MRI: A feasibility study

MRI: A feasibility study

G Model EURR-6744; No. of Pages 6 ARTICLE IN PRESS European Journal of Radiology xxx (2014) xxx–xxx Contents lists available at ScienceDirect Europ...

1MB Sizes 77 Downloads 80 Views

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS European Journal of Radiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study Amarnath Jena a , Sangeeta Taneja a , Reema Goel a,∗ , Pushpendranath Renjen b , Pradeep Negi a a b

Department of Molecular Imaging and Nuclear Medicine, Indraprastha Apollo Hospitals, Sarita Vihar, Mathura Road, New Delhi 110076, Delhi, India Department of Neurology, Indraprastha Apollo Hospitals, Sarita Vihar, Mathura Road, New Delhi 110076, Delhi, India

a r t i c l e

i n f o

Article history: Received 18 March 2014 Received in revised form 1 April 2014 Accepted 3 April 2014 Keywords: Simultaneous brain PET/MRI PET quantitation Ultrashort echo time sequence 18 F-FDG brain PET/CT Attenuation correction

a b s t r a c t Purpose: Simultaneous brain PET/MRI faces an important issue of validation of accurate MRI based attenuation correction (AC) method for precise quantitation of brain PET data unlike in PET/CT systems where the use of standard, validated CT based AC is routinely available. The aim of this study was to investigate the feasibility of evaluation of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI using ultrashort echo time (UTE) sequences for AC and to assess their agreement with those obtained from PET/CT examination. Methods: Sixteen patients (age range 18–73 years; mean age 49.43 (19.3) years; 13 men 3 women) underwent simultaneous brain PET/MRI followed immediately by PET/CT. Quantitative analysis of brain PET images obtained from both studies was undertaken using Scenium v.1 brain analysis software package. Twenty ROIs for various brain regions were system generated and 6 semiquantitative parameters including maximum standardized uptake value (SUV max), SUV mean, minimum SUV (SUV min), minimum standard deviation (SD min), maximum SD (SD max) and SD from mean were calculated for both sets of PET data for each patient. Intra-class correlation coefficients (ICCs) were determined to assess agreement between the various semiquantitative parameters for the two PET data sets. Results: Intra-class co-relation between the two PET data sets for SUV max, SUV mean and SD max was highly significant (p < 0.00) for all the 20 predefined brain regions with ICC > 0.9. SD from mean was also found to be statistically significant for all the predefined brain regions with ICC > 0.8. However, SUV max and SUV mean values obtained from PET/MRI were significantly lower compared to those of PET/CT for all the predefined brain regions. Conclusion: PET quantitation accuracy using the MRI based UTE sequences for AC in simultaneous brain PET/MRI is reliable in a clinical setting, being similar to that obtained using PET/CT. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Fluorine 18 fluorodeoxyglucose (FDG) PET or dual-modality PET and computed tomographic (CT) imaging (PET/CT) is currently the most accurate and widely available in vivo method for the investigation of regional human brain metabolism. Its diagnostic use has emerged as an important problem solving tool in neurology, neurosurgery and psychiatry [1–5]. Since magnetic resonance imaging (MRI) represents the first-line diagnostic imaging modality for

∗ Corresponding author. Tel.: +91 1129873079; mobile: +91 9654343326. E-mail addresses: [email protected] (A. Jena), s [email protected] (S. Taneja), [email protected] (R. Goel), [email protected] (P. Renjen), [email protected] (P. Negi).

several indications in the field of neuroimaging, fusion of PET and MR images has been attempted combining the high resolution of MR imaging with the low-resolution functional capability of PET. Fully integrated scanners capable of simultaneous acquisition of PET and MRI have now been developed [6–8], thus eliminating CT co registration for PET and reducing radiation exposure to the patient. An accurate attenuation-correction (AC) method is necessary for quantification of 18 F-FDG brain PET studies. However validation of a MRI-based AC method remains an important issue in simultaneous PET/MRI systems unlike in PET/CT for which standard, validated CT transmission scan for AC is routinely available. This is because the signal obtained from MRI reflects tissue hydrogen densities and relaxation times and not electron density, thus the MR images are not directly related to the tissue linear attenuation coefficients as opposed to the CT images 9. Although several attenuation maps

http://dx.doi.org/10.1016/j.ejrad.2014.04.008 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS

2

A. Jena et al. / European Journal of Radiology xxx (2014) xxx–xxx

have been previously estimated from segmented MR images 10, and atlas based methods [11], the most challenging task consists of differentiating bone tissue from air filled spaces which is especially important for obtaining accurate correction in head datasets. In this aspect new ultrashort echo time (UTE) sequence has been proposed and shown to be as accurate as the segmented CT method [10,12]. Precise quantitative measurements of pathologic processes in the brain such as in neurodegenerative diseases, epilepsy, tumors especially before and after treatment and vascular disturbances using simultaneous brain PET/MRI could improve the clinical evaluation of these various disorders of the central nervous system. In this work we investigated the feasibility of evaluation of semiquantitative 18 F-FDG PET parameters obtained from simultaneous brain PET/MRI using UTE sequences for AC and assessed their agreement with the parameters obtained from conventional PET/CT.

Fig. 1. Attenuation map of head as obtained from CT based AC (a) and MRI UTE sequence based AC (b).

2. Material and methods 2.1. Patient population It is a prospective study which included 16 patients, 12 of whom were referred for simultaneous brain PET/MRI for various neurological indications and 4 for simultaneous whole body PET/MRI for non neurological complaints. Of these twelve patients, three were suspected cases of Parkinson’s disease, three with dementia, one with a clinical suspicion of multiple system atrophy, one a suspected case of Creutzfeldt–Jakob disease, one with multiple cavernous hemangiomas while three presented with epilepsy. All patients signed a written informed consent for the brain PET/MR and PET/CT examinations and the scientific evaluation of the datasets. This study was approved by the local institutional review board. The inclusion criteria were written informed consent and ability to undergo another scan after the PET/MRI examination. Exclusion criteria were pregnancy, age below 18 years, and standard contraindications for MRI examinations (magnetic metal implants, pacemakers, etc.) 2.2. Instrumentation 2.2.1. PET/MR Simultaneous PET/MR was performed on the Biograph mMR (Siemens Co, Erlangen, Germany). This system consists of a 3-T MRI scanner harboring a fully functional PET system, equipped with the avalanche photodiode technology [13]. MR scanner features high-performance gradient systems (45 mT/m) with a slew rate of 200 T/(m s) and is equipped with total imaging matrix coil technology, covering the entire body with multiple integrated radiofrequency surface coils [14]. The PET scanner has a spatial resolution of 4.3 mm at 1 cm and 5.0 mm at 10 cm from the transverse field of view (FOV); its sensitivity being 1.47% at the center of the FOV and 1.38% at 10 cm. 2.2.2. PET/CT PET/CT was performed on BiographTM mCT (128 S, Siemens Co, Erlangen, Germany). The PET scanner has a spatial resolution of 4.4 mm at 1 cm, a sensitivity of 0.97% at the center of the FOV and 0.95% at 10 cm. 2.2.3. Imaging protocol Patients fasted for at least 6 h before intravenous tracer injection of 18 F-FDG. All patients underwent a dual-imaging protocol including brain PET/MRI followed by PET/CT with the smallest possible temporal delay, to allow for using the remaining activity of the initial 18 F-FDG injection, thus avoiding additional radiation exposure to the patients. Simultaneous PET/MRI examination comprised of a transversal T1-weighted UTE for attenuation correction (repetition

time (TR)/echo time (TE1) (TE2), 11.94/0.07/22.46; excitation angle, 10◦ ; matrix size, 192 × 192 × 192; resolution, 1.6 mm × 1.6 mm × 1.6 mm; bandwidth, 1532 Hz/pixel (Fig. 1)). MRI sequences for complete diagnostic evaluation of brain included transversal 2D-encoded fluid attenuated inversion recovery (FLAIR) sequence (TR/TE, 7000/94; inversion time (TI), 2215.2 s; matrix size, 418 × 512; resolution, 1 mm × 0.9 mm × 5 mm; slice thickness, 5 mm; and bandwidth, 260 Hz/pixel); T2-weighted 2Dencoded turbo spin-echo sequence in axial, sagittal, and coronal orientations depending on the disease pattern (TR/TE, 4300/100; matrix size, 278 × 512; resolution, 0.7 mm × 0.4 mm × 0.5 mm; slice thickness, 5 mm; and pixel bandwidth, 222 Hz/pixel); and sagittal 3D-encoded magnetization-prepared rapid-acquisition gradient-echo (MPRAGE) sequence (TR/TE, 1500/2.33; TI, 900 s; matrix size, 410 × 512; resolution, 1.2 mm × 1 mm × 1 mm; bandwidth, 180 Hz/pixel; and parallel imaging GRAPPA factor 2). Other optional sequences included diffusion weighted imaging (DWI) (TR/TE, 4600/101; matrix size 160 × 160; resolution, 1.4 mm × 1.4 mm × 5 mm; band-width, 1008 Hz/pixel; and parallel imaging GRAPPA factor 2), Susceptibility weighted imaging (SWI) (TR/TE, 26/20; matrix size 238 × 320; resolution, 0.8 mm × 0.7 mm × 4 mm; band-width, 170 Hz/pixel) and perfusion echo planar imaging (EPI) (TR/TE, 2550/31; matrix size 128 × 128; resolution, 1.8 mm × 1.8 mm × 5 mm; band-width, 1446 Hz/pixel). Total MRI acquisition time was 15 min with simultaneous PET data acquisition. All PET data were acquired in sinogram/frame mode. After the scan all coincidence data were sorted into a 2D PET sinogram, which was subsequently reconstructed into transaxial slices corresponding to a FOV of 25.8 cm using an iterative three-dimensional-ordered-subset expectation maximization (OSEM) algorithm with 3 iterations and 21 subsets, Gaussian smoothing of 4 mm in full width at half maximum, and a zoom of 1. The attenuation maps were computed from UTE sequences [9,12]. Subsequent to PET/MRI all patients underwent PET/CT according to standard clinical protocols. Non-contrast-enhanced CT scan (309 mA/120 kV) of the brain was acquired for all patients followed by PET data acquisition for 15 min. Subsequent reconstruction of the data into transaxial slices (matrix size, 400 × 400, corresponding to an axial FOV of 21.8 cm, voxel size 3.18 mm × 3.18 mm × 2.03 mm) was undertaken using the standard software of the scanner (Fourier rebinning, two-dimensional OSEM with 6 iterations, 21 subsets). Attenuation maps obtained from the CT data by bilinear transformation were used for attenuation correction of the PET/CT data. For both modalities, emission data were corrected for randoms, dead time, scatter, and attenuation.

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS A. Jena et al. / European Journal of Radiology xxx (2014) xxx–xxx

2.2.4. Data analysis Quantitative analysis of both sets of brain PET images obtained from the conventional PET/CT and PET/MRI was undertaken using Scenium v.1 brain analysis software package (Siemens Co.). This software compares the patient’s scan with a normal database on a voxel-by-voxel basis. Its normal database is derived from FDG-PET brain scans of 30 normal patients who underwent genetic testing and were all found to be APOE4 negative [15]. The regions of interest (ROI) used in Scenium have been defined on a high-resolution T1 MRI volume scan [16]. In order to be able to compare individual voxels, each patient scan was normalized first by registration. For PET/CT cases fusion with CT was done via rigid registration while for PET/MRI, MPRAGE MRI sequences were used. This was followed by fusion with the PET data of normal database using deformable registration. Subsequently smoothing using a 12 mm full width half maximum isotropic Gaussian kernel was done followed by intensity normalization to compensate for overall metabolism and imaging-related differences in general intensity relative to the top 25% of intensity values in the cerebellum. The extent to which each voxel differs from the normal is then expressed as a statistical Z score (the intensity difference expressed as the number of standard deviations from the mean). ROIs for bilateral frontal, parietal, temporal lobes, cingulate and paracingulate gyri, central region, occipital lobes, calcarine fissure and surrounding cortex, basal ganglia, mesial temporal lobe and both sides of cerebellum were system generated and 6 semiquantitative parameters including maximum standardized uptake value (SUV max), SUV mean, minimum SUV (SUV min), minimum standard deviation (SD min), maximum SD (SD max) and SD from mean were calculated for both sets of PET data for each patient. Qualitative analysis of both the data sets was also undertaken by two trained nuclear medicine physicians with more than 5 years of experience. PET images were visually inspected and evaluated for any disagreement with the statistical data generated by the software. For statistical analysis, SPSS software package (version 16.0, SPSS for Windows, 2007) was used. Intra-class correlation coefficients (ICCs) were determined to assess agreement between the various semiquantitative parameters for the two PET data sets; one obtained via AC by CT and another via AC by MRI. For evaluation of the statistical difference between SUV mean, SUV max

3

and SUV min values derived from the 2 imaging modalities, a nonparametric Wilcoxon matched-pairs signed rank test was used. For demographic data and time delay, mean and standard deviation are presented. Other data are presented as median and range (minimum–maximum value). p-Values < 0.05 were considered statistically significant.

3. Results 16 patients (age range 18–73 years; mean age 49.43 (±19.3) years; median age 68 years; 13 men 3 women) were examined. PET/MRI was acquired after a mean uptake period of 48.2 ± 12 min from the time of 18 F-FDG injection (mean dose 355.2 Mbq; 350–399 MBq depending on the body weight) whereas PET/CT was acquired after PET/MRI with a mean interval of 112.6 ± 16 min from the time of injection. The MRI datasets acquired during simultaneous PET acquisition showed diagnostic image quality, without any major artifacts or distortions. Regional semiquantitative parameters including SUV max, SUV mean, SUV min, SD min, SD max and SD from mean were evaluated for the 20 system generated ROIs in all sixteen patients with the use of Scenium software package (Figs. 2 and 3). Intra-class co-relation between the two PET data sets for SUV max, SUV mean and SD max was highly significant (p < 0.00) for all the 20 predefined brain regions with intra-class correlation co efficient (ICC) > 0.9. SD from mean was also found to be highly significant (p < 0.00) for all the predefined brain regions with ICC > 0.8 except for the right temporal lobe where ICC was found to be 0.590 and p < 0.047 (Table 1). Intra-class correlation between the 2 data sets for SUV min was found to be highly significant for cingulate and paracingulate gyri (ICC = 0.911; p < 0.000), basal ganglia (ICC = 0.870; p < 0.000), central region (ICC = 0.739; p < 0.007), mesial temporal lobe (ICC = 0.840; p < 0.001), left calcarine fissure and surrounding cortex (ICC = 0. 764; p < 0.004), left frontal (ICC = 0.643; p < 0.027) and right parietal lobe (ICC = 0.660; p < 0.022) whereas for SD min it was found to be highly significant (p < 0.00) for all the predefined brain regions with ICC > 0.8 except frontal lobe, right basal ganglia and left side of cerebellum. SUV max and SUV mean values obtained from PET/MRI were significantly lower compared to those of PET/CT for all the predefined brain regions (p < 0.001). However, there was no statistical

Fig. 2. ROIs for CT (a) and MRI (b) images as defined by Scenium software package for a normal study.

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS

4

A. Jena et al. / European Journal of Radiology xxx (2014) xxx–xxx

Fig. 3. Bilaterally symmetrical relative hypometabolism in the posterior parietal and temporal regions including bilateral occipital lobes and relatively preserved perfusion in the striatum seen on CT AC derived cortical maps (a) as well as MRI AC derived cortical maps (b) for a 73-year-old male patient presenting with long term memory loss, tremors and postural fall.

difference between SUV min values obtained from the PET/CT and PET/MRI for all the brain regions except for the frontal lobe (p < 0.001 for left frontal lobe and p < 0.04 for right frontal lobe) and right basal ganglia (p < 0.047). 4. Discussion Quantification of tracer uptake is critical for neurological studies especially neurocognition studies, epilepsy evaluation, vascular disturbances and therapy response assessment, where absolute quantification is required, and differences between healthy subjects and patient populations are often subtle. Accurate quantitation of brain PET data using MR based attenuation correction still needs validation. Several solutions for attenuation correction of the PET/MRI data have been proposed, such as atlas-based methods and MR-based segmentation of the scanned volume into specific tissue types (water, lungs, fat, background/air) with known attenuation factors

using Dixon MRI sequences [10,11,17]. Lesion based quantification of tracer uptake in brain PET/MRI has been attempted in recent studies using these sequences. Schwenzer et al., studied mean frontal/parieto-occipital ratio of FDG uptake and the mean asymmetry index (right − left) in the PET datasets of the PET/MR and the PET/CT and found a higher symmetry in the PET datasets of the PET/CT in 50 patients with intracranial masses, head and upper neck tumors or neurodegenerative diseases [18]. However their study did not include calculation of standardized uptake values which are highly desirable with regard to clinical applications. In another study by the same group tumor-to-gray matter ratios (T/G ratios) and tumor-to-white matter ratios (T/W ratios) were computed by ROI analysis in 10 patients with intracranial masses [19]. Valentina Garibotto et al., studied 15 patients with cognitive impairment of probable neurodegenerative origin, presurgical evaluation of drugrefractory epilepsy, and brain tumor staging and demonstrated the feasibility of PET/MRI of the brain focusing only on qualitative PET assessment, rather than quantitative measures [20].

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS A. Jena et al. / European Journal of Radiology xxx (2014) xxx–xxx

5

Table 1 Quantitative parameters with corresponding statistical correlates. Part 1 SUV max

Frontal lobe Temporal lobe Parietal lobe Cingulate and paracingulategyri Central region Occipital lobe Calcarine fissure Basal ganglia Mesial temporal lobe Cerebellum

SUV mean

Mean value

ICC

P value

Mean value

ICC

P value

7.3a ;12.4b 7.57;11.41 8.21;12.41 7.83;11.85 7.40;11.98 8.11;13.02 7.87;12.45 7.88;12.16 5.58;8.78 6.37;10.61

0.917 0.938 0.955 0.937 0.924 0.941 0.933 0.925 0.893 0.958

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

4.06a ;6.02b 4.30;6.10 4.36;6.26 4.73;6.20 4.47;6.05 4.77;6.38 5.32;6.90 5.01;6.51 3.64;4.80 3.80;5.29

0.967 0.959 0.963 0.972 0.978 0.974 0.979 0.963 0.968 0.975

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Part 2 SD max

Frontal lobe Temporal lobe Parietal lobe Cingulate and paracingulategyri Central region Occipital lobe Calcarine fissure Basal ganglia Mesial temporal lobe Cerebellum

SD from mean

Mean value

ICC

P value

Mean value

ICC

P value

−3.5a ;−2.8b −2.8;−1.7 −1.2;−1.3 0.57;−0.13 −0.57;−0.33 0.55;0.38 0.62;0.21 0.91;0.14 0.98;0.46 −1.5;−0.9

0.840 0.925 0.967 0.955 0.899 0.967 0.981 0.850 0.975 0.946

0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000

4.3a ;4.3b 3.8;4.2 4.4;4.3 5.08;4.0 3.9;3.5 5.2;5.2 3.5;3.1 6.4;5.1 5.2;4.3 4.2;4.1

0.914 0.590 0.825 0.955 0.918 0.927 0.946 0.947 0.924 0.905

0.000 0.047 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000

ICC, intra-class co-relation coefficient. a MRI derived values. b CT derived values

These MRI sequences, however, do not account for attenuation caused by bone for which accurate correction is needed in head datasets [21]. Other sequences, such as UTE MR sequences which allow an approximation of bone distribution have thus been proposed [9,12,22]. UTE sequence does not make any assumptions about patient anatomy and can be used in patients who have different anatomic features, giving an advantage over most other methods, which derive the location of the cortical bone in an indirect manner, mainly by making some assumption about the general anatomy of the patient [12]. In our study we evaluated brain PET data quantification using attenuation correction maps derived from MRI UTE sequences and assessed its agreement with the PET data obtained after CT based AC, considered to be the reference standard. Semiquantitative PET parameters, SUV max and SUV mean derived from MR AC brain PET images showed a high correlation with those obtained from CT AC images for the entire brain validating their use in a clinical setting. Highly significant agreement was also found between maximum standard deviation and standard deviation from mean of the two PET data sets signifying similar observer independent interpretation of the statistical data when compared to the normal database for all brain regions. However, consistently lower tracer uptake by means of SUVbased analysis was observed for data acquired on PET/MRI as compared to PET/CT. Such differences in the SUV values have also been reported in the past and are attributed to several factors, mainly biological factors such as time delay between tracer administration and scan acquisition, differences in scanner technology and data processing algorithms [8,23]. In our study there was a longer time delay between radiotracer administration and PET/CT acquisition which could have influenced the higher PET/CT SUV values. In addition the effect of the differences in PET scanner design between the two modalities

on SUV quantitation cannot be excluded [14]. Despite the observed differences in SUV values, excellent agreement between simultaneous brain PET/MRI and PET/CT data has been observed. This work demonstrates the use of this hybrid imaging modality with respect to quantifying brain PET studies allowing the benefits of PET/MRI to be realized clinically. Moreover all relevant imaging information can be collected in one single session which is not only more comfortable for the patients, especially when additional procedures, such as anesthesia or sedation, are required but also makes the study time shorter as compared to the total time that would be required if a PET/CT and a MR study were to be acquired separately. The versatility of MR imaging, superior soft-tissue contrast and no use of ionizing radiation renders PET/MR useful in answering clinical questions which require not only MR imaging but also PET such as in the preoperative evaluation of focal epilepsy, early diagnosis and differential diagnosis of neurodegenerative diseases or local tumor recurrence in gliomas, as were the referral indications in our study [24–27]. Simultaneous PET/MRI allows for both spatial and temporal correlation of the signals assuring that the fast changing processes in the brain can be evaluated by both modalities at the same time. Moreover, simultaneous measurements of micro vascular proliferation and permeability via MRI and specific tracer uptake via PET can help in exploration of several research areas such as quantification of tumor proliferation, its vascular properties, and antitumor effects of novel drugs, thus enabling better understanding of tumor biology and therapeutic response on a personalized basis. The present study involved a small mixed group of patients and was limited by a lack of clinical follow up of the patients. However, these limitations are of little concern for the present feasibility study.

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008

G Model EURR-6744; No. of Pages 6

ARTICLE IN PRESS

6

A. Jena et al. / European Journal of Radiology xxx (2014) xxx–xxx

5. Conclusion This feasibility study indicates that PET quantitation accuracy using the optimally co registered MRI data of simultaneous brain PET/MRI is similar to that using PET/CT and is reliable in a clinical setting. However, larger prospective studies are needed to further validate the use of simultaneous brain PET/MRI for clinical and research applications that require improved quantification. Conflict of interest The authors declare that they have no conflict of interest. Acknowledgement Dr Abhishek Jha for literature survey and proof reading of the manuscript. References [1] Schelbert HR, Hoh CK, Royal HD, et al. Procedure guideline for tumor imaging using fluorine-18-FDG. Society of Nuclear Medicine. J Nucl Med 1998;39:1302–5. [2] Ishii K. PET approaches for diagnosis of dementia. Am J Neuroradiol 2013 [Epub ahead of print]. [3] Silverman DHS. Brain F-18-FDG PET in the diagnosis of neurodegenerative dementias: comparison with perfusion SPECT and with clinical evaluations lacking nuclear imaging. J Nucl Med 2004;45:594–607. [4] Kosaka N, Tsuchida T, Uematsu H, Kimura H, Okazawa H, Itoh H. 18FFDG PET of common enhancing malignant brain tumors. Am J Roentgenol 2008;190(6):W365–9, http://dx.doi.org/10.2214/AJR.07.2660. [5] Van Paesschen W, Dupont P, Sunaert S, Goffin K, Van Laere K. The use of SPECT and PET in routine clinical practice in epilepsy. Curr Opin Neurol 2007;20:194–202. [6] Schlemmer HP, Pichler BJ, Schmand M, et al. Simultaneous MR/PET imaging of the human brain: feasibility study. Radiology 2008;248:1028–35. [7] Ratib O, Beyer T. Whole-body hybrid PET/MRI: ready for clinical use? Eur J Nucl Med Mol Imaging 2011;38:992–5. [8] Drzezga A, Souvatzoglou M, Eiber M, 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. [9] Catani C, Kouwe A, Benner T, et al. Toward implementing an MRI-based PET attenuation-correction method for neurologic studies on the MR-PET brain prototype. J Nucl Med 2010;51:1431–8. [10] Zaidi H, Montandon ML, Slosman DO. Magnetic resonance imaging-guided attenuation and scatter corrections in three-dimensional brain positron emission tomography. Med Phys 2003;30:937–48.

[11] Martinez-Moller A, Nekolla SG. Attenuation correction for PET/MR: problems, novel approaches and practical solutions. Z Med Phys 2012;22(4):299–310. [12] Keereman V, Fierens Y, Broux T, De Deene Y, Lonneux M, Vandenberghe S. MRIbased attenuation correction for PET/MRI using ultrashort echo time sequences. J Nucl Med 2010;51:812–8. [13] Pichler BJ, Judenhofer MS, Catana C, et al. Performance test of an LSO-APD detector in a 7-T MRI scanner for simultaneous PET/MRI. J Nucl Med 2006;47:639–47. [14] Delso G, Fürst S, Jakoby B, et al. Performance measurements of the Siemens mMR integrated whole-body PET/MR scanner. J Nucl Med 2011;52:1914–22. [15] Reiman EM, Caselli RJ, Yun LS, et al. Preclinical evidence of Alzheimer’s disease in persons homozygous for the epsilon 4 allele for apolipoprotein. N Engl J Med 1996;334(12):752–8. [16] Tzourio-Mazoyer N, Landeau B, Papathanassiou D, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 2002;15(1):273–89. [17] Malone IB, Ansorge RE, Williams GB, Nestor PJ, Carpenter TA, Fryer TD. Attenuation correction methods suitable for brain imaging with a PET/MRI scanner: a comparison of tissue atlas and template attenuation map approaches. J Nucl Med 2011;52:1142–9. [18] Schwenzer L, Stegger S, Bisdas, et al. Simultaneous PET/MR imaging in a human brain PET/MR system in 50 patients—current state of image quality. Eur J Radiol 2012;88:3472–8. [19] Boss A, Bisdas S, Kolb A, et al. Hybrid PET/MRI of intracranial masses: initial experiences and comparison to PET/CT. J Nucl Med 2010;51(8):1198–205. [20] Garibotto V, Heinzer S, Vulliemoz S, et al. Clinical applications of hybrid PET/MRI in neuroimaging. Clin Nucl Med 2013;38(1):e13–8, http://dx.doi.org/10.1097/RLU.0b013e3182638ea6. [21] Catana C, Kouwe AV, Benner T, et al. Is accurate bone segmentation required for MR-based PET attenuation correction? Proc Int Soc Mag Reson Med 2009;17:593. [22] Robson MD, Gatehouse PD, Bydder M, Bydder GM. Magnetic resonance: an introduction to ultrashort TE (UTE) imaging. J Comput Assist Tomogr 2003;27:825–46. [23] Wiesmüller M, Quick HH, Navalpakkam B, 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. [24] Kurth J, Sakretz M, Teipel S, Krause BJ. Molecular imaging of dementia. Geriatr Mental Health Care 2013;1:56–62. [25] Mortimer AM, Likeman M, Lewis TT. Neuroimaging in a practical guide. Pract Neurol 2013;13:92–103, dementia: http://dx.doi.org/10.1136/practneurol-2012-000337. [26] Shaffer JL, Petrella JR, Sheldon FC, et al. Predicting cognitive decline in subjects at risk for Alzheimer disease by using combined cerebrospinal fluid, MR imaging, and PET biomarkers. Radiology 2013;266(2):583–91, http://dx.doi.org/10.1148/radiol.12120010 [Epub 2012 Dec 11]. [27] Lee KK, Salamon N. [18F] fluorodeoxyglucose-positron-emission tomography and MR imaging coregistration for presurgical evaluation of medically refractory epilepsy. Am J Neuroradiol 2009;30(November (10)):1811–6, http://dx.doi.org/10.3174/ajnr.A1637 [Epub 2009 Jul 23].

Please cite this article in press as: Jena A, et al. Reliability of semiquantitative 18 F-FDG PET parameters derived from simultaneous brain PET/MRI: A feasibility study. Eur J Radiol (2014), http://dx.doi.org/10.1016/j.ejrad.2014.04.008