August 2015

Imaging Informatics: DICOM Dose Tracking

By David Yeager
Radiology Today
Vol. 16 No. 8 P. 32

The DICOM Radiation Dose Structured Report (RDSR) is a tool that can help track radiation dose, but many radiologists do not realize what types of information they can and cannot capture with RDSR, according to an article in the July issue of the Journal of the American College of Radiology. Ioannis Sechopoulos, PhD, DABR, an associate professor of radiology and imaging sciences, hematology, and medical oncology at the Emory University School of Medicine and one of the article’s coauthors, says most people think of DICOM as just an imaging standard, but it can do more than that. DICOM can also save information about how a particular study was produced.

Increasing attention has been paid to tracking radiation dose in the past few years, but those data need to be stored in a standardized way so they can be analyzed. Prior to RDSR’s release in 2005, there was no standard way to separate radiation exposure data from the image data. Accessing and storing dose data also required a lot of storage space because the information was attached to the images. If an image was duplicated, the exposure data were duplicated, too, which could result in overestimates of radiation dose. Also, if an image was rejected, due to technical issues or patient motion, or not recorded, such as fluoroscopic images or a scout image for mammography, those radiation exposures were not saved. RDSR addressed those shortcomings.

“If the image is rejected and deleted, that header will not exist anymore, but the RDSR will,” Sechopoulos says. “In addition, if it was a radiation event that did not generate an image, there’s no image header with dose information, but a radiation dose structured report will be generated. For example, with digital mammograms, there is a very low dose scout exposure that occurs before the mammogram is acquired. That information is recorded in the SR section of the patient’s mammogram and stored in PACS.”

Since its release, RDSR has become part of the DICOM standard for certain imaging modalities: X-ray angiography, CT, CR, DR, mammography, and nuclear medicine. It captures all data related to X-ray output that affect radiation dose, such as tube voltage, tube current, exposure time, and tube filtration. It also captures parameters that are either set by the modality or based on estimates from measurements, such as CT dose index (CTDI), dose length product, and average glandular dose.

Getting More Specific
But RDSR has its limits. It does not record patient-specific dose. CTDI and other parameters are based on measurements from standard-sized phantoms. They record the radiation emitted by the modality but not the radiation that’s absorbed by the patient, and they don’t reflect variations between patients, such as differences in size, body composition, and position during the exam.
“The same exam with the same CTDI on a very small patient vs a very large patient results in very different dose levels to the patients and their organs,” Sechopoulos says.

For this reason, the Working Group 28 (Physics) of the DICOM Standard Committee is working on a new information object definition called Patient RDSR (P-RDSR). P-RDSR will archive estimates of the radiation dose that a patient’s organ or entire body receives, based on information related to patient size and positioning, as well as archival data of the methodology that a medical physicist used to estimate the dose. Sechopoulos says potential uses include improving the accuracy and utility of data that are submitted to national dose registries and providing data for large, retrospective studies that measure the effectiveness of different imaging protocols or shielding techniques.

“Once you have that information, and it’s stored in a way that can be machine read in a standardized way, then it could be used to study, for example, the effectiveness of different types of shielding in reducing patient dose to particular organs,” Sechopoulos says. “With these structured reports, we hope it will be easier to characterize how well these techniques are doing. In the future, if you want to have a better assessment of how particular techniques affect dose in patients, we’ll have more accurate data and easier access to that data in a high number of patients.”

P-RDSR is being developed with input from the ACR, the American Association of Physicists in Medicine, the European Federation of Organisations for Medical Physics, and the European Society of Radiology, as well as suggestions from vendors. It will be submitted for approval next year, and if it’s approved, Sechopoulos estimates that it will take three to five years to be integrated with vendor offerings. He believes the standard is not only feasible but also highly beneficial for radiology practice, and he sees a lot of potential uses.

“If you have an automatic method to store data for each study for each patient, then it will be much easier to do Big Data analysis,” Sechopoulos says. “You can go back and look at the structured reports and see how the techniques have changed and, especially, if you have the patient dose structured report, how the organ dose has changed with respect to the exam and taking the patient size and other factors into account. That’s not easy to do without a systematic, automated system.”