We included 48 subjects that participated in two PET/MR research protocols in the context of neurodegenerative diseases. Thirty-four patients suspected for dementia were investigated with a static [¹⁸F]FDG PET/MR protocol between October 2016 and June 2017. Three patients were excluded from this comparison study due to dental implants which led to metal artifacts in the MR images. In one case the patient was positioned too low in the head coil which led to very low ZTE MR signal in the caudal end of the head due to low coil sensitivity in that region. This case was excluded as well. The mean age of the remaining 30 patients was 63 y (range 40–77 y). In addition, we analyzed 14 dynamic [¹⁸F]PE2I acquisitions of healthy controls (mean age 40.8 y, range 21–70 y). As in the case of the static acquisitions, three cases had to be excluded due to metal artifacts caused by dental implants.

All patients were examined on a SIGNA 3T TOF PET/MR (GE Healthcare, Chicago, US) [29]. The static [¹⁸F]FDG PET/MR protocol included a 25 min static PET acquisition 66±9 min after tracer injection (mean injected dose 144±31 MBq). For the 11 [¹⁸F]PE2I cases, 60 min of dynamic PET data were acquired directly after tracer injection (mean injected dose 153±15 MBq). During the PET acquisitions a LAVA flex MR [acquisition details: repetition time (T_R) 4 ms, echo time (T_E) 2.23 ms, flip angle (α) 5°, matrix 256 × 256 × 120, voxel size 1.95 mm × 1.95 mm × 2.6 mm, number of averages 0.7, acquisition time: 18 s], a ZTE MR (acquisition details: 3D radial acquisition, α.8°, matrix 110 × 110 × 116, voxel size 2.4 mm × 2.4 mm × 2.4 mm, number of averages 4, bandwidth ± 62.5 kHz, acquisition time 42 s) and other study-specific MR sequences were acquired. Among the study-specific MR sequences were a 3D volumetric sagittal T1-weighted BRAVO sequence [acquisition details: T_E 3.2 ms, T_R 8.5 ms, inversion time (T_I) 450 ms, α 12°, receiver bandwidth ± 31.2 kHz, number of exicatations (NEX) 1, voxel size 1 mm × 1 mm × 1 mm] and a 3D sagittal T2-weighted CUBE FLAIR sequence [acquisition details: T_E 137 ms, number of echoes 1, T_R 8,500 ms, T_I 50 ms, receiver bandwidth ±31.25 kHz, NEX 1, voxel size 1.2 mm × 1.3 mm × 1.4 mm] The T1-weighted and FLAIR MR sequences were used to define regions of interest in the brain in our analysis. In all cases, a standard head coil (8-channel HR brain, GE Healthcare) was used for the MR acquisitions. This coil was validated to ensure that the achievable transmit to receive switching time was short enough for zero echo time imaging purposes.

All subjects underwent a PET/CT acquisition before ([¹⁸F]FDG cases) or after ([¹⁸F]PE2I cases) the PET/MR acquisition. The PET/CT acquisitions were performed on a Siemens Biograph 16 or with a Siemens Biograph 40 (Siemens Healthcare, Erlangen, Germany) PET/CT. All PET/CT examinations included a low-dose CT acquisition (120 kV, 11 mAs) which was used to generate a CT-based attenuation image taken as the ground truth in the study.

The PET raw data from all PET/MR acquisitions were reconstructed with three different methods for attenuation correction, shown in Figure 1. First, a GE atlas-based attenuation image (default method in the SIGNA PET/MR software version MP24) was used to reconstruct PETAtlasAC. Subsequently, a GE ZTE-based attenuation image and a co-registered CT-based attenuation image were used to reconstruct PETZTEAC and PETCTAC, respectively. The generation of all attenuation images is described in detail in the following subsection. The reconstructions of the static PET data sets were performed offline with the GE reconstruction toolbox v.1.28 (GE Healthcare, Chicago, US) using time of flight ordered subset maximum likelihood expectation maximization (TOF OSEM) with 4 iterations and 28 subsets, a voxel size of 1.17 mm × 1.17 mm × 2.78 mm, and a Gaussian post-smoothing with an FWHM of 4 mm.

Reconstruction of the dynamic [¹⁸F]PE2I PET data sets was performed on the scanner (software version MP24.R03). The acquired listmode data were split into 32 frames (frame length 10–360 s). All frames were reconstructed with TOF OSEM with 4 iterations and 28 subsets, a voxel size of 1.56 mm × 1.56 mm × 2.78 mm and a Gaussian post-smoothing with an FWHM of 3 mm.

First, the atlas-based attenuation images were generated with the GE reconstruction toolbox v.1.28 which uses the same post-processing algorithm as implemented in the software release of the SIGNA PET/MR (MP24.R03). The algorithm uses a non-rigid registration of an input in-phase LAVA flex MR image to an atlas of predefined attenuation images [30, 31]. The resulting atlas-based attenuation images are post-smoothed with a Gaussian kernel with FWHM ca. 10 mm.

Second, the ZTE-based attenuation images were generated by post-processing the ZTE MR images with a research tool provided by GE (v.1.6.2). Software release MP26 of the SIGNA PET/MR contains an option to use this algorithm for ZTE-based AC. The ZTE post-processing algorithm identifies bone voxels based on the ZTE image intensity and assigns continuous bone attenuation values [19–21]. The bone segmentation in the ZTE post-processing is completely model-free. To avoid misclassifications of air, tissue and bone in the nasal region, the ZTE post-processing algorithm v.1.6.2 uses the sinus/edge correction evaluated in Yang et al. [24]. Details of the ZTE post-processing algorithm that was used in our analysis are given in Delso et al. [27].

The vendor-provided ZTE post-processing has several input parameters. For all parameters but one (the partial volume slope) we used the default values suggested by the vendor. We used a value of 2 for the parameter for the partial volume slope which was obtained based on an evaluation of the results of the first 15 static subjects. The main influence of the partial volume slope parameter that we observed was a change in the size of the outer contour of the head (transition between background air and soft tissue of the skin). By changing the partial volume slope we obtained better agreement with the size of the outer contours derived from the CT-based attenuation images. When using the default partial volume slope of 1, the outer contour of the head is dilated by 1 voxel (2.4 mm) compared to using a partial volume slope of 2. This in turn led to a small global positive bias of 3%. All ZTE-based attenuation images were post-smoothed with a Gaussian kernel with FWHM 4 mm.

Third, after automatically removing the patient bed and cushions, the low-dose CT images from the PET/CT acquisition were rigidly co-registered to the in-phase LAVA flex MR. Subsequently, the Hounsfield units of the co-registered CT were scaled to 511-keV attenuation coefficients by using the GE-provided multi-linear scaling. We verified that the Siemens and GE scaling for 120 kV are virtually identical up to 1,200 HU (where GE decreases the slope while Siemens does not). After adding the templates for the PET/MR patient table and the head coil, a CT-based attenuation image that could be used to reconstruct the PET/MR raw data was obtained. All CT-based attenuation images were also post-smoothed with a Gaussian kernel with FWHM 4 mm.

The axial field of view (FOV) of the ZTE MR (limited by the sensitivity of the head coil) and the one of the CT was slightly smaller than the axial FOV of the PET detector rings in the SIGNA PET/MR. To complete areas in the neck and shoulders where ZTE or CT image information was not available, a simple segmentation-based two class attenuation image based on the LAVA flex MR image was used.

For all static acquisitions the mean uptake in 85 anatomical volumes of interest (VOIs) was calculated in PET_AtlasAC, PET_ZTEAC, and PET_CTAC. The VOIs were defined in the neuro tool of PMOD v.3.8 (PMOD technologies LCC, Zurich, Switzerland) using the Hammers atlas [32]. In every VOI we calculated the fractional bias of the mean uptake as

where a_CTAC(VOI) is the mean uptake of the VOI in PETCTAC that was used as the gold standard and a_AtlasAC(VOI) and a_ZTEAC(VOI) are the mean uptake of the VOI in PET_AtlasAC and PET_ZTEAC, respectively. In three subjects (6, 7, and 21), the caudal end of the occipital skull was not completely in the FOV in the attenuation CT. In those subjects, the cerebellum VOIs were excluded from the analysis. All VOIs were grouped according to their anatomical location into the following groups: frontal cortex, temporal cortex, parietal cortex, occipital cortex, medial temporal cortex, striatum, thalamus, cerebellum, and cerebral white matter. All VOIs and the assigned groups (regions) are listed in Tables S1, S2. A Wilcoxon signed-rank test was used to test whether the subject averaged mean of a_AtlasAC and a_ZTEAC is different from a_CTAC in all VOIs and regions.

To analyze the robustness of the vendor-provided atlas-based and ZTE-based attenuation correction, we applied the metric proposed in the multi-center evaluation of Ladefoged et al. [17]. This metric calculates the fraction of subjects in which the MRAC-introduced voxel bias of at least a given fraction of brain voxels is within ±5, ±10, ±15%. As mentioned in Ladefoged et al. [17], for a perfect AC method, the bias in the whole brain of all subjects would be within ±0%. The results of this metric were visualized in a characteristic curve for the three bias thresholds ±5, ±10, ±15%. As in Ladefoged et al. [17], we also analyzed three subjects with the biggest fraction of voxels exceeding a bias of ± 10%.

Regional time activity curves (TACs) were extracted for the left and right caudate nucleus, left and right putamen, and the cortex of the cerebellum. All VOIs were defined based on the 3D T1 BRAVO MR image using the Freesurfer image analysis suite which is documented and freely available online (http://surfer.nmr.mgh.harvard.edu/) [33]. Subsequently, we used the simplified reference tissue model (SRTM) with the cerebellar grey matter as reference region to estimate binding potential values (BP_nd) in the four striatal VOIs.

As proposed in Lammertsma and Hume [34] and validated for [¹⁸F]PE2I in Sasaki et al. [35], the tissue response C_t(t) was modeled as

where C_r(t) is the TAC of the reference tissue (the cerebellum), R₁ is the ratio between K₁ of the tissue and reference tissue, and * denotes the convolution operator. The parameters R₁, k₂, and BP_nd were obtained with non-linear curve fitting using the python package lmfit (v.0.9.7).

Similar to Equations (1) and (2), we calculated the bias of BP_nd, R₁, and k₂ in the four striatal VOIs for PET_AtlasAC and PET_ZTEAC compared to PET_CTAC. In all VOIs, a Wilcoxon signed-rank test was used to test whether the subject averaged mean of BP_nd,AtlasAC and BP_nd,ZTEAC differ from BP_nd,CTAC.

Figures 2, 3 and Tables S3, S4 show the results for the regional bias in the static [¹⁸F]FDG reconstructions caused by ZTEAC and AtlasAC compared to CTAC on a subject and regional level, respectively. Globally the bias ranges from −31.6 to +16.6% with a mean of −0.4% and a standard deviation of 4.3% for PETAtlasAC. For PETZTEAC the bias ranges from −8.0 to +7.7% with a mean of 0.1% and a standard deviation of 2.0%. Excluding the outliers based on the boxplot shown in Figure 3 reduces the global bias range to −12 to +14% for PET_AtlasAC and to −5.5 to +5.5% for PET_ZTEAC.

On a subject level, Figure 2 and Table S3 demonstrate that ZTEAC strongly reduces the inter- and intra-subject variability in the bias. In subject 24 where the non-rigid alignment to the atlas failed, PET_AtlasAC showed severe negative bias of more than −25% in the orbitofrontal cortical VOIs (see Figure 5). In these VOIs of subject 24, the bias of PET_ZTEAC was <1.6%.

On a regional level, Figure 3 and Table S4 show that ZTEAC strongly reduces the inter- and intra-regional variability in the bias, as well as the mean bias in the frontal cortex, temporal cortex, parietal cortex, medial temporal cortex, cerebellum, and cerebral white matter. In all regions shown in Figure 2, the mean bias in PET_ZTEAC is between −1.2 and +0.6%. PETAtlasAC shows a distinct negative bias in the cerebellum (mean −4.2%) and distinct positive bias in the parietal cortex (mean +4%).

On a VOI level, Figure 4 and Tables S1, S2 demonstrate that ZTEAC strongly reduces the inter- and intra-VOI variability in the bias, as well as the mean bias in almost all VOIs. The mean VOI bias caused by ZTEAC ranges from −1.8% in the lateral remainder of the left occipital lobe to +2.2% in the left lateral ventricle. In PET_AtlasAC, a distinct gradient in the mean VOI bias in the cranio-caudal direction is visible. The mean VOI bias caused by AtlasAC ranges from −4.5% in the cerebral white matter to +6.6% in the left superior frontal gyrus. In PET_ZTEAC, only 1.4% of the analyzed VOIs in all subjects had a bias of more than 5% whereas in PET_AtlasAC 20.3% of all VOIs showed a bias of more than 5%.

Figure 6 shows the results of the outlier metric [17] for biases within (±5, ±10, ±15%). Again, the performance of ZTE-based attenuation correction is much better than the one of the atlas-based attenuation correction. At least 95/77% of all brain voxels in all subjects show a bias within ±10% for PET_ZTEAC/PET_AtlasAC. For a bias within ±5% the corresponding values are 82/46% and for a bias within ±15% the corresponding values are 97/89%. Table 1 shows the results of the three worst outliers in terms of subjects with highest voxel bias, highest VOI bias and highest fraction of the brain exceeding a bias of ±10%.

Figure 7 and Table S5 summarize the bias in the modeled BP_nd in four different regions of the striatum using the cerebellum as reference region and TACs derived from PET_AtlasAC and PET_ZTEAC compared to TACs from PET_CTAC. The bias in the BP_nd ranges from −0.6% (right putamen in subject 11) to +9.4% (left caudate nucleus subject 9) for PET_AtlasAC and from −0.8% (right putamen subject 3) to +4.8% (right caudate nucleus subject 8) for PET_ZTEAC. The right caudate nucleus shows the biggest subject averaged regional bias (5.1±2.6%, p = 0.003 for PET_AtlasAC and 2.0±1.5%, p = 0.006 for PET_ZTEAC). In addition, Figures S1–S3 show the bias in the time activity curves, and the R₁ and k₂ estimates in PET_AtlasAC and PET_ZTEAC, respectively.

GD is an employee of GE Healthcare, Cambridge, UK. GS, MK, AR, NM, RP, JN, and KV received travel grants from GE Healthcare for attending workshops. GS and KV received travel grants for presenting PET/MR results at GE user meetings. KU Leuven received a research collaboration grant for partial funding of this study. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.