Director of Gynecologic
Beth Israel Deaconess
Name: Stephen Cannistra
Position: Director of Gynecologic Medical Oncology, Beth Israel Deaconess Medical Center; Co-director of the Women's Cancer Center at BIDMC, 1998 — the present
Background: Program Director for Gynecologic Oncology at the Dana-Farber Cancer Institute, 1990 — 1998
Education: MD, Brown University School of Medicine — 1979
In light of the recent progress being made in prognostic indicators of chemotherapy and remission in women's health, by Genomic Health and others, Pharmacogenomics Reporter decided to speak to Steve Cannistra, the director of gynecologic medical oncology at Beth Israel Deaconess Medical Center.
Cannistra and his colleagues developed an Ovarian Cancer Prognostic Profile in December 2004 that has had some success in identifying patients who are likely to do well in first-line chemotherapy from those who may need more aggressive treatment. Here Cannistra speaks about an update on that profile, the Chemotherapy Response Profile, which appears in the Nov. 1 issue of The Journal of Clinical Oncology.
Are there any profiles on the market that resemble what your group is doing?
Well, nothing in ovarian cancer. Certainly in breast cancer — gene profiling to identify patients may respond better to chemotherapy in the adjuvant setting, for instance. That is something that is becoming more widely recognized as a possible avenue to take, but nothing in ovarian cancer, not at the moment. This is the first, that I am aware of, that addresses this particular issue.
Is the breast-cancer profile you're talking about similar to that marketed by Genomic Health — the Oncotype Dx profile?
Exactly. That's what I was referring to, and I think that certainly is a very nice paradigm that other diseases will hopefully be able to follow in the future.
It's similar, and it's not. It's not similar in the sense that we're not saying that the chemotherapy-response profile will lead to therapeutic options. You see, everyone needs to be treated with pacitaxil and carboplatin.
Their options are extremely limited?
That's correct, and you wouldn't deprive someone of a possible benefit from pacitaxil and carboplatin, even if the gene profile was not in the 'highly responsive' category. But I see a day where we may be able to have a better treatment option for patients who might not be as responsive to pacitaxil and carboplatin, based on the chemotherapy-response profile. We're just not there yet.
This new profile is an update on an existing ovarian cancer prognostic profile your group developed, correct?
That's correct. The Ovarian Cancer Prognostic Profile was originally identified on the basis of its ability to separate patient groups into those that did very well with first-line chemotherapy, versus those that didn't do as well as we would like. This involves patients with advanced ovarian cancer, where we don't have very precise clinical measures of prognosis, so we're always trying to identify better ways to determine how patients will fare with our chemotherapy, and the Ovarian Cancer Prognotic Profile provided us with that.
But what it didn't do was help us understand which patients responded the best to chemotherapy, versus which patients didn't respond as well. So for that reason, we tried to identify a gene-expression profile that actually provided insight into that question, namely, 'Who will derive the most benefit from chemotherapy, versus who won't?' And in order to do that, we used the results at the time of a second-look procedure. That means that patients have finished their chemotherapy, and are now taken back to the operating room, and either with the laparoscope or actually doing a laparotomy — where a surgeon looks inside the peritoneal cavity — we can determine very precisely whether they have achieved a complete remission. And by grouping the patients into those who have achieved a complete remission, versus those who unfortunately did not, we were able to identify a gene-expression profile that was quite characteristic of patients who tended to respond much better to chemotherapy, and therefore developed a complete remission after first-line treatment.
They showed this gene-expression profile prior to the first treatment?
That's correct. The profile was derived from tumor tissue obtained at the time of their original diagnosis. So, what we have, in essence, is a genetic snapshot from the very beginning of treatment that can predict how a patient will respond to six cycles of chemotherapy. And that's interesting to us, because it tells us that the die is cast, in a sense, at the very beginning, and that there's a genetic secret within that tumor that still remains to be unlocked. But at least it gives us a chance to further study this, and to understand why it is that those patients are responding better, based on the genes that their tumor expresses.
Beyond prognostic uses, does this kind of profile have a future in finding pharmacogenomic markers to identify patients likely to respond to specific drugs?
Yes. What we hope to do, in fact — and that's the reason for developing the Chemotherapy Response Profile — is to use it to hopefully identify new targets that might be amenable to therapeutic intervention. You see, the original profile, the Ovarian Cancer Prognostic Profile, simply provided us with prognostic information. But it didn't necessarily give us a handle on therapeutic targets.
This, on the other hand, gives us insight into certain genes. One of them is the BAX gene, for instance, that may well form the basis for developing new therapies to improve the outcome of patients with this disease. We're not there yet, though. But that's the ultimate intent.
In ovarian cancer, we have a set regimen that is used for all newly diagnosed patients. It involves two drugs, pacitaxil and carboplatin. And those are the best drugs that we have so far. It's not as if we currently have three or four different regimens that are equally effective in the first-line setting — that seems to be the best. But I envision a day, using gene-expression profiles like the CRP, where we may be able to be smart enough to study the patient's tumor at the time of diagnosis, and based on the genes that are expressed, tailor first-line therapy to that particular patient. This is a first step [in] that direction.
Have any pharma companies shown interest in licensing the profiles for use in drug development?
Not at the moment.
Is that the next step?
I don't know. We'll just have to take that one step at a time.