Implementacija sustava umjetne inteligencije u radiologiji: povijesni razvoj, trendovi i buduće perspektive

Professional paper - Review paper

• Title (english): Implementation of Artificial Intelligence Systems in Radiology: Historical Development, Trends, and Future Perspectives
• Author: Neva Coce
• Author: Maja Prutki
• Author: Luka Filipović-Grčić
• Author: Tinamarel Mandić
• Author: Ivana Kuhtić
• Author's institution: University of Zagreb; School of Medicine; (Department of Radiology)
• Scientific / art field, discipline and subdiscipline: BIOMEDICINE AND HEALTHCARE; Clinical Medical Sciences; Radiology

Abstract (croatian)

Za razliku od klasičnih računalnih programa koji funkcioniraju na temelju eksplicitno definiranih instrukcija koje im programer daje, umjetna inteligencija (engl. artificial intelligence – AI) temeljena na strojnom učenju (engl. machine learning) i korištenju neuralnih mreža (engl. neural networking) može samostalno poboljšavati svoje performanse kako dobiva više podataka, odnosno ima sposobnost učenja. Takvi sustavi u radiologiji trenirani su na ogromnim skupovima podataka kako bi prepoznali patološke promjene na radiološkim snimkama uključujući tumore, frakture i druge abnormalnosti. AI sustavi u medicini dramatično mijenjaju način kako se dijagnosticiraju i liječe bolesti. Strojno učenje omogućava naprednu analizu medicinskih slika, točnije identificirajući patologiju. Ovi sustavi, nadmašujući tradicionalne metode, ubrzavaju dijagnostički proces i unapređuju kvalitetu slika, što dovodi do bržih i preciznijih dijagnoza. Primjena AI-a omogućuje i bolje praćenje pacijenata te predviđanje tijeka bolesti, doprinoseći personaliziranom pristupu liječenju. Međutim, izazovi poput integracije u zdravstvene sustave, obrade privatnosti i podataka, te etičkih pitanja predstavljaju prepreke za širu primjenu. Postoji potreba za rigoroznom validacijom AI algoritama, kao i za obukom medicinskog osoblja u korištenju ovih tehnologija. Unatoč ovim izazovima, budućnost AI-a u radiologiji izgleda obećavajuće. Očekuje se daljnji razvoj preciznijih i efikasnijih AI algoritama te njihova integracija s drugim medicinskim inovacijama. To će voditi ka boljem praćenju pacijenata i pružanju personalizirane skrbi. AI u radiologiji ima potencijal ne samo za poboljšanje kvalitete zdravstvene skrbi, već i za transformaciju načina na koji se skrb pruža, s naglaskom na pažljivu regulaciju i kontinuiranu edukaciju kako bi se maksimizirale koristi i minimizirali rizici.

Abstract (english)

Unlike traditional computer programs that operate based on explicitly defined instructions given by a programmer, artificial intelligence (AI), based on machine learning and the use of neural networks, can independently improve its performance as it receives more data, exhibiting the ability to learn. Such systems in radiology are trained on vast datasets to recognize pathological changes in radiological images, including tumors, fractures, and other abnormalities. AI systems in medicine are dramatically changing as diseases are diagnosed and treated. Machine learning enables advanced analysis of medical images and more accurate identification of pathology. These systems, surpassing traditional methods, speed up the diagnostic process and enhance image quality, leading to quicker and more accurate diagnoses. The application of AI also allows for better patient monitoring and disease progression prediction, contributing to a personalized approach to treatment. However, challenges like integration into healthcare systems, data privacy, processing, and ethical issues present obstacles to broader adoption. There is a need for rigorous validation of AI algorithms, as well as training medical staff in the use of these technologies. Despite these challenges, the future of AI in radiology looks promising. Further development of more precise and efficient AI algorithms and their integration with other medical innovations are expected. This will lead to better patient monitoring and the provision of personalized care. AI in radiology has the potential not only to improve the quality of healthcare but also to transform the way care is delivered, with a focus on careful regulation and continuous education to maximize benefits and minimize risks.

• Keywords (croatian): umjetna inteligencija – trendovi; radiologija; algoritmi; mašinsko učenje; neuronske mreže
• Keywords (english): artificial intelligence – trends; radiology; algorithms; machine learning; neural networks, computer
• Language: croatian
• Language: english
• Publication type: Professional paper - Review paper
• Publication status: Published
• Peer review: Peer review
• Publication version: Published version
• Journal title: Liječnički vjesnik
• Numbering: vol. 146, no. 9-10, pp. 346-349
• p-ISSN: 0024-3477
• e-ISSN: 1849-2177
• DOI: https://doi.org/10.26800/LV-146-9-10-4
• URN:NBN: urn:nbn:hr:105:962371
• Publication: 2024
• Document URL: https://hrcak.srce.hr/322408
• Type of resource: Text
• Access conditions: Open access
• Created on: 2024-12-17 10:18:03