Generative Adversarial Network-Enhanced Ultra-Low-Dose [18F]-PI-2620 τ PET/MRI in Aging and Neurodegenerative Populations.

BACKGROUND AND PURPOSE: With the utility of hybrid τ PET/MR imaging in the screening, diagnosis, and follow-up of individuals with neurodegenerative diseases, we investigated whether deep learning techniques can be used in enhancing ultra-low-dose [18F]-PI-2620 τ PET/MR images to produce diagnostic-quality images. MATERIALS AND METHODS: Forty-four healthy aging participants and patients with neurodegenerative diseases were recruited for this study, and [18F]-PI-2620 τ PET/MR data were simultaneously acquired. A generative adversarial network was trained to enhance ultra-low-dose τ images, which were reconstructed from a random sampling of 1/20 (approximately 5% of original count level) of the original full-dose data. MR images were also used as additional input channels. Region-based analyses as well as a reader study were conducted to assess the image quality of the enhanced images compared with their full-dose counterparts. RESULTS: The enhanced ultra-low-dose τ images showed apparent noise reduction compared with the ultra-low-dose images. The regional standard uptake value ratios showed that while, in general, there is an underestimation for both image types, especially in regions with higher uptake, when focusing on the healthy-but-amyloid-positive population (with relatively lower τ uptake), this bias was reduced in the enhanced ultra-low-dose images. The radiotracer uptake patterns in the enhanced images were read accurately compared with their full-dose counterparts. CONCLUSIONS: The clinical readings of deep learning-enhanced ultra-low-dose τ PET images were consistent with those performed with full-dose imaging, suggesting the possibility of reducing the dose and enabling more frequent examinations for dementia monitoring.

Investigating Simultaneity for Deep Learning-Enhanced Actual Ultra-Low-Dose Amyloid PET/MR Imaging.

BACKGROUND AND PURPOSE: Diagnostic-quality amyloid PET images can be created with deep learning using actual ultra-low-dose PET images and simultaneous structural MR imaging. Here, we investigated whether simultaneity is required; if not, MR imaging-assisted ultra-low-dose PET imaging could be performed with separate PET/CT and MR imaging acquisitions. MATERIALS AND METHODS: We recruited 48 participants: Thirty-two (20 women; mean, 67.7 [SD, 7.9] years) were used for pretraining; 328 (SD, 32) MBq of [18F] florbetaben was injected. Sixteen participants (6 women; mean, 71.4 [SD. 8.7] years of age) were scanned in 2 sessions, with 6.5 (SD, 3.8) and 300 (SD, 14) MBq of [18F] florbetaben injected, respectively. Structural MR imaging was acquired simultaneously with PET (90-110 minutes postinjection) on integrated PET/MR imaging in 2 sessions. Multiple U-Net-based deep networks were trained to create diagnostic PET images. For each method, training was done with the ultra-low-dose PET as input combined with MR imaging from either the ultra-low-dose session (simultaneous) or from the standard-dose PET session (nonsimultaneous). Image quality of the enhanced and ultra-low-dose PET images was evaluated using quantitative signal-processing methods, standardized uptake value ratio correlation, and clinical reads. RESULTS: Qualitatively, the enhanced images resembled the standard-dose image for both simultaneous and nonsimultaneous conditions. Three quantitative metrics showed significant improvement for all networks and no differences due to simultaneity. Standardized uptake value ratio correlation was high across different image types and network training methods, and 31/32 enhanced image pairs were read similarly. CONCLUSIONS: This work suggests that accurate amyloid PET images can be generated using enhanced ultra-low-dose PET and either nonsimultaneous or simultaneous MR imaging, broadening the utility of ultra-low-dose amyloid PET imaging.

Amyloid vs FDG-PET in the differential diagnosis of AD and FTLD.

OBJECTIVE: To compare the diagnostic performance of PET with the amyloid ligand Pittsburgh compound B (PiB-PET) to fluorodeoxyglucose (FDG-PET) in discriminating between Alzheimer disease (AD) and frontotemporal lobar degeneration (FTLD). METHODS: Patients meeting clinical criteria for AD (n = 62) and FTLD (n = 45) underwent PiB and FDG-PET. PiB scans were classified as positive or negative by 2 visual raters blinded to clinical diagnosis, and using a quantitative threshold derived from controls (n = 25). FDG scans were visually rated as consistent with AD or FTLD, and quantitatively classified based on the region of lowest metabolism relative to controls. RESULTS: PiB visual reads had a higher sensitivity for AD (89.5% average between raters) than FDG visual reads (77.5%) with similar specificity (PiB 83%, FDG 84%). When scans were classified quantitatively, PiB had higher sensitivity (89% vs 73%) while FDG had higher specificity (83% vs 98%). On receiver operating characteristic analysis, areas under the curve for PiB (0.888) and FDG (0.910) were similar. Interrater agreement was higher for PiB (κ = 0.96) than FDG (κ = 0.72), as was agreement between visual and quantitative classification (PiB κ = 0.88-0.92; FDG κ = 0.64-0.68). In patients with known histopathology, overall classification accuracy (2 visual and 1 quantitative classification per patient) was 97% for PiB (n = 12 patients) and 87% for FDG (n = 10). CONCLUSIONS: PiB and FDG showed similar accuracy in discriminating AD and FTLD. PiB was more sensitive when interpreted qualitatively or quantitatively. FDG was more specific, but only when scans were classified quantitatively. PiB slightly outperformed FDG in patients with known histopathology.

Distinct clinical and metabolic deficits in PCA and AD are not related to amyloid distribution.

BACKGROUND/OBJECTIVE: Patients with posterior cortical atrophy (PCA) often have Alzheimer disease (AD) at autopsy, yet are cognitively and anatomically distinct from patients with clinical AD. We sought to compare the distribution of ß-amyloid and glucose metabolism in PCA and AD in vivo using Pittsburgh compound B (PiB) and FDG-PET. METHODS: Patients with PCA (n = 12, age 57.5 ± 7.4, Mini-Mental State Examination [MMSE] 22.2 ± 5.1), AD (n = 14, age 58.8 ± 9.6, MMSE 23.8 ± 6.7), and cognitively normal controls (NC, n = 30, age 73.6 ± 6.4) underwent PiB and FDG-PET. Group differences in PiB distribution volume ratios (DVR, cerebellar reference) and FDG uptake (pons-averaged) were assessed on a voxel-wise basis and by comparing binding in regions of interest (ROIs). RESULTS: Compared to NC, both patients with AD and patients with PCA showed diffuse PiB uptake throughout frontal, temporoparietal, and occipital cortex (p < 0.0001). There were no regional differences in PiB binding between PCA and AD even after correcting for atrophy. FDG patterns in PCA and AD were distinct: while both groups showed hypometabolism compared to NC in temporoparietal cortex and precuneus/posterior cingulate, patients with PCA further showed hypometabolism in inferior occipitotemporal cortex compared to both NC and patients with AD (p < 0.05). Patients with AD did not show areas of relative hypometabolism compared to PCA. CONCLUSIONS: Fibrillar amyloid deposition in PCA is diffuse and similar to AD, while glucose hypometabolism extends more posteriorly into occipital cortex. Further studies are needed to determine the mechanisms of selective network degeneration in focal variants of AD.

Episodic memory loss is related to hippocampal-mediated beta-amyloid deposition in elderly subjects.

Although beta-amyloid (Abeta) plaques are a primary diagnostic criterion for Alzheimer's disease, this pathology is commonly observed in the brains of non-demented older individuals. To explore the importance of this pathology in the absence of dementia, we compared levels of amyloid deposition (via 'Pittsburgh Compound-B' (PIB) positron emission tomography (PET) imaging) to hippocampus volume (HV) and episodic memory (EM) in three groups: (i) normal controls (NC) from the Berkeley Aging Cohort (BAC NC, n = 20); (ii) normal controls (NC) from the Alzheimer's disease neuroimaging initiative (ADNI NC, n = 17); and (iii) PIB+ mild cognitive impairment subjects from the ADNI (ADNI PIB+ MCI, n = 39). Age, gender and education were controlled for in each statistical model, and HV was adjusted for intracranial volume (aHV). In BAC NC, elevated PIB uptake was significantly associated with smaller aHV (P = 0.0016) and worse EM (P = 0.0086). Within ADNI NC, elevated PIB uptake was significantly associated with smaller aHV (P = 0.047) but not EM (P = 0.60); within ADNI PIB+ MCI, elevated PIB uptake was significantly associated with both smaller aHV (P = 0.00070) and worse EM (P = 0.046). To further understand these relationships, a recursive regression procedure was conducted within all ADNI NC and PIB+ MCI subjects (n = 56) to test the hypothesis that HV mediates the relationship between Abeta and EM. Significant correlations were found between PIB index and EM (P = 0.0044), PIB index and aHV (P < 0.0001), as well as between aHV and EM (P < 0.0001). When both aHV and PIB were included in the same model to predict EM, aHV remained significant (P = 0.0015) whereas PIB index was no longer significantly associated with EM (P = 0.50). These results are consistent with a model in which Abeta deposition, hippocampal atrophy, and EM occur sequentially in elderly subjects, with Abeta deposition as the primary event in this cascade. This pattern suggests that declining EM in older individuals may be caused by Abeta-induced hippocampus atrophy.

11C-PIB PET imaging in Alzheimer disease and frontotemporal lobar degeneration.

BACKGROUND: The PET tracer (11)C-labeled Pittsburgh Compound-B ((11)C-PIB) specifically binds fibrillar amyloid-beta (Abeta) plaques and can be detected in Alzheimer disease (AD). We hypothesized that PET imaging with (11)C-PIB would discriminate AD from frontotemporal lobar degeneration (FTLD), a non-Abeta dementia. METHODS: Patients meeting research criteria for AD (n = 7) or FTLD (n = 12) and cognitively normal controls (n = 8) underwent PET imaging with (11)C-PIB (patients and controls) and (18)F-fluorodeoxyglucose ((18)F-FDG) (patients only). (11)C-PIB whole brain and region of interest (ROI) distribution volume ratios (DVR) were calculated using Logan graphical analysis with cerebellum as a reference region. DVR images were visually rated by a blinded investigator as positive or negative for cortical (11)C-PIB, and summed (18)F-FDG images were rated as consistent with AD or FTLD. RESULTS: All patients with AD (7/7) had positive (11)C-PIB scans by visual inspection, while 8/12 patients with FTLD and 7/8 controls had negative scans. Of the four PIB-positive patients with FTLD, two had (18)F-FDG scans that suggested AD, and two had (18)F-FDG scans suggestive of FTLD. Mean DVRs were higher in AD than in FTLD in whole brain, lateral frontal, precuneus, and lateral temporal cortex (p < 0.05), while DVRs in FTLD did not significantly differ from controls. CONCLUSIONS: PET imaging with (11)C-labeled Pittsburgh Compound-B ((11)C-PIB) helps discriminate Alzheimer disease (AD) from frontotemporal lobar degeneration (FTLD). Pathologic correlation is needed to determine whether patients with PIB-positive FTLD represent false positives, comorbid FTLD/AD pathology, or AD pathology mimicking an FTLD clinical syndrome.