Enriching the structural MRI information by cross-scale associations with the diffusion-weighted MRI

dc.contributor.advisorRomero Castro, Eduardo
dc.contributor.authorMurcia Tapias, Al-yhuwert
dc.contributor.researchgroupComputer Imaging and Medical Aplications Laboratory - Cim@labspa
dc.contributor.supervisorGiraldo Franco, Diana Lorena
dc.date.accessioned2024-07-02T21:46:42Z
dc.date.available2024-07-02T21:46:42Z
dc.date.issued2024
dc.descriptionilustraciones (principalmente a color), diagramasspa
dc.description.abstractUnimodal MRI provides a unique channel of information specific to the organ under examination, but it tends to restrict the amount of information necessary for accurate diagnoses. Conversely, in multimodal MRI different tissue structures are highlighted, thereby enriching the information about processes affecting an organ and hence improving the diagnostic precision. Nevertheless, in clinical settings the availability of multiple MR modalities and the scanning time for every patient is limited. To address this challenge, image-to-image translation techniques can be used to synthesize different brain image modalities and provide enriched complementary information about the organ. The image translation task is often done with the use of Generative Adversarial Networks (GANs) which is a computationally expensive approach. This work presents the synthesis of diffusion-derived fractional anisotropy maps (FA) from T1-weighted brain Magnetic Resonance Images using a simplified GAN-based architecture that enrich the structural information while reducing the computational cost associated with training the generative model. Furthermore, to prove that the latent information of the generative network is inherently enriched by both input and target image modalities, a classification task of three stages of the Alzheimer’s disease spectrum (healthy, mild cognitive impairment and mild dementia) was performed. Brain magnetic resonance images from the ADNI database were employed. Paired T1 and FA slices in axial, coronal, and sagittal views were utilized for the synthesis task. For the classification task, T1 slices in the same orientations were used. We evaluated the synthesis task by comparing the performance of the proposed GAN architecture against two state-of-the-art networks: Pix2pix and CycleGAN. Using almost 70% less parameters than those used in Pix2pix, the proposed method showed competitive results in mean PSNR (20.21 ± 1.38) and SSIM (0.65 ± 0.07) when compared to Pix2pix (PSNR: 20.46 ± 1.46, SSIM: 0.66 ± 0.07), outperforming quality metrics achieved by CycleGAN (PSNR: 18.65 ± 1.31, SSIM: 0.61 ± 0.08). For the classification task, a Support Vector Machine (SVM) classifier was trained with the latent information of the proposed generative network. A boost in classification was demonstrated when comparing the enriched (multimodal) latent information with non-enriched (unimodal) information (Texto tomado de la fuente).eng
dc.description.abstractLa resonancia magnética unimodal proporciona un canal único de información especı́fica del órgano examinado, pero tiende a restringir la cantidad de información necesaria para diagnósticos precisos. En contraposición, en la resonancia magnética multimodal se destacan diferentes estructuras de tejido, lo cual enriquece la información sobre los procesos que afectan un órgano y por tanto mejora la precisión diagnóstica. Sin embargo, en entornos clı́nicos, la disponibilidad de múltiples modalidades de resonancia magnética y el tiempo de exploración para cada paciente son limitados. Para abordar este desafı́o, se pueden utilizar técnicas de traducción de imagen a imagen para sintetizar diferentes modalidades de imágenes cerebrales que proporcionen información complementaria enriquecida sobre el órgano. Esta tarea de generación de imágenes depende en gran medida del uso de Redes Generativas Adversarias (GAN), caracterizadas por su alta demanda computacional. Este trabajo presenta la sı́ntesis de mapas de anisotropı́a fraccional (FA) derivados de la imagen de difusión a partir de imágenes de resonancia magnética cerebral ponderada en T1, utilizando una arquitectura simplificada basada en GAN que enriquece la información estructural al tiempo que reduce el costo computacional asociado al entrenamiento del modelo generativo. Adicionalmente, con el fin de demostrar que la información latente de la red generativa está naturalmente enriquecida por ambas imágenes de entrada y de salida, se realizó una tarea de clasificación de tres de los estadios del espectro de la enfermedad de Alzheimer (sano, deterioro cognitivo moderado, demencia leve). Se utilizaron imágenes de resonancia magnética cerebral de la base de datos ADNI. Cortes pareados de T1 y FA en vistas axial, coronal y sagital fueron empleados en la tarea de sı́ntesis. Para la tarea de clasificación se usaron cortes de T1 en los mismos planos. La tarea de sı́ntesis fue evaluada comparando la arquitectura propuesta con dos redes del estado del arte: Pix2pix y CycleGAN. Utilizando casi un 70% menos de parámetros que los utilizados en Pix2pix, el método propuesto mostró resultados competitivos en PSNR (20.21 ± 1.38) y SSIM (0.65 ± 0.07) en comparación con Pix2pix (PSNR: 20.46 ± 1.46, SSIM: 0.66 ± 0.07), superando las métricas de calidad alcanzadas por CycleGAN (PSNR: 18.65 ± 1.31, SSIM: 0.61 ± 0.08). Para la tarea de clasificación, se entrenó un clasificador de máquina de soporte vectorial (SVM) utilizando la información latente de la red generativa propuesta. Se demostró una mejora en la clasificación cuando se comparó la información latente enriquecida (multimodal) con la información no enriquecida (unimodal) (Texto tomado de la fuente).spa
dc.description.curricularareaMedicina.Sede Bogotáspa
dc.description.degreelevelMaestríaspa
dc.description.degreenameMagister en Ingeniería Biomédicaspa
dc.description.researchareaImágenes médicasspa
dc.format.extentvii, 34 páginasspa
dc.format.mimetypeapplication/pdfspa
dc.identifier.instnameUniversidad Nacional de Colombiaspa
dc.identifier.repoRepositorio Institucional Universidad Nacional de Colombiaspa
dc.identifier.repourlhttps://repositorio.unal.edu.co/spa
dc.identifier.urihttps://repositorio.unal.edu.co/handle/unal/86355
dc.language.isoengspa
dc.publisherUniversidad Nacional de Colombiaspa
dc.publisher.branchUniversidad Nacional de Colombia - Sede Bogotáspa
dc.publisher.facultyFacultad de Medicinaspa
dc.publisher.placeBogotá, Colombiaspa
dc.publisher.programBogotá - Medicina - Maestría en Ingeniería Biomédicaspa
dc.relation.referencesOzlem Coskun. Magnetic resonance imaging and safety aspects. Toxicology and Industrial Health, 27(4):307–313, 2011. PMID: 21112927.spa
dc.relation.referencesM. Jenkinson and M. Chappell. Introduction to Neuroimaging Analysis. Oxford neuroimaging primers. Oxford University Press, 2018.spa
dc.relation.referencesM Symms, H R Jäger, K Schmierer, and T A Yousry. A review of structural magnetic resonance neuroimaging. Journal of Neurology, Neurosurgery & Psychiatry, 75(9):1235– 1244, 2004.spa
dc.relation.referencesClifford Jack, David Bennett, Kaj Blennow, Maria Carrillo, Billy Dunn, Samantha Haeberlein, David Holtzman, William Jagust, Frank Jessen, Jason Karlawish, Enchi Liu, Jose Molinuevo, Thomas Montine, Creighton Phelps, Katherine Rankin, Christopher Rowe, Philip Scheltens, Eric Siemers, Heather Snyder, and Nina Silverberg. Nia-aa research framework: Toward a biological definition of alzheimer’s disease. Alzheimer’s Dementia, 14:535–562, 04 2018.spa
dc.relation.referencesVijay Grover, Joshua Tognarelli, Mary Crossey, I. Cox, Simon Taylor-Robinson, and Mark McPhail. Magnetic resonance imaging: Principles and techniques: Lessons for clinicians. Journal of Clinical and Experimental Hepatology, 5, 08 2015.spa
dc.relation.referencesViviana Lo Buono, Rosanna Palmeri, Francesco Corallo, Cettina Allone, Deborah Pria, Placido Bramanti, and Silvia Marino. Diffusion tensor imaging of white matter degeneration in early stage of Alzheimer’s disease: a review. International Journal of Neuroscience, 130(3):243–250, 2020. PMID: 31549530.spa
dc.relation.referencesWoo-Suk Tae, Byung Ham, Sung-Bom Pyun, Shin-Hyuk Kang, and Byung-Jo Kim. Current clinical applications of diffusion-tensor imaging in neurological disorders. Journal of clinical neurology (Seoul, Korea), 14, 02 2018.spa
dc.relation.referencesChantel D. Mayo, Erin L. Mazerolle, Lesley Ritchie, John D. Fisk, and Jodie R. Gawryluk. Longitudinal changes in microstructural white matter metrics in alzheimer’s disease. NeuroImage: Clinical, 13:330–338, 2017.spa
dc.relation.referencesJosue Dalboni, Ivanei Bramati, Gabriel Coutinho, Fernanda Moll, and Ranganatha Sitaram. Fractional anisotropy changes in parahippocampal cingulum due to alzheimer’s disease. Scientific Reports, 10, 02 2020.spa
dc.relation.referencesVince D. Calhoun and Jing Sui. Multimodal fusion of brain imaging data: A key to finding the missing link(s) in complex mental illness. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 1(3):230–244, 2016. Brain Connectivity in Psychopathology.spa
dc.relation.referencesYun Wang, Chenxiao Xu, Ji-Hwan Park, Seonjoo Lee, Yaakov Stern, Shinjae Yoo, Jong Hun Kim, Hyoung Seop Kim, and Jiook Cha. Diagnosis and prognosis of Alzheimer’s disease using brain morphometry and white matter connectomes. NeuroImage: Clinical, 23:101859, January 2019.spa
dc.relation.referencesXianjin Dai, Yang Lei, Yabo Fu, Walter J. Curran, Tian Liu, Hui Mao, and Xiaofeng Yang. Multimodal MRI Synthesis Using Unified Generative Adversarial Networks. Medical physics, 47(12):6343–6354, December 2020.spa
dc.relation.referencesYingxue Pang, Jianxin Lin, Tao Qin, and Zhibo Chen. Image-to-image translation: Methods and applications. 2021.spa
dc.relation.referencesHenri Hoyez, Cédric Schockaert, Jason Rambach, Bruno Mirbach, and Didier Stricker. Unsupervised image-to-image translation: A review. Sensors, 22(21), 2022.spa
dc.relation.referencesJun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired image-to-image translation using cycle-consistent adversarial networks. In 2017 IEEE International Conference on Computer Vision (ICCV). IEEE, October 2017.spa
dc.relation.referencesPhillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-image translation with conditional adversarial networks. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5967–5976, 2017.spa
dc.relation.referencesLei Wang, Wei Chen, Wenjia Yang, Fangming Bi, and Fei Yu. A state-of-the-art review on image synthesis with generative adversarial networks. IEEE Access, PP:1–1, 03 2020.spa
dc.relation.referencesIan J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks, 2014.spa
dc.relation.referencesRongguang Wang, Vishnu Bashyam, Zhijian Yang, Fanyang Yu, Vasiliki Tassopoulou, Sai Spandana Chintapalli, Ioanna Skampardoni, Lasya P. Sreepada, Dushyant Sahoo, Konstantina Nikita, Ahmed Abdulkadir, Junhao Wen, and Christos Davatzikos. Applications of generative adversarial networks in neuroimaging and clinical neuroscience. NeuroImage, 269:119898, 2023.spa
dc.relation.referencesMehdi Mirza and Simon Osindero. Conditional generative adversarial nets, 2014.spa
dc.relation.referencesKarissa Chan, Pejman Jabehdar Maralani, Alan R. Moody, and April Khademi. Synthesis of diffusion-weighted mri scalar maps from flair volumes using generative adversarial networks. Frontiers in Neuroinformatics, 17, 2023.spa
dc.relation.referencesXuan Gu, Hans Knutsson, Markus Nilsson, and Anders Eklund. Generating diffusion mri scalar maps from t1 weighted images using generative adversarial networks. In Michael Felsberg, Per-Erik Forssén, Ida-Maria Sintorn, and Jonas Unger, editors, Image Analysis, pages 489–498, Cham, 2019. Springer International Publishing.spa
dc.relation.referencesHaoyu Lan, the Alzheimer Disease Neuroimaging Initiative, Arthur W. Toga, and Farshid Sepehrband. Three-dimensional self-attention conditional gan with spectral normalization for multimodal neuroimaging synthesis. Magnetic Resonance in Medicine, 86(3):1718–1733, 2021.spa
dc.relation.referencesJake McNaughton, Justin Fernandez, Samantha Holdsworth, Benjamin Chong, Vickie Shim, and Alan Wang. Machine learning for medical image translation: A systematic review. Bioengineering, 10(9), 2023.spa
dc.relation.referencesKarim Armanious, Chenming Jiang, Marc Fischer, Thomas Küstner, Tobias Hepp, Konstantin Nikolaou, Sergios Gatidis, and Bin Yang. Medgan: Medical image translation using gans. Computerized Medical Imaging and Graphics, 79:101684, 2020.spa
dc.relation.referencesLingke Kong, Chenyu Lian, Detian Huang, Zhenjiang Li, Yanle Hu, and Qichao Zhou. Breaking the dilemma of medical image-to-image translation, 2021.spa
dc.relation.referencesSeong-Jin Son, Bo yong Park, Kyoungseob Byeon, and Hyunjin Park. Synthesizing diffusion tensor imaging from functional mri using fully convolutional networks. Computers in Biology and Medicine, 115:103528, 2019.spa
dc.relation.referencesBenoit Anctil-Robitaille, Antoine Théberge, Pierre-Marc Jodoin, Maxime Descoteaux, Christian Desrosiers, and Hervé Lombaert. Manifold-aware synthesis of high-resolution diffusion from structural imaging. Frontiers in Neuroimaging, 1, 2022.spa
dc.relation.referencesOlaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation, 2015.spa
dc.relation.referencesDor Bank, Noam Koenigstein, and Raja Giryes. Autoencoders, 2021.spa
dc.relation.referencesAbel Gonzalez-Garcia, Joost van de Weijer, and Yoshua Bengio. Image-to-image translation for cross-domain disentanglement, 2018.spa
dc.relation.referencesApoorva Sikka, Skand Vishwanath Peri, and Deepti R. Bathula. MRI to FDG-PET: Cross-Modal Synthesis Using 3D U-Net for Multi-modal Alzheimer’s Classification. In Ali Gooya, Orcun Goksel, Ipek Oguz, and Ninon Burgos, editors, Simulation and Synthesis in Medical Imaging, pages 80–89, Cham, 2018. Springer International Publishing.spa
dc.relation.referencesYunbi Liu, Ling Yue, Shifu Xiao, Wei Yang, and Mingxia Liu. Assessing clinical progression from subjective cognitive decline to mild cognitive impairment with incomplete multi-modal neuroimages. Medical Image Analysis, 75:102266, 10 2021.spa
dc.relation.referencesAl yhuwert Murcia Tapias, Diana L. Giraldo, and Eduardo Romero. Synthesizing fractional anisotropy maps from T1-weighted magnetic resonance images using a simplified generative adversarial network. In Barjor S. Gimi and Andrzej Krol, editors, Medical Imaging 2024: Clinical and Biomedical Imaging, volume 12930, page 129302P. International Society for Optics and Photonics, SPIE, 2024.spa
dc.relation.referencesR. C. Petersen, P. S. Aisen, L. A. Beckett, M. C. Donohue, A. C. Gamst, D. J. Harvey, C. R. Jack, W. J. Jagust, L. M. Shaw, A. W. Toga, J. Q. Trojanowski, and M. W. Weiner. Alzheimer’s disease neuroimaging initiative (adni) clinical characterization. Neurology, 74(3):201–209, 2010.spa
dc.relation.referencesNicholas J Tustison, Brian B Avants, Philip A Cook, Yuanjie Zheng, Alexander Egan, Paul A Yushkevich, and James C Gee. N4itk: improved n3 bias correction. IEEE transactions on medical imaging, 29(6):1310—1320, June 2010.spa
dc.relation.referencesJuan Iglesias, Cheng-Yi Liu, Paul Thompson, and Z. Tu. Robust brain extraction across datasets and comparison with publicly available methods. IEEE transactions on medical imaging, 30:1617–34, 09 2011.spa
dc.relation.referencesWilliam Penny, Karl Friston, John Ashburner, Stefan Kiebel, and T. Nichols. Statistical Parametric Mapping: The Analysis of Functional Brain Images. 01 2007.spa
dc.relation.referencesJacob C Reinhold, Blake E Dewey, Aaron Carass, and Jerry L Prince. Evaluating the impact of intensity normalization on MR image synthesis. In Medical Imaging 2019: Image Processing, volume 10949, page 109493H. International Society for Optics and Photonics, 2019.spa
dc.relation.referencesJelle Veraart, Dmitry S. Novikov, Daan Christiaens, Benjamin Ades-aron, Jan Sijbers, and Els Fieremans. Denoising of diffusion mri using random matrix theory. NeuroImage, 142:394–406, 2016.spa
dc.relation.referencesJesper Andersson and Stamatios Sotiropoulos. An integrated approach to correction for off-resonance effects and subject movement in diffusion mr imaging. NeuroImage, 125, 10 2015.spa
dc.relation.referencesStephen M Smith, Mark Jenkinson, Mark W Woolrich, Christian F Beckmann, Timothy E J Behrens, Heidi Johansen-Berg, Peter R Bannister, Marilena De Luca, Ivana Drobnjak, David E Flitney, Rami K Niazy, James Saunders, John Vickers, Yongyue Zhang, Nicola De Stefano, J Michael Brady, and Paul M Matthews. Advances in functional and structural mr image analysis and implementation as fsl. NeuroImage, 23 Suppl 1:S208—19, 2004.spa
dc.relation.referencesJelle Veraart, Jan Sijbers, Stefan Sunaert, Alexander Leemans, and Ben Jeurissen. Weighted linear least squares estimation of diffusion mri parameters: Strengths, limitations, and pitfalls. NeuroImage, 81:335–346, 2013.spa
dc.relation.referencesP.J. Basser, J. Mattiello, and D. Lebihan. Estimation of the effective self-diffusion tensor from the nmr spin echo. Journal of Magnetic Resonance, Series B, 103(3):247–254, 1994.spa
dc.relation.referencesJ.-D. Tournier, R. E. Smith, D. Raffelt, R. Tabbara, T. Dhollander, M. Pietsch, E. Christiaens, B. Jeurissen, C. Yeh, R. F. N. van den Berg, D. A. Porter, M. P. J. van Osch, S. Jbabdi, A. Tax, T. E. Nichols, E. J. W. van Nunen, J. A. De Backer, D. L. Flanagan, G. Calamante, C. Connell, and A. Connelly. Mrtrix3: A fast, flexible and open software framework for medical image processing and visualisation. NeuroImage, 202:116137, 2019.spa
dc.relation.referencesMark Jenkinson, Peter Bannister, Michael Brady, and Stephen Smith. Improved optimization for the robust and accurate linear registration and motion correction of brain images. NeuroImage, 17(2):825–841, 2002.spa
dc.relation.referencesDouglas N. Greve and Bruce Fischl. Accurate and robust brain image alignment using boundary-based registration. NeuroImage, 48(1):63–72, 2009.spa
dc.relation.referencesKaren Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition, 2015.spa
dc.relation.referencesAlain Horé and Djemel Ziou. Image quality metrics: Psnr vs. ssim. In 2010 20th International Conference on Pattern Recognition, pages 2366–2369, 2010.spa
dc.relation.referencesF. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830, 2011.spa
dc.relation.referencesB. Efron. Bootstrap Methods: Another Look at the Jackknife. The Annals of Statistics, 7(1):1 – 26, 1979.spa
dc.relation.referencesSergey Kastryulin, Jamil Zakirov, Nicola Pezzotti, and Dmitry V. Dylov. Image quality assessment for magnetic resonance imaging. IEEE Access, 11:14154–14168, 2023.spa
dc.relation.referencesHuanqing Yang, Hua Xu, Qingfeng Li, Yan Jin, Weixiong Jiang, Jinghua Wang, Yina Wu, Wei Li, Cece Yang, Xia Li, Shifu Xiao, Feng Shi, and Tao Wang. Study of brain morphology change in alzheimer’s disease and amnestic mild cognitive impairment compared with normal controls. General Psychiatry, 32(2), 2019.spa
dc.rights.accessrightsinfo:eu-repo/semantics/openAccessspa
dc.rights.licenseAtribución-NoComercial 4.0 Internacionalspa
dc.rights.urihttp://creativecommons.org/licenses/by-nc/4.0/spa
dc.subject.ddc620 - Ingeniería y operaciones afines::629 - Otras ramas de la ingenieríaspa
dc.subject.lembResonancia magnética en imágenesspa
dc.subject.lembMagnetic resonance imagingeng
dc.subject.proposalMRI synthesiseng
dc.subject.proposalStructural MRIeng
dc.subject.proposalFractional anisotropyeng
dc.subject.proposalDiffusion weighted imagingeng
dc.subject.proposalImage-to-image translationeng
dc.subject.proposalGenerative adversarial networkseng
dc.subject.proposalSı́ntesis de resonancia magnéticaspa
dc.subject.proposalResonancia magnética estructuralspa
dc.subject.proposalAnisotropı́a fraccionalspa
dc.subject.proposalImágenes ponderadas en difusiónspa
dc.subject.proposalTraducción de imagen a imagenspa
dc.subject.proposalRedes generativas adversariasspa
dc.subject.umlsAnisotropíaspa
dc.subject.umlsAnisotropyeng
dc.subject.umlsModelos de Redes Neuralesspa
dc.subject.umlsNeural Network Simulationeng
dc.titleEnriching the structural MRI information by cross-scale associations with the diffusion-weighted MRIeng
dc.title.translatedEnriqueciendo la información de la resonancia magnética estructural mediante asociaciones interescala con la resonancia magnética ponderada por difusióneng
dc.typeTrabajo de grado - Maestríaspa
dc.type.coarhttp://purl.org/coar/resource_type/c_bdccspa
dc.type.coarversionhttp://purl.org/coar/version/c_970fb48d4fbd8a85spa
dc.type.contentTextspa
dc.type.driverinfo:eu-repo/semantics/masterThesisspa
dc.type.redcolhttp://purl.org/redcol/resource_type/TMspa
dc.type.versioninfo:eu-repo/semantics/publishedVersionspa
dcterms.audience.professionaldevelopmentBibliotecariosspa
dcterms.audience.professionaldevelopmentEstudiantesspa
dcterms.audience.professionaldevelopmentInvestigadoresspa
dcterms.audience.professionaldevelopmentMaestrosspa
dcterms.audience.professionaldevelopmentProveedores de ayuda financiera para estudiantesspa
oaire.accessrightshttp://purl.org/coar/access_right/c_abf2spa

Archivos

Bloque original

Mostrando 1 - 1 de 1
Cargando...
Miniatura
Nombre:
1101690075.2024.pdf
Tamaño:
3.41 MB
Formato:
Adobe Portable Document Format
Descripción:
Tesis de Maestría en Ingenieria-Biomédica - Enriching the Structural MRI information

Bloque de licencias

Mostrando 1 - 1 de 1
Cargando...
Miniatura
Nombre:
license.txt
Tamaño:
5.74 KB
Formato:
Item-specific license agreed upon to submission
Descripción: