Influence and Compensation of Truncation Artifacts in MR-Based Attenuation Correction in PET/MR

• Published in: IEEE transactions on medical imaging Vol. 32; no. 11; pp. 2056 - 2063
• Main Authors: Schramm, G., Langner, J., Hofheinz, F., Petr, J., Lougovski, A., Beuthien-Baumann, B., Platzek, I., van den Hoff, J.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01-11-2013; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Attenuation; Attenuation correction; Biomedical imaging; Cavity resonators; Compensation; Humans; Image reconstruction; Lesions; Magnetic Resonance Imaging - methods; Neoplasms - pathology; positron emission tomography/magnetic resonance imaging (PET/MRI); Positron-Emission Tomography - methods; reduced field-of-view (FOV); Retrospective Studies; truncation compensation; Whole Body Imaging - methods; Whole-body PET
• ISSN: 0278-0062; 1558-254X
• DOI: 10.1109/TMI.2013.2272660

The goal of this article is to quantify the influence of truncation artifacts in the magnetic resonance (MR)-based attenuation map (MRMap) on reconstructed positron emission tomography (PET) image volumes and to propose a new method for minimizing this influence. Methods: PET data sets of 20 patients investigated in a Philips Ingenuity PET/MR were reconstructed with and without applying two different methods for truncation compensation (TC1 vendor-provided, TC2 newly developed). In this patient group, the extent of truncation artifacts and quality of the truncation compensation (TC) was assessed visually in the MRMaps. In three additional patients MRMaps generated by algorithm TC2 could be compared to the ground truth of transmission-based attenuation maps obtained with a Siemens ECAT HR + scanner. The influence of truncation on regional SUVs in lesions, other hot structures (bladder, kidney, myocardium) and the arms was assessed in suitable volume of interests (VOI). Results: Truncation compensated MRMaps exhibited residual artifacts in the arms in 16 patients for algorithm TC1 and to a lesser extent in eight patients for algorithm TC2. Compared to the transmission-based attenuation maps algorithm TC2 slightly overestimated the size of the truncated arms by 0.3 cm in the radial direction. Without truncation compensation, VOIs located in the trunk showed an average SUV max underestimation of less than 5.4% relative to the results obtained with TC2. Inside the patients' arms underestimations up to 46.5% were found. Conclusion: In the trunk, standardized uptake values (SUV) underestimations due to truncation artifacts in the MRMap are rather small. Inside the arms, severe SUV underestimations can occur. Therefore, reliable TC is mandatory and can be achieved by applying the newly developed algorithm TC2 which has yielded promising results so far. Implementation of the proposed method is straightforward and should be easily adaptable to other PET/MR systems.