SINGAPORE – AI is already transforming medical imaging practices and further technological growth is expected, according to a plenary address given on May 4 at the International Society of Magnetic Resonance in Medicine (ISMRM) Annual Meeting.
Dr. Charlene Liew, Director of Cardiothoracic Imaging at Singapore's Changi General Hospital, discussed the current state and future direction of AI in radiology in a presentation titled “AI in Radiology: The past informs the future.” I have outlined it.
“Today's radiology workforce is at the forefront of medical innovation, leveraging cutting-edge technology to improve patient outcomes,” she said.
In the 2010s, Chinese-born American computer scientist Dr. Feifei Li built ImageNet, a dataset of 14 million annotated images that enabled advances in computer vision. More than a decade later, AI has taken root in healthcare companies, especially in the medical imaging field. In fact, in 2022, 77% of his AI tools approved by the U.S. Food and Drug Administration (FDA) were for radiology indications.
Liew said AI is now being implemented in every step of the radiology workflow, from patient preparation, scan preparation and execution, to image reading and result reporting. The impact on the radiologist's workflow extends to the detection of poor image quality in X-rays. Automation of CT imaging protocols. For MRI, streamline workflows to reduce scan time, image reconstruction, and use synthetic MRI sequences.
“The advent of AI heralds a new era in medicine, characterized by the automation of tasks traditionally performed by radiologists and radiology technicians,” she said. “AI systems will revolutionize healthcare delivery, providing unprecedented speed and accurate diagnosis.”
Going forward, AI will continue to improve preclinical radiology workflows by automating measurements, segmentation, and longitudinal comparisons. Generate heatmaps to detect anomalies on images. And it extends beyond detection and triage to prognosis, patient outcome prediction, and diagnosis, Liew said.
She listed the following ongoing issues that departments must address when integrating AI into their operations:
· Aggregate annotated big AI data in a cost-effective and secure manner.
· Integrate AI into clinical workflows and build trust in technology among departmental staff.and
· Measure the impact of technology on productivity and quality.
“Introduction of AI” [process in radiology needs to take into account] Seamless orchestration of AI, explainable AI and smart reporting, responsible AI,” said Liew.
The goal, Liu said, is to ethically integrate humans and artificial intelligence. She quotes Albert Einstein, who says, “If the creations of our minds are to be a blessing and not a curse, concern for man and his destiny must always be the main objective.” and encouraged session participants to engage in human integration. AI is boldly tackling medical image processing.
“Be brave. You write history,” she concluded.
