for translational research
of Texas MD Anderson Cancer Center
Name: Robert Bast
Position: Vice president for translational research and professor at the University of Texas MD Anderson Cancer Center since 2000.
Background: MD, Harvard Medical School, 1971; Assistant Professor of Medicine, Harvard Medical School, 1977-1983; Associate Professor of Medicine, Harvard Medical School, 1983‑1984; Professor of Medicine, Duke University Medical Center, 1984-1992; Professor of Medicine, University of Texas MD Anderson Center, 1994 to present; Harry Carothers Wiess Chair for Cancer Research, University of Texas, 1994-2004; Harry Carothers Wiess Distinguished University Chair for Cancer Research, University of Texas MD Anderson Cancer Center 2004 to present; Head, Division of Medicine, University of Texas MD Anderson Cancer Center 1994 to 2000; Head, Division of Medical Oncology, Department of Medicine University of Texas Health Science Center, Houston, 1994 to 2000;
Robert Bast is best known for developing the ovarian cancer 125 monoclonal antibody that led to the production of the CA-125 radioimmunoassay for the monitoring of epithelial ovarian cancer. Last month, Ciphergen — which is developing an ovarian cancer diagnostic of its own based on a panel of seven blood biomarkers — appointed him to its newly created scientific advisory board.
In addition to Ciphergen, a number of organizations, such as Correlogic and the University of Pittsburgh Medical Center, have identified ovarian cancer as their first target for proteomics-based diagnostics. ProteoMonitor caught up with Bast recently to discuss the current state of the field and the promise of proteomics for developing an alternative to the CA-125 test.
You helped developed the CA-125 test for ovarian cancer. Can you talk about your background that led to this test?
We had developed monoclonal antibodies against ovarian cancer cell lines in the late 1970s to try to develop a therapy for ovarian cancer. We had a mouse model that showed that rabbit antiserum and bacterial immunostimulants like corynebacterium parvum interacted and we were attempting to mimic that in clinical trials. And with the advent of the monoclonal technology, we developed a series of antibodies, [and the]125th attempt, OC-125 for ovarian cancer 125, was an antibody that turned out to recognize a shed antigen. It was not very effective for therapy but because the antigen was shed we thought that it might be effective for monitoring patients with ovarian cancer. In 1983, the monitoring assay was reported for cancer antigen 125 and that is widely used at the present time for patients with known ovarian cancer.
Early on it became apparent that CA-125 could be elevated at least in a fraction of patients prior to conventional diagnosis and that there’s been a lot of interest in several major trials in the UK to evaluate CA-125 either alone or as a trigger to do ultrasound in a small fraction of patients to detect ovarian cancer early.
Why is there a need to develop alternative diagnostics to CA-125?
Given the prevalence of ovarian cancer in the general population in the United States, even if you take women over 50, you’d be at higher risk for developing sporadic ovarian cancer. In order to get a positive predictive value of 10 percent, you need a sensitivity of greater than 75 percent for early-stage disease and a specificity of greater than 99.6 percent.
Currently, [in] a conventional diagnosis at stage 1 disease, only 50 to 60 percent of patients will have an abnormal CA-125. And the specificity, particularly in a pre-menopausal population, is just not adequate. In the post-menopausal population, it’s about 99 percent but not 99.6 percent. So people have looked at either combining CA-125 with ultrasound or looking at CA-125 over time as a more specific test.
You could argue that if you looked at values that were within the normal range — in most assays less than 35 units — with a rising value you could detect a higher fraction than the 50 to 60 percent of stage 1’s. But about 20 percent of ovarian cancers don’t make enough CA-125 to show up in the serum. So that particularly for early-stage disease you’d really need to have more markers than CA-125 and there are a number of groups, including our own, that are looking at multiple markers for that.
CA 125 has been around for more than 20 years. How would you characterize the effectiveness of it in detecting and monitoring ovarian cancer?
For monitoring, it’s about 90 percent accurate for patients with advanced disease. There are effusions that can confound it and there are occasion cases where it doesn’t track disease faithfully, but it’s a reasonably good marker for tracking known disease. What no one knows at the present time is whether CA-125 or other markers will pick up pre-clinical disease — that is before conventional diagnosis.
You recently joined the scientific advisory board of Ciphergen. What is your role there?
Basically, only to provide advice regarding some of their initiatives and interpretation of their data. Independent of my advisory role, Ciphergen is sponsoring research at MD Anderson as a collaboration where they are providing instruments and personnel and we are providing expertise and specimens.
There are a number of ovarian cancer tests in development aside from Ciphergen’s. There’s the one at the University of Pittsburgh, the Yale/Lab Corp test and one by LabCorp and Quest. Can you comment on these or others?
I think it’s fair to say that all of the panels of markers for ovarian cancer at the present time are a work in progress and all will require independent validation with large numbers of patients with early-stage disease. And that ultimately the test isn’t whether you can detect stage 1 disease but whether you can detect pre-clinical disease before it becomes stage 3.
I think there are two general approaches to utilizing proteomics. One is to look at the pattern of proteomics. The other is to try to identify just those peaks that have the greatest information and either to develop proteomic or more conventional assays for just those specific peaks. My own bias is that the identification of a limited number of well-defined antigens will be a more reliable test in the short run.
As the technology evolves, it may be possible to get stable signatures that are helpful as well. In the short run I think it’s probably going to be more straightforward to try to develop assays for the most important individual markers. One of the appealing things about what’s going on at the University of Pittsburgh is that they have managed to capture most of the known markers for ovarian cancer and miniaturize those with Luminex multiplex assays. And I think that will be of great interest to see how that progresses.
How long do you think it will take before these proteomic tests reach the market?
I think a lot may depend upon how the trial progresses in the UK with Ian Jacobs’s study with CA-125 alone [Ian Jacobs is director of the Department of Gynecological Oncology and the Institute of Women's Health at University College London. Currently, Jacobs is doing a trial with 200,000 women where 100,000 are receiving conventional examinations, 50,000 annual ultrasounds and 50,000 CA-125]. If that, in fact, does improve survival and it results in a larger fraction of cures because of earlier detection, that could become the standard on which other approaches are based. And at that point, if you can show either with multiple individual markers or proteomics or whatever that you can improve on CA-125 alone or replace it, hopefully that would be respected by the regulatory groups.
I think though that it’s going to require some retrospective validation with serum banks such as that from this trial in the UK or the PLCO [Prostate, Lung, Colorectal, and Ovarian cancer screening trial], and that it may also require prospective studies. And the time frame, if a prospective study is required, is going to be more substantial. Whatever, I think it’s going to be crucial that there be multiple data sets to validate biomarkers that are going to be used for widespread screening in the community.