Three-Dimensional Amyloid-Beta PET Synthesis from Structural MRI with Conditional Generative Adversarial Networks
2405.02109
0
0
🤯
Abstract
Motivation: Alzheimer's Disease hallmarks include amyloid-beta deposits and brain atrophy, detectable via PET and MRI scans, respectively. PET is expensive, invasive and exposes patients to ionizing radiation. MRI is cheaper, non-invasive, and free from ionizing radiation but limited to measuring brain atrophy. Goal: To develop an 3D image translation model that synthesizes amyloid-beta PET images from T1-weighted MRI, exploiting the known relationship between amyloid-beta and brain atrophy. Approach: The model was trained on 616 PET/MRI pairs and validated with 264 pairs. Results: The model synthesized amyloid-beta PET images from T1-weighted MRI with high-degree of similarity showing high SSIM and PSNR metrics (SSIM>0.95&PSNR=28). Impact: Our model proves the feasibility of synthesizing amyloid-beta PET images from structural MRI ones, significantly enhancing accessibility for large-cohort studies and early dementia detection, while also reducing cost, invasiveness, and radiation exposure.
Create account to get full access
Overview
- Alzheimer's disease is characterized by amyloid-beta deposits and brain atrophy, which can be detected using PET and MRI scans, respectively.
- PET scans are expensive, invasive, and expose patients to ionizing radiation, while MRI scans are cheaper, non-invasive, and radiation-free but limited to measuring brain atrophy.
- The goal of this research was to develop a 3D image translation model that can synthesize amyloid-beta PET images from T1-weighted MRI scans, leveraging the known relationship between amyloid-beta and brain atrophy.
Plain English Explanation
Alzheimer's disease is a brain disorder that causes memory and thinking problems. One of the key signs of Alzheimer's is the buildup of a protein called amyloid-beta in the brain. This can be detected using a type of medical imaging called PET scans. Another sign is the shrinking or 'atrophy' of the brain, which can be seen on MRI scans.
However, PET scans are quite expensive and require injecting a radioactive tracer into the patient's body, which some people may not feel comfortable with. MRI scans, on the other hand, are cheaper and don't use any radiation, but they can only measure the brain atrophy, not the amyloid-beta buildup.
The researchers in this study developed a 3D image translation model that can take an MRI scan of the brain and use it to create a simulated PET scan that shows the amyloid-beta levels. This allows doctors to get information about both the brain structure and the amyloid-beta buildup without having to do a separate PET scan.
This could be really helpful for large studies on Alzheimer's disease and for early detection of the disease, as it reduces the cost, discomfort, and radiation exposure for patients.
Technical Explanation
The researchers trained their 3D image translation model on a dataset of 616 paired PET and MRI scans, and then validated it on 264 additional pairs. The model was able to synthesize amyloid-beta PET images from the T1-weighted MRI scans with a high degree of similarity, as shown by the strong SSIM (>0.95) and PSNR (=28) metrics.
This demonstrates the feasibility of using structural MRI data to estimate the functional information about amyloid-beta levels that would normally require a separate, more invasive PET scan. By bridging this gap between structural and functional brain imaging, the model has the potential to significantly enhance the accessibility of amyloid-beta assessment for large-cohort studies and early Alzheimer's disease detection, while also reducing the cost, invasiveness, and radiation exposure for patients.
Critical Analysis
The paper provides a compelling proof-of-concept for this 3D image translation approach, but there are a few caveats and areas for further research that could be considered:
-
The model was trained and validated on a relatively small dataset of just over 800 PET/MRI pairs. Larger-scale validation on more diverse patient populations would help establish the robustness and generalizability of the approach.
-
The paper does not discuss potential sources of error or failure modes for the model, such as cases where there may be a poor correlation between brain atrophy and amyloid-beta deposition. Further analysis of the model's limitations would help contextualize the results.
-
While the model reduces the need for PET scans, patients would still require an MRI scan. Exploring ways to minimize the imaging burden on patients, such as multi-modal approaches that can extract information from a single scan, could further enhance the clinical utility.
Overall, this research represents an important step towards more accessible and patient-friendly methods for assessing Alzheimer's disease biomarkers, but additional work is needed to fully realize the potential of this approach.
Conclusion
This study presents a novel 3D image translation model that can synthesize amyloid-beta PET images from T1-weighted MRI scans, leveraging the known relationship between amyloid-beta deposition and brain atrophy. By bridging the gap between structural and functional brain imaging, this model has the potential to significantly enhance the accessibility of amyloid-beta assessment for large-scale Alzheimer's disease research and early detection efforts, while also reducing the cost, invasiveness, and radiation exposure for patients. While further validation and refinement is needed, this research represents an important step towards more patient-friendly and cost-effective methods for evaluating Alzheimer's disease biomarkers.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
Multimodal MRI-based Detection of Amyloid Status in Alzheimer's Disease Continuum
Giorgio Dolci (Department of Computer Science, University of Verona, Verona, Italy, Department of Engineering for Innovation Medicine, University of Verona, Verona, Italy, Tri-Institutional Center for Translational Research in Neuroimaging and Data Science), Charles A. Ellis (Tri-Institutional Center for Translational Research in Neuroimaging and Data Science), Federica Cruciani (Department of Engineering for Innovation Medicine, University of Verona, Verona, Italy), Lorenza Brusini (Department of Engineering for Innovation Medicine, University of Verona, Verona, Italy), Anees Abrol (Tri-Institutional Center for Translational Research in Neuroimaging and Data Science), Ilaria Boscolo Galazzo (Department of Engineering for Innovation Medicine, University of Verona, Verona, Italy), Gloria Menegaz (Department of Engineering for Innovation Medicine, University of Verona, Verona, Italy), Vince D. Calhoun (Tri-Institutional Center for Translational Research in Neuroimaging and Data Science)
0
0
Amyloid-$beta$ (A$beta$) plaques in conjunction with hyperphosphorylated tau proteins in the form of neurofibrillary tangles are the two neuropathological hallmarks of Alzheimer's disease (AD). In particular, the accumulation of A$beta$ plaques, as evinced by the A/T/N (amyloid/tau/neurodegeneration) framework, marks the initial stage. Thus, the identification of individuals with A$beta$ positivity could enable early diagnosis and potentially lead to more effective interventions. Deep learning methods relying mainly on amyloid PET images have been employed to this end. However, PET imaging has some disadvantages, including the need of radiotracers and expensive acquisitions. Hence, in this work, we propose a novel multimodal approach that integrates information from structural, functional, and diffusion MRI data to discriminate A$beta$ status in the AD continuum. Our method achieved an accuracy of $0.762pm0.04$. Furthermore, a textit{post-hoc} explainability analysis (guided backpropagation) was performed to retrieve the brain regions that most influenced the model predictions. This analysis identified some key regions that were common across modalities, some of which were well-established AD-discriminative biomarkers and related to A$beta$ deposition, such as the hippocampus, thalamus, precuneus, and cingulate gyrus. Hence, our study demonstrates the potential viability of MRI-based characterization of A$beta$ status, paving the way for further research in this domain.
6/21/2024
Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Image Synthesis: T1 MRI to Tau-PET
Symac Kim, Junho Moon, Haejun Chung, Ikbeom Jang
0
0
Alzheimer's Disease (AD) is the most common form of dementia, characterised by cognitive decline and biomarkers such as tau-proteins. Tau-positron emission tomography (tau-PET), which employs a radiotracer to selectively bind, detect, and visualise tau protein aggregates within the brain, is valuable for early AD diagnosis but is less accessible due to high costs, limited availability, and its invasive nature. Image synthesis with neural networks enables the generation of tau-PET images from more accessible T1-weighted magnetic resonance imaging (MRI) images. To ensure high-quality image synthesis, we propose a cyclic 2.5D perceptual loss combined with mean squared error and structural similarity index measure (SSIM) losses. The cyclic 2.5D perceptual loss sequentially calculates the axial 2D average perceptual loss for a specified number of epochs, followed by the coronal and sagittal planes for the same number of epochs. This sequence is cyclically performed, with intervals reducing as the cycles repeat. We conduct supervised synthesis of tau-PET images from T1w MRI images using 516 paired T1w MRI and tau-PET 3D images from the ADNI database. For the collected data, we perform preprocessing, including intensity standardisation for tau-PET images from each manufacturer. The proposed loss, applied to generative 3D U-Net and its variants, outperformed those with 2.5D and 3D perceptual losses in SSIM and peak signal-to-noise ratio (PSNR). In addition, including the cyclic 2.5D perceptual loss to the original losses of GAN-based image synthesis models such as CycleGAN and Pix2Pix improves SSIM and PSNR by at least 2% and 3%. Furthermore, by-manufacturer PET standardisation helps the models in synthesising high-quality images than min-max PET normalisation.
6/19/2024
TauAD: MRI-free Tau Anomaly Detection in PET Imaging via Conditioned Diffusion Models
Lujia Zhong, Shuo Huang, Jiaxin Yue, Jianwei Zhang, Zhiwei Deng, Wenhao Chi, Yonggang Shi
0
0
The emergence of tau PET imaging over the last decade has enabled Alzheimer's disease (AD) researchers to examine tau pathology in vivo and more effectively characterize the disease trajectories of AD. Current tau PET analysis methods, however, typically perform inferences on large cortical ROIs and are limited in the detection of localized tau pathology that varies across subjects. Furthermore, a high-resolution MRI is required to carry out conventional tau PET analysis, which is not commonly acquired in clinical practices and may not be acquired for many elderly patients with dementia due to strong motion artifacts, claustrophobia, or certain metal implants. In this work, we propose a novel conditional diffusion model to perform MRI-free anomaly detection from tau PET imaging data. By including individualized conditions and two complementary loss maps from pseudo-healthy and pseudo-unhealthy reconstructions, our model computes an anomaly map across the entire brain area that allows simply training a support vector machine (SVM) for classifying disease severity. We train our model on ADNI subjects (n=534) and evaluate its performance on a separate dataset from the preclinical subjects of the A4 clinical trial (n=447). We demonstrate that our method outperforms baseline generative models and the conventional Z-score-based method in anomaly localization without mis-detecting off-target bindings in sub-cortical and out-of-brain areas. By classifying the A4 subjects according to their anomaly map using the SVM trained on ADNI data, we show that our method can successfully group preclinical subjects with significantly different cognitive functions, which further demonstrates the effectiveness of our method in capturing biologically relevant anomaly in tau PET imaging.
5/24/2024
Cross-Modality Translation with Generative Adversarial Networks to Unveil Alzheimer's Disease Biomarkers
Reihaneh Hassanzadeh, Anees Abrol, Hamid Reza Hassanzadeh, Vince D. Calhoun
0
0
Generative approaches for cross-modality transformation have recently gained significant attention in neuroimaging. While most previous work has focused on case-control data, the application of generative models to disorder-specific datasets and their ability to preserve diagnostic patterns remain relatively unexplored. Hence, in this study, we investigated the use of a generative adversarial network (GAN) in the context of Alzheimer's disease (AD) to generate functional network connectivity (FNC) and T1-weighted structural magnetic resonance imaging data from each other. We employed a cycle-GAN to synthesize data in an unpaired data transition and enhanced the transition by integrating weak supervision in cases where paired data were available. Our findings revealed that our model could offer remarkable capability, achieving a structural similarity index measure (SSIM) of $0.89 pm 0.003$ for T1s and a correlation of $0.71 pm 0.004$ for FNCs. Moreover, our qualitative analysis revealed similar patterns between generated and actual data when comparing AD to cognitively normal (CN) individuals. In particular, we observed significantly increased functional connectivity in cerebellar-sensory motor and cerebellar-visual networks and reduced connectivity in cerebellar-subcortical, auditory-sensory motor, sensory motor-visual, and cerebellar-cognitive control networks. Additionally, the T1 images generated by our model showed a similar pattern of atrophy in the hippocampal and other temporal regions of Alzheimer's patients.
5/10/2024