In the early 2000s, several prognostic and predictive assays were introduced as a result of research on gene expression profiling. The purpose of these tests is to stratify breast cancer patients based on their risk of recurrence, and drive decisions about adjuvant therapies. Currently, multiple such tools are used in clinical practice globally. All prognostic and predictive tools in breast cancer use
machine learning methods that predict patient outcomes based on biomarkers, such as
transcriptomics, tumor morphology, or basic clinicopathologic features. The most commonly used are: • Oncotype DX (
Genomic Health) was developed in 2003 for use in
estrogen receptor (ER) positive tumors, and has been endorsed by the
ASCO and the
NCCN • EndoPredict/EPclin (
Myriad Genetics) is a 12-gene expression profiling assay for patients with ER-positive and HER2-negative breast cancer. EndoPredict/EPclin was found to be prognostic of distant recurrence in ABCSG 6/8 and TransATAC studies. The majority of these tests have been scientifically reviewed to compare them with other standard prognostic biomarkers, such as
grade and
receptor status. In these secondary analyses, patients with low Oncotype RS were found to derive minimal benefit from
adjuvant chemotherapy, suggesting it may be appropriate to choose to avoid side effects from that additional treatment. As an additional example, a
neoadjuvant clinical treatment program that included initial chemotherapy followed by surgery and subsequent additional chemotherapy,
radiotherapy, and
hormonal therapy found a strong correlation of the Oncotype
classification with the likelihood of a
complete response (CR) to the presurgical chemotherapy. Since high risk features may already be evident in many high risk cancers, for example hormone-receptor negativity or
HER-2 positive disease, the Oncotype test may especially improve the risk assessment that is derived from routine clinical variables in intermediate risk disease. Results from both the US and internationally suggest that Oncotype may assist in treatment decisions. Oncotype DX has been endorsed by the
American Society of Clinical Oncology (ASCO), developed in patients under age 55 years who had
lymph node negative
breast cancers
(N0). The commercial test is marketed for use in breast cancer irrespective of
estrogen receptor (ER) status. A summary of
clinical trials using MammaPrint is included in the
MammaPrint main article. MammaPrint was first approved for clinical use by the FDA in 2007. To be eligible for the MammaPrint
gene expression profile, a breast cancer should have the following characteristics:
stage 1 or 2, tumor size less than 5.0 cm,
estrogen receptor positive (ER+) or estrogen receptor negative (ER-). In the US, the tumor should also be
lymph node negative (N0), but internationally the test may be performed if the lymph node status is negative or positive with up to 3 nodes. MammaPrint has been validated in a prospective Phase III randomized controlled trial, MINDACT and a prospective registry study, FLEX. Several clinical trials are ongoing investigating the role of MammaPrint in additional treatment settings, including neoadjuvant therapy. One method of assessing the molecular subtype of a
breast cancer is by BluePrint, a commercial-stage 80-gene panel marketed by Agendia, either as a standalone test, or combined with the MammaPrint gene profile. BluePrint classifies breast tumors into luminal-type, HER2-type or basal-type.
Other prognostic and predictive tests Given the limitations of gene expression profiling tests, such as long turnaround times and high costs, statistical models and more complex
artificial intelligence models were developed to stratify recurrence risk and predict treatment response. In years 2000-2020, several online calculators and
nomograms were developed, including Adjuvant!,, Magee Equations, PREDICT, and CTS5. Recently, artificial intelligence tests analyzing
H&E-stained pathology slides were shown to potentially improve on the accuracy of gene expression profiling tests and expand clinical utility. In
real-world retrospective studies, Ataraxis Breast was found to have higher accuracy in predicting recurrence risk than Oncotype DX and refine selection of patients eligible for adjuvant
CDK4/6 inhibitors. == Other breast cancer treatment biomarkers ==