Changhee Han1, Leonardo Rundo2,3, Kohei Murao4, Tomoyuki Noguchi5, Yuki Shimahara6, Zoltán Ádám Milacski7, Saori Koshino8, Evis Sala2,3, Hideki Nakayama9,10, Shin'ichi Satoh4. 1. LPIXEL Inc., Tokyo, Japan. han@lpixel.net. 2. Department of Radiology, University of Cambridge, Cambridge, UK. 3. Cancer Research UK Cambridge Centre, University of Cambridge, Cambridge, UK. 4. Research Center for Medical Big Data, National Institute of Informatics, Tokyo, Japan. 5. National Center for Global Health and Medicine, Tokyo, Japan. 6. LPIXEL Inc., Tokyo, Japan. 7. Department of Artificial Intelligence, ELTE Eötvös Loránd University, Budapest, Hungary. 8. Department of Radiology, Juntendo University, Tokyo, Japan. 9. Graduate School of Information Science and Technology, The University of Tokyo, Tokyo, Japan. 10. International Research Center for Neurointelligence (WPI-IRCN), The University of Tokyo Institutes for Advanced Study, The University of Tokyo, Tokyo, Japan.
Abstract
BACKGROUND: Unsupervised learning can discover various unseen abnormalities, relying on large-scale unannotated medical images of healthy subjects. Towards this, unsupervised methods reconstruct a 2D/3D single medical image to detect outliers either in the learned feature space or from high reconstruction loss. However, without considering continuity between multiple adjacent slices, they cannot directly discriminate diseases composed of the accumulation of subtle anatomical anomalies, such as Alzheimer's disease (AD). Moreover, no study has shown how unsupervised anomaly detection is associated with either disease stages, various (i.e., more than two types of) diseases, or multi-sequence magnetic resonance imaging (MRI) scans. RESULTS: We propose unsupervised medical anomaly detection generative adversarial network (MADGAN), a novel two-step method using GAN-based multiple adjacent brain MRI slice reconstruction to detect brain anomalies at different stages on multi-sequence structural MRI: (Reconstruction) Wasserstein loss with Gradient Penalty + 100 [Formula: see text] loss-trained on 3 healthy brain axial MRI slices to reconstruct the next 3 ones-reconstructs unseen healthy/abnormal scans; (Diagnosis) Average [Formula: see text] loss per scan discriminates them, comparing the ground truth/reconstructed slices. For training, we use two different datasets composed of 1133 healthy T1-weighted (T1) and 135 healthy contrast-enhanced T1 (T1c) brain MRI scans for detecting AD and brain metastases/various diseases, respectively. Our self-attention MADGAN can detect AD on T1 scans at a very early stage, mild cognitive impairment (MCI), with area under the curve (AUC) 0.727, and AD at a late stage with AUC 0.894, while detecting brain metastases on T1c scans with AUC 0.921. CONCLUSIONS: Similar to physicians' way of performing a diagnosis, using massive healthy training data, our first multiple MRI slice reconstruction approach, MADGAN, can reliably predict the next 3 slices from the previous 3 ones only for unseen healthy images. As the first unsupervised various disease diagnosis, MADGAN can reliably detect the accumulation of subtle anatomical anomalies and hyper-intense enhancing lesions, such as (especially late-stage) AD and brain metastases on multi-sequence MRI scans.
BACKGROUND: Unsupervised learning can discover various unseen abnormalities, relying on large-scale unannotated medical images of healthy subjects. Towards this, unsupervised methods reconstruct a 2D/3D single medical image to detect outliers either in the learned feature space or from high reconstruction loss. However, without considering continuity between multiple adjacent slices, they cannot directly discriminate diseases composed of the accumulation of subtle anatomical anomalies, such as Alzheimer's disease (AD). Moreover, no study has shown how unsupervised anomaly detection is associated with either disease stages, various (i.e., more than two types of) diseases, or multi-sequence magnetic resonance imaging (MRI) scans. RESULTS: We propose unsupervised medical anomaly detection generative adversarial network (MADGAN), a novel two-step method using GAN-based multiple adjacent brain MRI slice reconstruction to detect brain anomalies at different stages on multi-sequence structural MRI: (Reconstruction) Wasserstein loss with Gradient Penalty + 100 [Formula: see text] loss-trained on 3 healthy brain axial MRI slices to reconstruct the next 3 ones-reconstructs unseen healthy/abnormal scans; (Diagnosis) Average [Formula: see text] loss per scan discriminates them, comparing the ground truth/reconstructed slices. For training, we use two different datasets composed of 1133 healthy T1-weighted (T1) and 135 healthy contrast-enhanced T1 (T1c) brain MRI scans for detecting AD and brain metastases/various diseases, respectively. Our self-attention MADGAN can detect AD on T1 scans at a very early stage, mild cognitive impairment (MCI), with area under the curve (AUC) 0.727, and AD at a late stage with AUC 0.894, while detecting brain metastases on T1c scans with AUC 0.921. CONCLUSIONS: Similar to physicians' way of performing a diagnosis, using massive healthy training data, our first multiple MRI slice reconstruction approach, MADGAN, can reliably predict the next 3 slices from the previous 3 ones only for unseen healthy images. As the first unsupervised various disease diagnosis, MADGAN can reliably detect the accumulation of subtle anatomical anomalies and hyper-intense enhancing lesions, such as (especially late-stage) AD and brain metastases on multi-sequence MRI scans.
Authors: R S Desikan; H J Cabral; B Fischl; C R G Guttmann; D Blacker; B T Hyman; M S Albert; R J Killiany Journal: AJNR Am J Neuroradiol Date: 2008-12-26 Impact factor: 3.825
Authors: Christian Salvatore; Antonio Cerasa; Petronilla Battista; Maria C Gilardi; Aldo Quattrone; Isabella Castiglioni Journal: Front Neurosci Date: 2015-09-01 Impact factor: 4.677
Authors: Halima Hamid N Alrashedy; Atheer Fahad Almansour; Dina M Ibrahim; Mohammad Ali A Hammoudeh Journal: Sensors (Basel) Date: 2022-06-06 Impact factor: 3.847
Authors: Weiyao Wang; Aniruddha Tamhane; Christine Santos; John R Rzasa; James H Clark; Therese L Canares; Mathias Unberath Journal: Front Digit Health Date: 2022-02-10
Authors: Maria Elena Laino; Pierandrea Cancian; Letterio Salvatore Politi; Matteo Giovanni Della Porta; Luca Saba; Victor Savevski Journal: J Imaging Date: 2022-03-23
Authors: Mauro Castelli; Leonardo Rundo; Erick Costa de Farias; Christian di Noia; Changhee Han; Evis Sala Journal: Sci Rep Date: 2021-11-01 Impact factor: 4.379