Challenges in Implementing the Local Node Infrastructure for a National Federated Machine Learning Network in Radiology DOI Open Access
Paul Jacobs,

Constantin Ehrengut,

Andreas Bucher

и другие.

Healthcare, Год журнала: 2023, Номер 11(17), С. 2377 - 2377

Опубликована: Авг. 23, 2023

Data-driven machine learning in medical research and diagnostics needs large-scale datasets curated by clinical experts. The generation of large can be challenging terms resource consumption time effort, while generalizability validation the developed models significantly benefit from variety data sources. Training algorithms on smaller decentralized through federated reduce but require implementation a specific ambitious infrastructure to share data, computing time. Additionally, it offers opportunity maintaining keeping locally. Thus, safety issues avoided because patient must not shared. Machine are trained local sharing model an established network. In addition commercial applications, there also numerous academic customized implementations network infrastructures available. configuration these networks primarily differs, yet adheres standard framework composed fundamental components. this technical note, we propose basic requirements for governance, science workflows, node set-up, report advantages experienced pitfalls implementing with German Radiological Cooperative Network initiative as use case example. We show how built upon some base components reflect they implemented considering both global requirements. After analyzing deployment process different settings scenarios, recommend integrating into existing IT infrastructure. This approach benefits maintenance effort compared external integration separate environment (e.g., radiology department). proposed groundwork taken exemplary development guideline future applications scientific environments.

Язык: Английский

Early and Rapid COVID-19 Diagnosis Using a Symptom-Based Machine Learning Model DOI Open Access
Abdul Samad, Muhammed Kürşad Uçar

International Journal of Innovative Science and Research Technology (IJISRT), Год журнала: 2024, Номер unknown, С. 1537 - 1543

Опубликована: Авг. 2, 2024

The COVID-19 pandemic has resulted in a significant global health crisis, claiming over 6.3 million lives. Rapid and accurate detection of symptoms is essential for effective public responses. This study utilizes machine learning algorithms to enhance the speed accuracy diagnosis based on symptom data. By employing Spearman feature selection algorithm, we identified most predictive features, thereby improving model performance reducing number features required. decision tree algorithm proved be effective, achieving an 98.57%, perfect sensitivity 1, high specificity 0.97. Our results indicate that combining various with AI-based techniques can accurately detect patients. These findings surpass previous studies, demonstrating superior across multiple evaluations. integration advanced models offers practical efficient tool early diagnosis, patient management approach holds promise enhancing healthcare delivery.

Язык: Английский

Процитировано

1

Enhancing Interoperability and Harmonisation of Nuclear Medicine Image Data and Associated Clinical Data DOI Creative Commons
Timo Fuchs, Lena Kaiser, Dominik Müller

и другие.

Nuklearmedizin - NuclearMedicine, Год журнала: 2023, Номер 62(06), С. 389 - 398

Опубликована: Окт. 31, 2023

Abstract Nuclear imaging techniques such as positron emission tomography (PET) and single photon computed (SPECT) in combination with (CT) are established modalities clinical practice, particularly for oncological problems. Due to a multitude of manufacturers, different measurement protocols, local demographic or workflow variations well various available reconstruction analysis software, very heterogeneous datasets generated. This review article examines the current state interoperability harmonisation image data related field nuclear medicine. Various approaches standards improve compatibility integration discussed. These include, example, structured history, standardisation acquisition standardised preparation evaluation. Approaches acquisition, storage will be presented. Furthermore, presented prepare way that they become usable projects applying artificial intelligence (AI) (machine learning, deep etc.). concludes an outlook on future developments trends AI medicine, including brief research commercial solutions.

Язык: Английский

Процитировано

2

Overview of medical analysis capabilities in radiology of current Artificial Intelligence models DOI Creative Commons
Paulina Kosiorowska, Karolina Pasieka, Helena Perenc

и другие.

Quality in Sport, Год журнала: 2024, Номер 20, С. 53933 - 53933

Опубликована: Авг. 19, 2024

Judgment is fundamental in medicine, particularly when combining complex data layers with detailed decision-making processes. Radiology processes present a distinct challenge for medical decisions due to the amount and shortage time personnel capable of analyzing images. Additionally, it's crucial consider each patient's specific situation, including their current state disease history. Despite advancements technology, there are still significant hurdles accurately radiology data. Radiographic assessments, which predominantly based on visual inspections, could greatly benefit from enhanced computational analyses. Artificial intelligence (AI) particular holds potential significantly improve qualitative interpretation imaging by experts - automating even replacing some parts work. This article will be an overview possibilities challenges associated introducing new technology into spaces. Doctors struggling it limits how much care they can show patient. The image marked most important parts, AI produce more user friendly version description, suggesting what might problem later human evaluation. Understanding or cutting down spend analyze allow faster deliver radiologic description doctors dealing patient treatment.

Язык: Английский

Процитировано

0

Development of a novel machine learning model based on laboratory and imaging indices to predict acute cardiac injury in cancer patients with COVID-19 infection: a retrospective observational study DOI

G. Wan,

Xuefeng Wu, Xiaowei Zhang

и другие.

Journal of Cancer Research and Clinical Oncology, Год журнала: 2023, Номер 149(19), С. 17039 - 17050

Опубликована: Сен. 25, 2023

Язык: Английский

Процитировано

1

Challenges in Implementing the Local Node Infrastructure for a National Federated Machine Learning Network in Radiology DOI Open Access
Paul Jacobs,

Constantin Ehrengut,

Andreas Bucher

и другие.

Healthcare, Год журнала: 2023, Номер 11(17), С. 2377 - 2377

Опубликована: Авг. 23, 2023

Data-driven machine learning in medical research and diagnostics needs large-scale datasets curated by clinical experts. The generation of large can be challenging terms resource consumption time effort, while generalizability validation the developed models significantly benefit from variety data sources. Training algorithms on smaller decentralized through federated reduce but require implementation a specific ambitious infrastructure to share data, computing time. Additionally, it offers opportunity maintaining keeping locally. Thus, safety issues avoided because patient must not shared. Machine are trained local sharing model an established network. In addition commercial applications, there also numerous academic customized implementations network infrastructures available. configuration these networks primarily differs, yet adheres standard framework composed fundamental components. this technical note, we propose basic requirements for governance, science workflows, node set-up, report advantages experienced pitfalls implementing with German Radiological Cooperative Network initiative as use case example. We show how built upon some base components reflect they implemented considering both global requirements. After analyzing deployment process different settings scenarios, recommend integrating into existing IT infrastructure. This approach benefits maintenance effort compared external integration separate environment (e.g., radiology department). proposed groundwork taken exemplary development guideline future applications scientific environments.

Язык: Английский

Процитировано

0