CHICAGO – Full-field digital mammography (FFDM) delivers a lower radiation dose to breast tissue than screen-film mammography (SFM), although it does require more exposures, according to results from the national Digital Mammographic Imaging Screening Trial (DMIST).
During a presentation at the RSNA on Tuesday, R. Edward Hendrick, Ph.D., from the Lynn Sage Comprehensive Breast Center at Northwestern University in Chicago discussed the technical performance of FFDM and SFM in the first 100 subjects from multiple DMIST sites.
“The goal of this specific project was to compare technical parameters that were used in DMIST and to assess possible technical problems on a site-by-site basis, to provide feedback to the sites, and to learn about differences between screen-film and FFDM,” Hendrick said.
The DMIST imaging sites performed FFDM and SFM on a total of 49,500 women between October 2001 and November 14, 2003. The 31 sites used breast imaging units from GE Medical Systems, Fischer Imaging, Fujifilm Medical Systems USA, and Hologic (Lorad).
For quality assurance, the first 100 subjects were included in the technical performance analysis. To date, technical assessment data has been complied from 18 sites on 1,610 subjects. Eventually, all 31 sites will be included.
“The data was checked for entry errors and re-checked by the site’s medical physicist,” Hendrick said.
The parameters that were compared included breast compression force, compressed breast thickness, mean glandular dose, and the number of additional views needed for complete breast coverage.
According to the combined results for 18 sites, the mean compression force was reported at 11.5 decaNewtons (dN) for both FFDM and SFM. The mean compressed breast thickness was 5.3 cm for SFM and 5.2 cm for FFDM.
Based on data from 16 sites, the mean glandular breast doses per view for SFM was 2.4 mGy and 1.9 mGy per view for FFDM -- 24% lower than screen-film.
“In terms of dose comparisons, this is where we started to
see some meaningful differences,” Hendrick said. “You
can see that for the Fischer slot scanning system that is using a manual
technique, the doses for digital were about a third lower than the doses for
screen-film. They were very similar for the
“For GE, using the automatic optimization parameters, the dose was 28% lower on digital than screen-film. The Lorad doses were quite similar,” he added. Also, data from the 100 patients at the Fischer sites also indicated that for larger breasts, dosage on screen-film rose while the dosage remained within a similar range on FFDM.
However, the decrease in dose did vary by manufacturer, he reported, ranging from 0%-52%. The great reduction was achieved with slot scanning systems.
Based on information from 18 sites, the number of cases requiring additional exposures beyond the normal four views was 215 (13.4%) for SFM and 292 (18.4%) for FFDM. Again, by manufacturer, the fraction of additional exposures needed for digital beyond SFM ranged from 0%-10.4%, with the highest percentage due to a 19.2 x 23.0 cm digital detector.
“The GE image receptor is 19 x 23 cm and therefore required considerably more exposures. About 15% of patients required extra exposures on the digital,” Hendrick said.
Given the significant differences between the manufacturers’ FFDM machines, one audience member questioned the validity of arriving at a single figure for FFDM dosage reduction or increase in exposures, saying that it was like “comparing CT to MR.”
“I understand your comment,” Hendrick replied. “It depends on the hypothesis being tested. The hypothesis that was decided on by consensus of the large number of people who designed DMIST is that we would test how digital compared to screen-film mammography in screening across the board.”
Dr. Etta Pisano, the DMIST lead investigator, added that a secondary analysis of the trial would make comparisons between the various systems. “In addition to that, we’re going to run a reader study for each manufacturer. So we’ll have a more powerful estimate of performance of each system.”
By Shalmali
Pal
AuntMinnie.com staff writer
December 3, 2003
