Measurement of inter- and intra-observer variability in the routine clinical interpretation of brain 18-FDG PET-CT

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.

Resistance Training Increases White Matter Density in Frail Elderly Women

We aimed to investigate the effects of maternal obesity on brain structure and metabolism in frail women, and their reversibility in response to exercise. We recruited 37 frail elderly women (20 offspring of lean/normal-weight mothers (OLM) and 17 offspring of obese/overweight mothers (OOM)) and nine non-frail controls to undergo magnetic resonance and diffusion tensor imaging (DTI), positron emission tomography with Fluorine-18-fluorodeoxyglucose (PET), and cognitive function tests (CERAD). Frail women were studied before and after a 4-month resistance training, and controls were studied once. White matter (WM) density (voxel-based morphometry) was higher in OLM than in OOM subjects. Exercise increased WM density in both OLM and OOM in the cerebellum in superior parietal regions in OLM and in cuneal and precuneal regions in OOM. OLM gained more WM density than OOM in response to intervention. No significant results were found from the Freesurfer analysis, nor from PET or DTI images. Exercise has an impact on brain morphology and cognition in elderly frail women.

Concordance between brain 18F-FDG PET and cerebrospinal fluid biomarkers in diagnosing Alzheimer's disease

OBJECTIVES

Cortical posterior hypometabolism on PET imaging with ¹⁸F-FDG (FDG-PET), and altered levels of Aß_1-42 peptide, total Tau (tTau) and phosphorylated Tau (pTau) proteins in cerebrospinal fluid (CSF) are established diagnostic biomarkers in Alzheimer's disease (AD). An evaluation has been made of the concordance and relationship between the results of FDG-PET and CSF biomarkers in symptomatic patients with suspected AD.

MATERIAL AND METHODS

A retrospective review was carried out on 120 patients with cognitive impairment referred to our Cognitive Neurology Unit, and who were evaluated by brain FDG-PET and a lumbar puncture for CSF biomarkers. In order to calculate their Kappa coefficient of concordance, the result of the FDG-PET and the set of the three CSF biomarkers in each patient was classified as normal, inconclusive, or AD-compatible. The relationship between the results of both methods was further assessed using logistic regression analysis, including the Aß_1-42, tTau and pTau levels as quantitative predictors, and the FDG-PET result as the dependent variable.

RESULTS

The weighted Kappa coefficient between FDG-PET and CSF biomarkers was 0.46 (95% CI: 0.35-0.57). Logistic regression analysis showed that the Aß_1-42 and tTau values together were capable of discriminating an FDG-PET result metabolically suggestive of AD from one non-suggestive of AD, with a 91% sensitivity and 93% specificity at the cut-off line Aß_1-42=44+1.3×tTau.

CONCLUSIONS

The level of concordance between FDG-PET and CSF biomarkers was moderate, indicating their complementary value in diagnosing AD. The Aß_1-42 and tTau levels in CSF help to predict the patient FDG-PET cortical metabolic status.