It’s disturbing enough when a mammogram prompts a biopsy. Now there’s increased concern that the interpretation of your biopsy may lead to a misdiagnosis. New research shows that pathologists who interpret biopsies tend to be very good at detecting straightforward invasive breast cancer…but poor at correctly identifying noninvasive breast cancer and a variety of other abnormalities. Sometimes they interpret results as more serious than they really are (overdiagnosis)…other times they declare biopsy results to be less serious than they really are (underdiagnosis). Either kind of mistake puts you at risk.


The new study gave breast biopsy slides to a large group of pathologists to interpret—and also to three top, internationally recognized breast pathology experts. The cases were randomly selected from cases already diagnosed and registered in a big database called the Breast Cancer Surveillance Consortium. Here’s what was discovered about the four main possible diagnoses…

Invasive breast cancer: Cancer that has spread beyond the milk ducts and glands. This is typically treated with surgery, radiation and chemotherapy. Result: The everyday pathologists and the experts agreed on the diagnosis 96% of the time. So far, so good!

Ductal carcinoma in situ (DCIS): Cancer that is confined within the milk ducts. DCIS may be treated with lumpectomy and radiation. Result: An agreement rate of only 84%. The everyday pathologists overdiagnosed it 3% of the time and underdiagnosed it 13% of the time.

Atypical hyperplasia: The presence of abnormal cells. This isn’t cancer but signals possible future cancer, so it calls for careful monitoring. Result: Agreement only 48% of the time! The everyday pathologists overdiagnosed it 17% of the time and underdiagnosed it 35% of the time.

Hyperplasia without atypia: The extra growth of normal, noncancerous cells. You’re safe—there is no cancer, just a breast condition such as a cyst or benign tumor. Result: 87% agreement. The everyday pathologists overdiagnosed it 13% of the time, and none of them underdiagnosed it.

The most disturbing of the findings is the one with atypical hyperplasia—35% of atypical hyperplasias were underdiagnosed as hyperplasia without atypia…nothing to worry about. Another 17% were overdiagnosed as DCIS, which often means surgery and radiation, when only monitoring was actually needed. As the researchers note, women diagnosed with DCIS when they have atypia may undergo unneeded surgery, radiation or hormonal therapy, which not only interrupts their lives and costs money but creates a burden of anxiety—all for no benefit!

It’s a pretty disturbing finding, especially since biopsies are the “gold standard” for diagnosis. Even though the study has limitations—the pathologists got to see only one slide rather than several as they would in normal practice and weren’t allowed to ask for more tests or a second opinion—it suggests that many thousands of women with low-grade breast cancer or precancerous breast disease are not being correctly treated. Indeed, it’s been known that diagnosis of conditions such as DCIS is prone to error. It’s worth noting that many of the pathologists admitted that some of the cases weren’t clear to them—and these tended to be the ones they got wrong.

For patients, it’s a wake-up call to not automatically accept the first diagnosis you receive. Both overdiagnosis and underdiagnosis cause harm, and there’s much we can do steer clear of both.


First and foremost, if you have a breast biopsy, consider getting a second pathology opinion, the study authors advised. If there’s disagreement between the first and second opinion, get a third pathologist to weigh in.

Scrutinize the qualifications of the pathologist: Make sure that he or she is board-certified by the American Board of Pathology, but also ask if he or she has undergone a one-year fellowship in breast pathology—in the study, only six had, while 109 hadn’t. Ask if the pathologist is affiliated with an academic institution—those whose diagnoses were off in this study typically didn’t have an academic affiliation. The ones who got it wrong were also more likely to be in small-sized practices (less than 10 pathologists) and didn’t see many breast cases each week (most saw fewer than 10 per week). About one-fifth had been interpreting breast pathology cases for only four or fewer years. When it comes to breast cancer diagnosis, especially in “grey” areas, experience and academic training in breast pathology really matter.