AIMed Radiology virtual conference took place last week.  Early on, at the keynote address, Dr. Geraldine McGinty, President of the American College of Radiology (ACR) shared her views on fostering a strong ecosystem for artificial intelligence (AI) in medical imaging. She said, overall, the radiology community is very enthusiastic towards the opportunity brought about by AI. They think of it the same way as industrial revolution took place in the 18th century and radiologists who use AI will eventually replace those who do not.

Understanding that AI is not perfect yet

In spite so, the present-day healthcare system is relatively fragile and fragmented. While physicians want to protect patients by providing them with better care and enhanced health outcomes, industrial partners are keener towards keeping their shareholder values. As such, managing this delicate relationship as the medical space continues to evolve for the benefits of patients becomes paramount.

As medicine becomes the epicenter for innovations, Dr. McGinty warned the danger of not paying attention to safety as many AI models are still not perfect at the moment. She cited x-ray as an example. It was a fast-adopted technology; used within four months of invention but without caution. Marie Curie brought a portable x-ray machine with her to the World War One battle field and trained about 150 staff to use it yet many of them fell ill due to high radiation exposure, including Marie Curie herself, who contracted aplastic anemia, a blood disorder years later.

Radiologists may have innovations in their DNA but they should be generous to themselves at the same time. Take an extra step on whether AI is trustworthy, particularly, the most complex Blackbox or the ones to be deployed in highly critical clinical scenarios. Think about how much should radiologists rely and the approach to use AI in their practices. Besides, AI had made mistakes in the past, those biases and structural racism that were once demonstrated indicate a need for active collaboration between clinicians and those outside the radiology space, as well as influencing regulators, to make sure AI is ready to move care forward.

Understand the reasons for using AI

Fundamentally, radiologists want to prevent illnesses; helping patients to get better as soon as possible, and stabilize the conditions of chronic patients. So, improving individual experiences of care, making healthcare more effective and reassured the health and wellbeing of those delivering care should be some of the primary reasons for using AI.

Personally, Dr. McGinty said in her own practice, she sees herself like an airport security check, she is constantly reading images, she never felt she had enough time to speak with her patients, and she will call people back for additional imaging. She believes all these are areas that AI can mitigate to make her a better imager.

She sees AI is a powerful tool; it is able to look at data in a way that’s not possible to be performed by human. In breast cancer screening, it means AI can help to minimize the need for sending patients for surgery which often creates a huge amount of anxiety. Likewise, by integrating imaging data, genomic data and the vast information coming from different channels, AI can probability predict how a patient may progress or whether individuals are at risks of certain conditions in the near future.

*

Author Bio

Hazel Tang A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.