Abstract
Background:
Simultaneous acquisition Positron Emission Tomography / Magnetic Resonance (PET/MR) is a new technology that has potential as a tool both in research and clinical diagnosis. However, cardiac PET acquisition has not yet been validated using MR imaging for attenuation correction (AC). The goal of this study is to evaluate the feasibility of PET imaging using a standard 2-point Dixon VIBE (Volume Interpolated Breathhold Examination) MR sequence for attenuation correction.
Methods and Results:
Evaluation was performed in both phantom and patient data. A chest phantom containing heart, lungs and a lesion insert was scanned by both PET/MR and PET/CT. In addition, 30 patients underwent whole-body 18F-Fluorodeoxyglucose PET/CT followed by simultaneous cardiac PET/MR. Phantom study showed 3% reduction of activity values in the myocardium due to the noninclusion of the phased-array coil in the attenuation correction. In patient scans, average standardized uptake values (SUVs) obtained by PET/CT and PET/MR showed no significant difference (n=30, 4.6 ± 3.5 vs. 4.7 ± 2.8, P=0.47). There was excellent per patient correlation between the values acquired by PET/CT and PET/MR (R2 =0.97).
Conclusions:
Myocardial SUVs PET imaging using MR for AC shows excellent correlation with myocardial SUVs obtained by standard PET/CT imaging. The 2-point Dixon VIBE MR technique can be used for AC in simultaneous PET/MR data acquisition.
Background
A simultaneous acquisition whole-body Positron Emission Tomography/Magnetic Resonance (PET/MR) system (Biograph mMR, Siemens Healthcare, Erlangen, Germany) was FDA approved in June, 2011 (1). This combined imaging technology has the potential to impact both research and clinical disciplines, offering the combined benefits of high spatial, temporal and contrast resolution of MR and metabolic and physiological information provided by PET (2,3,4). Early work in oncologic imaging suggests that PET/MR offers advantages over PET/CT for certain applications and decreases patient radiation dose (5,6,7,8).
The cardiac applications of PET/MR have yet to be fully identified. However, for myocardial imaging, PET/MR offers high contrast resolution MR images combined with quantitative PET measurements of perfusion and metabolism. For example, PET/MR using late gadolinium enhancement (LGE) and 18F-Fluorodeoxyglucose (18F-FDG) has the potential to provide mapping of anatomic scar and metabolic cardiac viability. (9,10). Moreover, 13N-ammonia PET/MR combines myocardial PET qualitative and quantitative perfusion with MR functional and anatomic information and may be useful in ischemia evaluation (11, SNMMI 2014 abstract No. 242). Accurate correction for photon attenuation is a prerequisite for precise quantification with PET (12,13,14). However, accurate attenuation correction (AC) using this integrated system poses certain challenges (15,16,17,18,19,20,21). Because of the small bore size and the strong magnetic field of the MR component, a rotating PET transmission source is not integrated into the combined PET/MR system (13). Therefore, unlike with PET/CT, with PET/MR there is no electron density information that can be used to create the attenuation map (μ-map) needed for AC. Currently, a MR-AC μ-map is derived from a 2-point Dixon VIBE sequence that generates a water only, fat only, in-phase, and opposed phase imaging series (19). Dixon-based MR AC is a tissue segmentation approach that permits segmentation of fat, soft tissue, lung and air. The ability of this MR method to accurately measure concentrations of PET tracers in tissues has been evaluated in both normal solid abdominal organs and various oncologic lesions where whole-body PET/MR measurements were compared with those obtained with PET/CT (7, 8, 16). However, a similar comparison has not been performed for the heart until recently (9) with a limited sample size (n=10). MR-AC of the heart poses unique challenges compared with abdominal organs or non-cardiac tumors due to the combination of cardiac and respiratory motion artifacts. Specifically, MR sequences that necessitate breath-holding may lead to spatial mis-registration with PET due to shifting heart position.
Accordingly, in the current study we compared myocardial 18F-FDG activity measured by PET/MR with values obtained by PET/CT in patients undergoing clinically indicated oncologic PET/CT imaging with 18F-FDG.
Materials and methods
Phantom Study
A chest phantom (Data Spectrum Inc.) containing lung and heart inserts was filled with 18F-FDG in ratio of 5:1 between left ventricle wall/sphere and the background area. The left ventricle cavity was filled with water but did not contain activity. The lung inserts were filled with wet Styrofoam beads. The total activity in the phantom was 66.6 MBq. The phantom was scanned in the Siemens mMR followed by scanning the phantom in a similar position in a Biograph-40 PET/CT, each for 10 minutes list mode acquisition. The phantom was scanned twice in the PET/MR, once using the body phased array coil, and second without the coil, using only the gantry body coil to capture MR signal..
Since MRI does not provide accurate MR attenuation correction (MR-AC) for phantoms, the CT scan from PET/CT was used to create an attenuation map for the PET data obtained in the PET/MR system.. For this purpose, an in-house IDL (Interactive Data Language, Exelis Inc) application was written to process the CT attenuation correction (CT-AC) scan. This application re-samples the CT-AC data into the MR μ-map pixel size, and performs a rigid-body registration of the CT-AC to the MR-AC μ-map using the Elastix (22) package. The parameters for the registration were: Mattes mutual information metric, Euler transform, B-Spline interpolator, and the Adaptive Stochastic Gradient Descent optimizer (22). The registered CT data were then scaled to PET attenuation coefficients using the bi-linear formulation of Burger et al. (23). A mask was applied to the CT image to remove the PET/CT patient couch. The modified MR-AC images were then re-inserted into the PET/MR acquired PET database for image reconstruction. Phantom images were reconstructed using the same algorithm as for the patient data.
Evaluation of Lung segmentation values in patient data
Unlike PET/CT where the actual lung density is measured, in PET/MR the lung segmentation value is set at a constant value throughout the lungs. In order to evaluate the effect of this value on the activity concentration on the heart, we measured the lung density (in Hounsfield units, HU) in different lobes of the lungs from 5 consecutive patients scanned for standard of care whole body PET/CT examination in our facility. These average lungs density values were used to establish typical lung density limits that would represent normal variants for patients undergoing examination. The measured lung density was then scaled to PET attenuation coefficients (23). The MR-AC of one 18F-FDG PET/MR patient was then modified using the limit lung density values derived from CT to segment the lungs, and images were reconstructed. ROIs on the myocardium and in the heart, heart left-ventricle cavity, lungs and liver were drawn and average SUV reported.
Patient study
30 patients with an oncologic history scheduled to undergo whole body clinical PET/CT were recruited prospectively to participate in a whole body PET/MR examination as part of a PET/MR-PET/CT comparison study (24). Myocardial SUV information from the thoracic component of these examinations was analyzed retrospectively. Patient characteristics and imaging data are shown in Table 1. PET/CT and PET/MR images were reviewed for the presence of myocardial uptake of 18F-FDG. A single dose (537± 81 MBq) of 18F-FDG was injected intravenously. The PET/CT scan was performed 59 ± 7 minutes after 18F-FDG injection. The PET/MR examination was performed 127 ± 23 minutes after 18F-FDG injection. Both PET/CT and PET/MR imaging were performed free-breathing and without EKG gating. Motion correction of PET data was not performed in either PET/CT or PET/MR. The study was approved by the Institutional Review Board at Washington University. All patients received a detailed informed consent discussion and gave written informed consent prior to the PET/MR imaging examination.
Table 1.
Activity concentration in various areas of the chest phantom with and without the phased-array coil.
Region | Phased-Array Coil CT-AC (Bq/mL) | Body Coil CT-AC (Bq/mL) | % change |
---|---|---|---|
Background | 8663 | 8886 | −2.5 |
Heart Wall | 28913 | 29801 | −3.0 |
Heart Cavity | 3623 | 2880 | 26 |
Lungs | 418 | 286 | 46 |
Spine insert | 1962 | 1477 | 33 |
18F-FDG PET/CT Imaging
PET/CT was performed on a Biograph 40 system (Siemens Healthcare, Knoxville, TN, USA). Standard AC of the acquired PET data was performed using a μ-map generated from CT information and reconstructed with 3D-OSEM (Ordered Subset Estimation Maximization) with 3 iterations, 21 subsets and post-Gaussian filter of 4 mm. The CT scan was performed at 120 kVp, 111 effective mAs and at pitch of 1, and images were reconstructed at a slice thickness of 4mm. The CT-AC scan duration was approximately 1 minute from ears to mid-thigh. Instructions were given to the patient to hold a relaxed breathing position during the CT scan over the chest.
18F-FDG PET/MR Imaging
PET/MR imaging was performed using a 3T field strength PET/MR system (Biograph mMR, Siemens Healthcare, Erlangen, Germany). A phased-array torso MR coil was placed on the chest and used during acquisition of the thoracic station component of the PET/MR examination.
A four-compartment-model AC map (μ-map) was created from fat-only and water-only components of a 19-second Dixon-based sequence by means of segmentation into background, lung, fat, and soft tissue (TE1/TE2 = 1.23 msec/2.46 msec, TR = 3.6 msec, left-right FOV = 500 mm, anterior-posterior FOV = 300 mm). The Dixon μ-map was then segmented into 4 compartments: air (0 cm-1), lungs (0.0224 cm-2), fat (0.0854 cm-1) and soft-tissue (0.096 cm-1) using the manufacturer provided software.
PET images on the mMR were reconstructed using 3D_OSEM, at 3 iterations, 24 subsets and a post-reconstruction filter of 4mm resulting in a similar voxel size as in the PET/CT.
Phantom Data Analysis
Phantom data were analyzed by drawing a circular region of interest (ROI) on the image along the margins of the lungs (ROI size ~ 5 cm3), the heart insert (~1 cm3), the heart cavity (~1 cm3), and in the background area (~20 cm3), in order to study the effect of body phased array coil on the attenuation correction. For the first scan on PET/MR, the images were acquired with the phased array coil, however the coil was not included in the attenuation correction. The second scan was acquired without the phased array coil and thus, the relative difference in intensity from these two scans represents the net effect of the phased array coil in the attenuation correction.
Patient data Analysis
Three-dimensional image co-registration and fusion of PET and MR imaging data were performed using dedicated postprocessing software for hybrid imaging (Syngo TRUE-D; Siemens Healthcare). Co-registration was performed by automatic algorithm registration using MR images as the source image for co-registration. Images were assessed blindly by a single reader (JML). Average myocardial SUVs were obtained by tracing the entire cross section of the left ventricular myocardium in the transaxial orientation of the heart at the mid-ventricular level (Figure 1). Average values with standard deviations are reported. Significance testing between the average myocardial 18F-FDG PET SUVs acquired on PET/CT and PET/MR imaging was performed by using paired 2-tailed t tests. A P value < 0.05 was considered significant. Agreement between SUVs acquired on PET/CT and PET/MR imaging was estimated with the Spearman correlation coefficient in a scatter plot, and coefficient of variation was graphically analyzed by a Bland-Altman plot.
Figure 1.
Demonstration of how myocardium was traced for average SUV measurements. The patient’s left ventricular myocardium was traced at the mid-ventricular level in the transaxial window, at the same anatomic level for the CT-attenuation-corrected PET and MR-attenuation-corrected PET images.
Of note, as per standard PET/CT and PET/MR imaging, the patients’ arms were raised during PET-CT data acquistion, whereas in PET/MR, the arms were to the side of the body. MLAA correction was performed to account for the difference in arm positions and tissue attenuation effect.
RESULTS
Phantom results
Figure 2 presents an axial image of the chest phantom with its associated μ-map obtained from scaling the CT images. Table 1 presents the measured activity concentration in different areas of the phantom for the different reconstructions. Data demonstrated a relative reduction of voxel intensity in the region of the heart insert and adjacent background of 2.5–3.0% caused by the presence of the phased array coil. Regions of interest traced in areas of low activity (lungs, heart cavity and bone insert) showed a large relative increase in voxel intensity secondary to scatter created by the coil. .
Figure 2.
Picture of the Chest Phantom with associated μ-map.
Effect of lung segmentation value on myocardium data
Lungs density (in HU) varies in different lobes. On PET/CT we observed average lung density (over the 5 patients) of −767, −711, −818 and −699 HU in the right upper, right lower, left upper, left lower lobes of the lungs, respectively. The corresponding PET attenuation coefficients are respectively 0.0224, 0.0277, 0.0175, and 0.0289 cm−1. MR-AC μ-maps were processed to segment the lungs at the minimum and maximum values of 0.0175 and 0.0289 cm−1 and images were reconstructed using these two modified u-map images. Percent change between minimum and maximum attenuation was calculated as ½ difference between 0.0175cm-1 and 0.0289cm-1 normalized to the PET activity values at 0.0224cm-1. These data are reported in Table 2.
Table 2.
SUV for various anatomical areas within the thorax for lung segmentation values. Percent change between minimum and maxiumum attenuation was calculated as 1/2 difference between 0.0175cm-1 and 0.0289cm-1 normalized to the PET activity values at 0.0224cm-1.
Region | Lungs @ 0.0224 cm−1 | Lungs @ 0.0289 cm−1 | Lungs @ 0.0175 cm−1 | % change |
---|---|---|---|---|
Liver | 2.00 | 2.00 | 2.00 | +/− 0.0 |
Myocardium | 4.36 | 4.28 | 4.47 | +/− 2.2 |
Left Ventricle | 1.31 | 1.30 | 1.32 | +/− 0.6 |
Vertebral body | 1.71 | 1.67 | 1.75 | +/− 2.5 |
Lungs | 0.40 | 0.35 | 0.47 | +/− 14.8 |
Aorta | 1.11 | 1.09 | 1.13 | +/− 1.5 |
Dual echo VIBE Dixon attenuation correction sequence
Thirty patients demonstrated 18F-FDG myocardial uptake on PET/CT and PET/MR imaging. Retrospective chart review confirmed that all selected patients had no significant cardiac history, including no history of myocardial infarction, percutaneous coronary interventions or coronary artery bypass surgery. A breath-held, 19-second, duel echo VIBE Dixon attenuation correction sequence was performed at the beginning of PET/MR data acquisition. Sample images of the coronal PET after CT-AC or MR-AC are shown in Figure 3.
Figure 3.
Coronal PET images after attenuation correction (AC) using CT (left column) or MR (right column) information. Top row: AC-PET images overlaid with CT μ-map (top left) and MR μ-map (top right). Bottom row: stand-alone AC-PET images. CT-based AC was performed using 3D-OSEM (Ordered Subset Estimation Maximization). MR AC was performed using a free-breathing dual echo VIBE Dixon sequence. Myocardial PET signal between CT AC and MR AC PET images are comparable. The major discrepancy between CT AC and MR AC PET images is in bone structures, which is secondary to poor delineation between air and bone on MRI which is reflected in MR AC segmentation methods.
SUV comparison between PET/CT and PET/MR
There was no significant difference between the average myocardial SUV obtained by PET/CT and PET/MR (4.6 ± 3.5 vs. 4.7 ± 2.8, P=0.47) (Figure 4A). Although 18F-FDG SUVs were highly variable among patients in this study group and one patient demonstrated an SUV difference of 1.4 between PET/MR and PET/CT, in general, there was excellent per patient correlation between PET/CT and PET/MR SUVs (R2 =0.97, Figure 4A). Bland-Altman analysis demonstrated the lower and upper limits of agreement in average SUVs to be −1.4 and +0.9 (Figure 4B). The slope of the regression line (slope, 1.24) suggests that for patients with substantial myocardial uptake, the measured FDG activity was slightly higher on our PET/MR images in comparison to PET/CT.
Figure 4.
A. Correlation plot between PET/CT and PET/MR SUVs. Spearman’s correlation coefficient R2 = 0.97. Red dotted line represents the line of unity (x=y). B. Bland-Altman plot for variability assessment. Dotted lines represent +/− two standard deviations, The lower and upper limits of agreement in average SUVs are −1.4 and +0.9, respectively. SUV: Standardized uptake value.
We further analyzed each patient for myocardial segmental uptake differences. ROIs of ~ 0.1 cm3 were drawn in the anterior, septal, inferior, and lateral walls of the mid left ventricle. Table 4 shows the average SUVs of the same 30 patients in the different left ventricular segments. There was no significant SUV difference among the wall segments (anterior wall SUV 4.68 +/− 3.47, septal wall 4.80 +/− 371, inferior wall 5.10 +/− 4.13, lateral wall 5.10 +/− 3.87, ANOVA test F=0.11, P=0.95).
Table 4.
No significant difference between PET/MR SUV in different wall segments (ANOVA F=0.1193, P=0.95).
PET/MR SUV | |
---|---|
Anterior Wall | 4.68 +/− 3.47 |
Septal Wall | 4.80 +/− 3.71 |
Inferior Wall | 5.10 +/− 4.13 |
Lateral Wall | 5.18 +/− 3.87 |
DISCUSSION
The results of the current study demonstrate that AC performed with MR permits PET measurements of myocardial activity that are comparable to those obtained with CT-based AC. These findings are consistent with those obtained in non-cardiac solid organ comparisons suggesting the Dixon method of MR—based AC is robust and accurate in various tissues. (7)
Nensa et al. (9) described, in a smaller sample size (n=10), that cardiac SUVmax obtained by PET/MR and PET/CT were comparable in patients who sustained acute myocardial infarction. The findings in our study differ in several ways: we evaluated a large sample size (n=30). We measured and compared mean SUV rather than max SUV in the myocardium ROI; and our SUV measurements were done in a non-infarcted myocardium, in an oncologic cohort with no known cardiac disease.
Since our patients did not undergo a glucose-loading protocol prior to PET/CT and PET/MR imaging, there is a substantial range of myocardial FDG uptake across patients, as manifested by the wide range of reported mean SUV values (from SUV value of 1 to ~15 amongst the thirty patients). Despite the wide range of myocardial FDG uptake, SUVs obtained by PET/CT and PET/MR correlated remarkably well in this retrospective analysis. Also, because our patients were recruited initially for oncologic purposes, we did not perform respiratory gating, which may limit the quality of cardiac images acquired. However, the PET/MR images superimposed well and we did not experience any issues with regard to PET and MR image fusion.
Our findings show that PET measurements of myocardial activity obtained with PET/MR are slightly higher than the values obtained on PET/CT (Figure 4A). This differs from the non-myocardial solid organ PET comparison reported by Drzezga et al. (7), which showed slightly higher activity on PET/CT in comparison to PET/MR as well as in Nensa et al. (9), where uptake in myocardium is statistically similar but slightly higher in PET/CT than PET/MR. Like Drzezga, et al., we performed the PET/CT scan first in all patients because our primary inclusion criterion was a patient undergoing FDG PET/CT for clinical oncologic evaluation. The most likely explanation for this activity difference is that 18F-FDG uptake in myocardium continues beyond the 50–60 min post-injection time point used for clinical PET/CT imaging. (26) While the MR phased array surface coil has been reported to have an attenuation effect of 5–10% (26), our phantom study shows a coil attenuation effect of 2.5–3.0%. This difference caused by the MR phased array surface coil attenuation is not accounted for in our patient study, which considered only attenuation from the patient and the couch/spine coil (hardware attenuation correction).
Our patient study was a retrospective analysis, thus randomization of scan order was not possible. To confirm the hypothesis that the greater myocardial activity on PET-MR was secondary to continued myocardial 18F-FDG uptake, future studies should randomize the order of the PET/CT and PET/MR image acquisition (26). Another limitation of this study is that neither our PET/CT nor PET/MR data were respiratory or cardiac motion corrected. This is because the PET/CT and PET/MR were both performed for whole-body imaging for oncologic evaluation. Researches assessing various methods of PET respiratory and cardiac gating techniques, including PET acquisition coupled to MR motion corrected algorithms, are ongoing and could be incorporated into future prospective research.
One of the clear advantages of PET/MR imaging is its ability to combine high sensitivity of radionuclide imaging with the relatively higher spatial resolution and excellent contrast resolution of MR. Since our study population was not a cardiovascular cohort (we reviewed the clinical charts and all of the recruited patients did not have documented clinical history of myocardial infarction), regional differences in myocardial SUV variation analysis could not be assessed. Additional future studies of patient populations with myocardial infarction could be performed in order to assess regional differences in disease processes.
The phantom data indicated a positive relative change in areas of low activity (lung insert, heart cavity and bone insert). This is due to the fact that non-attenuation of the phased −array coil not included but also consequently, the scatter correction does not model the additional scatter off the phased-array coil. Due to the low proton density in the lungs, the DIXON sequence employed for attenuation correction measurement is not optimized for accurate segmentation of the lungs. Therefore, the technique implemented by the manufacturer is to segment the low-intensity region in the chest and associate the PET attenuation coefficient of lung tissue. We demonstrated that this fixed lung segmentation value employed in PET/MR as opposed to the anatomically accurate value in PET/CT does not have a significant consequence on the measured activity levels in the heart. Similarly this suggests that lesions surrounding the lungs should not be affected as well. However, it remains to be seen if activity concentration measurements in the lungs themselves will be affected by the assumption of fixed attenuation coefficient. Certainly, the chosen value of 0.0224cm-1 represents an accurate average value but patients with parenchymal lung disease may show increased or reduced lung density. Thus PET imaging of lung disease may have limitation in PET/MR until optimized MR sequences for imaging the lung parenchyma are implemented for purposes of determining the local lung density.
New Knowledge Gained
Using clinical patient and phantom data, we have shown that PET measurements of myocardial activity using MR AC show excellent correlation with myocardial activity imaging using standard CT AC, laying the foundation for further clinical development and research in cardiac PET/MR imaging.
Table 3.
Patient characteristics and imaging parameters.
Patient Characteristics | |
---|---|
Total number (n) | 30 |
Sex (%) | |
Male | 11 (37%) |
Female | 19 (63%) |
Age (years) | 55.0 ± 12.4 |
Weight (kg) | 78.2 ± 20.2 |
Dose of 18F-FDG (MBq) | 537± 81 |
Time of PET/CT scan after 18F-FDG injection (min) | 59 ± 7 |
Time of PET/MR scan after 18F-FDG injection (min) | 127 ± 23 |
Abbreviations:
- PET/MR
Positron Emission Tomography / Magnetic Resonance
- PET/CT
Positron Emission Tomography / Computed Tomography
- AC
Attenuation correction
- SUV
Standardized uptake values
- LGE
Late gadolinium enhancement
- 18F-FDG
18F-Fluorodeoxyglucose
- MR-AC
Magnetic resonance attenuation correction
- CT-AC
Computed tomography attenuation correction
References:
- 1.Simultaneous PET/MRI Device - Siemens Biograph mMR System Approved by the FDA, www.medicalnewstoday.com/articles.228262
- 2.Torigian DA, Zaidi H, Kwee TC, Saboury B, Udupa JK, Cho ZH, et al. PET/MR imaging: technical aspects and potential clinical applications. Radiology. 2013;267:26–44. [DOI] [PubMed] [Google Scholar]
- 3.Catana C, Guimaraes AR, Rosen BR. PET and MR imaging: the odd couple or a match made in heaven? J Nucl Med. 2013;54:815–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jadvar H, Colletti PM. Competitive advantage of PET/MRI. Eur J Radiol. 2014. 83:84–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Souvatzoglou M, Eiber M, Takei T, Fürst S, Maurer T, Gaertner F, et al. Comparison of integrated whole-body 11C-choline PET/MR with PET/CT in patients with prostate cancer. Eur J Nucl Med Mol Imaging. 2013;40:1486–99. [DOI] [PubMed] [Google Scholar]
- 6.Wiesmüller M, Quick HH, Navalpakkam B, Lell MM, Uder M, Ritt P, et al. Comparison of lesion detection and quantitation of tracer uptake between PET from a simultaneously acquiring whole-body PET/MR hybrid scanner and PET from PET/CT. Eur J Nucl Med Mol Imaging. 2013;40:12–21. [DOI] [PubMed] [Google Scholar]
- 7.Drzezga A, Souvatzoglou M, Eiber M, Beer AJ, Fürst S, Martinez-Möller A, et al. First clinical experience with integrated whole-body PET/MR: comparison to PET/CT in patients with oncologic diagnoses. J Nucl Med. 2012;53:845–55. [DOI] [PubMed] [Google Scholar]
- 8.Kim JH, Lee JS, Song IC, Lee DS. Comparison of segmentation-based attenuation correction methods for PET/MRI: evaluation of bone and liver standardized uptake value with oncologic PET/CT data. J Nucl Med. 2012;53:1878–82. [DOI] [PubMed] [Google Scholar]
- 9.Rischpler C, Nekolla SG, Dregely I, Schwaiger M. Hybrid PET/MR imaging of the heart: potential, initial experiences, and future prospects. J Nucl Med. 2013;54:402–15. [DOI] [PubMed] [Google Scholar]
- 10.Nensa F, Poeppel TD, Beiderwellen K, Schelhorn J, Mahabadi AA, Erbel R, et al. Hybrid PET/MR Imaging of the Heart: Feasibility and Initial Results. Radiology. 2013;268:366–73. [DOI] [PubMed] [Google Scholar]
- 11.Lau J, Laforest R, Zheng J, Lesniak D, Priatna A, Gropler R, et al. (2014). 13N-Ammonia PET/MR myocardial stress perfusion imaging early experience. J Nucl Med. 2014; 55 (Supplement 1):242.24396030 [Google Scholar]
- 12.Schlosser T, Nensa F, Mahabadi AA, Poeppel TD. Hybrid MRI/PET of the heart: a new complementary imaging technique for simultaneous acquisition of MRI and PET data. Heart 2013;99:351–52. [DOI] [PubMed] [Google Scholar]
- 13.Martinez-Möller, Souvatzoglou M, Delso G, Bundschuh RA, Chefd’hotel C, Ziegler SI, et al. Tissue classification as a potential approach for attenuation correction in whole-body PET/MRI: evaluation with PET/CT data. J Nucl Med. 2009;50:520–6. [DOI] [PubMed] [Google Scholar]
- 14.Schulz V, Torres-Espallardo I, Renisch S, Hu Z, Ojha N, Börnert P, et al. Automatic, three-segment, MR-based attenuation correction for whole-body PET/MR data. Eur J Nucl Med Mol Imaging. 2011;38:138–52. [DOI] [PubMed] [Google Scholar]
- 15.Eiber M, Martinez-Möller A, Souvatzoglou M, Holzapfel K, Pickhard A, Löffelbein D, et al. Value of a Dixon-based MR/PET attenuation correction sequence for the localization and evaluation of PET-positive lesions. Eur J Nucl Med Mol Imaging. 2011;38:1691–701. [DOI] [PubMed] [Google Scholar]
- 16.Wagenknecht G, Kaiser HJ, Mottaghy FM, Herzog H. MRI for attenuation correction in PET: methods and challenges. MAGMA. 2013;26:99–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bezrukov I, Mantlik F, Schmidt H, Schölkopf B, Pichler BJ. MR-Based PET attenuation correction for PET/MR imaging. Semin Nucl Med. 2013;43:45–59. [DOI] [PubMed] [Google Scholar]
- 18.Keereman V, Mollet P, Berker Y, Schulz V, Vandenberghe S. Challenges and current methods for attenuation correction in PET/MR. MAGMA. 2013;26:81–98. [DOI] [PubMed] [Google Scholar]
- 19.Keller SH, Holm S, Hansen AE, Sattler B, Andersen F, Klausen TL, et al. Image artifacts from MR-based attenuation correction in clinical, whole-body PET/MRI. MAGMA. 2013;26:173–81. [DOI] [PubMed] [Google Scholar]
- 20.Berker Y, Franke J, Salomon A, Palmowski M, Donker HC, Temur Y, et al. MRI-based attenuation correction for hybrid PET/MRI systems: a 4-class tissue segmentation technique using a combined ultrashort-echo-time/Dixon MRI sequence. J Nucl Med. 2012;53:796–804. [DOI] [PubMed] [Google Scholar]
- 21.Martinez-Möller A, Nekolla SG. Attenuation correction for PET/MR: problems, novel approaches and practical solutions. Z Med Phys. 2012;22:299–310. [DOI] [PubMed] [Google Scholar]
- 22.Elastix: toolbox for rigid and nonrigid registration of images. http://elastix.isi.uu.nl.
- 23.Burger C, Goerres G, Schoenes S, Buck A, Lonn AH, Von Schulthess GK. PET attenuation coefficients from CT images: experimental evaluation of the transformation of CT into PET 511-keV attenuation coefficients. Eur J Nucl Med Mol Imaging. 2002;29:922–7. [DOI] [PubMed] [Google Scholar]
- 24.Lau J, Laforest R, Sharma S, Priatna A, McConathy J, Amado L, Gropler R, Woodard PK (2014). Feasibility of MRI attenuation correction in cardiac-gated FDG-PET. J Nucl Med. 2013; 54 (Supplement 2):27. [Google Scholar]
- 25.Kartmann R, Paulus DH, Braun H, Aklan B, Ziegler S, Navalpakkam BK, et al. Integrated PET/MR imaging: automatic attenuation correction of flexible RF coils. Med Phys. 2013;40:082301. [DOI] [PubMed] [Google Scholar]
- 26.Gambhir SS, Schwaiger M, Huang SC, Krivokapich J, Schelbert HR, Nienaber CA, Phelps ME. Simple noninvasive quantification method for measuring myocardial glucose utilization in humans employing positron emission tomography and fluorine-18 deoxyglucose. J Nucl Med. 1989;30:359–66. [PubMed] [Google Scholar]