Estruturamento de banco de imagens mamográficas
Data
2023-08-08
Autores
Cordeiro, Natália Monteiro 
Orientadores
Elias, Simone
Tipo
Dissertação de mestrado
Título da Revista
ISSN da Revista
Título de Volume
Resumo
Objetivo: Estruturar um banco de imagens mamográficas digitais com objetivo de
ensino e pesquisa em saúde, disponibilizado em plataforma online de acesso
gratuito. Métodos: Nosso banco de mamografias foi construído em uma plataforma
de gerenciamento de imagens médicas e composto a partir da seleção retrospectiva
de cem exames realizados pelo Hospital São Paulo - Universidade Federal de São
Paulo entre os anos de 2019 e 2023. O projeto tem por premissas segurança,
versatilidade e usabilidade e exigiu busca ampla da ferramenta adequada no
mercado. Características essas compatíveis com produto disponibilizado pela Ambra
Health, empresa americana que desenvolveu software baseado em nuvem e se
mostrou como a alternativa ideal. Sua plataforma atende a critérios internacionais de
segurança de dados e permitiu a customização criteriosa pretendida, além da
possibilidade de associar anexos e realizar anotações compatíveis com estudos em
inteligência artificial. Assim, o banco foi organizado de acordo com variáveis
diversas. Categorias foram criadas em conformidade com os descritores BI-RADS®,
cenários clínicos variáveis e materiais adicionais disponibilizados - e essas
categorias foram base para o recurso de pesquisa avançada, que filtra intuitivamente
os exames que atenderem simultaneamente aos critérios selecionados. A plataforma
foi integrada a um sistema de anonimização automática ao f, garantindo a
privacidade dos dados. Após envio, os exames permanecem em área restrita ao
administrador para conferência da anonimização, categorização, anotações e
manejo de anexos para posterior liberação para o usuário final. Almejando
abrangência geográfica, os descritores foram inseridos em inglês americano -
respeitando a origem do léxico BI-RADS® - enquanto para a estrutura do site,
optou-se por tradução automática para a língua padrão do navegador de acesso.
Reforçando a premissa da usabilidade, o sistema é responsivo, ou seja, pode ser
acessado em dispositivos móveis. Resultados: Nosso site está ativo e disponível
em www.mamografia.unifesp.br com acesso concedido após simples processo de
registro. Cem casos anonimizados, 62% com imagens em 3D, podem ser filtrados a
partir da combinação de 171 variáveis clínicas e imaginológicas, assim como da
disponibilidade de conteúdo anexo.
vii
O idioma base é adaptável à língua nativa do usuário e as buscas categorizadas
podem ser acessadas diretamente do navegador ou ter download em formato de
dataset personalizados. Além disso, funcionalidades como salvar buscas específicas
ou favoritar exames específicos também estão disponíveis. Conclusão:
Construímos banco online e gratuito de mamografias digitais anonimizadas em
densa, mas fluida, interface de busca que é compatível com tomossínteses e anexos
diversos, permitindo estreita correlação clínica dos casos. Essas características
tornam nosso projeto inovador em sua versatilidade: compatível com ampla gama de
perfis de usuário e com potencial para atender demandas tanto pedagógicas como
Objective: To structure a digital mammographic image database for teaching and research purposes in healthcare, available on a free online platform. Methods: Our mammography database was developed through a retrospective selection of one hundred exams performed by Hospital São Paulo - Federal University of São Paulo between the years 2019 and 2023. The project has premises of safety, versatility, and usability, and required an extensive search for the appropriate tool. Ambra Health, an American company, has developed cloud-based software for medical image management and stood out as a viable alternative. Their platform meets international data security criteria, and made possible the intended careful customization, as well as the possibility of associating attachments and region of interest annotations. The categories were created in accordance with the BI-RADS® descriptors, a wide variety of clinical scenarios and available additional materials, and served as the basis for the advanced search feature, which intuitively filters exams that simultaneously meet the selected criteria. The platform was integrated with an automatic anonymization system upon upload, ensuring data privacy. After submission, the exams are retained in a restricted area for anonymization verification, categorization, annotation and attachment management, before being released to the end-user. In order to achieve geographic coverage, the descriptors were entered in American English, respecting the origin of the BI-RADS® lexicon, while for the website structure, automatic translation to the standard language of the accessing browser was chosen. Results: Our website is active and available at www.mamografia.unifesp.br, with access granted after a simple registration process. one hundred anonymized cases, 62% of which include 3D images, can be filtered based on the combination of 171 clinical and imaging variables, as well as attachment availability. The language is adaptable to the user's native language, and categorized searches can be accessed directly from the browser or downloaded as customized datasets. Additionally, features such as saved searches or starred exams are also available. Conclusion: We have developed an online and free mammography database that is completely innovative by integrating various resources into a single platform. We provide high-resolution and 3D digital images that can be searched using an advanced search system. ix Additionally, we offer supplementary clinical information in various attachment formats, enabling a rich clinical correlation. In this way, we achieve the ambivalence of our goal, which is to promote education and research.
Objective: To structure a digital mammographic image database for teaching and research purposes in healthcare, available on a free online platform. Methods: Our mammography database was developed through a retrospective selection of one hundred exams performed by Hospital São Paulo - Federal University of São Paulo between the years 2019 and 2023. The project has premises of safety, versatility, and usability, and required an extensive search for the appropriate tool. Ambra Health, an American company, has developed cloud-based software for medical image management and stood out as a viable alternative. Their platform meets international data security criteria, and made possible the intended careful customization, as well as the possibility of associating attachments and region of interest annotations. The categories were created in accordance with the BI-RADS® descriptors, a wide variety of clinical scenarios and available additional materials, and served as the basis for the advanced search feature, which intuitively filters exams that simultaneously meet the selected criteria. The platform was integrated with an automatic anonymization system upon upload, ensuring data privacy. After submission, the exams are retained in a restricted area for anonymization verification, categorization, annotation and attachment management, before being released to the end-user. In order to achieve geographic coverage, the descriptors were entered in American English, respecting the origin of the BI-RADS® lexicon, while for the website structure, automatic translation to the standard language of the accessing browser was chosen. Results: Our website is active and available at www.mamografia.unifesp.br, with access granted after a simple registration process. one hundred anonymized cases, 62% of which include 3D images, can be filtered based on the combination of 171 clinical and imaging variables, as well as attachment availability. The language is adaptable to the user's native language, and categorized searches can be accessed directly from the browser or downloaded as customized datasets. Additionally, features such as saved searches or starred exams are also available. Conclusion: We have developed an online and free mammography database that is completely innovative by integrating various resources into a single platform. We provide high-resolution and 3D digital images that can be searched using an advanced search system. ix Additionally, we offer supplementary clinical information in various attachment formats, enabling a rich clinical correlation. In this way, we achieve the ambivalence of our goal, which is to promote education and research.