Breast cancer: Reducing the risk of unnecessary chemo

Identifying whether a patient's cancer is at low or high risk of recurring would help doctors reduce unnecessary treatments for low risk patients. This could have a huge impact on a patient's quality of life and also significantly reduce the cost of health care.

Currently, most doctors assess a patient's prognosis using their age and "tumour grade," but this approach doesn't work very well. Now, NRC researchers have developed a tool to determine which breast cancer patients have little risk of their disease recurring. The tool -- an algorithm that identifies "gene expression signatures" or biomarkers that can predict low risk tumours with 87-100 percent accuracy in different groups of patients -- has the potential to virtually eliminate unnecessary chemotherapy.

To conduct their study, which appeared in a recent issue of Nature Communications, Dr. Edwin Wang and his colleagues at the NRC Biotechnology Research Institute in Montreal (NRC-BRI) used published data on gene expression profiles from more than 1000 breast cancer samples. "Every tumour has a gene expression profile, which indicates how the patient's genes have changed," he explains. "We combined this data with information on the patient's outcome -- such as whether the original tumour spread and how long the person survived -- to develop our algorithm."

The NRC team now hopes to see its algorithm applied in a clinical setting. "We have a provisional patent on the intellectual property and we would like to get a Canadian company to license it and turn it into a kit format," says Dr. Maureen O'Connor of NRC-BRI, who co-authored the study. "We've had interest expressed from more than one company so far."

Dr. O'Connor adds that the NRC algorithm could be adapted to other types of cancer where over-treatment is common, such as prostate cancer. "Prostate cancer in particular is usually not an aggressive disease, but the treatment can be extreme," she says. "We would like to develop a test that can predict with 99 percent accuracy whether a patient's cancer is not aggressive, so they can make the best decision about whether to proceed with treatment right away."

In future, the algorithm may also help pave the way toward personalized therapy for cancer patients. "On average, every cancer patient has 14-16 mutated genes," says Dr. Wang. "Based on their unique genetic signature, we hope to figure out which mutations to target to block the cancer process in each patient."