NCI Cancer Bulletin: A Trusted Source for Cancer Research NewsNCI Cancer Bulletin: A Trusted Source for Cancer Research News
October 24, 2006 • Volume 3 / Number 41 E-Mail This Document  |  Download PDF  |  Bulletin Archive/Search  |  Subscribe

NCI Cancer Bulletin Archive

Page Options

  • Print This Page
  • Print This Document
  • View Entire Document
  • Email This Document
  • View/Print PDF

The information and links on this page are no longer being updated and are provided for reference purposes only.


Translational Breast Cancer Research Gets Personal

In recent decades, revolutionary advances in our knowledge of the molecular and cellular biology of cancer have emerged from the laboratory. The resulting challenge has been how to turn these advances in knowledge into advances in the clinic as rapidly and practically as possible.

The idea of using translational research teams - multidisciplinary groups of investigators who work together to move an idea from "bench to bedside" - to achieve this goal has been gaining momentum. In 1992, NCI funded the first eight Specialized Programs of Research Excellence (SPOREs) translational research grants awarded to institutions and designed to promote such interactions between basic, clinical, and applied scientists. Since then, the program has expanded to include 59 grants focusing on 14 cancer types.

The success of this program is evident in the output, explained Dr. Jorge Gomez, chief of NCI's Organ Systems Branch, which oversees the SPORE grants. "Back in 1998, there were maybe 15 to 17 clinical trials in the SPORE program. As of 2005, we had 255 clinical studies, all based on molecular pathways of cancer. The SPORE program has created an environment that is very conducive to translational research, whether the clinical application is in the area of detection, diagnosis, treatment, or prevention."

One area of research that has proven particularly amenable to the translational approach is applying the knowledge of differential gene expression within a specific cancer type to customize treatment for different patients, with the ultimate goal of "personalized medicine" - treatment selected to match the unique molecular characteristics of each patient's tumor.

In a promising recent example in this field, work from the breast cancer SPORE at UNC-Chapel Hill has shown that several experimental gene-expression-based predictors for breast cancer outcome show strong concordance in their predictions, even though there was little overlap in the gene sets used to create the different predictive models.

The UNC SPORE paper, published in the August 10 New England Journal of Medicine, provides an important step in validating these models for future clinical use, explained Dr. Charles M. Perou, assistant professor of Genetics and Pathology at UNC. "There are a dizzying number of prognostic and predictive profiles for breast cancer," said Dr. Perou. "We kept saying to ourselves, 'They can't all be saying different things,' and they're not. They do tend to point to the same results."

Two of the models used in the North Carolina study are currently being tested in large-scale clinical trials: the European Organization for Research and Treatment of Cancer's Microarray in Node-Negative Disease May Avoid Chemotherapy (MINDACT) trial and the Eastern Cooperative Oncology Group's Trial Assigning Individualized Options for Treatment (TAILORx) study. These trials will test if these predictive models can be used prospectively to determine which women will benefit from adjuvant chemotherapy and which can be spared unnecessary treatment.

The SPORE investigators examined a total of five well-known gene-expression-based models, all based on different sets of genes and developed in different laboratories. The 5 models were each tested in tissue samples taken from 295 women with breast cancer and known relapse-free and overall survival times.

Patients were classified using each model, and the assignments of good or bad prognosis, or low or high recurrence probability were compared. The investigators also performed analyses that included the models and variables currently used to make treatment decisions in the clinic, such as estrogen-receptor status, tumor grade, and whether or not tumor cells had spread to the lymph nodes.

For all 295 patients, 4 out of 5 models "were significant predictors of relapse-free survival and overall survival," and "added new and important prognostic information beyond that provided by the standard clinical predictors," stated the authors, led by Cheng Fan and Dr. Daniel Oh of UNC. In addition, they explained, three of the models - including those being tested in the MINDACT and TAILORx trials - were more predictive of outcome than traditional pathological data, such as tumor size and grade.

A greater understanding of the molecular subtypes of breast cancer may lead not only to more targeted use of available treatments, explained Dr. Perou, but also to the development of new therapies specifically aimed at certain subtypes.

"I think one of the exciting things about identifying these subtypes is that it's not just prognostic information," said Dr. Perou. "Going along with the subtypes are all these potential therapeutic targets, present in some tumors and absent in others. We know the importance of [the] ER and HER2 [proteins], but there are others embedded in our profiles."

The investigators are further using their translational research relationships and have initiated an inter-SPORE clinical trial to test a targeted therapy against the protein HER1, which was found to be overexpressed in the basal-like breast cancer subtype.