Inconsistency of HER2 Test Raises Questions
Inconsistency of HER2 Test Raises Questions
The tests that determine who gets the powerful breast cancer drug trastuzumab (Herceptin) may not be as reliable as previously thought, researchers reported at the annual meeting of the American Society of Clinical Oncology. That means some women who should be getting trastuzumab treatment are not, while other women who will not benefit are unnecessarily exposed to a drug that can cause heart problems.
"This is a very big deal," said Larry Norton, M.D., deputy physician in chief for breast cancer programs at Memorial Sloan-Kettering Cancer Center, who was not involved with either of the new studies that uncovered the problem.
Trastuzumab is often cited as the one of the few successful examples of personalized medicine, because only patients whose tumors express the HER2/neu protein are given the drug. It has been shown to reduce the risk of relapse and improve overall survival in those patients.
Currently, two tests are used in clinical practice to determine whether a given tumor is HER2 positive or HER2 negative. The new research appears to show that some patients who have tested negative for HER2 on both tests still benefit from trastuzumab. In a series of interviews, experts agreed that the problem lies not in a misunderstanding of the mechanism of action but with the tests themselves. However, exactly what is wrong with the tests or what should be done about it is hard to agree on.
The U.S. Food and Drug Administration approved trastuzumab for use in women with early breast cancer after three large randomized trials showed that patients treated with chemotherapy plus trastuzumab had an improved progression-free and overall survival than those treated with chemotherapy alone. Because the drug was designed to work directly on the HER2/neu protein, researchers enrolled only patients with a positive HER2 tumor, based on a local laboratory test. The local laboratories were allowed to use either of the standard HER2 tests: immunohistochemistry (IHC), in which a 3+ score is considered positive, or fluorescent in situ hybridization (FISH), which shows evidence of HER2 gene amplification. The trials were designed so that a central laboratory could review or retest the tumor samples−and that is when the current problems appeared. Many tumors that had appeared to be HER2 positive at the time of patient enrollment now appear to be negative.
Soonmyung Paik, M.D., director of pathology at the National Surgical Adjuvant Breast and Bowel Project (NSABP) in Pittsburgh, reported that their central lab retested all the available tumor blocks from the NSABP B-31 trial, using both tests whenever possible. Of 1,662 tumors reassessed by FISH, 207 now appeared negative. Of 1,795 blocks retested by IHC, 255 appeared negative. Overall, 161 (9.7%) of the 1,662 tumor blocks tested by both methods appeared HER2 negative in both IHC and FISH.
However, Paik found that the drug still appeared to be working in at least some of these negative patients. Specifically, the relapse rate in the 87 doubly negative patients treated with chemotherapy alone was two- to threefold higher than the rate of relapse in the 74 doubly negative patients treated with chemotherapy plus trastuzumab (21% versus 8% events). If the tests were working correctly, you would expect no difference between the two groups. He found that the decreased risk of relapse in this subgroup was 64%, which was statistically significant. By comparison, the overall trial showed a 52% risk of relapse, which given the number of patients involved in the trial is likely to be a more accurate measure of benefit than that seen in the few patients whose retest appeared doubly negative.
Edith Perez, M.D., director of the breast cancer program at the Mayo Clinic in Jacksonville, Fla., reported similar findings when her group retested the tumor blocks from the Intergroup N9831 trial. Of 1,842 blocks tested, 103 (6%) appeared HER2 negative by both tests. Yet, as in Paik's analysis, the relapse rate in the 44 doubly negative women treated with chemotherapy alone was higher at 14 (32%) than the relapse rate for the 59 doubly negative women in the chemotherapy-plus-trastuzumab arm who had nine events (15%). The difference in response rate in Perez's analysis did not reach statistical significance, but the trend resembled that of the overall study population, with a benefit for the addition of trastuzumab.
"If the data are correct, it is significant," said Dennis Slamon, M.D., Ph.D., professor of hematology and oncology at the University of California, Los Angeles, School of Medicine, who helped develop trastuzumab. He suggests a series of tests using the samples already available from the four phase III trials of trastuzumab to determine whether the tests are as inconsistent as might be supposed by the newly reported data. In this study, some samples from each study−including the ones reported negative using both tests−should be blinded and retested by each cooperative group's central laboratory. "The answer can be had, I promise you, in the datasets we already have, without a single additional patient," he said.
But not everyone is convinced that the problem is as simple as false negatives and false positives. Paik, for one, thinks one problem may be with the cutoff points that scientists have established to delineate between HER2 negative and positive in IHC testing. During his presentation, Paik reviewed data from several preclinical and clinical studies suggesting that trastuzumab activity does not strictly require HER2 overexpression or gene amplification, as is currently thought. Instead, even tumor cells that express a lower level of the protein might respond to a trastuzumab−chemotherapy combination. Currently, tumors that have moderate amounts of HER2 protein would be scored as 1+ or 2+ on IHC tests−and would be called negative because they are below the predetermined cutoff value for the test.
Even FISH testing, which is considered the "gold standard" for detecting gene amplification, has problems. The test measures the ratio between the area surrounding the HER2 gene on chromosome 17 and other parts of chromosome 17. That means a cell that has extra copies of chromosome 17, called polysomy 17, appears negative by FISH but actually has extra copies of HER2 and probably expresses more protein than a wild-type cell would.
Not everyone is having trouble with the tests, however. Michael Press, M.D., Ph.D., a professor of pathology at the University of Southern California in Los Angeles and a lead pathologist for the Breast Cancer International Research Group (BCIRG), said that his group has retested samples from the BCIRG-006 trial, which examined trastuzumab's effectiveness in early stage breast cancer. They only found 10 discrepancies with FISH of the 3,001 tumor samples retested. "If there are lots of errors with this test, it is not something we've experienced."
To learn whether this factor really is a problem in HER2 assessment and trastuzumab treatment, Peter Kaufman, M.D., associate professor of hematology and oncology at the Norris Cotton Cancer Center at Dartmouth Medical School in New Hampshire, led a retrospective reexamination of tumor samples from the Cancer and Leukemia Group B 9840 trial, which he also presented at ASCO. Patients in the trial with HER2-negative tumors, as assessed by local laboratory testing with either FISH or IHC, were randomly assigned to paclitaxel with or without trastuzumab. The primary results of the trial, reported several years ago, showed a trend toward increased response in patients treated with the combination versus those treated with chemotherapy alone, but the finding was not statistically significant.
When Kaufman retested 157 tumor blocks from HER2-negative patients, he found that 25 (19%) had extra copies of chromosome 17, which would not appear as extra copies of the HER2 gene by FISH testing. But these patients were more likely to respond when treated with trastuzumab than when given chemotherapy alone. When they were tested by IHC, nearly all these cases were scored as IHC 2+, which would not meet the threshold to be treated. Hence, these patients would be scored HER2 negative by both tests−even though they may benefit from the drug. Previous studies report that chromosome 17 polysomy occurs in 8%−27% of all breast cancers, and thus this could be a bigger problem than previously considered.
Perez is also concerned that the inconsistency in the test may stem, in part, from heterogeneity of HER2 expression in the tumor sample. If so, then multiple tests on the same tumor could yield different results, even though the tests are working as designed. HER2 is thought to instigate tumor formation, however, and thus the tumor is unlikely to lose HER2 expression. Slamon thinks that such tumor heterogeneity comes from pathologists − not paying strict enough attention to the boundary between tumor and normal tissue. With all these concerns, Slamon said, "the tests should be done by a board certified pathologist−not a technician."
The testing problem has been a matter of discussion among clinical researchers for some time, Slamon said. "It is coming to a head now because of the newly released data, and that is a good thing." How it will be resolved remains unclear, however. Norton and colleagues at Cold Spring Harbor Laboratory in New York are trying to devise a new test altogether, the details of which they are not ready to discuss. Paik and others have suggested using an RNA-based test, with the idea that HER2 messenger RNA may be more predictive of protein overexpression than DNA. But RNA is notoriously unstable and likely to create problems of its own in a clinical laboratory.
While the testing problem is being worked out in research labs, a key question is which patients should receive trastuzumab. Perez said that given the data she and Paik presented, she will be tempted to retest patients who appear negative on a first pass−just to be sure that every patient who might benefit has access to the drug.
"I am worried about excluding patients who would benefit," she said.
Content
The tests that determine who gets the powerful breast cancer drug trastuzumab (Herceptin) may not be as reliable as previously thought, researchers reported at the annual meeting of the American Society of Clinical Oncology. That means some women who should be getting trastuzumab treatment are not, while other women who will not benefit are unnecessarily exposed to a drug that can cause heart problems.
"This is a very big deal," said Larry Norton, M.D., deputy physician in chief for breast cancer programs at Memorial Sloan-Kettering Cancer Center, who was not involved with either of the new studies that uncovered the problem.
Trastuzumab is often cited as the one of the few successful examples of personalized medicine, because only patients whose tumors express the HER2/neu protein are given the drug. It has been shown to reduce the risk of relapse and improve overall survival in those patients.
Currently, two tests are used in clinical practice to determine whether a given tumor is HER2 positive or HER2 negative. The new research appears to show that some patients who have tested negative for HER2 on both tests still benefit from trastuzumab. In a series of interviews, experts agreed that the problem lies not in a misunderstanding of the mechanism of action but with the tests themselves. However, exactly what is wrong with the tests or what should be done about it is hard to agree on.
The U.S. Food and Drug Administration approved trastuzumab for use in women with early breast cancer after three large randomized trials showed that patients treated with chemotherapy plus trastuzumab had an improved progression-free and overall survival than those treated with chemotherapy alone. Because the drug was designed to work directly on the HER2/neu protein, researchers enrolled only patients with a positive HER2 tumor, based on a local laboratory test. The local laboratories were allowed to use either of the standard HER2 tests: immunohistochemistry (IHC), in which a 3+ score is considered positive, or fluorescent in situ hybridization (FISH), which shows evidence of HER2 gene amplification. The trials were designed so that a central laboratory could review or retest the tumor samples−and that is when the current problems appeared. Many tumors that had appeared to be HER2 positive at the time of patient enrollment now appear to be negative.
Soonmyung Paik, M.D., director of pathology at the National Surgical Adjuvant Breast and Bowel Project (NSABP) in Pittsburgh, reported that their central lab retested all the available tumor blocks from the NSABP B-31 trial, using both tests whenever possible. Of 1,662 tumors reassessed by FISH, 207 now appeared negative. Of 1,795 blocks retested by IHC, 255 appeared negative. Overall, 161 (9.7%) of the 1,662 tumor blocks tested by both methods appeared HER2 negative in both IHC and FISH.
However, Paik found that the drug still appeared to be working in at least some of these negative patients. Specifically, the relapse rate in the 87 doubly negative patients treated with chemotherapy alone was two- to threefold higher than the rate of relapse in the 74 doubly negative patients treated with chemotherapy plus trastuzumab (21% versus 8% events). If the tests were working correctly, you would expect no difference between the two groups. He found that the decreased risk of relapse in this subgroup was 64%, which was statistically significant. By comparison, the overall trial showed a 52% risk of relapse, which given the number of patients involved in the trial is likely to be a more accurate measure of benefit than that seen in the few patients whose retest appeared doubly negative.
Edith Perez, M.D., director of the breast cancer program at the Mayo Clinic in Jacksonville, Fla., reported similar findings when her group retested the tumor blocks from the Intergroup N9831 trial. Of 1,842 blocks tested, 103 (6%) appeared HER2 negative by both tests. Yet, as in Paik's analysis, the relapse rate in the 44 doubly negative women treated with chemotherapy alone was higher at 14 (32%) than the relapse rate for the 59 doubly negative women in the chemotherapy-plus-trastuzumab arm who had nine events (15%). The difference in response rate in Perez's analysis did not reach statistical significance, but the trend resembled that of the overall study population, with a benefit for the addition of trastuzumab.
"If the data are correct, it is significant," said Dennis Slamon, M.D., Ph.D., professor of hematology and oncology at the University of California, Los Angeles, School of Medicine, who helped develop trastuzumab. He suggests a series of tests using the samples already available from the four phase III trials of trastuzumab to determine whether the tests are as inconsistent as might be supposed by the newly reported data. In this study, some samples from each study−including the ones reported negative using both tests−should be blinded and retested by each cooperative group's central laboratory. "The answer can be had, I promise you, in the datasets we already have, without a single additional patient," he said.
But not everyone is convinced that the problem is as simple as false negatives and false positives. Paik, for one, thinks one problem may be with the cutoff points that scientists have established to delineate between HER2 negative and positive in IHC testing. During his presentation, Paik reviewed data from several preclinical and clinical studies suggesting that trastuzumab activity does not strictly require HER2 overexpression or gene amplification, as is currently thought. Instead, even tumor cells that express a lower level of the protein might respond to a trastuzumab−chemotherapy combination. Currently, tumors that have moderate amounts of HER2 protein would be scored as 1+ or 2+ on IHC tests−and would be called negative because they are below the predetermined cutoff value for the test.
Even FISH testing, which is considered the "gold standard" for detecting gene amplification, has problems. The test measures the ratio between the area surrounding the HER2 gene on chromosome 17 and other parts of chromosome 17. That means a cell that has extra copies of chromosome 17, called polysomy 17, appears negative by FISH but actually has extra copies of HER2 and probably expresses more protein than a wild-type cell would.
Not everyone is having trouble with the tests, however. Michael Press, M.D., Ph.D., a professor of pathology at the University of Southern California in Los Angeles and a lead pathologist for the Breast Cancer International Research Group (BCIRG), said that his group has retested samples from the BCIRG-006 trial, which examined trastuzumab's effectiveness in early stage breast cancer. They only found 10 discrepancies with FISH of the 3,001 tumor samples retested. "If there are lots of errors with this test, it is not something we've experienced."
To learn whether this factor really is a problem in HER2 assessment and trastuzumab treatment, Peter Kaufman, M.D., associate professor of hematology and oncology at the Norris Cotton Cancer Center at Dartmouth Medical School in New Hampshire, led a retrospective reexamination of tumor samples from the Cancer and Leukemia Group B 9840 trial, which he also presented at ASCO. Patients in the trial with HER2-negative tumors, as assessed by local laboratory testing with either FISH or IHC, were randomly assigned to paclitaxel with or without trastuzumab. The primary results of the trial, reported several years ago, showed a trend toward increased response in patients treated with the combination versus those treated with chemotherapy alone, but the finding was not statistically significant.
When Kaufman retested 157 tumor blocks from HER2-negative patients, he found that 25 (19%) had extra copies of chromosome 17, which would not appear as extra copies of the HER2 gene by FISH testing. But these patients were more likely to respond when treated with trastuzumab than when given chemotherapy alone. When they were tested by IHC, nearly all these cases were scored as IHC 2+, which would not meet the threshold to be treated. Hence, these patients would be scored HER2 negative by both tests−even though they may benefit from the drug. Previous studies report that chromosome 17 polysomy occurs in 8%−27% of all breast cancers, and thus this could be a bigger problem than previously considered.
Perez is also concerned that the inconsistency in the test may stem, in part, from heterogeneity of HER2 expression in the tumor sample. If so, then multiple tests on the same tumor could yield different results, even though the tests are working as designed. HER2 is thought to instigate tumor formation, however, and thus the tumor is unlikely to lose HER2 expression. Slamon thinks that such tumor heterogeneity comes from pathologists − not paying strict enough attention to the boundary between tumor and normal tissue. With all these concerns, Slamon said, "the tests should be done by a board certified pathologist−not a technician."
The testing problem has been a matter of discussion among clinical researchers for some time, Slamon said. "It is coming to a head now because of the newly released data, and that is a good thing." How it will be resolved remains unclear, however. Norton and colleagues at Cold Spring Harbor Laboratory in New York are trying to devise a new test altogether, the details of which they are not ready to discuss. Paik and others have suggested using an RNA-based test, with the idea that HER2 messenger RNA may be more predictive of protein overexpression than DNA. But RNA is notoriously unstable and likely to create problems of its own in a clinical laboratory.
While the testing problem is being worked out in research labs, a key question is which patients should receive trastuzumab. Perez said that given the data she and Paik presented, she will be tempted to retest patients who appear negative on a first pass−just to be sure that every patient who might benefit has access to the drug.
"I am worried about excluding patients who would benefit," she said.
Source...