Open source applications for imaging informatics have come a long way in the last decade.

During a recent Society for Imaging Informatics in Medicine webinar, moderated by SIIM Chair Paul Nagy, a panel of contributors to the June issue of the Journal of Digital Imaging discussed the new tools available in imaging informatics.

Steve Langer, a professor of diagnostic physics and imaging informatics at the Mayo Clinic, spoke about how he and colleagues have developed a way to access data on multiple imaging devices for research purposes.

“Suppose you want to know what the state of a given patient exam is, or you want to find certain types of studies for a research protocol,” Langer said. “You would have to go from machine to machine to find it.”

The team developed what Langer called a Standardized Operational Log of Events (SOLE), which allows machines to export system logs in a standard format to a central server, to make it easier to do clinical workflow data mining or research.

Langer also discussed how the SIIM Committee on Machine Learning has worked on reproducing machine learning study results using code and data, or the models derived from machine learning papers, comparing results of different models on publicized data sets.

Brad Genereaux, manager of product management and user experience for Agfa HealthCare, spoke about DICOMweb, and its importance for patient-centric and value-based care.

“There’s an ever-increasing demand for being able to access any image at any time, anywhere I happen to be, whether it’s at my desktop, or if I’m at home, or if I’m on the go in a subway or an airport with my mobile phone,” Genereaux said. “It’s also critically important that it is secure.”

Genereaux noted DICOMweb will be discussed at the upcoming SIIM annual meeting in National Harbor, Maryland.

Dr. Judy Gichoya, a radiology resident at the Indiana University School of Medicine and co-chair of the SIIM Global Outreach Committee, spoke about the Open MRS Open-Source Radiology Information System she is involved with.

Gichoya, who attended medical school in Kenya, showed a photograph of boxes filled with files of images and records, which she said displays the importance of electronic records systems. Open source projects can help make an impact in parts of the world where there isn’t access to advanced technology, she noted.

Dr. Paras Lakhani, clinical director for imaging informatics and assistant professor in the department of radiology at Thomas Jefferson University Hospital, discussed the article he co-authored for the Journal of Digital Imaging titled “Hello World Deep Learning in Medical Imaging”, which is a deep learning tutorial, complete with Python code.

“The idea was that people who are publishing in the future could sort of share this concept or model this concept … and encourage reproducible research,” Lakhani said.

Lakhani noted that there will be further opportunity to explore artificial intelligence and machine learning at the upcoming SIIM annual meeting.

Health IT Homepage