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

• Published in: IEEE transactions on radiation and plasma medical sciences Vol. 6; no. 6; p. 1
• Main Authors: Leynes, Andrew P., Ahn, Sangtae, Wangerin, Kristen A., Kaushik, Sandeep S., Wiesinger, Florian, Hope, Thomas A., Larson, Peder E. Z.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.07.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Anatomy; Artificial neural networks; Attenuation; Attenuation coefficients; Bayesian analysis; Bayesian deep learning; Bone density; Computed tomography; Crosstalk; Deep learning; Implants; Machine learning; Magnetic resonance imaging; Maximum likelihood estimation; Metals; MLAA; MRAC; Neural networks; Positron emission; Positron emission tomography; Resonance; Root-mean-square errors; synthetic CT; Tomography; Transplants & implants; Uncertainty; deep learning; synthetic CT; Bayesian deep learning; magnetic resonance-based attenuation correction (MRAC); maximum-likelihood estimation of activity and attenuation (MLAA)
• ISSN: 2469-7311; 2469-7303
• DOI: 10.1109/TRPMS.2021.3118325

A major remaining challenge for magnetic resonance-based 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 (UpCT-MLAA), 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. UpCT-MLAA estimated attenuation coefficients of metal implants alongside accurate anatomic depiction outside of implant regions.