Accurate quantification of positron emission tomography (PET) uptake depends on accurate attenuation correction in reconstruction. Current magnetic resonance-based attenuation correction methods (MRAC) for body PET imaging use a fat/water map derived from a two-echo Dixon magnetic resonance imaging (MRI) sequence, where bone is neglected. Ultrashort echo-time and zero echo-time (ZTE) pulse sequences can capture bone information. We propose the use of patient-specific multi-parametric MRI consisting of Dixon MRI and proton-density-weighted ZTE MRI to directly synthesize pseudoCT images with the use of a deep learning model: we name this method Zero echo-time and Dixon Deep pseudoCT (ZeDD-CT). Methods: Twenty-six patients were scanned using an integrated 3 Tesla time-of-flight PET/MRI system. Helical x-ray computed tomography (CT) images of the patients were acquired separately. A deep convolutional neural network was trained to transform ZTE and Dixon MRI into synthetic CT images (ZeDD-CT). Ten patients were used for model training and sixteen patients were used for evaluation. Bone and soft tissue lesions were identified and the SUVmax was measured. The root-mean-squared-error (RMSE) was used to compare the MRAC methods with the ground-truth CTAC. Results: A total of 30 bone lesions and 60 soft tissue lesions were evaluated. For bone lesions, there was a factor of 4 reduction of RMSE in PET quantification (RMSE were 10.24% for Dixon PET, and 2.68% for ZeDD PET); for soft tissue lesions, there was a factor of 1.5 reduction of RMSE (RMSE were 6.24% for Dixon PET, and 4.07% for ZeDD PET). Conclusion: The ZeDD-CT produces natural-looking and quantitatively accurate pseudoCT images and reduces error in pelvis PET/MRI attenuation correction compared to standard methods.
from #MedicinebyAlexandrosSfakianakis via xlomafota13 on Inoreader http://ift.tt/2yYOJHt
via IFTTT
Δεν υπάρχουν σχόλια:
Δημοσίευση σχολίου