By current projections from the American Cancer Society, one out of every eight women in the United States will be diagnosed with breast cancer this year. Caught early, most forms of the disease are treatable, and to do so, oncologists have relied on biopsies, sampling tissues for cancerous cells.
Researchers from the University of Washington, however, published the results of a study on March 17 to quantify the magnitude of diagnostic disagreement among pathologists when it comes to accurately categorizing biopsies and how well they work. As it turns out, they don’t.
The study found that in one out of five cases, participating pathologists disagreed on the diagnosis of borderline cases of breast cancer. Now, cancer researchers are pressing in other directions to make up the ground it lost in the fight.
“In the U.S. alone, 1.6 million women every year are undergoing breast biopsies,” said Joann Elmore, the author of the study and a professor of medicine and epidemiology at the University of Washington. “These are challenging to interpret, and the goal of our study was to evaluate the accuracy of doctors when they’re interpreting these biopsied tissues.”
Doctors in clinical practices from eight states interpreted slides of 60 breast biopsies that had been previously evaluated by a panel of highly qualified breast pathology experts. The 60 slides included a mixture of cases of varying prognoses.
“When the case had invasive breast cancer, there was near perfect agreement on the diagnosis, so that was very reassuring,” Elmore said. “When we got into the diagnosis of DCIS [ductal carcinoma in situ], four out of five cases agreed to the diagnosis, but one out of five cases was either over or under-interpreted.”
Elmore said even experienced doctors often misinterpret borderline cases like DCIS and atypia, both earlier, less severe forms of the cancer. Though the pathologists who participated in the study were able to agree on almost every case of invasive cancer, the incongruence of results regarding the less clear cases provided insight into the average accuracy of diagnoses in this difficult middle ground. Elmore said while typically a pathologist would look at many slides of the same biopsy before making a final decision, the study only provided one slide per case. After all, atypia and DCIS are often difficult to discern, even given expert knowledge and multiple biopsy slides.
“Doctors are looking at the structure and organization of cells visually and trying to classify them into categories,” Elmore said. “There might just be some cases where we truly are not able to provide a single specific diagnosis, due to the underlying biology.”
Women diagnosed with atypia or DCIS do not need to rush into treatment, she said. Rather, they have time to learn more and often get a second opinion on the biopsy before proceeding.
“We hope to study the impact of obtaining a second opinion on some or all of these cases,” she said. “We can now also digitize the images from the glass slides and send them over the Internet to experienced pathologists.”
That ability for doctors to digitize biopsy slides to send them to experts for a second opinion bodes well for the diagnoses of borderline cases going forward. It’s not much, given how much pathologists have depended on biopsies up to now, but it’s a start.
Until it’s perfected though, there are other means of diagnosis and monitoring, especially from Boston institutions. At Boston University, researchers such as Darren Roblyer, assistant professor of biomedical engineering, are working toward better cancer monitoring with wearable devices. At Massachusetts Institute of Technology’s David H. Koch Institute for Integrative Cancer Research, researchers across countless disciplines are attacking the disease from nearly every angle.
“In particular, our current research focus areas are developing nanotechnology-based cancer therapeutics, creating novel devices for detection and monitoring, exploring the molecular and cellular basis of metastasis, advancing personalized medicine by studying cancer pathways and drug resistance and engineering the immune system to fight cancer,” Kevin Leonardi, Koch Institute communications coordinator, wrote in an email. Of course, that’s just to name a few.
In the case of biopsies, though Elmore’s study showed a lack of accuracy in DCIS and atypia cases, the high percentage of accuracy with invasive cases of breast cancer reflected astute knowledge among the participating pathologists, even without providing multiple slides to reference. In most cases, that counts for more than one would think.
“Every year, millions of women undergo breast biopsies,” Elmore said. “There is an art to medicine.”
Olivia Deng contributed to the reporting of this story.