Comments on AI in Radiological Imaging

Share

The Coalition for Healthcare AI says that a strong regulatory framework that ensures the safety, efficacy, and equity of healthcare artificial intelligence (AI) is essential to ensure patient safety and allow clinical deployment. In a letter commenting on the agency’s earlier public workshop on the evolving role of AI in radiological imaging, the coalition says it agrees with FDA’s outline of opportunities and challenges in this area and in particular the unique questions raised regarding patient-derived data responsibility; ethics, including the risks of undesired bias; explainability; liability; access to evidence and experience; and augmenting healthcare.

In its comment, the Medical Imaging & Technology Alliance offers a list of questions to consider when evaluating potential risks posed by an AI solution: 

  •          What does the AI solution do?
  •          What is the potential for review of the AI output and any intervention?
  •          What are the qualifications of the user?
  •          Where is the AI solution used?
  •          What is the potential patient harm of an incorrect output?
  •          How are updates to the AI algorithm managed? and
  •          What is the breadth of use?

Read more