The use of artificial intelligence (AI) in medical imaging tests is becoming increasingly popular. AI can be used to detect cancerous lesions, identify thickening of certain muscle structures, and monitor changes in blood flow through the heart and associated arteries. AI-based algorithms can also be used to combine several non-contrast magnetic resonance images with information about heart movement, improve the pathological signals they contain, and reveal scars in a similar way to conventional magnetic resonance imaging with contrast. To ensure accuracy, the Food and Drug Administration (FDA) must approve any algorithm that includes images, and it must be 80 to 90% accurate of the time. In order to promote standardized, safe and effective AI for diagnosis and clinical decision support, the Data Science Institute of the American College of Radiology (ACR DSI) has published a series of high-value use cases for artificial intelligence in medical imaging, which will be continuously updated as new opportunities arise.
They are creating large scale image repositories with automatic extraction of biomarkers from images to characterize the condition of patients. AI tools are more likely to detect subtle variations in images that could indicate instability that requires surgery. For example, when a patient enters the emergency department with a complaint such as shortness of breath, “a chest x-ray is often the first available imaging study” according to the ACR DSI. AI-PHI uses the QUIBIM precision platform to centrally manage, store and quantitatively analyze images and medical algorithms. In recent decades, medical images have evolved from projection images such as x-rays or flat scintigrams to tomographic images. Researchers are still exploring the best ways to interpret and display the information in these T1 maps without contrast, which is one of the reasons why doctors are not yet using them widely.
Just like our natural intelligence, AI algorithms analyze medical images to identify patterns after they have been trained through a large number of exams and images. The use of AI in medical imaging tests has many potential benefits. It can help detect cancerous lesions, identify thickening of certain muscle structures, monitor changes in blood flow through the heart and associated arteries, and detect subtle variations in images that could indicate instability that requires surgery. However, it is important to ensure accuracy by having algorithms approved by the FDA before they are used in clinical settings.
