Accreditation guides CT dose reduction for community radiologists

May 20, 2009

Like Goldilocks testing the bears’ porridge, the American College of Radiology and other professional societies are using diagnostic reference level (DRL) data to tell radiologists if the patient dose radiation from their CT scanners is too hot or just right.

Like Goldilocks testing the bears' porridge, the American College of Radiology and other professional societies are using diagnostic reference level (DRL) data to tell radiologists if the patient dose radiation from their CT scanners is too hot or just right.

Forging standards is not easy, said Dr. Cynthia McCollough, Ph.D., director of the CT Clinical Innovation Center at Mayo Clinic in Rochester, MN. The happy medium varies by patient age, size, and gender and by organ to be scanned, slice thickness, pitch, and other factors. She explained in a Tuesday morning lecture at the International Multidetector-Row CT conference in San Francisco that identifying a bad radiation dose is easier than defining a good one.

"The worst radiation dose is one that does not provide a diagnostic image," she said.

So, beyond that, how can radiologists determine if they are exposing patients to inappropriate radiation?

DRLs have helped the ACR, the European Commission, and other agencies offer guidance. They are an important part of the ACR's CT accreditation program, and they offer an alternative to effective-dose measurements that are easy to misinterpret, McCollough said.

DRLs do not specify ideal doses or mandate absolute upper limits to exposures, McCollough said, but they identify facilities where the dosage is often unusually high.

The benefits of DRLs can be seen in the U.K., where they have guided radiation reduction efforts since the early 1990s. They, along with improved equipment design, are credited with cutting the mean radiographic dose by 30% from 1984 to 1995 and another 50% from 1996 to 2000, she said.

"It shows how facilities in the highest quadrant of services in terms of radiation dose can be encouraged to reduce their dosages," McCollough said. "Over time, the range of dose distribution in the U.K. has narrowed, and the mean dose has declined."

DRLs produce a mean dosage for an average patient for the head, abdomen, or other general anatomic regions. Eighty mAs (a measure of tube current and x-ray photon intensity) may represent the mean, but it can be dialed down to 20 mAs, an ideal setting for infants, and up to 400 mAs for extra-large adults.

But mAs alone cannot serve as a standard because other variables frustrate attempts at simple measurement, McCollough said.

The ACR CT accreditation program adopted as its standard for measuring dosage CT dose indexes (CTDI) for measures obtained with its phantom. The method is able to boil everything down to a few statistics. CTDI is the dose measure at a specific location in the phantom. CTDIw is the weighted average of CTDI measures at the center and periphery of the phantom. And CTDIvol is an even more complicated measure, useful for calculating the dose of spiral scans.

The ACR's CTDIw reference dose for an adult head is 60 to 75 milligrays. For an adult abdomen, it is 25 to 35 mGy, and for a five-year-old child's abdomen, the reference dose is 20 to 25 mGy.

For CTDIvol, however, McCollough cautioned the audience that it is not synonymous with patient dose.

"You need to know how the equipment interacts with the patient," she said.

Still, the ACR reference system has been good enough to help cut radiation dosage at accredited sites. The mean dose of an adult head scan fell from 66.7 mGy in 2002 to 54.6 mGy in 2004.

DRLs are still needed for many more types of CT exams, including pulmonary emboli diagnosis and coronary CT angiography, McCollough said. The ACR is sponsoring a pilot project to collect CTDI volume information from DICOM header data in order to identify typical dosages. Dosage questions have been included in the ACR Imaging Network's Protection I coronary CTA trial.

For community-based radiologists, dosage calculation is still not as easy as sampling porridge, but McCollough believes that the ACR accreditation program provides a scale that can warn radiologists when their CT scanners are too hot.