A 3D convolutional neural network to classify subjects as Alzheimer's disease, frontotemporal dementia or healthy controls using brain 18F-FDG PET.

With the arrival of disease-modifying drugs, neurodegenerative diseases will require an accurate diagnosis for optimal treatment. Convolutional neural networks are powerful deep learning techniques that can provide great help to physicians in image analysis. The purpose of this study is to introduce and validate a 3D neural network for classification of Alzheimer's disease (AD), frontotemporal dementia (FTD) or cognitively normal (CN) subjects based on brain glucose metabolism. Retrospective [18F]-FDG-PET scans of 199 CE, 192 FTD and 200 CN subjects were collected from our local database, Alzheimer's disease and frontotemporal lobar degeneration neuroimaging initiatives. Training and test sets were created using randomization on a 90 %-10 % basis, and training of a 3D VGG16-like neural network was performed using data augmentation and cross-validation. Performance was compared to clinical interpretation by three specialists in the independent test set. Regions determining classification were identified in an occlusion experiment and Gradient-weighted Class Activation Mapping. Test set subjects were age- and sex-matched across categories. The model achieved an overall 89.8 % accuracy in predicting the class of test scans. Areas under the ROC curves were 93.3 % for AD, 95.3 % for FTD, and 99.9 % for CN. The physicians' consensus showed a 69.5 % accuracy, and there was substantial agreement between them (kappa = 0.61, 95 % CI: 0.49-0.73). To our knowledge, this is the first study to introduce a deep learning model able to discriminate AD and FTD based on [18F]-FDG PET scans, and to isolate CN subjects with excellent accuracy. These initial results are promising and hint at the potential for generalization to data from other centers.

Usefulness of 18F-Fluorodeoxyglucose Positron Emission Tomography in Diagnosing Polymyalgia Rheumatica and Large-Vessel Vasculitis: A Case-Control Study.

INTRODUCTION: We aimed to evaluate the utility of FDG-PET/CT in diagnosing polymyalgia rheumatica (PMR) and associated large-vessel vasculitis (LVV). METHODS: We analyzed FDG-PET/CT completed between 2015 and 2019 on patients diagnosed with PMR. For comparisons, patients with PMR were matched 1:1 to controls based on age and gender. FDG-PET/CT had been completed on the controls over the same period. The FDG uptake was scored visually for 17 articular or periarticular sites and 13 vascular sites using a semi-quantitative scoring system (score of 0-3). RESULTS: Eighty-one patients with PMR and eighty-one controls were included (mean age 70.7 (9.8) years; 44.4% women). Significant differences between the PMR and control groups were found at all articular and periarticular sites for the following: (i) the FDG uptake score (p < 0.001 for all locations); (ii) the number of patients per site with significant FDG uptake (score ≥ 2); (iii) the global FDG articular uptake scores (31 [IQR, 21 to 37] versus 6 [IQR, 3 to 10], p < 0.001); and (iv) the number of sites with significant FDG uptake (score ≥ 2) (scores of 0-17) (11 [IQR, 7 to 13] versus 1 [IQR, 0 to 2], p < 0.001). No significant differences in the global FDG vascular uptake scores were found between the patients who were considered isolated PMR and the control groups. CONCLUSIONS: The FDG uptake score and the number of sites with significant FDG uptake could be pertinent criteria for the diagnosis of PMR. Unlike others, we did not confirm the presence of vascular involvement in patients with isolated PMR.

Dopamine transporter imaging for the diagnosis of multiple system atrophy cerebellar type.

INTRODUCTION: The added value of dopamine transporter SPECT (DAT-SPECT) for the diagnosis of "possible" multiple system atrophy of the cerebellar type (MSA-C) remains unknown. METHODS: We reviewed retrospectively the charts of 128 consecutive patients with a clinical diagnosis of MSA-C who were seen between 2007 and 2016 at the French Reference Center for MSA. The main objective was to evaluate the proportion of patients for whom the diagnosis of "possible" MSA-C was made because of a positive DAT-SPECT. RESULTS: Seventy-eight MSA-C patients had at least one DAT-SPECT. Fifty-nine of them were considered for the final analysis. In these, 22 had "possible" MSA-C and 23 "probable" MSA-C before DAT-SPECT, while 14 did not reach diagnosis criteria at that time. In those with "possible" MSA-C, DAT-SPECT was positive in 64%. In patients with "probable" MSA-C, 83% showed nigrostriatal denervation. Six out of 14 (43%) received a diagnosis of "possible" MSA-C because of positive DAT-SPECT. These patients had mean disease duration of 2.3 years at the time of DAT-SPECT compared to 3.5 years of the entire cohort of MSA-C patients with DAT-SPECT. Of the eight remaining, one had positive DAT-SPECT but also pons atrophy on magnetic resonance imaging, and seven progressed to "probable" MSA based on clinical features. CONCLUSION: Our results suggest that DAT-SPECT significantly contributes to the diagnosis of "possible" MSA-C (43% of patients not reaching consensus diagnosis criteria before DAT-SPECT). DAT-SPECT seems especially useful in patients with shorter disease duration, while a negative result does not exclude a diagnosis of MSA.