DICOM data more accurate, reliable than RIS for MRI workflow improvements

DICOM metadata offers more accurate study information, according to a Jan. 8 study published in the Journal of Digital Imaging. This may ultimately help increase the efficiency of MRI exams while reducing their associated costs.

“By retrieving and parsing DICOM metadata, we can gain valuable insight into the MRI examination process far beyond that provided by the (radiology information system) RIS and identify opportunities for intervention and improvement,” wrote Ish A. Talati, of Georgetown University School of Medicine in Washington, and colleagues.

For their study, Talati et al. identified and extracted timestamped metadata from MedStar Georgetown University Hospital’ s prostate, abdomen MRCP and brain MRI exam database. In total, 443 exams consisting of the eight most frequent study descriptions were included.

Keying in on the abdomen, brain and prostate, the group found “substantial” differences between the exam times of each part and noted RIS tracking times were “consistently and substantially” lower compared to DICOM data.

“This confirmed our suspicion that manually tracked RIS times are unreliable for this type of detailed and granular analysis,” the group wrote.

Additionally, Talati et al. found shorter median exam times were associated with technologists who conducted more exams, suggesting experience and repetition are beneficial for improved efficiency, they wrote.

“Our experience shows that extracting DICOM metadata in a systematic way can provide valuable process improvement recommendations for site directors and technologists,” the authors wrote. “We also confirmed that this data is more accurate, reliable, and detailed than correlative RIS information.”

Overall, the DICOM MRI exam data allowed the authors to easily identify opportunities for improvement across various areas of their system. With the help of deep learning, the authors wrote, the manual process of obtaining data could become automated for further efficiency insights.