The use of artificial intelligence (AI) in radiology can help doctors analyze patient data and facilitate the diagnosis of Covid-19, according to an analysis by the European Parliament think tank. However, in addition to its benefits, it brings with it the risk of error and legal and ethical challenges.
European Parliament experts noted that artificial intelligence-supported medical imaging is successfully used to detect serious diseases. In their opinion, imaging diagnostics can also play an important role in the fight against the coronavirus.
"In situations where there is a suspicion of an error with a negative Covid-19 test result, medical imaging offers additional diagnostic options and helps in assessing treatment outcomes, disease development and prognosis," the study indicated.
Medical SI imaging allows for much faster image processing. With the support of artificial intelligence, it takes about 10 seconds compared to 15 minutes for a regular computed tomography reading.
Experts say medical imaging powered by artificial intelligence can automate the search of extensive databases and more accurately isolate the infection on x-ray and computed tomography images. Consequently, it makes it easier to evaluate photos and identify Covid-19.
As indicated, physicians can also use machine learning algorithms to evaluate medical images, which provide better opportunities for locating and quantifying disease features. Thanks to this, you can diagnose the disease earlier and more accurately, and at the same time achieve a more accurate prognosis.
According to the think tank, AI medical imaging could be crucial for the rapid detection and classification of Covid-19. This system enables the immediate indication of a chest tomography scan with suspected Covid-19. Image recognition algorithms can also be used to predict deterioration or improvement in patients. Consequently, these forecasts can help with hospital planning. Data provided by artificial intelligence may also enable better categorization of patients.
The advantages of using artificial intelligence in medical imaging also include increasing the safety of patients taking X-rays and computed tomography. Thanks to the use of artificial intelligence, the exposure time to X-rays is reduced and low-dose tomography is performed. Algorithms speed up scanning and automate risk assessment.
The analysis of the PE think tank noted that SI medical imaging models are already used in many hospitals around the world. As noted, the US Food and Drug Administration allowed the widespread use of Covid-19 detection algorithms for lung imaging. In turn, the European Union has implemented and is funding a project to increase the use of AI-supported computed tomography in the diagnosis of coronavirus.
However, experts said there is still a lot of uncertainty about using artificial intelligence techniques in radiology. The authors of the report pointed to the lack of sufficiently large databases and knowledge about the long-term effects of Covid-19. In their opinion, this influences the development of the large-scale forecasting process.
The report emphasizes that the effectiveness of tools supported by artificial intelligence depends on the accuracy of the data, and the use of uncertain information may affect the accuracy and reliability of the results, which in turn may lead to misdiagnosis.
The authors of the publication also drew attention to the doubts concerning the legal procedures concerning the collection and processing of medical data of patients. Under the EU's General Data Protection Regulation, patients must consent to the use of their data, including medical scans and photos, in developing AI algorithms. Moreover, such permission must be renewed each time the algorithms are updated.
As highlighted in the report, the use of AI-assisted image recognition in medical diagnostics currently ranges from 1%. up to 20% depending on the disease. Most hospitals still lack the infrastructure, staff and knowledge to use AI-powered systems effectively. As a consequence, experts say, greater use of AI imaging remains one of the main challenges to be addressed in the Covid-19 pandemic. (PAP)
author: Mateusz Mikowski