Imaging Informatics Insider

Dear Imaging Informatics Insider,

For all the progress that has been made in PACS technology, there is always room for improvement. Indeed, PACS software developers still need to address important technological limitations that can have a negative impact on user experience, according to researchers from the U.K. and Kuwait.

After monitoring and analyzing posts in several online PACS discussion groups over a two-year period, the researchers concluded that the current generation of PACS has a number of limitations, including issues with storage, the number of windows on PACS workstations, and integrating PACS from different vendors. Click here for their opinions on how these and other PACS problems could be addressed.

The difficulty in gathering large, well-curated datasets is one of the big challenges in training artificial intelligence (AI) algorithms for imaging applications. An AI algorithm that can produce synthetic abnormal images could help solve that problem, however. Recent research found that synthetic brain MR images could be utilized to augment a small dataset or even used on their own to train a deep-learning model. Click here to learn more.

Speaking of AI, Swiss researchers have determined that an AI algorithm can be highly sensitive and specific for identifying acute findings on abdominal CT scans, enabling radiologists to prioritize reading these urgent exams. Click here for all the details on this workflow triage application.

In other news, AI can predict how cancer patients will respond to immunotherapy by analyzing a radiomic-based biomarker on CT scans, according to French researchers. Click here to access our report.

AI is often overhyped, though, says Dr. Paul McCoubrie of Southmead Hospital in Bristol, U.K., in an update to his popular "20 golden rules of radiology" articles on What else is on his list of new golden rules for radiology? Click here to find out.

Bad or unethical use of AI in radiology could be dangerous, and Italian researchers believe that regulation is needed to ensure AI is used safely, ethically, and appropriately with protection of patient privacy. Patients, radiologists, and regulatory authorities will have to collaborate to accomplish this goal, according to the researchers. Click here for our coverage of their analysis.

If you have any tips or suggestions for topics you'd like to see covered in the Imaging Informatics Community, please feel free to drop me a line.

Page 1 of 1247
Next Page