Major US Cancer Institutions Unite to Revolutionize Oncology with AI
Four leading cancer medical centres in the US have partnered to identify new ways to use AI tools to transform cancer research and care. Backed by $40 million in funding, the Cancer AI Alliance (CAIA) aims to become a milestone in leveraging the potential of technology to assist oncology development.
The alliance, formed by Fred Hutch Cancer Center – the coordinating body that secured its initial funding-, Dana-Farber Cancer Institute, Memorial Sloan Kettering Cancer Center, and The Sidney Kimmel Comprehensive Cancer Center and Whiting School of Engineering at Johns Hopkins, will rely on records from all participating centres while focusing on preserving data confidentiality. To achieve this, it will be joining forces with global technology giants AWS, Deloitte, Microsoft and NVIDIA.
AI As a Tool For Data Processing Acceleration
One of the main challenges that modern cancer researchers face when it comes to utilising AI devices is acquiring the computing resources needed to evaluate huge amounts of data quickly.
To address this issue, CAIA will provide members of the alliance with computer capacity to handle large amounts of information generated during normal cancer care, such as electronic health records, pathology imaging, medical images, and genomic sequences.
With AI, the data examination process is faster and more effective, and could potentially lead to significant advancements in tumor biology, treatment resistance, and the identification of new therapeutic targets.
The Downside of AI
AI-powered technologies often need to rely on real patient data to strengthen the learning mechanisms that enable them to function correctly.
However, the increasing application of AI tools in healthcare has made evident severe data privacy concerns. AI systems often acquire, handle and analyze sensitive data, including real medical information, and these procedures are on some occasions carried out without prior patient consent.
CAIA proposes a potential solution to this issue. The goal of the partnership is utilizing responsible AI to harness the combined data from the participating cancer centres, aiming to uncover valuable insights while ensuring data security, privacy, and compliance with regulatory and ethical norms.
CAIA will employ a federated AI learning architecture with rigorous governance, in which each cancer centre retains its own data and AI models are applied to the data to create results. The data are then averaged across participating members to reveal valuable insights while maintaining confidentiality.
As Thomas J. Lynch Jr., MD, Fred Hutch’s president and director and holder of the Raisbeck Endowed Chair stated, “Collectively, the data held by the nation’s leading cancer centers has been an untapped source of new cancer discoveries that has been out of reach. This alliance helps solve the key technical challenges that will enable us to securely use both AI and massive computational power to find these breakthrough insights and save more lives.”
What’s Next?
CAIA is planned to begin operations by the end of 2024 and produce its first insights by the end of 2025.
If successful, this cooperation could represent a potential step forward in overcoming the recurring privacy hurdle in the use of AI in medicine, guaranteeing that AI technologies are used in cancer development while protecting patient privacy.
As NVIDIA’s CEO Jensen Huang said, “The convergence of AI, multimodal healthcare data and federated learning will usher in a new era for cancer research.”
Although AI has already proven to be an invaluable tool to enhance research and treatment for different types of cancer, CAIA must overcome significant obstacles before these tools can be considered reliable in revolutionizing cancer care. Therefore, ensuring the accuracy and robustness of AI models, addressing biases in the data, and establishing transparent and explainable AI systems will be critical for technology’s long-term success in healthcare.