Artificial Intelligence-Empowered Radiology-Current Status and Critical Review

Area of Science:

• Medical Imaging and Artificial Intelligence
• Radiology and Diagnostic Technologies
• Health Informatics and AI Applications

Background:

• Artificial intelligence (AI) is transforming fields reliant on human cognition, with neural networks advancing data processing.
• Convolutional neural networks (CNNs) and GPU technology have significantly improved image recognition accuracy and computational efficiency.
• AI applications in radiology offer enhanced diagnostic accuracy and efficiency for pattern detection and classification.

Purpose of the Study:

• To review advancements in AI-based imaging applications for diagnostic support.
• To focus on the role of CNNs, regulatory challenges, and labor impacts in diagnostic imaging.
• To analyze AI product development trends and key application areas in medical imaging.

Main Methods:

• Analysis of AI product development trends from 2017 to 2020.
• Identification of primary focus areas for AI products in medical imaging (neuroimaging, chest imaging, CT, MRI).
• Review of AI tool applications in high-prevalence diseases (lung cancer, stroke, breast cancer).
• Examination of regulatory compliance (MDD, MDR Class IIa/I) for AI medical devices.

Main Results:

• Neuroimaging, chest imaging, CT, and MRI are primary areas for AI products.
• AI tools target high-prevalence diseases like lung cancer, stroke, and breast cancer.
• AI product development surged from 2017-2020, peaking in 2020, followed by stabilization.
• Most AI products comply with moderate-risk standards under MDD and MDR.

Conclusions:

• AI, particularly CNNs, holds transformative potential for enhancing diagnostic imaging and support.
• Regulatory compliance is crucial for AI product development in radiology.
• The integration of AI in diagnostic imaging presents both opportunities and potential challenges to human labor.