Evaluating the characteristics of a breast cancer -- hormone-receptor status and HER2 status, for example -- helps doctors figure out which breast cancer treatments make the most sense for a specific person.
Research has shown that each breast cancer characteristic listed in the pathology report has one or more variations, called subtypes. The study reviewed here evaluated a test that uses new technology to identify all a breast cancer's subtypes. The new technology, subtype classification model (SCM), was more reliable than methods that tested for single breast cancer subtypes (called single sample predictor or SSP).
Right now, doctors usually don't use breast cancer subtype analysis to help make treatment decisions because there aren't specific treatments available for different subtypes. For example, HER2-positive breast cancers can be treated with the targeted therapies Herceptin (chemical name: trastuzumab) and Tykerb (chemical name: lapatinib). Even though there are several subtypes of HER2-positive breast cancer, there are currently no treatment options or approaches that have been shown to work on breast cancers with a specific HER2 subtype.
Still, having a way to quickly and accurately determine breast cancer subtypes can help researchers as they work to develop new treatments. Much more research is needed before tests that determine breast cancer subtypes will be used routinely to develop highly individualized breast cancer treatment plans based on subtype.
Stay tuned to Breastcancer.org's Research News to learn more about lab research that may lead to better ways to diagnose and treat breast cancer.
BRUSSELS (MedPage Today) -- A simple, three-gene test robustly distinguishes breast cancer subtypes at the individual patient level, researchers reported here.
The SCMGENE model included only the estrogen receptor gene ESR1, the HER2 receptor gene ERBB2, and the proliferation gene AURKA, explained Benjamin Haibe-Kains, PhD, of the Dana-Farber Cancer Institute in Boston.
But adding expression of more genes using the same Subtype Classification Model (SCM) approach increased the robustness in a study his group presented at the IMPAKT Breast Cancer Conference.
Models of this type were more reliable (P<0.001) and more consistent with clinical parameters than models of the Single Sample Predictor (SSP) type, they found.
Clifford A. Hudis, MD, of Memorial Sloan-Kettering Cancer Center in New York City and chair of the session, called the results important, given the many competing technologies and approaches jockeying for a role in the division of breast cancer into subsets.
Oncologists have yet to commit to a single technology, or even a single classification scheme in this regard, he noted.
But this may be moot at the moment, since there are still just three broad types of therapy for breast cancer, and no reliable evidence upon which to base risk-benefit decisions in individual breast cancer subtypes, Hudis added.
"Our reach is exceeding our grasp here," he said in an interview. "We can subdivide breast cancer right now and even prognosticate to a pretty fine degree, but it's far more than we can respond to clinically."
Nevertheless, that day is coming, he speculated. "Someday it's going to matter that we can classify very finely, and that robustness question is going to be important."
Because it was a question that hadn't been asked in a trial, Haibe-Kains' group ran gene expression profiles of 4,607 patients from 32 publicly-available datasets through six different models:
The SCMOD2 performed best in determining molecular subtypes that correlated with traditional clinical variables, including ER and HER2 status determined by FISH and immunohistochemistry, tumor size and histological grade, and age at diagnosis.
In comparison, the other two SCM versions more consistently identified the same tumor as the same type across classification models than did the SSP models.
Haibe-Kains called the concordance between SCM versions "strong" but between the SSPs "substantial."
All the models had good concordance in identifying tumor subtype as basal (estrogen receptor [ER]-negative, HER2-negative).
But tumors in the HER2+ subtype were only consistently identified by the SCM models.
Concordance for luminal A (ER+) and luminal B (ER-) depended more on the specific model than whether it was SCM or SSP, Haibe-Kains said.
Notably, the SSP models identified only a few normal-type tumors, which the SCM classified predominantly as luminal A.
In a cohort of 1,400 untreated, node-negative breast tumors, all the models except SSP2006 produced similar survival curves.
Robustness -- defined as ability to assign the same tumors to the same subtypes whatever the gene expression data used to fit them -- was significantly greater for the SCMS than SSPs.
This held whether investigators were considering just the main subtypes (basal, HER2+, and luminal) or distinguishing between additional luminal subtypes.
Together these findings suggest that SCM-type models are most promising for translation into the clinic, Haibe-Kains said at the session.
Study discussant Fabrice Andre, MD, of the Institut Gustave Roussy in Villejuif, France,, called the robustness and concordance of the three-gene SCMGENE model "acceptable," making it possibly the optimal approach.
Haibe-Kains noted that this technology is quite challenging and may take a few years for validation and feasibility studies, followed by development into a commercial assay chip.
The researchers reported no conflicts of interest.
Andre and Hudis reported no conflicts of interest.
Primary source: IMPAKT Breast Cancer Conference Source reference: Haibe-Kains B, et al "Robustness of Breast Cancer Molecular Subtypes Identification" IMPAKT 2010; Abstract 98O.
Breastcancer.org is a non-profit organization dedicated to providing information and community to those touched by this disease. Learn more about our commitment to providing complete, accurate, and private breast cancer information.
Breastcancer.org 7 East Lancaster Avenue, 3rd Floor Ardmore, PA 19003
©2011 Breastcancer.org - All rights reserved.
