Attenuation Coefficient Estimation for PET/MRI With Bayesian Deep Learning Pseudo-CT and Maximum-Likelihood Estimation of Activity and Attenuation

A major remaining challenge for magnetic resonancebased attenuation correction methods (MRAC) is their susceptibility to sources of MRI artifacts (e.g. implants, motion) and uncertainties due to the limitations of MRI contrast (e.g. accurate bone delineation and density, and separation of air/bone). We propose using a Bayesian deep convolutional neural network that, in addition to generating an initial pseudo-CT from MR data, also produces uncertainty estimates of the pseudo-CT to quantify the limitations of the MR data. These outputs are combined with MLAA reconstruction that uses the PET emission data to improve the attenuation maps. With the proposed approach (UpCTMLAA), we demonstrate accurate estimation of PET uptake in pelvic lesions and show recovery of metal implants. In patients without implants, UpCT-MLAA had acceptable but slightly higher RMSE than Zero-echo-time and Dixon Deep pseudo-CT when compared to CTAC. In patients with metal implants, MLAA recovered the metal implant; however, anatomy outside the implant region was obscured by noise and crosstalk artifacts. Attenuation coefficients from the pseudo-CT from Dixon MRI were accurate in normal anatomy; however, the metal implant region was estimated to have attenuation coefficients of air. UpCTMLAA estimated attenuation coefficients of metal implants alongside accurate anatomic depiction outside of implant regions.

[1]  A. McMillan,et al.  Deep learning Mr imaging–based attenuation correction for PeT/Mr imaging 1 , 2017 .

[2]  Jae Sung Lee,et al.  Improving the Accuracy of Simultaneously Reconstructed Activity and Attenuation Maps Using Deep Learning , 2018, The Journal of Nuclear Medicine.

[3]  Keith A. Johnson,et al.  MR-Based PET Attenuation Correction using a Combined Ultrashort Echo Time/Multi-Echo Dixon Acquisition. , 2020, Medical physics.

[4]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[5]  Peder E. Z. Larson,et al.  Synthetic CT Generation Using MRI with Deep Learning: How Does the Selection of Input Images Affect the Resulting Synthetic CT? , 2018, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[6]  Meher R. Juttukonda,et al.  Deep learning‐based T1‐enhanced selection of linear attenuation coefficients (DL‐TESLA) for PET/MR attenuation correction in dementia neuroimaging , 2021, Magnetic resonance in medicine.

[7]  Jae Sung Lee A Review of Deep-Learning-Based Approaches for Attenuation Correction in Positron Emission Tomography , 2021, IEEE Transactions on Radiation and Plasma Medical Sciences.

[8]  Josef A. Lundman,et al.  Zero TE‐based pseudo‐CT image conversion in the head and its application in PET/MR attenuation correction and MR‐guided radiation therapy planning , 2018, Magnetic resonance in medicine.

[9]  A. Soricelli,et al.  Dixon-VIBE Deep Learning (DIVIDE) Pseudo-CT Synthesis for Pelvis PET/MR Attenuation Correction , 2018, The Journal of Nuclear Medicine.

[10]  G. Delso,et al.  PET–MR imaging using a tri-modality PET/CT–MR system with a dedicated shuttle in clinical routine , 2013, Magnetic Resonance Materials in Physics, Biology and Medicine.

[11]  Seongho Seo,et al.  Accurate Transmission-Less Attenuation Correction Method for Amyloid-β Brain PET Using Deep Neural Network , 2021, Electronics.

[12]  Ciprian Catana,et al.  PET/MRI in the Presence of Metal Implants: Completion of the Attenuation Map from PET Emission Data , 2017, The Journal of Nuclear Medicine.

[13]  Ciprian Catana,et al.  Transmission imaging for integrated PET-MR systems , 2016, Physics in medicine and biology.

[14]  Craig S. Levin,et al.  Pseudo CT Image Synthesis and Bone Segmentation From MR Images Using Adversarial Networks With Residual Blocks for MR-Based Attenuation Correction of Brain PET Data , 2021, IEEE Transactions on Radiation and Plasma Medical Sciences.

[15]  A. Buck,et al.  Clinical Evaluation of Zero-Echo-Time Attenuation Correction for Brain 18F-FDG PET/MRI: Comparison with Atlas Attenuation Correction , 2016, The Journal of Nuclear Medicine.

[16]  Jae Sung Lee,et al.  Generation of PET Attenuation Map for Whole-Body Time-of-Flight 18F-FDG PET/MRI Using a Deep Neural Network Trained with Simultaneously Reconstructed Activity and Attenuation Maps , 2019, The Journal of Nuclear Medicine.

[17]  Andrew P. Leynes,et al.  Zero-Echo-Time and Dixon Deep Pseudo-CT (ZeDD CT): Direct Generation of Pseudo-CT Images for Pelvic PET/MRI Attenuation Correction Using Deep Convolutional Neural Networks with Multiparametric MRI , 2017, The Journal of Nuclear Medicine.

[18]  Steven G. Ross,et al.  Application and Evaluation of a Measured Spatially Variant System Model for PET Image Reconstruction , 2010, IEEE Transactions on Medical Imaging.

[19]  Christine DeLorenzo,et al.  Synthesis of Patient-Specific Transmission Data for PET Attenuation Correction for PET/MRI Neuroimaging Using a Convolutional Neural Network , 2018, The Journal of Nuclear Medicine.

[20]  S. Vandenberghe,et al.  MRI-Based Attenuation Correction for PET/MRI Using Ultrashort Echo Time Sequences , 2010, Journal of Nuclear Medicine.

[21]  Francesco C Stingo,et al.  An evaluation of three commercially available metal artifact reduction methods for CT imaging , 2015, Physics in medicine and biology.

[22]  Ilja Bezrukov,et al.  MRI-Based Attenuation Correction for Whole-Body PET/MRI: Quantitative Evaluation of Segmentation- and Atlas-Based Methods , 2011, The Journal of Nuclear Medicine.

[23]  Willem Waegeman,et al.  Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methods , 2019, Machine Learning.

[24]  Florian Wiesinger,et al.  Joint estimation of activity and attenuation for PET using pragmatic MR-based prior: application to clinical TOF PET/MR whole-body data for FDG and non-FDG tracers , 2018, Physics in medicine and biology.

[25]  F. Wiesinger,et al.  Evaluation of Sinus/Edge-Corrected Zero-Echo-Time–Based Attenuation Correction in Brain PET/MRI , 2017, The Journal of Nuclear Medicine.

[26]  Habib Zaidi,et al.  Joint Estimation of Activity and Attenuation in Whole-Body TOF PET/MRI Using Constrained Gaussian Mixture Models , 2015, IEEE Transactions on Medical Imaging.

[27]  Alex Kendall,et al.  What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? , 2017, NIPS.

[28]  Zoubin Ghahramani,et al.  Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , 2015, ICML.

[29]  G. Hermosillo,et al.  Whole-Body PET/MR Imaging: Quantitative Evaluation of a Novel Model-Based MR Attenuation Correction Method Including Bone , 2015, The Journal of Nuclear Medicine.

[30]  Yasheng Chen,et al.  MR-based attenuation correction for PET/MRI neurological studies with continuous-valued attenuation coefficients for bone through a conversion from R2* to CT-Hounsfield units , 2015, NeuroImage.

[31]  Patrick Dupont,et al.  Simultaneous maximum a posteriori reconstruction of attenuation and activity distributions from emission sinograms , 1999, IEEE Transactions on Medical Imaging.

[32]  Stefan Förster,et al.  MR-Based Attenuation Correction Using Ultrashort-Echo-Time Pulse Sequences in Dementia Patients , 2015, The Journal of Nuclear Medicine.

[33]  Liselotte Højgaard,et al.  AI-driven attenuation correction for brain PET/MRI: Clinical evaluation of a dementia cohort and importance of the training group size , 2020, NeuroImage.

[34]  Jian Sun,et al.  Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[35]  A. Buck,et al.  Evaluation of Atlas-Based Attenuation Correction for Integrated PET/MR in Human Brain: Application of a Head Atlas and Comparison to True CT-Based Attenuation Correction , 2016, The Journal of Nuclear Medicine.

[36]  H. Quick,et al.  Magnetic Resonance–Based Attenuation Correction for PET/MR Hybrid Imaging Using Continuous Valued Attenuation Maps , 2013, Investigative radiology.

[37]  Young Lee,et al.  Deep Generative Model for Synthetic-CT Generation with Uncertainty Predictions , 2020, MICCAI.

[38]  Yusheng Li,et al.  Attenuation correction in emission tomography using the emission data--A review. , 2016, Medical physics.

[39]  G. Delso,et al.  Evaluation of an Atlas-Based PET Head Attenuation Correction Using PET/CT & MR Patient Data , 2012, IEEE Transactions on Nuclear Science.

[40]  Gaspar Delso,et al.  Design Features and Mutual Compatibility Studies of the Time-of-Flight PET Capable GE SIGNA PET/MR System , 2016, IEEE Transactions on Medical Imaging.

[41]  Seongho Seo,et al.  Data-driven respiratory phase-matched PET attenuation correction without CT , 2021, Physics in medicine and biology.

[42]  S. Holm,et al.  Region specific optimization of continuous linear attenuation coefficients based on UTE (RESOLUTE): application to PET/MR brain imaging , 2015, Physics in medicine and biology.

[43]  Luca Antiga,et al.  Automatic differentiation in PyTorch , 2017 .

[44]  Thomas Beyer,et al.  X-ray-based attenuation correction for positron emission tomography/computed tomography scanners. , 2003, Seminars in nuclear medicine.

[45]  Quanzheng Li,et al.  MR-Based Attenuation Correction for Brain PET Using 3-D Cycle-Consistent Adversarial Network , 2021, IEEE Transactions on Radiation and Plasma Medical Sciences.

[46]  I. Burger,et al.  PET/MR imaging of bone lesions – implications for PET quantification from imperfect attenuation correction , 2012, European Journal of Nuclear Medicine and Molecular Imaging.

[47]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[48]  Harini Veeraraghavan,et al.  Patch-Based Generative Adversarial Neural Network Models for Head and Neck MR-Only Planning. , 2019, Medical physics.

[49]  Habib Zaidi,et al.  One registration multi-atlas-based pseudo-CT generation for attenuation correction in PET/MRI , 2016, European Journal of Nuclear Medicine and Molecular Imaging.

[50]  A. Rahmim,et al.  Deep-JASC: joint attenuation and scatter correction in whole-body 18F-FDG PET using a deep residual network , 2020, European Journal of Nuclear Medicine and Molecular Imaging.

[51]  Y. Ouchi,et al.  Deep learning-based attenuation correction for brain PET with various radiotracers , 2021, Annals of Nuclear Medicine.

[52]  G. Hermosillo,et al.  Dixon Sequence with Superimposed Model-Based Bone Compartment Provides Highly Accurate PET/MR Attenuation Correction of the Brain , 2016, The Journal of Nuclear Medicine.

[53]  J. Gee,et al.  The Insight ToolKit image registration framework , 2014, Front. Neuroinform..

[54]  Habib Zaidi,et al.  Deep learning-guided estimation of attenuation correction factors from time-of-flight PET emission data , 2020, Medical Image Anal..

[55]  S. Thiruvenkadam,et al.  Comparison of 4-class and continuous fat/water methods for whole-body, MR-based PET attenuation correction , 2012, 2012 IEEE Nuclear Science Symposium and Medical Imaging Conference Record (NSS/MIC).

[56]  Andrew P. Leynes,et al.  Hybrid ZTE/Dixon MR‐based attenuation correction for quantitative uptake estimation of pelvic lesions in PET/MRI , 2017, Medical physics.