Project description:Background Deep convolutional neural networks have demonstrated robust and reliable PET attenuation correction (AC) as an alternative to conventional AC methods in integrated PET/MRI systems. However, its whole-body implementation is still challenging due to anatomical variations and the limited MRI field of view. The aim of this study is to investigate a deep learning (DL) method to generate voxel-based synthetic CT (sCT) from Dixon MRI and use it as a whole-body solution for PET AC in a PET/MRI system. Materials and methods Fifteen patients underwent PET/CT followed by PET/MRI with whole-body coverage from skull to feet. We performed MRI truncation correction and employed co-registered MRI and CT images for training and leave-one-out cross-validation. The network was pretrained with region-specific images. The accuracy of the AC maps and reconstructed PET images were assessed by performing a voxel-wise analysis and calculating the quantification error in SUV obtained using DL-based sCT (PETsCT) and a vendor-provided atlas-based method (PETAtlas), with the CT-based reconstruction (PETCT) serving as the reference. In addition, region-specific analysis was performed to compare the performances of the methods in brain, lung, liver, spine, pelvic bone, and aorta. Results Our DL-based method resulted in better estimates of AC maps with a mean absolute error of 62 HU, compared to 109 HU for the atlas-based method. We found an excellent voxel-by-voxel correlation between PETCT and PETsCT (R2 = 0.98). The absolute percentage difference in PET quantification for the entire image was 6.1% for PETsCT and 11.2% for PETAtlas. The regional analysis showed that the average errors and the variability for PETsCT were lower than PETAtlas in all regions. The largest errors were observed in the lung, while the smallest biases were observed in the brain and liver. Conclusions Experimental results demonstrated that a DL approach for whole-body PET AC in PET/MRI is feasible and allows for more accurate results compared with conventional methods. Further evaluation using a larger training cohort is required for more accurate and robust performance. Supplementary Information The online version contains supplementary material available at 10.1186/s40658-022-00486-8.

Project description:Respiratory motion causes misalignments between positron emission tomography (PET) and magnetic resonance (MR)-derived attenuation maps (µ-maps) in addition to artifacts on both PET and MR images in simultaneous PET/MRI for organs such as liver that can experience motion of several centimeters. To address this problem, we developed an efficient MR-based attenuation correction (MRAC) method to generate phase-matched µ-maps for quiescent period PET (PETQ) in abdominal PET/MRI. MRAC data was acquired with CIRcular Cartesian UnderSampling (CIRCUS) sampling during 100?s in free-breathing as an accelerated data acquisition strategy for phase-matched MRAC (MRACPM-CIRCUS). For comparison, MRAC data with raster (Default) k-space sampling was also acquired during 100?s in free-breathing (MRACPM-Default), and used to evaluate MRACPM-CIRCUS as well as un-matched MRAC (MRACUM) that was un-gated. We purposefully oversampled the MRACPM data to ensure we had enough information to capture all respiratory phases to make this comparison as robust as possible. The proposed MRACPM-CIRCUS was evaluated in 17 patients with 68Ga-DOTA-TOC PET/MRI exams, suspected of having neuroendocrine tumors or liver metastases. Effects of CIRCUS sampling for accelerating a data acquisition were evaluated by simulating the data acquisition time retrospectively in increments of 5?s. Effects of MRACPM-CIRCUS on PETQ were evaluated using uptake differences in the liver lesions (n??=??35), compared to PETQ with MRACPM-Default and MRACUM. A Wilcoxon signed-rank test was performed to compare lesion uptakes between the MRAC methods. MRACPM-CIRCUS showed higher image quality compared to MRACPM-Default for the same acquisition times, demonstrating that a data acquisition time of 30?s was reasonable to achieve phase-matched µ-maps. Lesion update differences between MRACPM-CIRCUS (30?s) versus MRACPM-Default (reference, 100?s) were 0.1%??±??1.4% (range of??-2.7% to 3.2%) and not significant (P??>??.05); while, the differences between MRACUM versus MRACPM-Default were 0.6%??±??11.4% with a large variation (range of??-37% to 20%) and significant (P??<??.05). In conclusion, we demonstrated that a data acquisition of 30?s achieved phase-matched µ-maps when using specialized CIRCUS data sampling and phase-matched µ-maps improved PETQ quantification significantly.

Project description:BackgroundThis study aims to compare proton density weighted magnetic resonance imaging (MRI) zero echo time (ZTE) and head atlas attenuation correction (AC) to the reference standard computed tomography (CT) based AC for 11C-methionine positron emission tomography (PET)/MRI.MethodsA retrospective cohort of 14 patients with suspected or confirmed brain tumour and 11C-Methionine PET/MRI was included in the study. For each scan, three AC maps were generated: ZTE-AC, atlas-AC and reference standard CT-AC. Maximum and mean standardised uptake values (SUV) were measured in the hotspot, mirror region and frontal cortex. In postoperative patients (n?=?8), SUV values were additionally obtained adjacent to the metal implant and mirror region. Standardised uptake ratios (SUR) hotspot/mirror, hotspot/cortex and metal/mirror were then calculated and analysed with Bland-Altman, Pearson correlation and intraclass correlation reliability in the overall group and subgroups.ResultsZTE-AC demonstrated narrower SD and 95% CI (Bland-Altman) than atlas-AC in the hotspot analysis for all groups (ZTE overall???2.84, - 1.41 to 1.70; metal???1.67, - 3.00 to 2.20; non-metal???3.04, - 0.96 to 3.38; Atlas overall???4.56, - 1.05 to 3.83; metal???3.87, - 3.81 to 4.64; non-metal???4.90, - 1.68 to 5.86). The mean bias for both ZTE-AC and atlas-AC was???2.4% compared to CT-AC. In the metal region analysis, ZTE-AC demonstrated a narrower mean bias range-closer to zero-and narrower SD and 95% CI (ZTE 0.21-0.48,???2.50, - 1.70 to 2.57; Atlas 0.56-1.54,???4.01, - 1.81 to 4.89). The mean bias for both ZTE-AC and atlas-AC was within 1.6%. A perfect correlation (Pearson correlation) was found for both ZTE-AC and atlas-AC compared to CT-AC in the hotspot and metal analysis (ZTE ? 1.00, p?<?0.0001; atlas ? 1.00, p?<?0.0001). An almost perfect intraclass correlation coefficient for absolute agreement was found between Atlas-, ZTE and CT maps for maxSUR and meanSUR values in all the analyses (ICC?>?0.99).ConclusionsBoth ZTE and atlas-AC showed a good performance against CT-AC in patients with brain tumour.

Project description:PurposeAttenuation correction and scatter compensation (AC/SC) are two main steps toward quantitative PET imaging, which remain challenging in PET-only and PET/MRI systems. These can be effectively tackled via deep learning (DL) methods. However, trustworthy, and generalizable DL models commonly require well-curated, heterogeneous, and large datasets from multiple clinical centers. At the same time, owing to legal/ethical issues and privacy concerns, forming a large collective, centralized dataset poses significant challenges. In this work, we aimed to develop a DL-based model in a multicenter setting without direct sharing of data using federated learning (FL) for AC/SC of PET images.MethodsNon-attenuation/scatter corrected and CT-based attenuation/scatter corrected (CT-ASC) 18F-FDG PET images of 300 patients were enrolled in this study. The dataset consisted of 6 different centers, each with 50 patients, with scanner, image acquisition, and reconstruction protocols varying across the centers. CT-based ASC PET images served as the standard reference. All images were reviewed to include high-quality and artifact-free PET images. Both corrected and uncorrected PET images were converted to standardized uptake values (SUVs). We used a modified nested U-Net utilizing residual U-block in a U-shape architecture. We evaluated two FL models, namely sequential (FL-SQ) and parallel (FL-PL) and compared their performance with the baseline centralized (CZ) learning model wherein the data were pooled to one server, as well as center-based (CB) models where for each center the model was built and evaluated separately. Data from each center were divided to contribute to training (30 patients), validation (10 patients), and test sets (10 patients). Final evaluations and reports were performed on 60 patients (10 patients from each center).ResultsIn terms of percent SUV absolute relative error (ARE%), both FL-SQ (CI:12.21-14.81%) and FL-PL (CI:11.82-13.84%) models demonstrated excellent agreement with the centralized framework (CI:10.32-12.00%), while FL-based algorithms improved model performance by over 11% compared to CB training strategy (CI: 22.34-26.10%). Furthermore, the Mann-Whitney test between different strategies revealed no significant differences between CZ and FL-based algorithms (p-value > 0.05) in center-categorized mode. At the same time, a significant difference was observed between the different training approaches on the overall dataset (p-value < 0.05). In addition, voxel-wise comparison, with respect to reference CT-ASC, exhibited similar performance for images predicted by CZ (R2 = 0.94), FL-SQ (R2 = 0.93), and FL-PL (R2 = 0.92), while CB model achieved a far lower coefficient of determination (R2 = 0.74). Despite the strong correlations between CZ and FL-based methods compared to reference CT-ASC, a slight underestimation of predicted voxel values was observed.ConclusionDeep learning-based models provide promising results toward quantitative PET image reconstruction. Specifically, we developed two FL models and compared their performance with center-based and centralized models. The proposed FL-based models achieved higher performance compared to center-based models, comparable with centralized models. Our work provided strong empirical evidence that the FL framework can fully benefit from the generalizability and robustness of DL models used for AC/SC in PET, while obviating the need for the direct sharing of datasets between clinical imaging centers.

Project description:Several factors have to be considered for implementing an accurate attenuation-correction (AC) method in a combined MR-PET scanner. In this work, some of these challenges were investigated, and an AC method based entirely on the MRI data obtained with a single dedicated sequence was developed and used for neurologic studies performed with the MR-PET human brain scanner prototype.The focus was on the problem of bone-air segmentation, selection of the linear attenuation coefficient for bone, and positioning of the radiofrequency coil. The impact of these factors on PET data quantification was studied in simulations and experimental measurements performed on the combined MR-PET scanner. A novel dual-echo ultrashort echo time (DUTE) MRI sequence was proposed for head imaging. Simultaneous MR-PET data were acquired, and the PET images reconstructed using the proposed DUTE MRI-based AC method were compared with the PET images that had been reconstructed using a CT-based AC method.Our data suggest that incorrectly accounting for the bone tissue attenuation can lead to large underestimations (>20%) of the radiotracer concentration in the cortex. Assigning a linear attenuation coefficient of 0.143 or 0.151 cm(-1) to bone tissue appears to give the best trade-off between bias and variability in the resulting images. Not identifying the internal air cavities introduces large overestimations (>20%) in adjacent structures. On the basis of these results, the segmented CT AC method was established as the silver standard for the segmented MRI-based AC method. For an integrated MR-PET scanner, in particular, ignoring the radiofrequency coil attenuation can cause large underestimations (i.e., <or=50%) in the reconstructed images. Furthermore, the coil location in the PET field of view has to be accurately known. High-quality bone-air segmentation can be performed using the DUTE data. The PET images obtained using the DUTE MRI- and CT-based AC methods compare favorably in most of the brain structures.A DUTE MRI-based AC method considering all these factors was implemented. Preliminary results suggest that this method could potentially be as accurate as the segmented CT method and could be used for quantitative neurologic MR-PET studies.

Project description:This study introduces a new hybrid ZTE/Dixon MR-based attenuation correction (MRAC) method including bone density estimation for PET/MRI and quantifies the effects of bone attenuation on metastatic lesion uptake in the pelvis.Six patients with pelvic lesions were scanned using fluorodeoxyglucose (18F-FDG) in an integrated time-of-flight (TOF) PET/MRI system. For PET attenuation correction, MR imaging consisted of two-point Dixon and zero echo-time (ZTE) pulse sequences. A continuous-value fat and water pseudoCT was generated from a two-point Dixon MRI. Bone was segmented from the ZTE images and converted to Hounsfield units (HU) using a continuous two-segment piecewise linear model based on ZTE MRI intensity. The HU values were converted to linear attenuation coefficients (LAC) using a bilinear model. The bone voxels of the Dixon-based pseudoCT were replaced by the ZTE-derived bone to produce the hybrid ZTE/Dixon pseudoCT. The three different AC maps (Dixon, hybrid ZTE/Dixon, CTAC) were used to reconstruct PET images using a TOF-ordered subset expectation maximization algorithm with a point-spread function model. Metastatic lesions were separated into two classes, bone lesions and soft tissue lesions, and analyzed. The MRAC methods were compared using a root-mean-squared error (RMSE), where the registered CTAC was taken as ground truth.The RMSE of the maximum standardized uptake values (SUVmax ) is 11.02% and 7.79% for bone (N = 6) and soft tissue lesions (N = 8), respectively, using Dixon MRAC. The RMSE of SUVmax for these lesions is significantly reduced to 3.28% and 3.94% when using the new hybrid ZTE/Dixon MRAC. Additionally, the RMSE for PET SUVs across the entire pelvis and all patients are 8.76% and 4.18%, for the Dixon and hybrid ZTE/Dixon MRAC methods, respectively.A hybrid ZTE/Dixon MRAC method was developed and applied to pelvic regions in an integrated TOF PET/MRI, demonstrating improved MRAC. This new method included bone density estimation, through which PET quantification is improved.

Dataset Information

A multi-centre evaluation of eleven clinically feasible brain PET/MRI attenuation correction techniques using a large cohort of patients.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets