Longitudinal changes in 18F-Flutemetamol amyloid load in cognitively intact APOE4 carriers versus noncarriers: Methodological considerations

PURPOSE

Measuring longitudinal changes in amyloid load in the asymptomatic stage of Alzheimer's disease is of high relevance for clinical research and progress towards more efficacious, timely treatments. Apolipoprotein E ε4 (APOE4) has a well-established effect on the rate of amyloid accumulation. Here we investigated which region of interest and which reference region perform best at detecting the effect of APOE4 on longitudinal amyloid load in individuals participating in the Flemish Prevent Alzheimer's Disease Cohort KU Leuven (F-PACK).

METHODS

Ninety cognitively intact F-PACK participants (baseline age: 68 (52-80) years, 46 males, 42 APOE4 carriers) received structural MRI and ¹⁸F-Flutemetamol PET scans at baseline and follow-up (6.2 (3.4-10.9) year interval). Standardised uptake value ratios (SUVRs) and Centiloids (CLs) were calculated in a composite cortical volume of interest (SUVR_comp/CL) and in the precuneus (SUVR_prec), and amyloid rate of change derived: (follow-up amyloid load - baseline amyloid load) / time interval (years). Four reference regions were used to derive amyloid load: whole cerebellum, cerebellar grey matter, eroded subcortical white matter, and pons.

RESULTS

When using whole cerebellum or cerebellar grey matter as reference region, APOE4 carriers had a significantly higher SUVR_comp amyloid rate of change than non-carriers (p_corr = 0.004, t = 3.40 (CI 0.005-0.018); p_corr = 0.036, t = 2.66 (CI 0.003-0.018), respectively). Significance was not observed for eroded subcortical white matter or pons (p_corr = 0.144, t = 2.13 (CI 0.0003-0.008); p_corr = 0.116, t = 2.22 (CI 0.005-0.010), respectively). When using CLs as the amyloid measurement, and whole cerebellum, APOE4 carriers had a higher amyloid rate of change than non-carriers (p_corr = 0.012, t = 3.05 (CI 0.499-2.359)). Significance was not observed for the other reference regions. No significance was observed with any of the reference regions and amyloid rate of change in the precuneus (SUVR_prec).

CONCLUSION

In this cognitively intact cohort, a composite neocortical volume of interest together with whole cerebellum or cerebellar grey matter as reference region are the methods of choice for detecting APOE4-dependent differences in amyloid rate of change.

Influence of common reference regions on regional tau patterns in cross-sectional and longitudinal [18F]-AV-1451 PET data

Tau PET has allowed for critical insights into in vivo patterns of tau accumulation and change in individuals early in the Alzheimer's disease (AD) continuum. A key methodological step in tau PET analyses is the selection of a reference region, but there is not yet consensus on the optimal region especially for longitudinal tau PET analyses. This study examines how reference region selection influences results related to disease stage at baseline and over time. Longitudinal flortaucipir ([18F]-AV1451) PET scans were examined using several common reference regions (e.g., eroded subcortical white matter, inferior cerebellar gray matter) in 62 clinically unimpaired amyloid negative (CU A-) individuals, 73 CU amyloid positive (CU A+) individuals, and 64 amyloid positive individuals with mild cognitive impairment (MCI A+) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Cross-sectionally, both reference regions resulted in robust group differences between CU A-, CU A+, and MCI A+ groups, along with significant associations with CSF phosphorylated tau (pTau-181). However, these results were more focally specific and akin to Braak Staging when using eroded white matter, whereas effects with inferior cerebellum were globally distributed across most cortical regions. Longitudinally, utilization of eroded white matter revealed significant accumulation greater than zero across more regions whereas change over time was diminished using inferior cerebellum. Interestingly, the inferior temporal target region seemed most robust to reference region selection with expected cross-sectional and longitudinal signal across both reference regions. With few exceptions, baseline tau did not significantly predict longitudinal change in tau in the same region regardless of reference region. In summary, reference region selection deserves further evaluation as this methodological step may lead to disparate findings. Inferior cerebellar gray matter may be more sensitive to cross-sectional flortaucipir differences, whereas eroded subcortical white matter may be more sensitive for longitudinal analyses examining regional patterns of change.

Selecting software pipelines for change in flortaucipir SUVR: Balancing repeatability and group separation

Since tau PET tracers were introduced, investigators have quantified them using a wide variety of automated methods. As longitudinal cohort studies acquire second and third time points of serial within-person tau PET data, determining the best pipeline to measure change has become crucial. We compared a total of 415 different quantification methods (each a combination of multiple options) according to their effects on a) differences in annual SUVR change between clinical groups, and b) longitudinal measurement repeatability as measured by the error term from a linear mixed-effects model. Our comparisons used MRI and Flortaucipir scans of 97 Mayo Clinic study participants who clinically either: a) were cognitively unimpaired, or b) had cognitive impairments that were consistent with Alzheimer's disease pathology. Tested methods included cross-sectional and longitudinal variants of two overarching pipelines (FreeSurfer 6.0, and an in-house pipeline based on SPM12), three choices of target region (entorhinal, inferior temporal, and a temporal lobe meta-ROI), five types of partial volume correction (PVC) (none, two-compartment, three-compartment, geometric transfer matrix (GTM), and a tau-specific GTM variant), seven choices of reference region (cerebellar crus, cerebellar gray matter, whole cerebellum, pons, supratentorial white matter, eroded supratentorial WM, and a composite of eroded supratentorial WM, pons, and whole cerebellum), two choices of region masking (GM or GM and WM), and two choices of statistic (voxel-wise mean vs. median). Our strongest findings were: 1) larger temporal-lobe target regions greatly outperformed entorhinal cortex (median sample size estimates based on a hypothetical clinical trial were 520-526 vs. 1740); 2) longitudinal processing pipelines outperformed cross-sectional pipelines (median sample size estimates were 483 vs. 572); and 3) reference regions including supratentorial WM outperformed traditional cerebellar and pontine options (median sample size estimates were 370 vs. 559). Altogether, our results favored longitudinally SUVR methods and a temporal-lobe meta-ROI that includes adjacent (juxtacortical) WM, a composite reference region (eroded supratentorial WM + pons + whole cerebellum), 2-class voxel-based PVC, and median statistics.

In vitro and In vivo Assessment of Suitable Reference Region and Kinetic Modelling for the mGluR1 Radioligand [11C]ITDM in Mice

PURPOSE

This study aimed at investigating binding specificity, suitability of reference region-based kinetic modelling, and pharmacokinetics of the metabotropic glutamate receptor 1 (mGluR1) radioligand [¹¹C]ITDM in mice.

PROCEDURES

We performed in vivo blocking as well as displacement of [¹¹C]ITDM during positron emission tomography (PET) imaging using the specific mGluR1 antagonist YM-202074. Additionally, we assessed in vitro blocking of [³H]ITDM at two different doses of YM-202074. As an alternative to reference region models, we validated the use of a noninvasive image-derived input function (IDIF) compared to an arterial input function measured with an invasive arteriovenous (AV) shunt using a population-based curve for radiometabolite correction and characterized the pharmacokinetic modelling of [¹¹C]ITDM in the mouse brain. Finally, we also assessed semi-quantitative approaches.

RESULTS

In vivo blocking with YM-202074 resulted in a decreased [¹¹C]ITDM binding, ranging from - 35.8 ± 8.0 % in pons to - 65.8 ± 3.0 % in thalamus. Displacement was also markedly observed in all tested regions. In addition, in vitro [³H]ITDM binding could be blocked in a dose-dependent manner. The volume of distribution (V_T) based on the noninvasive IDIF (V_{T (IDIF)}) showed excellent agreement with the V_T values based on the metabolite-corrected plasma input function regardless of the metabolite correction (r² > 0.943, p < 0.0001). Two-tissue compartmental model (2TCM) was found to be the preferred model and showed optimal agreement with Logan plot (r² > 0.960, p < 0.0001). A minimum scan duration of 80 min was required for proper parameter estimation. SUV was not reliable (r² = 0.379, p = 0.0011), unlike the SUV ratio to the SUV of the input function, which showed to be a valid approach.

CONCLUSIONS

No suitable reference region could be identified for [¹¹C]ITDM as strongly supported by in vivo and in vitro evidence of specific binding in all brain regions. However, by applying appropriate kinetic models, [¹¹C]ITDM PET imaging represents a promising tool to visualize mGluR1 in the mouse brain.

Reduced [18F]flortaucipir retention in white matter hyperintensities compared to normal-appearing white matter

PURPOSE

Recent research has suggested the use of white matter (WM) reference regions for longitudinal tau-PET imaging. However, tau tracers display affinity for the β-sheet structure formed by myelin, and thus WM lesions might influence tracer retention. Here, we explored whether the tau-sensitive tracer [¹⁸F]flortaucipir shows reduced retention in WM hyperintensities (WMH) and how this retention changes over time.

METHODS

We included 707 participants from the Alzheimer's Disease Neuroimaging Initiative with available [¹⁸F]flortaucipir-PET and structural and FLAIR MRI scans. WM segments and WMH were automatically delineated in the structural MRI and FLAIR scans, respectively. [¹⁸F]flortaucipir standardized uptake value ratios (SUVR) of WMH and normal-appearing WM (NAWM) were calculated using the inferior cerebellar grey matter as reference region, and a 3-mm erosion was applied to the combined NAWM and WMH masks to avoid partial volume effects. Longitudinal [¹⁸F]flortaucipir SUVR changes in NAWM and WMH were estimated using linear mixed models. The percent variance of WM-referenced cortical [¹⁸F]flortaucipir SUVRs explained by longitudinal changes in the WM reference region was estimated with the R² coefficient.

RESULTS

Compared to NAWM, WMH areas displayed significantly reduced [¹⁸F]flortaucipir SUVR, independent of cognitive impairment or Aβ status (mean difference = 0.14 SUVR, p < 0.001). Older age was associated with lower [¹⁸F]flortaucipir SUVR in both NAWM (- 0.002 SUVR/year, p = 0.005) and WMH (- 0.004 SUVR/year, p < 0.001). Longitudinally, [¹⁸F]flortaucipir SUVR decreased in NAWM (- 0.008 SUVR/year, p = 0.03) and even more so in WMH (- 0.02 SUVR/year, p < 0.001). Between 17% and 66% of the variance of longitudinal changes in cortical WM-referenced [¹⁸F]flortaucipir SUVRs were explained by longitudinal changes in the reference region.

CONCLUSIONS

[¹⁸F]flortaucipir retention in the WM decreases over time and is influenced by the presence of WMH, supporting the hypothesis that [¹⁸F]flortaucipir retention in the WM is partially myelin-dependent. These findings have implications for the use of WM reference regions for [¹⁸F]flortaucipir-PET imaging.

Automatic delineation algorithm of reference region for amyloid imaging based on kinetics

OBJECTIVE

This study aims to develop an algorithm named AutoRef to delineate a reference region for quantitative PET amyloid imaging.

METHODS

AutoRef sets the reference region automatically using a distinguishing feature in the kinetics of reference region. This is reflected in the shapes of the tissue time activity curve. A statistical shape recognition algorithm of the gaussian mixture model is applied with considering spatial and temporal information on a reference region. We evaluate the BP_ND with manually set reference region and AutoRef using 86 cases (43 positive cases, 10 equivocal cases, and 33 negative cases) of dynamically scanned ¹¹C-Pittsburgh Compound-B.

RESULTS

From the Bland-Altman plot, the difference between two BP_ND is 0.099 ± 0.21 as standard deviation, and no significant systematic error is observed between the BP_ND with AutoRef and with manual definition of a reference region. Although a proportional error is detected, it is smaller than the 95% limits of agreement. Therefore, the proportional error is negligibly small.

CONCLUSIONS

AutoRef presents the same performance as the manual definition of the reference region. Further, since AutoRef is more algorithmic than the ordinary manual definition of the reference region, there are few operator-oriented uncertainties in AutoRef. We thus conclude that AutoRef can be applied as an automatic delineating algorithm for the reference region in amyloid imaging.

Reference region extraction by clustering for the pharmacokinetic analysis of dynamic contrast-enhanced MRI in prostate cancer

PURPOSE

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) measures changes in the concentration of an administered contrast agent to quantitatively evaluate blood circulation in a tumor or normal tissues. This method uses a pharmacokinetic analysis based on the time course of a reference region, such as muscle, rather than arterial input function. However, it is difficult to manually define a homogeneous reference region. In the present study, we developed a method for automatic extraction of the reference region using a clustering algorithm based on a time course pattern for DCE-MRI studies of patients with prostate cancer.

METHODS

Two feature values related to the shape of the time course were extracted from the time course of all voxels in the DCE-MRI images. Each voxel value of T1-weighted images acquired before administration were also added as anatomical data. Using this three-dimensional feature vector, all voxels were segmented into five clusters by the Gaussian mixture model, and one of these clusters that included the gluteus muscle was selected as the reference region.

RESULTS

Each region of arterial vessel, muscle, and fat was segmented as a different cluster from the tumor and normal tissues in the prostate. In the extracted reference region, other tissue elements including scattered fat and blood vessels were removed from the muscle region.

CONCLUSIONS

Our proposed method can automatically extract the reference region using the clustering algorithm with three types of features based on the time course pattern and anatomical data. This method may be useful for evaluating tumor circulatory function in DCE-MRI studies.

Test-retest reliability and convergent validity of (R)-[11C]PK11195 outcome measures without arterial input function

Purpose

The PET radioligand (R)-[¹¹C]PK11195 is used to quantify the 18-kDa translocator protein (TSPO), a marker for glial activation. Since there is no brain region devoid of TSPO, an arterial input function (AIF) is ideally required for quantification of binding. However, obtaining an AIF is experimentally demanding, is sometimes uncomfortable for participants, and can introduce additional measurement error during quantification. The objective of this study was to perform an evaluation of the test-retest reliability and convergent validity of techniques used for quantifying (R)-[¹¹C]PK11195 binding without an AIF in clinical studies.

Methods

Data from six healthy individuals who participated in two PET examinations, 6 weeks apart, were analyzed. Regional non-displaceable binding potential (BP_ND) values were calculated using the simplified reference tissue model, with either cerebellum as reference region or a reference input derived using supervised cluster analysis (SVCA). Standardized uptake values (SUVs) were estimated for the time interval of 40–60 min.

Results

Test-retest reliability for BP_ND estimates were poor (80% of ICCs < 0.5). BP_ND estimates derived without an AIF were not correlated with BP_ND, total or specific distribution volume from the 2TCM using an AIF (all R² < 12%). SUVs showed moderate reliability but no correlation to any other outcome measure.

Conclusions

Caution is warranted when interpreting patient-control comparisons employing (R)-[¹¹C]PK11195 outcome measures obtained without an AIF.

Electronic supplementary material

The online version of this article (10.1186/s13550-018-0455-8) contains supplementary material, which is available to authorized users.

Accuracy and reliability of [11C]PBR28 specific binding estimated without the use of a reference region

[¹¹C]PBR28 is a positron emission tomography radioligand used to examine the expression of the 18 kDa translocator protein (TSPO). TSPO is located in glial cells and can function as a marker for immune activation. Since TSPO is expressed throughout the brain, no true reference region exists. For this reason, an arterial input function is required for accurate quantification of [¹¹C]PBR28 binding and the most common outcome measure is the total distribution volume (V_T). Notably, V_T reflects both specific binding and non-displaceable binding. Therefore, estimates of specific binding, such as binding potential (e.g. BP_ND) and specific distribution volume (V_S) should theoretically be more sensitive to underlying differences in TSPO expression. It is unknown, however, if unbiased and accurate estimates of these outcome measures are obtainable for [¹¹C]PBR28. The Simultaneous Estimation (SIME) method uses time-activity-curves from multiple brain regions with the aim to obtain a brain-wide estimate of the non-displaceable distribution volume (V_ND), which can subsequently be used to improve the estimation of BP_ND and V_S. In this study we evaluated the accuracy of SIME-derived V_ND, and the reliability of resulting estimates of specific binding for [¹¹C]PBR28, using a combination of simulation experiments and in vivo studies in healthy humans. The simulation experiments, based on data from 54 unique [¹¹C]PBR28 examinations, showed that V_ND values estimated using SIME were both precise and accurate. Data from a pharmacological competition challenge (n = 5) showed that SIME provided V_ND values that were on average 19% lower than those obtained using the Lassen plot, but similar to values obtained using the Likelihood-Estimation of Occupancy technique. Test-retest data (n = 11) showed that SIME-derived V_S values exhibited good reliability and precision, while larger variability was observed in SIME-derived BP_ND values. The results support the use of SIME for quantifying specific binding of [¹¹C]PBR28, and suggest that V_S can be used in complement to the conventional outcome measure V_T. Additional studies in patient cohorts are warranted.

Cerebellar heterogeneity and its impact on PET data quantification of 5-HT receptor radioligands

In the quantification of positron emission tomography (PET) radiotracer binding, a commonly used method is reference tissue modeling (RTM). RTM necessitates a proper reference and a ubiquitous choice for G-protein coupled receptors is the cerebellum. We investigated regional differences in uptake within the grey matter of the cerebellar hemispheres (CH), the cerebellar white matter (CW), and the cerebellar vermis (CV) for five PET radioligands targeting the serotonin system. Furthermore, we evaluated the impact of choosing different reference regions when quantifying neocortical binding. The PET and MR images are part of the Cimbi database: 5-HT1AR ([11C]CUMI-101, n = 8), 5-HT1BR ([11C]AZ10419369, n = 36), 5-HT2AR ([11C]Cimbi-36, n = 29), 5-HT4R ([11C]SB207145, n = 59), and 5-HTT ([11C]DASB, n = 100). We employed SUIT and FreeSurfer to delineate CV, CW, and CH and quantified mean standardized uptake values (SUV) and nondisplaceable neocortical binding potential (BPND). Statistical difference was assessed with paired nonparametric two-sided Wilcoxon signed-rank tests and multiple comparison corrected via false discovery rate. We demonstrate significant radioligand specific regional differences in cerebellar uptake. These differences persist when using different cerebellar regions for RTM, but the influence on the neocortical BPND is small. Nevertheless, our data highlight the importance of validating each radioligand carefully for defining the optimal reference region.

Use of white matter reference regions for detection of change in florbetapir positron emission tomography from completed phase 3 solanezumab trials

INTRODUCTION

We compared subject-specific white matter (SSWM) and whole cerebellum (CBL) reference regions for power to detect longitudinal change in amyloid positron emission tomography signal.

METHODS

Positive florbetapir positron emission tomography scans were analyzed from participants (66 placebo treated and 63 solanezumab treated) with mild dementia caused by Alzheimer's disease from the EXPEDITION and EXPEDITION2 studies. For comparison to CBL, a second normalization was performed on longitudinal data using an SSWM correction factor (SSWM normalization ratio [SSWMnr]). Analysis of covariance assessed baseline to 18-month change between treatment with solanezumab and placebo. Sample and effect size estimations provided magnitude of observed treatment changes.

RESULTS

Longitudinal percent change between placebo and solanezumab using CBL was not significant (P = .536) but was significant for SSWMnr (P = .042). Compared with CBL, SSWMnr technique increased the power to detect a treatment difference, more than tripling the effect size and reducing the sample size requirements by 85% to 90%.

DISCUSSION

Adjusting longitudinal standardized uptake value ratios with an SSWM reference region in these antiamyloid treatment trials increased mean change detection and decreased variance resulting in the substantial improvement in statistical power to detect change.

Reference region selection and the association between the rate of amyloid accumulation over time and the baseline amyloid burden

Relative quantitative analysis of amyloid plaque burden in Alzheimer's disease (AD) patients can be reported as standardized uptake value ratio (SUVR) from positron emission tomography (PET). Here, the SUVR is the ratio of the mean amyloid radioligand retention in a composite (COMP) neocortical volume of interest (VOI) to that in a reference VOI, such as the cerebellum, brainstem (BST)/pons, or white matter (WM). Some longitudinal PET investigations show that the rate of amyloid accumulation to follow-up has an inverted U relationship with baseline amyloid SUVR relative to cerebellar or brainstem/pons reference VOIs. The corresponding association with SUVR relative to WM is unknown. To test the possible benefits of WM normalization, we analyzed [¹⁸F]-AV45 PET data from 404 subjects in the AD Neuroimaging Initiative (ADNI) database at baseline and 2-year follow-up (144 cognitively normal controls, 225 patients with mild cognitive impairment, and 35 AD patients). Reference regions included subcortical WM as well as conventional cerebellar gray matter (CBL), and BST. We tested associations between each subject's inter-session change (∆) of SUVR and their baseline SUVR by applying linear, logarithmic, and quadratic regression analyses. Unscaled standardized uptake values (SUVs) were correlated between VOIs at baseline and follow-up, and within VOIs in the longitudinal run. The association between ∆SUVR and baseline SUVR relative to WM reference was best described by an inverted U-shaped function. Correlation analyses demonstrated a high regional and temporal correlation between COMP and WM VOI SUVs. For WM normalization, we confirm that the rate of amyloid accumulation over time follows an inverted U-shaped function of baseline amyloid burden. Reference region selection, however, has substantial effects on SUVR results. This reflects the extent of covariance between SUVs in the COMP VOI and those in the various reference VOIs. We speculate that PET labeling of amyloid deposition within target regions is partially confounded by effects of longitudinal changes of cerebral blood flow (CBF) on tracer delivery. Indeed, CBF may be the leading factor influencing longitudinal SUV changes. We suggest that SUVR relative to WM may be more robust to changes in CBF, and thus fitter for sensitive detection of amyloid accumulation in intervention studies.

Estimation of in vivo nonspecific binding in positron emission tomography studies without requiring a reference region

Estimation of outcome measures in in vivo neuroreceptor mapping with positron emission tomography (PET) commonly depends on an assumption of uniform nondisplaceable binding throughout the brain. In many cases, this can be estimated based on data from a "reference region," an area of the brain devoid of the receptor of interest. However, often such a region does not exist, as there are some receptors everywhere throughout the brain. Erroneously designating a region as a "reference" can lead to biased estimation, and furthermore, if the level of specific binding in the purported reference region differs between comparison groups, the validity of resulting conclusions may be called into question. We present a method for estimation of all common PET outcome measures that can provide good estimates even when no reference region exists. Our aim is to use information from several regions simultaneously to estimate the information common to all regions. By not requiring specification (or validation) of a reference region, such an approach can provide an automated, objective approach for kinetic modeling of PET data. We illustrate the performance of these methods on simulated data, human [(11)C]WAY-100635 data, and [(11)C]CUMI-101 blocking data in baboons. We show close agreement between estimates obtained by using the proposed method (which does not require a reference region) and estimates based on either a reference region or a blocking study.

Improved longitudinal [(18)F]-AV45 amyloid PET by white matter reference and VOI-based partial volume effect correction

Amyloid positron-emission-tomography (PET) offers an important research and diagnostic tool for investigating Alzheimer's disease (AD). The majority of amyloid PET studies have used the cerebellum as a reference region, and clinical studies have not accounted for atrophy-based partial volume effects (PVE). Longitudinal studies using cerebellum as reference tissue have revealed only small mean increases and high inter-subject variability in amyloid binding. We aimed to test the effects of different reference regions and PVE-correction (PVEC) on the discriminatory power and longitudinal performance of amyloid PET. We analyzed [(18)F]-AV45 PET and T1-weighted MRI data of 962 subjects at baseline and two-year follow-up data of 258 subjects. Cortical composite volume-of-interest (VOI) values (COMP) for tracer uptake were generated using either full brain atlas VOIs, gray matter segmented VOIs or gray matter segmented VOIs after VOI-based PVEC. Standard-uptake-value ratios (SUVR) were calculated by scaling the COMP values to uptake in cerebellum (SUVRCBL), brainstem (SUVRBST) or white matter (SUVRWM). Mean SUV, SUVR, and changes after PVEC were compared at baseline between diagnostic groups of healthy controls (HC; N=316), mild cognitive impairment (MCI; N=483) and AD (N=163). Receiver operating characteristics (ROC) were calculated for the discriminations between HC, MCI and AD, and expressed as area under the curve (AUC). Finally, the longitudinal [(18)F]-AV45-PET data were used to analyze the impact of quantitation procedures on apparent changes in amyloid load over time. Reference region SUV was most constant between diagnosis groups for the white matter. PVEC led to decreases of COMP-SUV in HC (-18%) and MCI (-10%), but increases in AD (+7%). Highest AUCs were found when using PVEC with white matter scaling for the contrast between HC/AD (0.907) or with brainstem scaling for the contrast between HC/MCI (0.658). Longitudinal increases were greatest in all diagnosis groups with application of PVEC, and inter-subject variability was lowest for the white matter reference. Thus, discriminatory power of [(18)F]-AV45-PET was improved by use of a VOI-based PVEC and white matter or brainstem rather than cerebellum reference region. Detection of longitudinal amyloid increases was optimized with PVEC and white matter reference tissue.