Enhancing magnetic resonance imaging-driven Alzheimer’s disease classification performance using generative adversarial learning

Abstract

Background Generative adversarial networks (GAN) can produce images of improved quality but their ability to augment image-based classification is not fully explored. We evaluated if a modified GAN can learn from magnetic resonance imaging (MRI) scans of multiple magnetic field strengths to enhance Alzheimer’s disease (AD) classification performance. Methods T1-weighted brain MRI scans from 151 participants of the Alzheimer’s Disease Neuroimaging Initiative (ADNI), who underwent both 1.5-Tesla (1.5-T) and 3-Tesla imaging at the same time were selected to construct a GAN model. This model was trained along with a three-dimensional fully convolutional network (FCN) using the generated images (3T*) as inputs to predict AD status. Quality of the generated images was evaluated using signal to noise ratio (SNR), Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE) and Natural Image Quality Evaluator (NIQE). Cases from the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL, n = 107) and the National Alzheimer’s Coordinating Center (NACC, n = 565) were used for model validation. Results The 3T*-based FCN classifier performed better than the FCN model trained using the 1.5-T scans. Specifically, the mean area under curve increased from 0.907 to 0.932, from 0.934 to 0.940, and from 0.870 to 0.907 on the ADNI test, AIBL, and NACC datasets, respectively. Additionally, we found that the mean quality of the generated (3T*) images was consistently higher than the 1.5-T images, as measured using SNR, BRISQUE, and NIQE on the validation datasets. Conclusion This study demonstrates a proof of principle that GAN frameworks can be constructed to augment AD classification performance and improve image quality.

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

“…The task-induced discriminator focused not only on the quality of the generated images but also on whether AD pathological information was retained. In addition, the results of the downstream classification task were fed back to the generator and discriminator during training in the study by Zhou X. et al (2021) . This training may ensure the classification performance of the generated images.…”

“…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 ).…”

“…Regarding the data modality, some studies used MRI data, while others used PET data. Zhou X. et al (2021) developed a GAN model to generate 3-T MRI scans from 1.5-T scans. Then, researchers trained a fully convolutional network (FCN) using generated 3-T MRI as inputs to complete the task of AD vs. CN classification.…”

“…Fake images generated by this model, which highly resemble the real images, might exercise the same function as real images in disease diagnosis. In recent years, GAN has shown application value in diagnosing AD by providing image processing support, including quality improvement for low-dose PET images or 1.5-T MRIs ( Wang et al, 2018 ; Ouyang et al, 2019 ; Zhou X. et al, 2021 ), predicting brain images at a future time point ( Wegmayr et al, 2019 ; Zhao et al, 2021 ), data augmentation for network training ( Islam and Zhang, 2020 ; Sajjad et al, 2021 ), and interconversion of PET and MRI data ( Gao et al, 2021 ; Lin W. et al, 2021 ). With the GAN-based deep-learning classification framework, a more accurate diagnosis of AD is promising and may be achieved.…”

Deep Convolutional Neural Networks With Ensemble Learning and Generative Adversarial Networks for Alzheimer’s Disease Image Data Classification

“…The performance of AD classification models is highly correlated to the advancements of the scanners. Zhou et al (2021) recently explored the relationship between GAN performance using T1-weighted MRIs of various quality and AD classification accuracy. Both 1.5-Tesla (1.5-T) and 3-Tesla (3-T) scans were produced during the same patient visit and were available for this study.…”