Insufficient evidence to recommend gene-expression profiles in breast cancer
Posted Jan 23 2009 12:00am
Just out are the findings of a working group for the CDC looking at the three tumor gene-expression profiles for breast cancer which are reported in Genetics in Medicine:
There is insufficient evidence to make a recommendation either for or
against the use of 3 tumor gene-expression profile products to improve
outcomes in defined populations of women with breast cancer, concludes
a working group supported by the US Centers for Disease Control and
The group evaluated 3 products: MammaPrint (Agendia), Oncotype DX
(Genomic Health), and the Breast Cancer Gene Expression Ratio Assay —
known as the Quest H:I Ratio test (developed by Quest Laboratories,
based on technology from AviaraDx). All claim to provide prognostic
information (about recurrence and survival rates) in specific
populations of women with early-stage breast cancer and to identify
women most likely to benefit from chemotherapy.
The group found no evidence of clinical utility for the MammaPrint
and Quest H:I ratio tests, and inadequate evidence for Oncotype DX.
My experience at my institution is limited compared with the seemingly more routine ordering of OncotypeDX at other places, but I find that, in general, it has not been particularly helpful in managing our breast cancer patients. Some of this may be selection bias since it has been usually ordered in "intermediate"/borderline patients to begin with and the results I've seen have all come back as "intermediate risk."
More importantly, as Dr. Noel Weidner discussed in a recent USCAP presentation, "Multigene expression profiling as a prognosticator and/or predictor of therapeutic response in breast cancer patients," there are serious methodological problems in the landmark papers on breast including small and biased patient cohorts, inappropriate statistical analysis, and lack of methodology validation (See Reis-FIlho JS, Westbury C, Pierga J-Y. The impact of expression profiling on prognostic and predictive testing in breast cancer. J Clin Oncol 2006;24:1665-1671. Not to mention that there is marked author overlap of the reports establishing OncotypeDX with clear-cut financial interests in Genomic Health.
Finally, we as pathologists have an ethical obligation to our patients to be critically skeptical of such tests and not fall for the hype either from sales people or other clinical colleagues.