Molecular Profiling in Breast Cancer

Molecular profiling refers to the use of gene expression patterns to predict the outcomes from breast cancer and its therapy. It is a developing area and newer tests are likely to become available over the coming years. The two main players in the field are Oncotype DX (21 genes) and the Amsterdam Mammoprint (70 genes).

General Features

The tests are linked by the use of gene expression patterns to predict outcomes. The standard approach is to identify likely candidate genes (either from the literature or through DNA microarray analysis of tumour specimens) and then examine the DNA profiles of a known patient group. The pattern of gene expression can then be correlated with the clinical outcomes. The second step is to validate the predictive ability of the molecular profiling test on a different patient population.
At present there are no completed prospective studies examining the utility of these tests.

Oncotype DX

The Oncotype DX is the most well studied molecular profiling test. The genes used were identified on literature review and then checked against known outcomes from a randomised controlled trial. From this process, 16 predictive genes and 5 control genes were identified. Results were validated in a different randomised controlled trial. The Oncotype DX is notable for the ability to the run the test on paraffin fixed tissue, as opposed to the Mammoprint which needs fresh frozen tissue to run DNA analysis on.
Oncotype DX is only used for patients with hormone receptor positive, node negative disease. In this setting, it is used to give a low or high risk of recurrence score. Patients with a high risk of recurrence (despite having essentially low risk disease) are offered systemic therapy with anthracyclines (eg. FEC or TAC). Patients with low risk of recurrence are adequately treated with endocrine therapy alone and chemotherapy provides no additional benefit.
The test is controversial because it is expensive (~ $4000) and there is some debate as to whether more accessible pathology findings (eg. tumour grade, lymphovascular invasion etc) may have a similar predictive value. It is being evaluated in a randomised controlled trial in a prospective fashion.
Although it is not used in node positive disease there is a suggestion that it may provide some prognostic information in this setting as well; this needs further evaluation.


Mammoprint was developed in Amsterdam by researchers evaluating gene expression patterns in fresh frozen tissue taken from the malignancies in women who had or had not developed distant metastasis. Seventy genes were determined to have predictive value, with distant metastatic rates of 30% in the high risk versus 10% in the low risk group. It has been validated in an independent patient group.
The limitations of Mammoprint are that it provides prognostic information only (as opposed to Oncotype DX which predicts chemotherapy effect) and was developed in young women (<55) who had hormone receptor positive disease. This makes the findings difficult to extrapolate to other patient groups.