Testing for estrogen and progesterone biomarkers following a breast cancer diagnosis is critically important to treatment decisions, yet an international panel that included a University of Rochester physician found the tests can be inaccurate in 20 percent of cases due to inconsistent laboratory practices, lack of standardization, and other failures.
The findings prompted the College of American Pathologists (CAP) and the American Society of Clinical Oncology (ASCO) task force to release new standards and recommendations for pathologists and oncologists worldwide, and to begin a campaign to educate women to demand better practices.
“You can have the best surgeon in the world and the best medical oncologist, but if your pathology test was not done properly or is inaccurate, it can jeopardize the outcome of your case,” said David G. Hicks, M.D., director of Surgical Pathology at the University of Rochester Medical Center, a member of the James P. Wilmot Cancer Center at URMC, and also a member of the prestigious CAP/ASCO task force.
“What we have tried to do is understand the factors that contribute to variation in the accuracy of these tests, and then make recommendations to improve quality and reproducibility,” he said. “With these new guidelines we will be emphasizing best practices, re-training for pathologists and clinicians, and urging women to ask questions and insist upon testing from laboratories that meet the new standards so they can make the best possible treatment decisions.”
Breast cancer is the most commonly diagnosed cancer among American women, aside from skin cancer. The estrogen (ER) and/or progesterone (PgR) biomarkers, which have been known for 30 years and are present in nearly two-thirds of all breast cancers, usually have a direct connection to disease prognosis and treatment options.
Hormonally sensitive tumors tend to be less aggressive, and women with ER and/or PgR-positive cancers have the option of taking well-studied medications such as tamoxifen or aromatase inhibitors, which block the action of estrogen receptors in the breast tissue and might prevent a recurrence.
Currently, though, ER/PgR testing is decentralized and variable in the United States. Two laboratories in the same community might use different methods to handle the tissue, analyze it and make a determination, and they might also employ different criterion to interpret and validate the results. This problem is rarely acknowledged among patients, Hicks said, and unfortunately has left some women with an incomplete picture of her cancer.
“No test is perfect, but we should do everything we can to help make this important testing as accurate as possible,” Hicks said. “We need to get that variability as low as possible so that all women can receive the correct treatments.”
Disagreement over what constitutes an ER-positive or PgR-positive tissue sample is a central problem. Just as airline investigators study the black box for clues, the pathologist’s job is to pore over the cancerous tissue and measure important parameters such as tumor grade, size, and other biologic characteristics that might help guide the selection of treatment.
Pathology laboratories also routinely perform an immunohistochemistry stain on the sample to detect what percentage of the cancer cells stain positive for the ER or PgR markers. In some places a 10-percent threshold is used to report a positive result; other labs report a tumor positive if only one percent of the cells stain for ER/PgR antibodies.
The discrepancy in the threshold means there may be women who received an ER-negative result, when in fact their tumor might have been ER-positive and they might have derived benefits from therapies that target estrogen. On the other hand, a false positive can mean that some women are being over-treated, and needlessly exposed to the side effects of drug therapy that lasts for years.
Any therapy that targets a biomarker, the task force said, is only as good as the test used to detect the biomarker.
Therefore, the new guidelines set the ER/PgR threshold at the lowest end, requiring that a test be considered positive if at least one percent of the cells in the sample test positive. The panel also recommended that even weakly positive samples should prompt a doctor-patient discussion of anti-estrogen treatment options.
Tissue handling is another area that must be improved, Hicks said. One of the most surprising aspects of the task force literature review emerged from data in Asia and Africa, which showed an alarmingly high rate of ER-negative breast cancers. Those cancers tend to have a poorer prognosis. The panel discovered, however, that improper handling of the biopsy and surgical samples might explain the high ER-negative rate.
Once a surgeon cuts the blood supply to the tissue and removes a tumor, the cells begin to degrade almost immediately. The longer the sample sits at room temperature, the more degradation, which may contribute to inaccurate test results. If a breast tumor sits long enough without being preserved, Hicks said, degradation can cause a loss of estrogen receptor from the tissue, which could change the test score from ER-positive to ER-negative.
Previously scientists believed that a sample could sit for two hours before it was seriously compromised. The new guidelines suggest that one hour or less should elapse from tumor removal to when the pathologist cuts and fixes the sample with a solution called formalin to stop the degradation process.
It will take a lot of education to bring all hospitals up to the same consistent, high standards, Hick said. For example, a common practice at many institutions is for the surgical team to leave tumor samples sitting on a table in the back of the operating room until the case is complete, the panel discovered.
The task force also found that ER and PgR test results varied by day of the week, with Friday and Saturday cases showing the highest rate of ER-negative results. This suggests that tumors removed on those days are not attended to as quickly and might be falsely negative due to variability in tissue handling and degradation.
“The whole concept of personalized medicine, which is the cornerstone of scientific advancement, will require that we pay very close attention to the quality of information needed for making personalized treatment decisions,” Hicks said.
CAP and ASCO recently published the guidelines on their websites and in the Journal of Clinical Oncology. The task force leader and co-author of the report is Elizabeth H. Hammond, M.D., professor of Pathology at Intermountain Healthcare in Salt Lake City, Utah.
Education is the next step, according to Hicks, a professor of Pathology and Laboratory Medicine at URMC. On May 1, CAP sponsored an initial three-day training program to help pathologists understand the factors that contribute to testing variability and to minimize those factors. The program was sold out. (http://www.cap.org/apps/docs/membership/transformation/new/institute_certificate_BPFT.html). Another program for pathologists will take place this fall.
# # #
For Media Inquiries:
Email Leslie Orr