Zobrazit v EDS

Evaluating the use of synthetic T1-w images in new T2 lesion detection in multiple sclerosis

Uloženo v:
Podrobná bibliografie
Název: Evaluating the use of synthetic T1-w images in new T2 lesion detection in multiple sclerosis
Autoři: Liliana Valencia, Albert Clèrigues, Sergi Valverde, Mostafa Salem, Arnau Oliver, Àlex Rovira, Xavier Lladó
Přispěvatelé: Agencia Estatal de Investigación, Institut Català de la Salut, [Valencia L, Clèrigues A, Oliver A, Lladó X] Research Institute of Computer Vision and Robotics, University of Girona, Girona, Spain. [Valverde S] Tensor Medical, Girona, Spain. [Salem M] Research Institute of Computer Vision and Robotics, University of Girona, Girona, Spain. Department of Computer Science, Faculty of Computers and Information, Assiut University, Asyut, Egypt. [Rovira À] L’Institut de Diagnòstic per la Imatge (IDI), Servei de Radiologia, Vall d'Hebron Hospital Universitari, Barcelona, Spain, Vall d'Hebron Barcelona Hospital Campus
Zdroj: Front Neurosci
Frontiers in Neuroscience, 2022, vol. 16, art.núm. 954662
Articles publicats (D-ATC)
DUGiDocs – Universitat de Girona
instname
Scientia
Scientia. Dipòsit d'Informació Digital del Departament de Salut
Frontiers in Neuroscience, Vol 16 (2022)
Informace o vydavateli: Frontiers Media SA, 2022.
Rok vydání: 2022
Témata: Artificial intelligence, Metric (unit), Economics, Image Processing, Esclerosi múltiple, 02 engineering and technology, Imatges -- Processament, multiple sclerosis, Pattern recognition (psychology), DISEASES::Nervous System Diseases::Autoimmune Diseases of the Nervous System::Demyelinating Autoimmune Diseases, CNS::Multiple Sclerosis, synthetic images, Engineering, 0302 clinical medicine, Pathology, 0202 electrical engineering, electronic engineering, information engineering, Lesion, ANATOMÍA::sistema nervioso::sistema nervioso central::encéfalo, Psychiatry, Otros calificadores::Otros calificadores::Otros calificadores::/diagnóstico por imagen, Other subheadings::Other subheadings::Other subheadings::/diagnostic imaging, ANATOMY::Nervous System::Central Nervous System::Brain, Otros calificadores::Otros calificadores::/terapia, Operations management, ENFERMEDADES::enfermedades del sistema nervioso::enfermedades autoinmunitarias del sistema nervioso::enfermedades autoinmunes desmielinizantes del SNC::esclerosis múltiple, Physical Sciences, Medicine, Computer Vision and Pattern Recognition, Radiology, MRI, RC321-571, Imaging systems in medicine, Autofocusing in Microscopy and Photography, brain, Other subheadings::Other subheadings::/therapy, Fluid-attenuated inversion recovery, Neurosciences. Biological psychiatry. Neuropsychiatry, Pathology and Forensic Medicine, Multiple sclerosis, 03 medical and health sciences, Magnetic resonance imaging, Image processing, Health Sciences, Media Technology, deep learning, Cervell - Imatgeria, Computer science, Diagnosis and Pathogenesis of Multiple Sclerosis, Computer Science, Nuclear medicine, Imatgeria mèdica, Imatgeria per ressonància magnètica, Image Denoising Techniques and Algorithms, Esclerosi múltiple - Tractament, Neuroscience
Popis: The assessment of disease activity using serial brain MRI scans is one of the most valuable strategies for monitoring treatment response in patients with multiple sclerosis (MS) receiving disease-modifying treatments. Recently, several deep learning approaches have been proposed to improve this analysis, obtaining a good trade-off between sensitivity and specificity, especially when using T1-w and T2-FLAIR images as inputs. However, the need to acquire two different types of images is time-consuming, costly and not always available in clinical practice. In this paper, we investigate an approach to generate synthetic T1-w images from T2-FLAIR images and subsequently analyse the impact of using original and synthetic T1-w images on the performance of a state-of-the-art approach for longitudinal MS lesion detection. We evaluate our approach on a dataset containing 136 images from MS patients, and 73 images with lesion activity (the appearance of new T2 lesions in follow-up scans). To evaluate the synthesis of the images, we analyse the structural similarity index metric and the median absolute error and obtain consistent results. To study the impact of synthetic T1-w images, we evaluate the performance of the new lesion detection approach when using (1) both T2-FLAIR and T1-w original images, (2) only T2-FLAIR images, and (3) both T2-FLAIR and synthetic T1-w images. Sensitivities of 0.75, 0.63, and 0.81, respectively, were obtained at the same false-positive rate (0.14) for all experiments. In addition, we also present the results obtained when using the data from the international MSSEG-2 challenge, showing also an improvement when including synthetic T1-w images. In conclusion, we show that the use of synthetic images can support the lack of data or even be used instead of the original image to homogenize the contrast of the different acquisitions in new T2 lesions detection algorithms.
Druh dokumentu: Article
Other literature type
Popis souboru: application/pdf
ISSN: 1662-453X
DOI: 10.3389/fnins.2022.954662
DOI: 10.60692/3101n-4xx23
DOI: 10.60692/eqmf1-g8a35
Přístupová URL adresa: https://pubmed.ncbi.nlm.nih.gov/36248650
http://hdl.handle.net/10256/21722
https://hdl.handle.net/11351/8516
https://doaj.org/article/7e0f2376f74244c986854de0acba638d
Rights: CC BY
Přístupové číslo: edsair.doi.dedup.....a3255e591c6227d3ccf8e6af8e83f092
Databáze: OpenAIRE
Popis
Abstrakt:The assessment of disease activity using serial brain MRI scans is one of the most valuable strategies for monitoring treatment response in patients with multiple sclerosis (MS) receiving disease-modifying treatments. Recently, several deep learning approaches have been proposed to improve this analysis, obtaining a good trade-off between sensitivity and specificity, especially when using T1-w and T2-FLAIR images as inputs. However, the need to acquire two different types of images is time-consuming, costly and not always available in clinical practice. In this paper, we investigate an approach to generate synthetic T1-w images from T2-FLAIR images and subsequently analyse the impact of using original and synthetic T1-w images on the performance of a state-of-the-art approach for longitudinal MS lesion detection. We evaluate our approach on a dataset containing 136 images from MS patients, and 73 images with lesion activity (the appearance of new T2 lesions in follow-up scans). To evaluate the synthesis of the images, we analyse the structural similarity index metric and the median absolute error and obtain consistent results. To study the impact of synthetic T1-w images, we evaluate the performance of the new lesion detection approach when using (1) both T2-FLAIR and T1-w original images, (2) only T2-FLAIR images, and (3) both T2-FLAIR and synthetic T1-w images. Sensitivities of 0.75, 0.63, and 0.81, respectively, were obtained at the same false-positive rate (0.14) for all experiments. In addition, we also present the results obtained when using the data from the international MSSEG-2 challenge, showing also an improvement when including synthetic T1-w images. In conclusion, we show that the use of synthetic images can support the lack of data or even be used instead of the original image to homogenize the contrast of the different acquisitions in new T2 lesions detection algorithms.
ISSN:1662453X
DOI:10.3389/fnins.2022.954662