News Archive
Second opinions notably reduce breast-biopsy misdiagnoses
Breast biopsy samples are often difficult to interpret. A sample that appears benign may be malignant, in which case the diagnosis of breast cancer may be missed. And a sample that appears malignant may be benign, in which case the error in diagnosis may cause the patient to undergo unnecessary treatment.
Because of this uncertainty, most pathology laboratories have policies requiring that certain findings are always reviewed by another pathologist for a second opinion, particularly when the initial diagnosis is invasive carcinoma. The criteria that calls for a second opinion can vary considerably, however, from lab to lab.
In a study published June 22 by the BMJ, U.S. researchers evaluated 12 such policies to gauge their effectiveness in reducing errors. Each strategy reduced errors in interpretation, they found, but none eliminated errors entirely.
The most effective strategy, they found, required second opinions for all biopsies.
DCIS-ductal carcinoma in situ
“The millions of women undergoing breast biopsy each year rely on us for a correct diagnosis,” said Dr. Joann Elmore, a professor of general internal medicine at the University of Washington and one of the study’s lead authors. “Our ultimate goal is to improve health and healthcare for these women, which begins with a correct diagnosis.”
In the study, the researchers used a simulation to evaluate the possible impact of 12 strategies for obtaining a second opinion. They used data from an earlier study in which U.S. pathologists were asked to interpret breast-biopsy samples whose diagnosis had been agreed upon by a panel of three experts. That study, published in 2015, found that the study pathologists disagreed with the consensus of the expert panel about 25 percent of the time; most disagreement involved samples from difficult-to-interpret conditions such as atypia, when the cells appear abnormal but are not cancerous, and ductal carcinoma in situ (DCIS), considered the earliest form of breast cancer.
In the current study, researchers scrutinized a range of strategies: One required all biopsies to receive a second opinion. Others required second opinions after initial interpretations of DCIS or invasive cancer. Others necessitated second opinions if the initial pathologist deemed the case to be difficult to interpret.
The researchers also examined the effect of having interpretations by low-volume pathologists (fewer than 10 breast biopsies a week) reviewed for a second opinion by a high-volume pathologist (more than 10 biopsies a week), and the effect of having the high-volume pathologist’s interpretation reviewed by another high-volume pathologist for a second opinion.
Elmore-biopsy
Except in cases of invasive cancer, all the second-opinion strategies reduced misclassification rates significantly. Obtaining a second opinion in cases of invasive cancers had little effect because rarely did the second opinion differ from the initial interpretation. Requiring second opinions for all biopsies was most effective, reducing the overall misclassification rate from 24.7 percent to 18.1 percent.
The lowest misclassification rate emerged when both the initial interpretation and second opinion came from high-volume pathologists. In these cases, only 14.3 percent of the pathologists’ interpretations differed from that of the expert consensus panel.
"The results support the common belief among clinicians that second opinions should be sought ideally from those with greater clinical experience and especially where the primary reviewing pathologist is uncertain,” the researchers wrote.
However, they add, “[n]one of the strategies completely eliminated diagnostic variability, especially for cases of breast atypia, suggesting that approaches beyond obtaining a second opinion should be investigated for these challenging cases.”
The study was published online June 22 in the BMJ. Dr. Anna Tosteson, professor of medicine at Dartmouth College, and Elmore were the paper’s lead authors. Donald Weaver, professor of pathology at the University of Vermont, was senior author.
This work was supported by the National Cancer Institute (R01 CA140560, R01 CA172343, and K05 CA104699) and by the National Cancer Institute funded Breast Cancer Surveillance Consortium (HHSN261201100031C).