PT - JOURNAL ARTICLE AU - Xiao Zhou AU - Shangran Qiu AU - Prajakta S. Joshi AU - Chonghua Xue AU - Ronald J. Killiany AU - Asim Mian AU - Sang P. Chin AU - Rhoda Au AU - Vijaya B. Kolachalama AU - for the Alzheimer’s Disease Neuroimaging Initiative TI - Enhancing MR imaging driven Alzheimer’s disease classification performance using generative adversarial learning AID - 10.1101/2020.07.22.20159814 DP - 2020 Jan 01 TA - medRxiv PG - 2020.07.22.20159814 4099 - http://medrxiv.org/content/early/2020/07/25/2020.07.22.20159814.short 4100 - http://medrxiv.org/content/early/2020/07/25/2020.07.22.20159814.full AB - Background Generative adversarial networks (GAN) can generate images of improved quality but their ability to augment image-based classification tasks is not fully explored.Purpose We evaluated if a modified GAN can learn from MRI scans of multiple magnetic field strength to enhance Alzheimer’s disease (AD) classification performance.Materials and methods T1-weighted brain MRI scans from 151 participants of the Alzheimer’s Disease Neuroimaging Initiative (ADNI), who underwent both 1.5 Tesla (1.5T) 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 (1.5T*) 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). Data from the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL, n=107), and the National Alzheimer’s Coordinating Center (NACC, n=565) was used for model validation.Results The mean quality of the generated (1.5T*) images was consistently higher than the 1.5T images, as measured using SNR, BRISQUE and NIQE on the validation datasets. The 1.5T*-based FCN classifier performed better than the FCN model trained using the 1.5T 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.Conclusion This study demonstrates that GAN frameworks can be constructed to simultaneously improve image quality and augment classification performance.Key pointsOur proposed generative adversarial network used 1.5 and 3 Tesla MR scans of the brain to generate images of improved quality, as estimated using no-reference image quality algorithms.Classification models of Alzheimer’s disease risk developed using the generated images had higher classification performance than the models developed using the original 1.5 Tesla scans.Competing Interest StatementThe authors have declared no competing interest.Funding StatementThis project was supported in part by the National Center for Advancing Translational Sciences, National Institutes of Health, through BU-CTSI Grant (1UL1TR001430), a Scientist Development Grant (17SDG33670323) and a Strategically Focused Research Network (SFRN) Center Grant (20SFRN35460031) from the American Heart Association, and a Hariri Research Award from the Hariri Institute for Computing and Computational Science & Engineering at Boston University, Framingham Heart Study National Heart, Lung and Blood Institute contract (N01-HC-25195; HHSN268201500001I) and NIH grants (R56-AG062109, R01-AG062109, AG008122, R01-AG016495, and R01-AG033040). Additional support was provided by Boston University Affinity Research Collaboratives program and Boston University Alzheimers Disease Center (P30-AG013846).Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:Not applicableAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesPython scripts are provided on the GitHub repository. https://github.com/vkola-lab/gan2020 List of abbreviationsMRIMagnetic Resonance ImagingGANGenerative Adversarial NetworkCNNConvolutional Neural NetworkGPUGraphical Processing UnitSNRSignal to Noise RatioSSIMStructural Similarity Index MetricBRISQUEBlind/Referenceless Image Spatial Quality EvaluatorNIQENatural Image Quality Evaluator