4D attenuation map generation in PET/MR imaging using 4D PET derived motion fields

Respiratory motion produces differences in PET attenuation maps from different phases of a respiratory-gated cardiac transmission acquisition. The effect of the contraction of the heart on the attenuation map is assumed to be small, but has not yet been investigated. The authors acquired list mode transmission data from two normally breathing humans while monitoring both respiration and the cardiac waveform. The data were retrospectively analyzed; first dividing them into separate time frames to check for gross patient motion, then into respiratory frames to isolate respiratory motion. A large, nearly motionless respiratory dataset is available near end expiration. These data were divided into 100 msec cardiac frames and reconstructed. Though up to 20 mm of respiratory motion was seen in the attenuation map, no discernible motion was detected between cardiac frames. The authors' finding implies that cardiac gating of the PET transmission map is unnecessary, and that only compensation for respiratory motion need be considered when correcting double-gated emission data for attenuation effects.

CT images acquired during deep-inspiration breath-holding (DIBH) have higher image quality than free-breathing (FB) CT images as they are free from respiratory motion artifacts. However, there exists significant mismatch between DIBH CT images and FB PET images since the displacement of organs in the thorax and abdomen at DIBH is significantly larger than that during regular FB. Our goal is to correct for the large mismatch between FB PET and DIBH CT in PET/CT imaging. The list-mode PET data were divided into four equal-count frames using amplitude based data-driven respiratory gating technique. Frame 1 and Frame 4 corresponded to end-expiration and end-inspiration respectively, and they were reconstructed using OSEM image reconstruction algorithm. To avoid attenuation-activity mismatch artifacts, the attenuation map used in attenuation correction was modified by filling the lung region with attenuation coefficient of soft tissue. 3D motion vector field from Frame 4 to Frame 1 was calculated from the reconstructed PET images using B-spline image registration algorithm. Assuming the respiratory motion during FB is consistent during FB and DIBH, the motion vector field from DIBH to end-inspiration phase is modeled by scaling the calculated respiratory motion from Frame 4 to Frame 1. The scalar was determined by maximizing the mutual information between transformed DIBH attenuation map and PET image at Frame 4. DIBH attenuation map was transformed using the scaled motion vector field from Frame 4 to Frame 1 to obtain attenuation map at Frame 4. Attenuation maps for other frames were obtained by transforming attenuation map at Frame 4 using corresponding motion vector fields. The method was applied to clinical patient data acquired using DIBH CT and FB PET protocol. While DIBH attenuation map creates significant mismatch artifacts in PET images, the proposed method was able to generate attenuation maps matching FB PET frames and reduce mismatch artifacts.

Current PET/MR imaging systems use methods based on MR image segmentation with subsequent assignment of empirical attenuation coefficients for attenuation correction in PET image reconstruction. Delineation of bone in MR images has been challenging, especially in the head and neck areas, due to the difficulty of separating bone from air. In this work, we study deep learning techniques to automatically generate attenuation maps directly from MR images, with focus on the head and neck areas. We use a generative adversarial network (GAN) in a conditional setting for this image translation task. GANs separate the deep learning network into a generator, which tries to generate fake examples, and a discriminator, which learns to distinguish between real and fake examples, and train the two networks simultaneously. The objective function of the conditional GAN is a combination of the generator loss, discriminator loss, and the L1 distance between the label image and the output image from the generator. Image pairs of PET/MR image (input) and corresponding PET/CT based 511 keV photon attenuation map (label) are used to train the network. The network is trained for 6k iterations. The generator loss is trained from 2.0 to 1.4. The discriminator loss is trained from 0.6 to 1.0. The L1 loss is trained from 0.2 to 0.1. In our previous work with a basic autoencoder network for the conversion from MR images to corresponding attenuation maps, if we convert the defined L2 loss for the network to the RMS pixel value prediction error, the autoencoder network, after training, has a pixel prediction error of around 0.2 when all pixel values are scaled from 0 to 1. In this study, the L1 loss also represents this pixel prediction error, which is around 0.1 after the network is trained. This indicates an average reduction of 50% of the pixel prediction error.

Definitions of fetal microcephaly and macrocephaly are debatable. A better understanding of their long-term prognoses would help guide parental education and counseling. This study aimed to explore the correlation between 2D and 3D fetal brain MR imaging biometry results and the long-term neurodevelopmental outcomes. This analysis is a historical cohort study. Fetal brain biometry was measured on 2D and 3D MR imaging using a volumetric MR imaging semiautomated algorithm. We measured and assessed the following brain structures: the supratentorial brain volume and cerebellar volume and cerebellar volume/supratentorial brain volume ratio, in addition to commonly used 2D brain MR imaging biometric variables, including occipitofrontal diameter, biparietal diameter, and transcerebellar diameter. Microcephaly was defined as ≤ 3rd percentile; and macrocephaly, as ≥ 97th percentile, corresponding to -2 SDs and +2 SDs. The neurodevelopmental outcome of this study cohort was evaluated using the Vineland-II Adaptive Behavior Scales, and the measurements were correlated to the Vineland standard scores. A total of 70 fetuses were included. No significant correlation was observed between the Vineland scores and either the supratentorial brain volume, cerebellar volume, or supratentorial brain volume/cerebellar volume ratio in 3D or 2D MR imaging measurements, after correction for multiple comparisons. No differences were found among fetuses with macrocephaly, normocephaly, or microcephaly regarding the median Vineland standard scores. Provided there is normal brain structure on MR imaging, the developmental milestone achievements in early years are unrelated to 2D and 3D fetal brain MR imaging biometry, in the range of measurements depicted in this study.

Respiratory motion and partial-volume effects are the two main sources of image degradation in whole-body PET imaging. Simultaneous PET-MR allows measurement of respiratory motion using MRI while collecting PET events. Improved PET images may be obtained by modeling respiratory motion and point spread function (PSF) within the PET iterative reconstruction process. In this study, the authors assessed the relative impact of PSF modeling and MR-based respiratory motion correction in phantoms and patient studies using a whole-body PET-MR scanner. An asymmetric exponential PSF model accounting for radially varying and axial detector blurring effects was obtained from point source acquisitions performed in the PET-MR scanner. A dedicated MRI acquisition protocol using single-slice steady state free-precession MR acquisitions interleaved with pencil-beam navigator echoes was developed to track respiratory motion during PET-MR studies. An iterative ordinary Poisson fully 3D OSEM PET reconstruction algorithm modeling all the physical effects of the acquisition (attenuation, scatters, random events, detectors efficiencies, PSF), as well as MR-based nonrigid respiratory deformations of tissues (in both emission and attenuation maps) was developed. Phantom and(18)F-FDG PET-MR patient studies were performed to evaluate the proposed quantitative PET-MR methods. The phantom experiment results showed that PSF modeling significantly improved contrast recovery while limiting noise propagation in the reconstruction process. In patients with soft-tissue static lesions, PSF modeling improved lesion contrast by 19.7%-109%, enhancing the detectability and assessment of small tumor foci. In a patient study with small moving hepatic lesions, the proposed reconstruction technique improved lesion contrast by 54.4%-98.1% and reduced apparent lesion size by 21.8%-34.2%. Improvements were particularly important for the smallest lesion undergoing large motion at the lung-liver interface. Heterogeneous tumor structures delineation was substantially improved. Enhancements offered by PSF modeling were more important when correcting for motion at the same time. The results suggest that the proposed quantitative PET-MR methods can significantly enhance the performance of tumor diagnosis and staging as compared to conventional methods. This approach may enable utilization of the full potential of the scanner in oncologic studies of both the lower abdomen, with moving lesions, as well as other parts of the body unaffected by motion.

Attenuation correction (AC) is a critical step to ensure accurate quantitative PET imaging. To eliminate the radiation dose from CT, deep learning (DL)-based methods have been extensively investigated to generate the CT-equivalent attenuation map (μ-CT) directly from the PET signal. However, almost all previous studies only focus on 18F-FDG due to its extensive data availability which is suitable for DL model training. For other less common tracer types, it is generally believed that new models must be trained separately on these tracer-specific data to ensure reasonable performance. In this work, we explored the cross-tracer generalizability of μ-DL generation DL models - primarily focusing on whether a model trained on a commonly used tracer like 18F-FDG can be effectively applied to less common tracers such as 68Ga-DOTATE and 18F-Fluciclovine, and vice versa. Unlike methods that directly generate attenuation-corrected (AC) PET images from non-attenuation corrected (NAC) PET images or MLAA reconstructions, we generate the CT-based deep learning attenuation maps (μ-DL) using MLAA reconstruction with the combined input of attenuation maps (μ-MLAA) and tracer activity (λ-MLAA). This μ-DL is then used for attenuation correction to obtain the final AC PET image. Our comprehensive evaluations on both μ-CT generation and the PET reconstruction found that the DL model trained on one specific tracer can be adapted to other tracers with competitive performance when compared to the tracer-specific trained DL model. The 18F-FDG-trained model demonstrated the best generalizability when applied to less common tracer types which often have relatively fewer available data for training. Additionally, we investigated the role of the μ-MLAA and λ-MLAA as inputs for the network performance. We found that combining both inputs resulted in the best performance, but the μ-MLAA contributed more significantly compared to the λ-MLAA.

Simultaneous PET and MR imaging is a promising new technique allowing the fusion of functional (PET) and anatomic/functional (MR) information. In the thoracic-abdominal regions, respiratory motion is a major challenge leading to reduced quantitative and qualitative image accuracy. Correction methodologies include the use of gated frames that lead to low signal-to-noise ratio considering the associated low statistics. More advanced correction approaches, previously developed for PET/CT imaging, consist of either registering all the reconstructed gated frames to the reference frame or incorporating motion parameters into the iterative reconstruction process to produce a single motion-compensated PET image. The goal of this work was to compare these two—previously implemented in PET/CT—correction approaches within the context of PET/MR motion correction for oncology applications using clinical 4-dimensional PET/MR acquisitions. Two different correction approaches were evaluated comparing the incorporation of elastic transformations extracted from 4-dimensional MR imaging datasets during PET list-mode image reconstruction to a postreconstruction image-based approach. Eleven patient datasets acquired on a PET/MR system were used. T1-weighted 4D MR images were registered to the end-expiration image using a nonrigid B-spline registration algorithm to derive deformation matrices accounting for respiratory motion. The derived matrices were subsequently incorporated within a PET image reconstruction of the original emission list-mode data (reconstruction space [RS] method). The corrected images were compared with those produced by applying the deformation matrices in the image space (IS method) followed by summing the realigned gated frames, as well as with uncorrected motion-averaged images. Both correction techniques led to significant improvement in accounting for respiratory motion artifacts when compared with uncorrected motion-averaged images. These improvements included signal-to-noise ratio (mean increase of 28.0% and 24.2% for the RS and IS methods, respectively), lesion size (reduction of 60.4% and 47.9%, respectively), lesion contrast (increase of 70.1% and 57.2%, respectively), and lesion position (changes of 60.9% and 46.7%, respectively). Our results demonstrate significant respiratory motion compensation using both methods, with superior results from a 4D PET RS approach.

Respiratory motion (RM) and cardiac motion (CM) degrade the quality and resolution in cardiac PET scans. We have developed non-rigid motion estimation methods to estimate both RM and CM based on 4D cardiac gated PET data alone, and compensate the dual respiratory and cardiac (R&C) motions after (MCAR), during (MCDR), and before (MCBR) image reconstruction. In all three R&C motion correction methods, attenuation-activity mismatch effect was modeled by using transformed attenuation maps using the estimated RM. The difference of using activity preserving and non-activity preserving models in R&C correction was also studied. Realistic Monte Carlo simulated 4D cardiac PET data using the 4D XCAT phantom and accurate models of the scanner design parameters and performance characteristics at different noise levels were employed as the known truth and for method development and evaluation. Results from the simulation study suggested that all three dual R&C motion correction methods provide substantial improvement in the quality of 4D cardiac gated PET images as compared with no motion correction. Specifically, the MCDR method yields the best performance for all different noise levels compared with the MCAR and MCBR methods. While MCBR reduces computational time dramatically but the resultant 4D cardiac gated PET images has overall inferior image quality when compared to that from the MCAR and MCDR approaches in the ‘almost’ noise free case. Also, the MCBR method has better noise handling properties when compared with MCAR and provides better quantitative results in high noise cases. When the goal is to reduce scan time or patient radiation dose, MCDR and MCBR provide a good compromise between image quality and computational times.

Given the ability of positron emission tomography (PET) imaging to localize malignancies in heterogeneous tumors and tumors that lack an X-ray computed tomography (CT) correlate, combined PET/CT-guided biopsy may improve the diagnostic yield of biopsies. However, PET and CT images are naturally susceptible to problems due to respiratory motion, leading to imprecise tumor localization and shape distortion. To facilitate PET/CT-guided needle biopsy, we developed and investigated the feasibility of a workflow that allows to bring PET image guidance into interventional CT suite while accounting for respiratory motion. The performance of PET/CT respiratory motion correction using registered and summed phases method was evaluated through computer simulations using the mathematical 4D extended cardiac-torso phantom, with motion simulated from real respiratory traces. The performance of PET/CT-guided biopsy procedure was evaluated through operation on a physical anthropomorphic phantom. Vials containing radiolabeled 18F-fluorodeoxyglucose were placed within the physical phantom thorax as biopsy targets. We measured the average distance between target center and the simulated biopsy location among multiple trials to evaluate the biopsy localization accuracy. The computer simulation results showed that the RASP method generated PET images with a significantly reduced noise of 0.10 ± 0.01 standardized uptake value (SUV) as compared to an end-of-expiration image noise of 0.34 ± 0.04 SUV. The respiratory motion increased the apparent liver lesion size from 5.4 ± 1.1 to 35.3 ± 3.0cc. The RASP algorithm reduced this to 15.7 ± 3.7cc. The distances between the centroids for the static image lesion and two moving lesions in the liver and lung, when reconstructed with the RASP algorithm, were 0.83 ± 0.72mm and 0.42 ± 0.72mm. For the ungated imaging, these values increased to 3.48 ± 1.45mm and 2.5 ± 0.12mm, respectively. For the ungated imaging, this increased to 1.99 ± 1.72mm. In addition, the lesion activity estimation (e.g., SUV) was accurate and constant for images reconstructed using the RASP algorithm, whereas large activity bias and variations (± 50%) were observed for lesions in the ungated images. The physical phantom studies demonstrated a biopsy needle localization error of 2.9 ± 0.9mm from CT. Combined with the localization errors due to respiration for the PET images from simulations, the overall estimated lesion localization error would be 3.08mm for PET-guided biopsies images using RASP and 3.64mm when using ungated PET images. In other words, RASP reduced the localization error by approximately 0.6mm. The combined error analysis showed that replacing the standard end-of-expiration images with the proposed RASP method in PET/CT-guided biopsy workflow yields comparable lesion localization accuracy and reduced image noise. The RASP method can produce PET images with reduced noise, attenuation artifacts and respiratory motion, resulting in more accurate lesion localization. Testing the PET/CT-guided biopsy workflow using computer simulation and physical phantoms with respiratory motion, we demonstrated that guided biopsy procedure with the RASP method can benefit from improved PET image quality due to noise reduction, without compromising the accuracy of lesion localization.

One of the challenges in PET/MRI is the derivation of an attenuation map to correct the PET image for attenuation. Different methods have been suggested for deriving the attenuation map from an MR image. Because the low signal intensity of cortical bone on images acquired with conventional MRI sequences makes it difficult to detect this tissue type, these methods rely on some sort of anatomic precondition to predict the attenuation map, raising the question of whether these methods will be usable in the clinic when patients may exhibit anatomic abnormalities. We propose the use of the transverse relaxation rate, derived from images acquired with an ultrashort echo time sequence to classify the voxels into 1 of 3 tissue classes (bone, soft tissue, or air), without making any assumptions on patient anatomy. Each voxel is assigned a linear attenuation coefficient corresponding to its tissue class. A reference CT scan is used to determine the voxel-by-voxel accuracy of the proposed method. The overall accuracy of the MRI-based attenuation correction is evaluated using a method that takes into account the nonlocal effects of attenuation correction. As a proof of concept, the head of a pig was used as a phantom for imaging. The new method yielded a correct tissue classification in 90% of the voxels. Five human brain PET/CT and MRI datasets were also processed, yielding slightly worse voxel-by-voxel performance, compared to a CT-derived attenuation map. The PET datasets were reconstructed using the segmented MRI attenuation map derived with the new method, and the resulting images were compared with segmented CT-based attenuation correction. An average error of around 5% was found in the brain. The feasibility of using the transverse relaxation rate map derived from ultrashort echo time MR images for the estimation of the attenuation map was shown on phantom and clinical brain data. The results indicate that the new method, compared with CT-based attenuation correction, yields clinically acceptable errors. The proposed method does not make any assumptions about patient anatomy and could therefore also be used in cases in which anatomic abnormalities are present.

A critical question in the development of combined MR/PET scanners is whether MR can provide the tissue attenuation data required for PET reconstruction. Unfortunately, MR images are often unable to encompass the entire patient. The resulting truncation in the transverse plane leads to incomplete attenuation maps, causing artifacts in the reconstructed PET image. This article describes the experiments performed to quantify these artifacts. A method to compensate the missing data was evaluated to determine whether software correction is possible or whether additional transmission hardware has to be included in the scanner. Three studies were made. First, simulated PET data were used to quantify the bias due to an incomplete attenuation map. A set of spherical lesions was simulated in the lungs and mediastinum of a patient. The data were reconstructed with complete and partial attenuation maps and the uptake differences were evaluated. Second, clinical data from PET/CT oncology patients were used. To reproduce the expected conditions in an MR/PET scanner, only patients scanned with the arms resting along the body were considered. These scans were then used to create maps of the reconstruction bias due to field of view (FOV) limitations. Lastly, a PET reconstruction with incomplete attenuation data was evaluated as a means to obtain attenuation information beyond the MR FOV. The patient outline was automatically segmented with a three-dimensional snake algorithm and used to fill the truncated data in the attenuation map. Average bias up to 15% and local biases up to 50% were estimated when PET data were reconstructed with incomplete attenuation information. Completing the attenuation map with data extracted from a PET prereconstruction globally reduced these biases to below 10%. This correction proved to be tolerant to inaccuracies in positioning and attenuation values. However, local artifacts up to 20% could still be found near the edges of the MR FOV. MR FOV restrictions can indeed make the reconstructed PET data unacceptable for diagnostic purposes. Biases can be globally compensated by automatic preprocessing of the attenuation map. However, inaccuracies in the correction will result in small artifacts near the periphery of the image that could lead to false-positive findings.

Background and purpose:Patients with unruptured intracranial aneurysms (UIAs) routinely undergo surveillance imaging to monitor growth. Angiography is the gold standard for aneurysm diagnosis, but is invasive. This study aims to evaluate the accuracy and reproducibility of a 3D non-contrast black-blood (BB) MR technique for UIA measurement, in comparison with 3D TOF- and CE-MRA, using 3D rotational angiography (3DRA) as a reference standard.Methods:64 patients (age 57.3±10.9 years, 41 female) with 68 saccular UIAs were recruited. Patients underwent 3T MRI with 3D TOF-MRA, 3D BB MRI and CE-MRA, and patients had a 3DRA within 2 weeks. The neck, width, height of the UIAs were measured by two radiologists independently on 3DRA and three MRI sequences. The accuracy and reproducibility was evaluated by Bland-Altman plots, coefficient of variance (CV), and intra-class coefficient (ICC).Results:3D BB MRI demonstrates the best agreement with DSA, with the smallest limit of agreement (LOA) and measurement error (CVs range 5.87% to 7.04%). 3D TOF-MRA had the largest LOA and measurement error (CVs range 12.73% to 15.78%). The average CV was 6.26% for 3D BB MR, 7.03% for CE-MRA and 15.54% for TOF-MRA. No bias was found among three MRI sequences compared with 3DRA. All three MRI sequences had excellent inter-reader agreement (ICCs>0.95). 3D BB MRI performed the best for patients with intraluminal thrombus (n=10).Conclusions:3D BB MRI achieves better accuracy for aneurysm size measurements compared to 3D TOF, using 3DRA as a gold standard. This non-contrast technique is promising for surveillance of UIA.

PurposeTo develop a framework for efficient and simultaneous acquisition of motion‐compensated whole‐heart coronary MR angiography (CMRA) and left ventricular function by MR and myocardial integrity by PET on a 3T PET‐MR system.MethodsAn acquisition scheme based on a dual‐phase CMRA sequence acquired simultaneously with cardiac PET data has been developed. The framework is integrated with a motion‐corrected image reconstruction approach, so that non‐rigid respiratory and cardiac deformation fields estimated from MR images are used to correct both the CMRA (respiratory motion correction for each cardiac phase) and the PET data (respiratory and cardiac motion correction). The proposed approach was tested in a cohort of 8 healthy subjects and 6 patients with coronary artery disease. Left ventricular (LV) function estimated from motion‐corrected dual‐phase CMRA was compared to the gold standard estimated from a stack of 2D CINE images for the healthy subjects. Relative increase of signal in motion‐corrected PET images compared to uncorrected images was computed for standard 17‐segment polar maps for each patient.ResultsMotion‐corrected dual‐phase CMRA images allow for visualization of the coronary arteries in both systole and diastole for all healthy subjects and cardiac patients. LV functional indices from healthy subjects result in good agreement with the reference method, underestimating stroke volume by 3.07 ± 3.26 mL and ejection fraction by 0.30 ± 1.01%. Motion correction improved delineation of the myocardium in PET images, resulting in an increased 18F‐FDG signal of up to 28% in basal segments of the myocardial wall compared to uncorrected images.ConclusionThe proposed motion‐corrected dual‐phase CMRA and cardiac PET produces co‐registered good quality images in both modalities in a single efficient examination of ~13 min.

Beat-to-beat respiratory motion correction with near 100% efficiency: a quantitative assessment using high-resolution coronary artery imaging

Summary To investigate the performance of a novel algorithm for correcting respiratory-induced heart motion for wholeheart coronary MRI. Background A right hemi-diaphragm (RHD) respiratory navigator is commonly used to suppress the respiratory motion of the heart in coronary MRI [1]. Typically, a small 5mm end expiratory gating window (GW) is used to gate data. While this technique successfully suppresses respiratory motion, it prolongs scan time due to navigator efficacies of only 30-50%. Increasing the GW to 15mm would increase the navigator efficiency. In this study, we present a novel respiratory motion correction algorithm which allows increasing the GW to 15mm. Methods Whole-heart coronary MRI with isotropic 1.3mm 3 resolution was acquired using partial Fourier and a navigator with 15mm GW on a 1.5T Philips CMR scanner from 10 healthy subjects (4 males; 29 ± 13 yr). The navigator RHD positions were used to sort the acquired k-space segments into 15 separate 1mm bins (based on their displacement). To compensate for the respiratory motion of the heart, all k-space segments acquired in each bin were assigned a 3D translation parameter. The 3D translation parameter was estimated using an iterative gradient descent optimization algorithm to correct the kspace segments such that the sharpness of the image, reconstructed using the corrected k-space segments, is maximized. The variance of the gradient of the image was used as the measure for the image sharpness [2]. For comparison, another whole-heart coronary MRI dataset was acquired using the same sequence and a navigator with a 5mm GW.