resistant epilepsy (DRE), however, concerns over PET bias from MRI-based attenuation correction (MRAC) have limited clinical adoption of PET/MRI. This study evaluated the diagnostic equivalency and potential clinical value of PET/MRI against PET/CT in DRE.

MATERIALS AND METHODS

MRI, FDG-PET and CT images (n = 18) were acquired using a hybrid PET/MRI and a CT scanner. To assess diagnostic equivalency, PET was reconstructed using MRAC (RESOLUTE) and CT-based attenuation correction (CTAC) to generate PET/MRI and PET/CT images, respectively. PET/MRI and PET/CT images were compared qualitatively through visual assessment and quantitatively through regional standardized uptake value (SUV) and z-score assessment. Diagnostic accuracy and sensitivity of PET/MRI and PET/CT for SOZ detection were calculated through comparison to reference standards (clinical hypothesis and histopathology, respectively).

RESULTS

Inter-reader agreement in visual assessment of PET/MRI and PET/CT images was 78 % and 81 %, respectively. PET/MRI and PET/CT were strongly correlated in mean SUV (r = 0.99, p < 0.001) and z-scores (r = 0.92, p < 0.001) across all brain regions. MRAC SUV bias was <5% in most brain regions except the inferior temporal gyrus, temporal pole, and cerebellum. Diagnostic accuracy and sensitivity were similar between PET/MRI and PET/CT (87 % vs. 85 % and 83 % vs. 83 %, respectively).

CONCLUSION

We demonstrate here that PET/MRI with optimal MRAC can yield similar diagnostic performance as PET/CT. Nevertheless, further exploration of the potential added value of PET/MRI is necessary before clinical adoption of PET/MRI for epilepsy imaging.

3D rigid-body motion information from spherical Lissajous navigators at small k-space radii: A proof of concept

PURPOSE

To demonstrate, for the first time, the feasibility of obtaining low-latency 3D rigid-body motion information from spherical Lissajous navigators acquired at extremely small k-space radii, which has significant advantages compared with previous techniques.

THEORY AND METHODS

A spherical navigator concept is proposed in which the surface of a k-space sphere is sampled on a 3D Lissajous curve at a radius of 0.1/cm. The navigator only uses a single excitation and is acquired in less than 5 ms. Rotation estimations were calculated with an algorithm from computer vision that exploits a rotation theorem of the spherical harmonics transform and has minimal computational cost. The effectiveness of the concept was investigated with phantom and in vivo measurements on a commercial 3T MRI scanner.

RESULTS

Scanner-induced in vivo motion was measured with maximum absolute errors of 0.58° and 0.33 mm for rotations and translations, respectively. In the case of real, in vivo motion, the proposed method showed good agreement with motion information from FSL image registrations (mean/maximum deviations of 0.37°/1.24° and 0.44 mm/1.35 mm). In addition, phantom measurements indicated precisions of 0.014° and 0.013 mm. The computations for complete motion information took, on average, 24 ms on an ordinary laptop.

CONCLUSIONS

This work demonstrates a proof of concept for obtaining accurate motion information from small-radius spherical navigators. The method has the potential to overcome several previously reported problems and could help increase the utility of navigator-based motion correction both in research and in the clinic.