Development of Artificial Intelligence-Based Programs for the Diagnosis of Myocarditis in COVID-19 Using Chest Computed Tomography Data» DOI Creative Commons
Євген Настенко, Maksym O. Honcharuk, Віталій Бабенко

и другие.

Ukrainian journal of cardiovascular surgery, Год журнала: 2024, Номер 32(3), С. 58 - 65

Опубликована: Сен. 27, 2024

It has been established that 7.2% of patients hospitalized with coronavirus disease (COVID-19) exhibit signs heart disease, 23% these experiencing failure. Currently, there is a lack data on chest computed tomography (CT) for diagnosing myocarditis associated COVID-19. The aim. To justify the feasibility and develop classification models in COVID-19 based CT processing. Materials methods. A retrospective analysis from 140 was conducted. Chest scans were analyzed using DRAGONFLY software, permission Object Research Systems. COVID-CT-MD database, which includes 169 confirmed cases SARS-CoV-2 infection, used to build models. regions interest fragments images. Texture methods employed create diagnostic Results. shown average density myocardium patient diagnosis infection according Hounsfield scale does not essentially differ densitometric indicators healthy person. Therefore, research focused finding structural changes images their use constructing different algorithms had little effect accuracy, probably due informational content input data. However, obtained accuracy acceptable allows them be support medical decision-making regarding treatment. Conclusions. Using classic methods, diagnosed 7.1% severe pneumonia caused by coronavirus. global closely aligns results our clinical studies. allowed assessing characteristic acute form infection. constructed indicate specific during can identified CT. highest test samples reached 74%. implementation developed programs texture artificial intelligence technologies enables assessment long-term treatment efficiency. Creation significantly simplifies work radiologists improves efficiency

Язык: Английский

Development of Artificial Intelligence-Based Programs for the Diagnosis of Myocarditis in COVID-19 Using Chest Computed Tomography Data» DOI Creative Commons
Євген Настенко, Maksym O. Honcharuk, Віталій Бабенко

и другие.

Ukrainian journal of cardiovascular surgery, Год журнала: 2024, Номер 32(3), С. 58 - 65

Опубликована: Сен. 27, 2024

It has been established that 7.2% of patients hospitalized with coronavirus disease (COVID-19) exhibit signs heart disease, 23% these experiencing failure. Currently, there is a lack data on chest computed tomography (CT) for diagnosing myocarditis associated COVID-19. The aim. To justify the feasibility and develop classification models in COVID-19 based CT processing. Materials methods. A retrospective analysis from 140 was conducted. Chest scans were analyzed using DRAGONFLY software, permission Object Research Systems. COVID-CT-MD database, which includes 169 confirmed cases SARS-CoV-2 infection, used to build models. regions interest fragments images. Texture methods employed create diagnostic Results. shown average density myocardium patient diagnosis infection according Hounsfield scale does not essentially differ densitometric indicators healthy person. Therefore, research focused finding structural changes images their use constructing different algorithms had little effect accuracy, probably due informational content input data. However, obtained accuracy acceptable allows them be support medical decision-making regarding treatment. Conclusions. Using classic methods, diagnosed 7.1% severe pneumonia caused by coronavirus. global closely aligns results our clinical studies. allowed assessing characteristic acute form infection. constructed indicate specific during can identified CT. highest test samples reached 74%. implementation developed programs texture artificial intelligence technologies enables assessment long-term treatment efficiency. Creation significantly simplifies work radiologists improves efficiency

Язык: Английский

Процитировано

0