The Laboratory of cancer bioinformatics and functional genomics from the Cancer Research Centre, a joint centre of the CSIC and the University of Salamanca (USAL), has designed a risk predictor for breast cancer patients using machine learning techniques. It allows the identification of genes associated with survival and risk of patients. The new test could be easily translated into the clinical practice, given that its development cost is similar to other routine techniques.

The researchers have obtained a gene signature linked to molecular biomarkers which are routinely measured for breast cancer patients, which allows for more accurate diagnosis and decision-making.

In the medical practice, diagnosis is crucial, not only as identifying the disease, but also making a prognosis of how each patient's disease will evolve. In cancer cases, prognosis indicates the hope of disease remission and survival or the risk of relapse throughout the course of the disease.

"Diagnosis and prognosis depend on the biology of the cancer and the tissue analysed and can vary greatly between cancer types, although they all rely on techniques for the detection of different types of biomarkers. Therefore, identifying new biomarkers through research allows us to improve the diagnosis and prognosis of diseases and provide a more personalized and appropriate treatment for each patient," explains Javier De las Rivas, leader of the laboratory responsible for the new design.

Specifically, the researchers have obtained a gene signature linked to molecular biomarkers which are routinely measured in the clinic (the oestrogen receptor, the progesterone receptor and the HER2 proto-oncogene) for breast cancer patients. These biomarkers are key to determine what type of tumour each patient has and to guide oncologists' clinical decisions.

Besides, this gene signature has been compared with the gene signatures included in the commercial platforms Prosigna, which offers services to hospitals in Castilla y León (Spain), and Oncotype. The new genetic signature not only improves the results of those that already exist, but also shows how risk is calculated and based on which factors, something that commercial platforms don’t offer.

Another improvement over these platforms is that it does not only assign a rate risk to the development of breast cancer, high or low, but estimates the risk from zero to 100, therefore providing much more accurate information. The predictor specifies the influence of the selected genes and their association with standard biomarkers as, for example, it shows the importance of each gene in estimating the risk.

The predictor specifies the influence of the selected genes and their association with standard biomarkers as, for example, it shows the importance of each gene in estimating the risk.

De Las Rivas explains that "in this research, the group has identified a set of biomarker genes for a cohort of approximately 500 patients and the results have been validated in another similar cohort". He also points out that "the advantage over current commercial platforms is that the risk prediction has been calculated by associating it with standard tumour biomarkers, which are measured by histopathology when a breast cancer diagnosis is made".

It is highly viable the translation of the new test into the oncology clinical practice, given that the cost for its development is similar to other routine techniques: around €3,000. Moreover, the improvements incorporated by this new test would have a positive impact on the clinical practice, because the criteria used to choose one treatment or another, depending on the diagnosis and prognosis obtained, would be more precise.

Contact:
This email address is being protected from spambots. You need JavaScript enabled to view it.

 

CIC Cáncer (USAL-CSIC) Comunicación