Check out the latest from Cancer Currents, an NCI cancer research blog, to learn more about artificial intelligence (AI) in cancer imaging and what it could mean for the future of cancer research.

Over the past several years, scientists have made exciting enhancements to AI, that is, algorithms using data to make decisions and predictions. Studies suggest that AI tools could make cancer imaging faster, more accurate, and more informative if they are allowed into clinics and doctors’ offices. But are these tools ready for real-world implementation?

In the research and development setting, AI tools have shown incredible potential. Studies suggest that AI and machine learning can take cancer imaging to the next level, increasing both speed and accuracy for doctors in cancer diagnosis, growth tracking, and treatment. Perhaps even more exciting is the potential of AI to go beyond human capabilities. In some cases, AI can detect patterns and complex relationships in data that humans cannot.

But despite the bright future, there are still a multitude of unanswered questions about the practical applications of these tools. Many algorithms, though proven accurate in preliminary testing, have never transitioned to the next testing phase. This phase, known as independent validation, determines if an AI tool can operate beyond the data that it was trained on. Additionally, questions concerning regulation, transparency, and potential data bias are ongoing.