Generation of Amyloid PET Images via Conditional Adversarial Training for Predicting Progression to Alzheimer’s Disease

Abstract

New positron emission tomography (PET) tracers could have a substantial impact on early diagnosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI) progression, particularly if they are accompanied by optimised deep learning methods. To realize the full potential of deep learning for PET imaging, large datasets are required for training. However, dataset sizes are restricted due to limited availability. Meanwhile, most of the AD classification studies have been based on structural MRI rather than PET. In this paper, we propose a novel application of conditional Generative Adversarial Networks (cGANs) to the generation of 𝐹 18 -florbetapir PET images from corresponding MRI images. Furthermore, we show that generated PET images can be used for synthetic data augmentation, and improve the performance of 3D Convolutional Neural Networks (CNN) for predicting progression to AD. Our method is applied to a dataset of 79 PET images, obtained from Alzheimer's Disease Neuroimaging Initiative (ADNI) database. We generate high quality PET images from corresponding MRIs using cGANs, and we evaluate the quality of generated PET images by comparison to real images. We then use the trained cGANs to generate synthetic PET images from additional MRI dataset. Finally we build a 152-layer ResNet to compare the MCI classification performance using both traditional data augmentation method and our proposed synthetic data augmentation method. Mean Structural Similarity (SSIM) index was 0.95± 0.05 for generated PET and real PET. For MCI progression classification, the traditional data augmentation method showed 75% accuracy while the synthetic data augmentation improved this to 82%.

Diagnostic Performance of Generative Adversarial Network-Based Deep Learning Methods for Alzheimer’s Disease: A Systematic Review and Meta-Analysis

“…Regarding the publication year, all the included articles were published between 2018 and 2021, and more than half of them (8/14) were published in 2021 ( Figure 2A ; Baydargil et al, 2021 ; Gao et al, 2021 ; Han et al, 2021 ; Kang et al, 2021 ; Lin W. et al, 2021 ; Sajjad et al, 2021 ; Zhao et al, 2021 ; Zhou X. et al, 2021 ). Regarding the data source, neuroimaging data analyzed in 13 studies were mainly from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) ( Pan et al, 2018 ; Yan et al, 2018 ; Wegmayr et al, 2019 ; Islam and Zhang, 2020 ; Kim et al, 2020 ; Shin et al, 2020 ; Baydargil et al, 2021 ; Gao et al, 2021 ; Kang et al, 2021 ; Lin W. et al, 2021 ; Sajjad et al, 2021 ; Zhao et al, 2021 ; Zhou X. et al, 2021 ), and some data were from the Open Access Series of Imaging Studies (OASIS) ( Han et al, 2021 ; Zhao et al, 2021 ), the Australian Imaging, Biomarker and Lifestyle Flagship Study of Aging (AIBL) and the National Alzheimer’s Coordinating Center (NACC) databases ( Figure 2B ; Zhou X. et al, 2021 ). Two studies established a test set from the collection of clinical data ( Wegmayr et al, 2019 ; Kim et al, 2020 ).…”

“…Two studies established a test set from the collection of clinical data ( Wegmayr et al, 2019 ; Kim et al, 2020 ). Regarding the data modality, 36 percent (5/14) of studies used data from two modalities ( Figure 2C ; Pan et al, 2018 ; Yan et al, 2018 ; Shin et al, 2020 ; Gao et al, 2021 ; Lin W. et al, 2021 ). One study used MRI and other clinical data (age, sex, education level, and other parameters) ( Zhao et al, 2021 ).…”

GANDALF: Generative Adversarial Networks with Discriminator-Adaptive Loss Fine-tuning for Alzheimer's Disease Diagnosis from MRI

“…For in-stance, Yan et. al [25] use the CGAN to generate AV45-PET from T1-weighted MRI to supplement the training dataset with additional synthetic PET-MRI image pairs. While for generating an image of different modality may be an endgoal for computer vision domain, in medical domain we often want to diagnose a disease, such as AD, using the generated image.…”

“…A combination of T1-weighted MRI and FDG-PET with three-dimensional convolutional neural network (CNN) was used to demonstrate binary classification of CN/AD, CN/pMCI, sMCI/pMCI in [10]. GAN was used to generate additional PET images from T1-weighted MRI that do not have AV45-PET image pairs in [25]. MRI and real-/synthetic-PET image pairs are subsequently used to train CNN to perform binary classification of stable-MCI/progressive-MCI.…”

GANBERT: Generative Adversarial Networks with Bidirectional Encoder Representations from Transformers for MRI to PET synthesis

“…PET synthesis from MRI using GAN was demonstrated in [13] for AD classification. A total of 108 amyloid (AV45) PET and the corresponding MRI image pairs with 58 early Mild Cognitive Impairment (EMCI) and 50 stable Mild Cognitive MRI images with different magnetic pulse sequences, namely T1 and FLAIR, were synthesized using unpaired image-to-image translation GAN [16].…”