MRI-Based Attenuation Correction for Whole-Body PET/MRI: Quantitative Evaluation of Segmentation- and Atlas-Based Methods

Abstract

PET/MRI is an emerging dual-modality imaging technology that requires new approaches to PET attenuation correction (AC). We assessed 2 algorithms for whole-body MRI-based AC (MRAC): a basic MR image segmentation algorithm and a method based on atlas registration and pattern recognition (AT&PR). Methods: Eleven patients each underwent a whole-body PET/ CT study and a separate multibed whole-body MRI study. The MR image segmentation algorithm uses a combination of image thresholds, Dixon fat-water segmentation, and component analysis to detect the lungs. MR images are segmented into 5 tissue classes (not including bone), and each class is assigned a default linear attenuation value. The AT&PR algorithm uses a database of previously aligned pairs of MRI/CT image volumes. For each patient, these pairs are registered to the patient MRI volume, and machine-learning techniques are used to predict attenuation values on a continuous scale. MRAC methods are compared via the quantitative analysis of AC PET images using volumes of interest in normal organs and on lesions. We assume the PET/CT values after CT-based AC to be the reference standard. Results: In regions of normal physiologic uptake, the average error of the mean standardized uptake value was 14.1% 6 10.2% and 7.7% 6 8.4% for the segmentation and the AT&PR methods, respectively. Lesion-based errors were 7.5% 6 7.9% for the segmentation method and 5.7% 6 4.7% for the AT&PR method. Conclusion: The MRAC method using AT&PR provided better overall PET quantification accuracy than the basic MR image segmentation approach. This better quantification was due to the significantly reduced volume of errors made regarding volumes of interest within or near bones and the slightly reduced volume of errors made regarding areas outside the lungs.

Comparison of atlas-based techniques for whole-body bone segmentation

“…Since bone tissue generates a void signal when using common MR sequences, it is indistinguishable from air. As such, bony structures are commonly replaced by soft-tissue in current methods, thus leading to significant underestimation of tracer uptake in the vicinity of bony structures ( Bezrukov et al, 2013 ;Hofmann et al, 2011 ).…”

“…Basically two categories have emerged: atlas-guided attenuation map generation approaches ( Hofmann et al, 2011 ;Bezrukov et al, 2013 ;Arabi and Zaidi, 2016a ;Marshall et al, 2013 ;Arabi and Zaidi, 2016b ) and emission-based approaches ( Rezaei et al, 2012 ;Mehranian and Zaidi, 2015 ). Atlas-guided methods primarily rely on prior information provided by registration of an atlas into target image coordinates to allow classification of bone tissues.…”

The Use of Anatomical Information for Molecular Image Reconstruction Algorithms: Attenuation/Scatter Correction, Motion Compensation, and Noise Reduction

“…Segmentation is performed on T1-weighted or T2-weighted MR images [29,85,86] or on ultrashort echo time (UTE) sequence MR images for better imaging bones with very short T 2 values [87,88]. The other is to use atlasbased methods with machine learning techniques [27,89]. For more information about MR-based attenuation correction methods, see [55].…”

Novel Quantitative Techniques in Hybrid (PET-MR) Imaging of Brain Tumors

“…Moreover, it is not clear to what extent global anatomy from an atlas could realistically predict an individual patient's attenuation map. Hofmann and colleagues 33 studied an MR-guided attenuation correction technique using image segmentation and a method based on an atlas registration and pattern recognition (AT&PR) algorithm in 11 patients and reported that the MR-guided technique using AT&PR provided better overall PET quantification accuracy than the basic MR image segmentation approach because of the significantly reduced volume of errors made regarding volumes of interest within or near bones and the slightly reduced volume of errors made regarding areas outside the lungs. Marshall and colleagues 34 developed a technique wherein variable lung density was taken into account in the attenuation correction of whole-body PET-MR imaging.…”