More proof that US radiologists must abandon the ‘single-reader paradigm’

Results from a high-profile breast cancer screening program in Ireland are adding more ammunition to the argument that U.S. radiologists should shift away from the field’s “single-reader paradigm.”

Experts recently made a call for this philosophy change in the latest issue of Radiology, published Tuesday. Cambridge University Hospitals radiologist Nuala Healy, MD, and colleagues explored the use of consensus meetings to discuss discordant mammography findings over a five-year period.

The gatherings occurred biweekly among three to five radiologists, sometimes including the original reader. Healy and colleagues reported promising results from the Irish national breast screening program, with consensus meetings improving screening accuracy, and radiologists missing just eight interval cancers during the five-year study period, researchers noted.  

In a corresponding editorial, two experts said Healy’s investigation is further proof that the U.S. needs additional eyeballs on images. (In addition to consensus meetings, the Irish study also had two readers view every study.) Whether it’s artificial intelligence or another caregiver, something needs to change, wrote Solveig Hofvind, PhD, and Christoph Lee, MD.

“While many would argue that double reading remains too costly in the United States, the clear benefits observed from consensus meetings should give pause,” noted Hofvind and Lee, with Oslo Metropolitan University in Norway and the University of Washington in Seattle, respectively. “In the U.S., where there is still a single-reader paradigm, more sets of eyes would likely benefit patients and would prevent harms from false-positive screenings,” they added.

All told, radiologists discussed discordant mammography findings for 2,565 women at the consensus meetings in the Healy study. Of that, 40% were referred for further assessment, with cancer detected in 108 instances. Of the 1,285 women who came back for a biennial screening, radiologists detected malignancy at the original site of concern in 12 individuals at a further round of screening, according to the study. The retrospective analysis occurred from 2010 to 2014, with Healy and colleagues reviewing mammograms from a single unit of the Irish screening program, following a switch to full-field digital mammography.

Hofvind and Lee noted that single reading in the U.S. has typically produced an average recall rate of 10%. Most European countries, however, utilize an independent double reading, coupled with consensus or arbitration, tallying a 5% recall rate. The editorialists added that some in the U.S. are already experimenting with this shift. One study conducted at a major academic institution found that double reading all DBT callbacks led to “significant” decreases in recall rates and positive predictive value, without negatively affecting cancer detection rates. Radiologists spent just 2.3 minutes, on average, discussing each potential recall.

“Studies like that of Healy et al. should encourage us to question the single-reader paradigm,” the editorial writers concluded. “We need to think of new and innovative ways to get multiple sets of eyes looking at screening mammograms.”