At A Glance
Name: Dr. Marc van de Vijver
Background: 1997 — present Head, Division of Diagnostic Oncology, Netherlands Cancer Institute, Amsterdam; 1996 — present Head, Department of Pathology, NCI Amsterdam; 1993 — present Pathologist, dept. of Pathology, Leiden University Medical Center, The Netherlands
Education: 1989-1993 — Residency in Pathology, dept. of Pathology, Leiden University Medical Center, The Netherlands; 1989 — Ph.D. University of Amsterdam; 1988 — M.D. Medical School, University of Amsterdam
A recent collaboration including the Netherlands Cancer Institute in Amsterdam produced the 70 gene expression signature that was later developed by Agendia into a diagnostic tool called MammaPrint which tests for the recurrence of breast cancer.
Now the Netherlands Cancer Institute has teamed up with researchers at Stanford University to discover a new gene-expression signature that can be used to distinguish which patients may be in need of further treatment after tumor removal.
In an article published in the Feb. 8 online edition of Proceedings of the National Academy of Sciences entitled “Robustness, scalability, and integration of a wound-response gene expression signature in predicting breast cancer survival,” the researchers, including Patrick Brown and Howard Chang of Stanford and Marc van de Vijver from the Netherlands Cancer Institute, announced that they had discovered a new profile that could be developed into a new prognostic test similar to MammaPrint. BioArray News this week spoke with Marc van de Vijver, the head of pathology at the Netherlands Cancer Institute, about the results from the study.
How did the Netherlands Cancer Institute and the Stanford Group link up for this study?
The Stanford group really pioneered the microarray gene expression profiling in breast cancer. So they were the first to look at gene expression profiling in breast cancer in a relatively large series of tumors and published on that before we started our work. They are probably the other main group in the world that works on breast cancer. This was one of the first projects where we really collaborated quite extensively.
Do you print your own chips there or do you have a relationship with a company that produces chips?
Both. We have a collaboration with Rosetta Inpharmatics in Seattle and in addition to that we also printed until recently cDNA arrays, but recently that has been changed to printing oligonucleotide arrays.
How did you share data with Stanford?
They have been using cDNA arrays and we have been using oligonucleotide arrays that are produced by Rosetta or Agilent, so to identify the wound signature genes in our data set we had to look up the genes that had been identified in the Stanford cDNA platform in our data set and for that we linked through Unigene. At Stanford University they also have software called Source, and with that you can link gene annotation through Unigene, so that’s what we did.
What is attractive about studying breast cancer? Are there any factors that make breast cancer an ideal cancer to study?
Breast cancer is a relatively frequent tumor in the western world. Patients are operated on, so for the type of research I do which is looking at genetic alterations in tumor tissue and gene expression profiling in tumor tissue the fact that these tumors are surgically removed after the diagnosis of breast cancer has made it possible to obtain tumor tissue to do the research with so that’s certainly something that helps.
Where did you get your sample of 295 early breast cancer patients from?
We’ve been collecting frozen tumor tissues since 1984, so we have a relatively big tumor bank, and also our institute treats a lot of breast cancer patients, so breast cancer is one of the tumor types that the clinical part of our institute specializes in. We’ve been trying to collaborate with other institutes to obtain tumor series from which tumor material is available from patients that have been treated also starting in 1985 or so to have a clinical follow up on those patients and that has turned out to be very hard, so there are not that many institutes around the world that have frozen down their tumor material.
What is the link between wound response and cancer progression?
This wound signature was in fact discovered by looking at fibroblasts, and then the gene expression signature that was associated with the response of fibroblasts to serum, that same set of genes was then subsequently studied in the breast cancers that we have been studying. So then we found out that the same set of genes that are activated or inactivated in response to serum in fibroblasts; that same set of genes appears to have a prognostic predictive power in breast cancer.
Have you designated a way to assign prognostic scores, or is that for further research?
Well, our first discovery was the 70-gene signature, which is a prognostic signature. The wound signature can be used to further define subgroups of breast cancer patients with different outcome. Based on the 70-gene profile we can distinguish what we call a good prognosis group and a poor prognosis group. Then the most important additional information you can get by using the wound signature is that within the poor prognosis group of patients or of tumors you can distinguish the quiescent wound signature, which is a group of patients with a relatively good prognosis, and the activated wound signature tumors, which define relatively poor prognosis tumors.
Explain why the wound-response signature is better at risk prediction than earlier models.
If you compare it to clinical factors, I think that these genetic profiles add to the clinical knowledge and will not replace it but in breast cancer it is very important to predict the risk of developing metastases as quickly as possible because based on that risk patients are advised to receive hormonal or chemotherapy treatment. If you can tell a patient that the risk of developing distant metastases is very small, then the advice of not treating with hormonal or chemotherapy treatment can be given. And none of the tests can give you 100 percent accuracy at the moment, but some of the tests are getting pretty good I think, and hopefully by further refining these tests, the prediction will become more accurate in the coming years.
Are you looking to move the results of your study to the market anytime soon?
At the moment we are not, or rather Agendia, which is the company that is producing the MammaPrint, [is not]. I am not part of Agendia, but there are no plans to include the wound signature in the prognosis test. I do think in the near future Agendia, and probably also other providers of diagnostic tests in breast cancer, will start using the new knowledge that’s emerging all the time.
In your study you say that methods will have to be developed to simplify the evaluation of the molecular signatures for clinical decision making. How can this be accomplished and who is up to the task?
We will certainly continue to work on this but I think the limiting factor is the availability of well-categorized clinical series of patients for whom frozen tumor material is available and for which clinical outcome data is available. Because to get reliable profiles in breast cancer I think ultimately will require the analysis of many hundreds or even thousands of patients, and that’s going to be a limiting factor.
How would a test using the signature work?
If you wanted to use the wound signature as part of your prognostic test, then the easiest way would be to print all the genes that are part of the wound signature on the diagnostic chip you are using, and then you can assign your tumors to wound-activated or wound-quiescent based on the gene expression pattern that you get on your array.
You created your own decision tree to make prognoses in your study. Could this be a model for a future test?
Yes. Maybe not exactly this algorithm, but certainly this type of approach can be used in the future.
Do you plan on continuing your collaboration with Stanford and do you expect more results in the near future?
Yes, we are continuing our collaboration at the moment. We are looking at other gene expression profiles as well. [The] paper that now came out is on the wound signature, and we are also collaborating on other gene expression signatures that have been identified by various approaches in Stanford.