Real-world radiology data for artificial intelligence-driven cancer support systems and biomarker development

Area of Science:

• Radiology
• Artificial Intelligence
• Oncology

Background:

• The convergence of artificial intelligence (AI) and real-world data (RWD) presents a transformative opportunity in medical imaging.
• Vast digital radiology datasets from routine clinical practice can be leveraged with AI for advanced diagnostic and therapeutic support systems.

Purpose of the Study:

• To explore the advantages and challenges of integrating AI with real-world radiology data for enhanced cancer patient care.
• To highlight the potential of AI in automating clinical processes, identifying novel biomarkers, and supporting cancer treatment decisions.

Main Methods:

• Review of current literature on AI applications in radiology utilizing real-world data.
• Analysis of deep learning models' data requirements and the role of RWD in accelerating assay development.
• Discussion of challenges including data access, quality, processing, and the need for standardization.

Main Results:

• AI applied to radiology data shows promise in automating tumor detection and identifying new cancer biomarkers.
• Real-world data provides diverse, extensive datasets crucial for training deep learning models and understanding real-world clinical practices.
• Standardized protocols and trust-building are essential for clinical applicability and implementation.

Conclusions:

• AI holds significant potential for analyzing real-world radiology imaging and reports to extract valuable information and advance biomarker discovery.
• Addressing challenges in data access, quality, and processing is key to realizing the practical benefits of AI in radiology.
• Successful implementation requires collaboration and trust among clinicians, patients, and regulatory bodies.