HSC Looks at Genomics for Cancer Therapy


--UNM and Exagen's initial validation study demonstrates accuracy of prognostic tests that detect DNA changes with application to all patients --

      An initial validation study presented by the University of New Mexico and Exagen Diagnostics, Inc. at the San Antonio Breast Cancer Symposium suggests it may soon be possible to distinguish good prognosis from poor prognosis in any newly diagnosed breast cancer patient, based on DNA changes in the patient's tumor. Those patients identified as having a very good prognosis may do well without chemotherapy or hormonal therapy after their tumors have been removed. 

      In a retrospective study of 308 patients conducted at the University of New Mexico Health Sciences Center, the Department of Pathology and Exagen Diagnostics reported the discovery of two, 3-gene sets of markers that were prognostic in testing archived specimens from hormone receptor (HR)-positive and HR-negative patients. These two sets of markers form a panel for use in testing tumor tissue from breast cancer patients, providing a same- or next-day result.

      In independent test sets, each of the 3-gene markers accurately identified 91 percent of HR-negative and HR-positive specimens from patients that did not experience recurrence of disease.  In patients that were also node negative, the negative predictive value was 100 percent (e.g., 100 percent of patients identified by these tests had a good prognosis clinically).

      The study population consisted of white and Hispanic patients with invasive ductal carcinoma that were diagnosed between 1986 and 1999 at the University of New Mexico Health Sciences Center.  A minimum of four years of follow-up clinical information was available for each of the patients, with an average follow-up of nine years. Poor prognosis was clinically defined as development of recurrence, as evidenced by either distant metastasis or death from breast cancer. Good prognosis was defined as the absence of recurrence (or death from breast cancer) as of the last date of follow-up.

      Three UNM Department of Pathology faculty members were involved in the study: John Hozier, Ph.D., Thomas Williams, M.D., and Therese Bocklage, M.D.

      "I think there are a couple of reasons why this project has gone well", said Dr. Hozier, principal investigator at UNM. "First, the root cause of cancer, and probably tumor progression, is in the chromosomes. So DNA is the best place to look for markers of recurrence. Second, our approach keeps the tumor tissue, and the pathologist, in the picture. We look at the tumor cells and the DNA markers at the same time, so that we are not influenced by the surrounding normal cells. That's not possible with other technologies where the first step is to grind up the tissue."   

      "Today, we are treating almost any woman as though her cancer is aggressive, which puts patients who don't actually need treatment at risk for side-effects of therapy that is not necessary," explains Ian Rabinowitz, M.D., of the UNM Cancer Research and Treatment Center, and the primary clinician involved in the initial study. "For early stage breast cancer there are about 70 to 80 percent of patients who are cancer-free and don't actually require therapy after they receive a lumpectomy. At this point in time, we have no way of identifying which patients fall into that category. So, the advantage of a technology such as Exagen's prognostic tests, once it is further validated, is that we can better identify patients that have a very good prognosis and potentially don't actually require therapy," he said.

      The study was conducted using a proprietary "pattern of genomic amplification" FISH (e.g., fluorescent in situ hybridization) test or PGA FISH™. Due to the small number of markers in each reagent set, Exagen, which sponsored the project and provided the computational platform for data analysis, will be able to implement these tests with PGA FISH technology, making them readily usable by any FISH-testing laboratory.  Exagen's discovery technology is based on a mathematical approach that enables the company to examine 25,000 genes with no "preference" for one gene over another to define the best, small combination of genes for a specific testing purpose.

Contact: Angela Heisel, 272-3322

Related Stories