Breast imaging research team calls for raising the acceptable recall rate

Breast imaging practices that have screening recall rates of at least 12 percent but less than 14 percent are in the “sweet spot” for optimal detection of breast cancer. Practices that call back more than 14 percent of screening patients see little incremental benefit, while those that come in under 10 percent risk missing many cancers.

That’s according to the authors of a study running in the January edition of the American Journal of Roentgenology.

Paula Grabler, MD, of Rush University Medical Center in Chicago, and colleagues arrived at these ranges after reviewing the institution’s outcome audit data over a five-year period ending in 2012.

The researchers assessed cancer detection rates, along with percentages of minimally invasive and early-stage cancers, based on their analysis of 984 cancers identified in 188,959 digital screening mammograms.

They reviewed cases involving 19 breast radiologists and calculated 75 annual recall rates, which they analyzed in two ways. First they separated the rates into recall groups to assess cancer detection rate variation by the recall categories using escalating rate ratios (less than 10 percent, 10 percent to less than 12 percent, 12 percent to less than 14 percent, and 14 percent or higher).

Then they performed a linear regression with bootstrap bias correction to assess changes in cancer detection rate with each unit increase in the recall rate up to 20 percent. Here they used the recall category of less than 7 percent as reference.

Analyzing the results, the team found cancer detection rates were significantly higher for recall rates 12 percent or higher, with rate ratios of 1.75 and 2.06 for the recall groups of 12 percent to less than 14 percent and of 14 percent and higher, respectively, compared with the less than 10 percent group.

Considering the category “12 percent to less than 14 percent” as a reference, the researchers found no statistically significant differences between recall groups “12 percent to less than 14 percent” and “14 percent or higher” in cancer detection rate.

Additionally, they saw a statistically significant increase in the cancer detection rate with each unit increase in the recall rate only for recall rates 12 percent or higher.

In contextualizing their findings, Grabler et al. reference previous studies on recall rates that used data from the Breast Cancer Screening Consortium and that set the target range between 5 and 12 percent.

These recall ranges “were recommended on the basis of the trade-off between the number of additional women who needed to be recalled to detect one additional cancer,” the authors point out. “However, the use of sensitivity measures to set recall guidelines may be problematic because they do not take into account the size and stage of the breast cancer detected, which influence both prognosis and treatment.”

They further underscore that the data from which the Breast Cancer Screening Consortium recall recommendations were based are now 10 to 19 years old and predate the widespread adoption of digital mammography.

Moreover, while the likelihood of additional workups and false-positive results from higher recall rates is sure to raise objections, studies of women’s preferences show that “many women prefer the anxiety of a higher recall rate if it results in a higher breast cancer detection rate,” the authors write.  

Grabler and team name as a limitation their study’s focus on just one academic practice, acknowledging that this may affect the generalizability of the findings. Their study’s dimensions were substantial—75 radiologist years and 188,959 screening mammograms—but its findings could be strengthened by comparisons across a number of similar expert practices, they write.

“We recommend raising the acceptable recall rate to 14 percent to optimize breast cancer detection and better reflect expert practice,” the authors conclude. “Understanding that increasing recall rates does not guarantee improved cancer detection rates, breast imagers must compare their own practice numbers with these reference standards to assess the effect.”