Currently, radiologists base mammography radiation exposure estimates on a breast that could never exist in real life.

A nearly $1 million grant from Susan G. Komen may soon change that, as Emory University researchers work to create software that can determine patient-specific dose using tomosynthesis.

The hypothetical breast used in dose models has a homogeneous make-up of glandular and fat tissue, which does not occur in nature. The new technique will use the 3-D images, which tomosynthesis produces, to create an accurate map of a patient's glandular tissue, fat and skin distribution, to use tissue volume to get an accurate estimate of dose for the actual patient.

"All breasts are different in terms of internal texture, the amount of glandular tissue and where it is," medical physicist and lead researcher Ioannis Sechopoulos said to DOTmed News. "Our estimates based on a homogeneous breast are fine to get an idea and to compare mammography with tomo, or compare different vendors. But patient dose estimates can vary a lot based on thickness or the amount of glandular tissue."

Sechopoulos said his team did a study last year measuring estimated dose to true dose of mammography, using CT to create a 3-D image of the breast. The study looked at 20 patients and found that in some cases dose was overestimated by 30 to 50 percent. While those numbers are anecdotal, Sechopoulos said dose absorption varies somewhat widely based on glandular composition and size.

He hopes tomosynthesis dose monitoring will eventually help give patients better information about their individual dose, and allow medical physicists to improve quality control by tracking dose.

"Hopefully, eventually we can have dose registries for mammography screening like there is for CT dose," said Sechopoulos. "That's way down the road, but this will help us get there."

More Industry Headlines