UNC scientist reports ways to improve breast cancer biomarker test

American and Spanish researchers report potential ways for doctors to improve the treatment of hormone receptor-positive breast cancer even if they lack access to costly multi-gene tests.

Because breast cancer is a biologically and clinically varied disease, doctors aim to choose appropriate treatments based on the characteristics of each patient’s individual tumor. In the past, this has been done using pathology-based biomarkers; however, these do not capture the full diversity of cancers.

“Multi-gene tests have been shown to provide valuable information beyond the information provided by the current pathology-based biomarkers,” says Aleix Prat, MD, a research fellow at UNC Lineberger Comprehensive Cancer Center in the laboratory of Charles Perou, PhD, May Goldman Shaw Distinguished Professor of Molecular Oncology Research.  “However, multi-gene tests are not readily available in most of the world due to cost, assay turnaround times, and other logistic issues.”

Dr. Prat is presenting his findings at the best abstracts session of the 4th IMPAKT Breast Cancer Conference in Brussels, Belgium on May 3, 2012. IMPAKT is an annual conference that was launched in 2009 by the Breast International Group (BIG) and the European Society for Medical Oncology, in collaboration with a multidisciplinary alliance of European breast cancer organizations and patient groups - referred to as partners.

Dr. Prat and colleagues addressed this problem by trying to improve the current pathology-based biomarkers to better represent data coming from a particular multi-gene test known as the PAM50 breast cancer intrinsic classifier.

“The PAM50 breast cancer intrinsic classifier identifies two major groups of hormonal receptor-positive breast cancer known as the Luminal A and Luminal B subtypes. These two molecular entities have different risks of relapse and responses to chemotherapy,” Dr. Prat said.

Alongside the development of this multi-gene assay, clinicians have devised pathology-based surrogate assays for the identification of the Luminal A and B subtypes. “In the absence of multi-gene assays, these pathology-based assays are clinically valuable,” Dr. Prat explained. “However, we observed that current pathology-based definitions of the Luminal A and B subtypes still show a 30-40 percent discordance rate compared to multi-gene tests such as the PAM50 breast cancer intrinsic classifier.”

The researchers examined differences in gene expression between Luminal A and B tumors using the PAM50 test. They also collected clinical-pathological features from 2,950 primary tumors across four independent studies. Using statistical methods, they tested the independent prognostic significance of those features.

They found that the expression of progesterone receptor was one of the most discriminatory molecules. “Addition of the quantitative scoring of the progesterone receptor into the current pathology-based Luminal A definition appears to better identify the subgroup of patients that have an outstanding survival when treated with endocrine therapy alone, and therefore do not need systemic chemotherapy,” Dr. Prat said. “This subpopulation of patients is likely to represent around 30 percent of the patients with low-risk pathology-defined Luminal A tumors.”

“Current pathology-based definitions of the Luminal A and B subtypes are valuable, but can be improved for the management of hormonal receptor-positive breast cancer,” the researcher concluded. “We believe that we have an improvement based upon the progesterone receptor, and given that progesterone receptor is widely used, our improvement could be widely and quickly adopted if further validated.”

Study co-author is Joel Parker, PhD, director of bioinformatics at UNC Lineberger.

This work was supported by funds from the National Cancer Institute’s Breast SPORE program (P50-CA58223-09A1), by RO1-CA138255, by the Breast Cancer Research Foundation, and the Sociedad Española de Oncología Médica (SEOM).

“Multi-gene tests have been shown to provide valuable information beyond the information provided by the current pathology-based biomarkers,” says Aleix Prat, MD, a research fellow at UNC Lineberger Comprehensive Cancer Center in the laboratory of Charles Perou, PhD, May Goldman Shaw Distinguished Professor of Molecular Oncology Research.  “However, multi-gene tests are not readily available in most of the world due to cost, assay turnaround times, and other logistic issues.”

Dr. Prat is presenting his findings at the best abstracts session of the 4th IMPAKT Breast Cancer Conference in Brussels, Belgium on May 3, 2012. IMPAKT is an annual conference that was launched in 2009 by the Breast International Group (BIG) and the European Society for Medical Oncology, in collaboration with a multidisciplinary alliance of European breast cancer organizations and patient groups - referred to as partners.

Dr. Prat and colleagues addressed this problem by trying to improve the current pathology-based biomarkers to better represent data coming from a particular multi-gene test known as the PAM50 breast cancer intrinsic classifier.

“The PAM50 breast cancer intrinsic classifier identifies two major groups of hormonal receptor-positive breast cancer known as the Luminal A and Luminal B subtypes. These two molecular entities have different risks of relapse and responses to chemotherapy,” Dr. Prat said.

Alongside the development of this multi-gene assay, clinicians have devised pathology-based surrogate assays for the identification of the Luminal A and B subtypes. “In the absence of multi-gene assays, these pathology-based assays are clinically valuable,” Dr. Prat explained. “However, we observed that current pathology-based definitions of the Luminal A and B subtypes still show a 30-40 percent discordance rate compared to multi-gene tests such as the PAM50 breast cancer intrinsic classifier.”

The researchers examined differences in gene expression between Luminal A and B tumors using the PAM50 test. They also collected clinical-pathological features from 2,950 primary tumors across four independent studies. Using statistical methods, they tested the independent prognostic significance of those features.

They found that the expression of progesterone receptor was one of the most discriminatory molecules. “Addition of the quantitative scoring of the progesterone receptor into the current pathology-based Luminal A definition appears to better identify the subgroup of patients that have an outstanding survival when treated with endocrine therapy alone, and therefore do not need systemic chemotherapy,” Dr. Prat said. “This subpopulation of patients is likely to represent around 30 percent of the patients with low-risk pathology-defined Luminal A tumors.”

“Current pathology-based definitions of the Luminal A and B subtypes are valuable, but can be improved for the management of hormonal receptor-positive breast cancer,” the researcher concluded. “We believe that we have an improvement based upon the progesterone receptor, and given that progesterone receptor is widely used, our improvement could be widely and quickly adopted if further validated.”

Study co-author is Joel Parker, PhD, director of bioinformatics at UNC Lineberger.

This work was supported by funds from the National Cancer Institute’s Breast SPORE program (P50-CA58223-09A1), by RO1-CA138255, by the Breast Cancer Research Foundation, and the Sociedad Española de Oncología Médica (SEOM).