Personalized medicine may sound futuristic, but tumor biology is already influencing treatment decisions for some patients with breast cancer. Drugs are being developed that attack specific forms of breast cancer rather than the entire spectrum. Some prognostic and predictive tests have become standards of care in breast cancer management.¹ Emerging tests promise to predict before or soon after chemotherapy has begun whether a prescribed drug treatment is right for the patient.²

“We are right in the middle of tumor biology becoming the most important thing to breast cancer management,” said Prof. Peter Barrett-Lee, a consultant oncologist in the breast unit at the Velindre Hospital in Cardiff, U.K.

Predictive tests based on cancer biology are commonly used when deciding which women should receive trastuzumab (Herceptin). This agent is recommended for women with breast cancer whose tumors have amplification of the cancer gene HER2/neu or overexpression of its protein product.

HER2/neu amplification can be determined by fluorescence in situ hybridization (FISH) and overexpression can be determined from immunohistochemistry (IHC), both of which can be performed on biopsied or resected tumor samples. Immunohistochemistry can also be used to measure the expression of estrogen and progesterone receptors, other important breast cancer biomarkers.

But these markers are sometimes not enough. Women who are recommended for Herceptin may find that their cancer does not respond to the anti-HER2 therapy. Treatment is expensive, and the side effects may be unpleasant. A test that was able to weed out more nonresponders in advance would be welcomed by healthcare providers and patients alike.

“We are already starting to look for markers that are involved in the HER2 pathway and, if downregulated or mutated, might predict resistance to Herceptin,” Barrett-Lee said. “It's quite likely that, over the next few years, we'll need to look at another two or three markers from the same molecular pathway.”

Genetic fingerprinting is another option that is under investigation. Tests such as Mammoprint and Oncotype DX, which look for a number of specific genes in tissue samples, may predict the likelihood of tumor recurrence before any form of adjuvant chemotherapy is started and may also play a role in predicting the tumor's likelihood of responding to a particular treatment strategy.

“This could be a very interesting tool for breast cancer,” said Prof. Jean-Pierre Lotz, chief medical oncologist at Tenon Hospital, University of Pierre and Marie Curie, in Paris. “Today we are giving chemotherapy to a lot of patients, although we are not always sure they need it. Perhaps they have been cured by surgery alone.”

This type of genetic test may also be used in the future to decide which breast cancer patients should proceed to radiotherapy.

“If you don't give radiotherapy to 100 patients who have had surgery for breast cancer, the cancer will recur in just 40 cases. We need to be able to identify the patients who actually need radiotherapy,” said Dr. Alphonse Taghian, chief of radiation oncology breast services at Massachusetts General Hospital Cancer Center and an associate professor at Harvard Medical School in Boston.

An alternative approach to predictive testing is looking for groups of cells in the blood rather than genes in tumor tissue. Researchers are exploring whether this type of test could determine which localized breast cancers are most likely to spread, again in advance of chemotherapy.³

RAPID RESPONSE

Predictive testing prior to therapy is one thing. Early warning that a chosen treatment strategy is not working is another matter. At the moment, practitioners must wait for imaging results after a period of treatment to check that regrowth has been prevented and metastatic spread curbed.

One option for flagging failing therapies at an earlier stage is to perform functional imaging earlier in treatment. PET scanning with the radiotracer fluorodeoxythymidine has shown promise in this role.⁴ Another option in routine clinical practice is to monitor the levels of molecular markers in patients' blood; for example, carcinoembryonic antigen and the cancer antigen 15-3. This strategy is helpful in the subset of patients with advanced breast cancer whose tumors secrete elevated levels of these markers, which can be followed during therapy.

More specific blood tests are now being developed for monitoring purposes as well. The cancer gene HER2/neu, for example, is known to shed its extracellular domain (ECD) into the circulation. Sustained elevation of HER2 ECD levels (15 ng/mL or higher) during HER2-targeted therapy may indicate that the treatment is not working well.⁵ A test developed by Siemens is the only FDA-cleared Serum HER-2/neu blood test that can indicate whether metastatic breast cancer is progressing or responding to treatment.

“It is important to know if the patient is improving because chemotherapy is efficient. You will see this because the level of the marker in serum will be decreasing. If the treatment is not efficient, you will see the marker serum levels increasing. You don't need to wait for two or three months to see this on imaging,” said Dr. Rafael Molina, a senior consultant at the Hospital Clinic Barcelona in Spain.

Lotz agrees that the HER2 ECD test is a good way of following a patient's response to treatment. He is also using the blood test to decide whether a patient should start Herceptin chemotherapy when tumor tissue samples are not available.

Full acceptance of the HER2 ECD test will take time and more prospective studies. Medical oncologists will want to be sure that the results can be used to inform treatment decisions.

“We need more evidence so that this marker can be included in guidelines on breast cancer management,” Molina said.

The use of biomarkers to guide therapy decisions will be dependent on the quality, cost, and convenience of tests used to measure them, Barrett-Lee said. At the moment, tissue testing for the Oncotype DX gene fingerprint can be performed only by the U.S. company that developed the procedure (Genomic Health), at a cost of approximately $1400 per sample.

New methods of testing for multiple biomarkers—for example, from DNA and/or RNA—will need to be standardized, Barrett-Lee said. IHC and FISH tests might be dismissed as “old” technology, but they are robust and trustworthy methods. Agreement will be needed on appropriate thresholds. Practitioners will also need to know what these results mean in terms of patient management.

“The need to validate and find the appropriate cutoff for assays always takes much longer than you think. You can't really start altering the way you treat breast cancer patients based on a study of just 100 cases,” he said.