DR systems have brought much benefit to radiology practice. However, they also come with significant challenges, such as technology and system complexity; vendor differences, including how to display exposure indication; dose creep; and inconsistency in the appearance of processed images, according to presenter Zaiyang Long, PhD. Studies in the literature also have shown disparities in technologist training.
The Mayo Clinic has a large number of DR systems from multiple vendors, as well as more than 100 technologists, and several thousand images are produced each day, Long noted.
"Consequently, it is a complex task to achieve image quality standardization and optimization," Long told AuntMinnie.com. "That's why we started mining the clinical DR image DICOM header to gain a better understanding of the practice patterns and to obtain data that could help to answer questions encountered in the process of standardization and optimization."
The data revealed complex practice patterns, such as detector usage, how the images were acquired, and whether processing parameters deviated from default settings, according to Long.
"[This] information helped us to understand causes for inconsistency, design education plans, improve image-processing parameters, and monitor changes," he said. "In addition, it provided data for purchase decisions. Overall, it's been very beneficial in our process of image quality standardization and optimization."
What else did the team find? Check out this late morning talk on Thursday to get all the details.
