Development of Amyloid PET Analysis Pipeline Using Deep Learning-Based Brain MRI Segmentation—A Comparative Validation Study

Deep learning-based binary classification of beta-amyloid plaques using 18 F florapronol PET

PURPOSE

This study aimed to investigate a deep learning model to classify amyloid plaque deposition in the brain PET images of patients suspected of Alzheimer's disease.

METHODS

A retrospective study was conducted on patients who were suspected of having a mild cognitive impairment or dementia, and brain amyloid 18 F florapronol PET/computed tomography images were obtained from 2019 to 2022. Brain PET images were visually assessed by two nuclear medicine specialists, who classified them as either positive or negative. Image rotation was applied for data augmentation. The dataset was split into training and testing sets at a ratio of 8 : 2. For the convolutional neural network (CNN) analysis, stratified k-fold ( k  = 5) cross-validation was applied using training set. Trained model was evaluated using testing set.

RESULTS

A total of 175 patients were included in this study. The average age at the time of PET imaging was 70.4 ± 9.3 years and included 77 men and 98 women (44.0% and 56.0%, respectively). The visual assessment revealed positivity in 62 patients (35.4%) and negativity in 113 patients (64.6%). After stratified k-fold cross-validation, the CNN model showed an average accuracy of 0.917 ± 0.027. The model exhibited an accuracy of 0.914 and an area under the curve of 0.958 in the testing set. These findings affirm the model's high reliability in distinguishing between positive and negative cases.

CONCLUSION

The study verifies the potential of the CNN model to classify amyloid positive and negative cases using brain PET images. This model may serve as a supplementary tool to enhance the accuracy of clinical diagnoses.

Effects of strategic white matter hyperintensities of cholinergic pathways on basal forebrain volume in patients with amyloid-negative neurocognitive disorders

BACKGROUND

The cholinergic neurotransmitter system is crucial to cognitive function, with the basal forebrain (BF) being particularly susceptible to Alzheimer's disease (AD) pathology. However, the interaction of white matter hyperintensities (WMH) in cholinergic pathways and BF atrophy without amyloid pathology remains poorly understood.

METHODS

We enrolled patients who underwent neuropsychological tests, magnetic resonance imaging, and 18F-florbetaben positron emission tomography due to cognitive impairment at the teaching university hospital from 2015 to 2022. Among these, we selected patients with negative amyloid scans and additionally excluded those with Parkinson's dementia that may be accompanied by BF atrophy. The WMH burden of cholinergic pathways was quantified by the Cholinergic Pathways Hyperintensities Scale (CHIPS) score, and categorized into tertile groups because the CHIPS score did not meet normal distribution. Segmentation of the BF on volumetric T1-weighted MRI was performed using FreeSurfer, then was normalized for total intracranial volume. Multivariable regression analysis was performed to investigate the association between BF volumes and CHIPS scores.

RESULTS

A total of 187 patients were enrolled. The median CHIPS score was 12 [IQR 5.0; 24.0]. The BF volume of the highest CHIPS tertile group (mean ± SD, 3.51 ± 0.49, CHIPSt3) was significantly decreased than those of the lower CHIPS tertile groups (3.75 ± 0.53, CHIPSt2; 3.83 ± 0.53, CHIPSt1; P = 0.02). In the univariable regression analysis, factors showing significant associations with the BF volume were the CHIPSt3 group, age, female, education, diabetes mellitus, smoking, previous stroke history, periventricular WMH, and cerebral microbleeds. In multivariable regression analysis, the CHIPSt3 group (standardized beta [βstd] = -0.25, P = 0.01), female (βstd = 0.20, P = 0.04), and diabetes mellitus (βstd = -0.22, P < 0.01) showed a significant association with the BF volume. Sensitivity analyses showed a negative correlation between CHIPS score and normalized BF volume, regardless of WMH severity.

CONCLUSIONS

We identified a significant correlation between strategic WMH burden in the cholinergic pathway and BF atrophy independently of amyloid positivity and WMH severity. These results suggest a mechanism of cholinergic neuronal loss through the dying-back phenomenon and provide a rationale that strategic WMH assessment may help identify target groups that may benefit from acetylcholinesterase inhibitor treatment.

A Comparative Analysis of Two Automated Quantification Methods for Regional Cerebral Amyloid Retention: PET-Only and PET-and-MRI-Based Methods

Accurate quantification of amyloid positron emission tomography (PET) is essential for early detection of and intervention in Alzheimer's disease (AD) but there is still a lack of studies comparing the performance of various automated methods. This study compared the PET-only method and PET-and-MRI-based method with a pre-trained deep learning segmentation model. A large sample of 1180 participants in the Catholic Aging Brain Imaging (CABI) database was analyzed to calculate the regional standardized uptake value ratio (SUVR) using both methods. The logistic regression models were employed to assess the discriminability of amyloid-positive and negative groups through 10-fold cross-validation and area under the receiver operating characteristics (AUROC) metrics. The two methods showed a high correlation in calculating SUVRs but the PET-MRI method, incorporating MRI data for anatomical accuracy, demonstrated superior performance in predicting amyloid-positivity. The parietal, frontal, and cingulate importantly contributed to the prediction. The PET-MRI method with a pre-trained deep learning model approach provides an efficient and precise method for earlier diagnosis and intervention in the AD continuum.

Development of a prediction model for cognitive impairment of sarcopenia using multimodal neuroimaging in non-demented older adults

INTRODUCTION

Despite prior research on the association between sarcopenia and cognitive impairment in the elderly, a comprehensive model that integrates various brain pathologies is still lacking.

METHODS

We used data from 528 non-demented older adults with or without sarcopenia in the Catholic Aging Brain Imaging (CABI) database, containing magnetic resonance imaging scans, positron emission tomography scans, and clinical data. We also measured three key components of sarcopenia: skeletal muscle index (SMI), hand grip strength (HGS), and the five times sit-to-stand test (5STS).

RESULTS

All components of sarcopenia were significantly correlated with global cognitive function, but cortical thickness and amyloid-beta (Aβ) retention had distinctive relationships with each measure. In the path model, brain atrophy resulting in cognitive impairment was mediated by Aβ retention for SMI and periventricular white matter hyperintensity for HGS, but directly affected by the 5STS.

DISCUSSION

Treatments targeting each sub-domain of sarcopenia should be considered to prevent cognitive decline.

HIGHLIGHTS

We identified distinct impacts of three sarcopenia measures on brain structure and Aβ. Muscle mass is mainly associated with Aβ and has an influence on the brain atrophy. Muscle strength linked with periventricular WMH and brain atrophy. Muscle function associated with cortical thinning in specific brain regions. Interventions on sarcopenia may be important to ease cognitive decline in the elderly.

Generative AI unlocks PET insights: brain amyloid dynamics and quantification

Introduction

Studying the spatiotemporal patterns of amyloid accumulation in the brain over time is crucial in understanding Alzheimer's disease (AD). Positron Emission Tomography (PET) imaging plays a pivotal role because it allows for the visualization and quantification of abnormal amyloid beta (Aβ) load in the living brain, providing a powerful tool for tracking disease progression and evaluating the efficacy of anti-amyloid therapies. Generative artificial intelligence (AI) can learn complex data distributions and generate realistic synthetic images. In this study, we demonstrate for the first time the potential of Generative Adversarial Networks (GANs) to build a low-dimensional representation space that effectively describes brain amyloid load and its dynamics.

Methods

Using a cohort of 1,259 subjects with AV45 PET images from the Alzheimer's Disease Neuroimaging Initiative (ADNI), we develop a 3D GAN model to project images into a latent representation space and generate back synthetic images. Then, we build a progression model on the representation space based on non-parametric ordinary differential equations to study brain amyloid evolution.

Results

We found that global SUVR can be accurately predicted with a linear regression model only from the latent representation space (RMSE = 0.08 ± 0.01). We generated synthetic PET trajectories and illustrated predicted Aβ change in four years compared with actual progression.

Discussion

Generative AI can generate rich representations for statistical prediction and progression modeling and simulate evolution in synthetic patients, providing an invaluable tool for understanding AD, assisting in diagnosis, and designing clinical trials. The aim of this study was to illustrate the huge potential that generative AI has in brain amyloid imaging and to encourage its advancement by providing use cases and ideas for future research tracks.

A Review of Machine Learning Approaches for Brain Positron Emission Tomography Data Analysis

Positron emission tomography (PET) imaging has moved forward the development of medical diagnostics and research across various domains, including cardiology, neurology, infection detection, and oncology. The integration of machine learning (ML) algorithms into PET data analysis has further enhanced their capabilities of including disease diagnosis and classification, image segmentation, and quantitative analysis. ML algorithms empower researchers and clinicians to extract valuable insights from complex big PET datasets, which enabling automated pattern recognition, predictive health outcome modeling, and more efficient data analysis. This review explains the basic knowledge of PET imaging, statistical methods for PET image analysis, and challenges of PET data analysis. We also discussed the improvement of analysis capabilities by combining PET data with machine learning algorithms and the application of this combination in various aspects of PET image research. This review also highlights current trends and future directions in PET imaging, emphasizing the driving and critical role of machine learning and big PET image data analytics in improving diagnostic accuracy and personalized medical approaches. Integration between PET imaging will shape the future of medical diagnosis and research.

Associations between Education Years and Resting-state Functional Connectivity Modulated by APOE ε4 Carrier Status in Cognitively Normal Older Adults

Objective

: Cognitive reserve has emerged as a concept to explain the variable expression of clinical symptoms in the pathology of Alzheimer's disease (AD). The association between years of education, a proxy of cognitive reserve, and resting-state functional connectivity (rFC), a representative intermediate phenotype, has not been explored in the preclinical phase, considering risk factors for AD. We aimed to evaluate whether the relationship between years of education and rFC in cognitively preserved older adults differs depending on amyloid-beta deposition and APOE ε4 carrier status as effect modifiers.

Methods

: A total of 121 participants underwent functional magnetic resonance imaging, [18F] flutemetamol positron emission tomography-computed tomography, APOE genotyping, and a neuropsychological battery. Potential interactions between years of education and AD risk factors for rFC of AD-vulnerable neural networks were assessed with whole-brain voxel-wise analysis.

Results

: We found a significant education years-by-APOE ε4 carrier status interaction for the rFC from the seed region of the central executive (CEN) and dorsal attention networks. Moreover, there was a significant interaction of rFC between right superior occipital gyrus and the CEN seed region by APOE ε4 carrier status for memory performances and overall cognitive function.

Conclusion

: In preclinical APOE ε4 carriers, higher years of education were associated with higher rFC of the AD vulnerable network, but this contributed to lower cognitive function. These results contribute to a deeper understanding of the impact of cognitive reserve on sensitive functional intermediate phenotypic markers in the preclinical phase of AD.

Modulation of associations between education years and cortical volume in Alzheimer's disease vulnerable brain regions by Aβ deposition and APOE ε4 carrier status in cognitively normal older adults

Background

Education years, as a measure of cognitive reserve, have been shown to affect the progression of Alzheimer's disease (AD), both pathologically and clinically. However, inconsistent results have been reported regarding the association between years of education and intermediate structural changes in AD-vulnerable brain regions, particularly when AD risk factors were not considered during the preclinical phase.

Objective

This study aimed to examine how Aβ deposition and APOE ε4 carrier status moderate the relationship between years of education and cortical volume in AD-vulnerable regions among cognitively normal older adults.

Methods

A total of 121 participants underwent structural MRI, [18F] flutemetamol PET-CT imaging, and neuropsychological battery assessment. Multiple regression analysis was conducted to examine the interaction between years of education and the effects of potential modifiers on cortical volume. The associations between cortical volume and neuropsychological performance were further explored in subgroups categorized based on AD risk factors.

Results

The cortical volume of the left lateral occipital cortex and bilateral fusiform gyrus demonstrated a significant differential association with years of education, depending on the presence of Aβ deposition and APOE ε4 carrier status. Furthermore, a significant relationship between the cortical volume of the bilateral fusiform gyrus and AD-nonspecific cognitive function was predominantly observed in individuals without AD risk factors.

Conclusion

AD risk factors exerted varying influences on the association between years of education and cortical volume during the preclinical phase. Further investigations into the long-term implications of these findings would enhance our understanding of cognitive reserves in the preclinical stages of AD.