Not ready for clinical primetime: 3D tomosynthesis images rendered in 2D

Comparing the quality of synthesized 2D images created using 3D digital breast tomosynthesis (DBT) with the quality of images captured in 2D full field digital mammography (FFDM), a research team at Duke has found that the former is indeed better than the latter at depicting some visual features.

However, the synthesized approach yields poorer overall resolution and noise properties, leading the team to conclude that its use in clinical settings calls for careful consideration—especially among imaging sites thinking about eventually discontinuing FFDM outright as a way to reduce or eliminate radiation exposure.

Lead author Jeffrey Nelson, MHP, and colleagues note in their study report that, when the FDA approved the use of DBT in 2011, the agency stipulated that it may only be used as an adjunct to FFDM. Plus all DBT images must be paired with a 2D image to assure the availability of adequate interpretative information.

Specifically looking at Hologic’s C-View software, which renders a 2D image from 3D tomosynthesis data, the team set out to compare image quality between the two modalities on the bases of resolution, contrast and noise.

Their findings ended up underscoring Hologic’s user instructions, which include an explicit warning:“Use the C-View images as an aid while examining the tomosynthesis images. Do not make a clinical decision or diagnosis from the C-View images alone.”

Foiled by fiber breaks

The authors used the American College of Radiology’s mammography accreditation phantom and a novel 3D printed anthropomorphic breast phantom, imaging both using a Hologic Selenia Dimensions 3D system.

Using ACR-defined scoring criteria for the ACR phantom, the FFDM images “scored statistically higher than C-View according to both the average observer and automated scores,” the authors reported.

They also found that between 50 percent and 70 percent of C-View images failed to meet the nominal minimum ACR accreditation requirements, primarily due to fiber breaks.

“Software analysis demonstrated that C-View provided enhanced visualization of medium and large microcalcification objects; however, the benefits diminished for smaller high contrast objects and all low contrast objects,” Nelson et al. wrote.

In addition:

  • Visual analysis of the anthropomorphic phantom showed “a measureable loss of resolution in the C-View image and loss in detection of small microcalcification objects.”
  • Spectral analysis of the anthropomorphic phantom showed higher total noise magnitude in the FFDM image compared with C-View.
  • Whereas the FFDM image contained white noise texture, the C-View exhibited “marked noise reduction at midfrequency and high frequency with far less noise suppression at low frequencies, resulting in a mottled noise appearance.”

Beneficial benchmarking

The potential radiation-dose reduction aspect of C-View “is advantageous, but caution must be exercised in determining the extent of its clinical utility,” the authors wrote.

The researchers further point out that, while the physical specifications of C-View did not match up to FFDM, only a clinical trial of reader studies can determine whether tomosynthesis plus C-View is equivalent to tomosynthesis plus FFDM or FFDM alone—a question that was beyond the practical scope of their study.

“However, we should recall that requirements for mammographic images (contrast, resolution and noise) are meaningful to provide solid practical benchmarks to assure adequate image quality for effective mammographic interpretation of fibrous structures, microcalcifications and masses,” concluded Nelson et al. “As such, these criteria should be carefully considered in the context of any new breast imaging technology.”

The study is running in Medical Physics, which has posted the full paper.