The deep learning model can identify parasites that cause the disease in blood sample images as accurately as human experts, but in a fraction of the time, researchers wrote in the Journal of Digital Imaging.
Two researchers found that using certain pixel dimensions helped tailor algorithms to detect specific abnormalities, and pushed radiology to keep this in mind when using such approaches.
Japan-based researchers believe the algorithm can illuminate "hidden" information contained in imaging exams, and help radiologists in their clinical decision-making.
That's according to a Jan. 17 white paper published by the American Society of Radiologic Technologists.
Experts believe their approach will allow specialists to pinpoint brain-related pathologies—such as physical injuries, cancer or language disorders, among other things—with improved accuracy.
This most recent approval marks the fourth of its kind for Tel Aviv, Israel-based Aidoc.
AI trained and tested on more than 8,000 biopsies was nearly perfect at spotting differences in samples with or without cancer.
By combining AI with coronary artery calcium scoring and other cardiac measurements, the team would have prevented 73 unnecessary scans.
The new approach can diagnose brain tumors similarly to humans, but in a fraction of the time.
The Silicon Valley company's DeepMind AI beat out six expert readers and with further clinical testing could change the face of early breast cancer detection.
It’s no surprise that AI dominated the landscape this past year, but there were still a number of important stories that will likely become trending topics as radiology continues to evolve.
Radiomic analysis can extract mounds of information from MRIs and help researchers determine if a patient’s cancer is likely to return 10 years after treatment.