DR. AGHAJANIAN: Just a comment for Dr. Bookman. You know, one of the nice parts of the 170 series, in terms of being able to accrue them rapidly is, the eligibility is sort of easy. It is one or two prior chemos. No matter busy we are, we could all count to two.

When you start getting to the other trials that are in the 126 or 146 series, which are zero to six and six to 12 months that they have to develop measurable disease, it is becoming increasingly complicated with consolidation therapies and other things going on with the patient, to sort of figure out if a patient exactly meets the eligibility criteria.

Sometimes they don't quite seem to. It seems that the ability to really easily put patients on those studies has become more difficult, and sort of going to a model of sort of screening by just saying one to two priors might be faster than getting caught up in the platinum-sensitive/resistant on our first pass.

I think that those terms were defined a long time ago before we have all the things we have now, and the long extended treatments that sometimes are done on various protocols. Our research nurse throws her hands up and we can't figure out where to put those patients.

DR. BOOKMAN: We actually discussed that at the July meeting, and developed a simplified eligibility regarding the platinum treatment free interval, which had become part of new trials.

In addition, because of the upcoming randomized trial 213, that Dr. Coleman is going to rewrite again, I think, once that opens, we are basically closing the 146 platinum-sensitive queue and focusing on the platinum-resistant population.

Even with that, most of the new agents coming forward are more suited to the 170, the biologic queue that already is a little more easy to enroll patients.

I think it is something that we are definitely sensitive to. It has gotten tricky as well from a group perspective, of how do you write eligibility that are very clear, so that the person on both ends of the phone knows what the right answer is, but I think we will get there.

DR. COLEMAN: I think one other idea we are going to have to come up with relatively soon, or decision I should say, to come up with very soon is, what are we going to do about patients previously exposed to either antibody or TKI-type agents that will obviously influence our subsequent development.

Right now, all the protocols are written to exclude patients with prior bev and RGF exposure, but a time is coming relatively quickly where our patients in the recurrent setting are going to be previously exposed or, when they go onto second or third line therapies or biologics, they are going to be previously exposed.

DR. SPRIGGS: Do you have a proposal?

DR. COLEMAN: I will tell you, the way I originally wrote 213 was to include those patients, because I have yet to hear data to the contrary.

We haven't formally studied it in our center, but I am not sure that I know that we are going to see the same types of previous exposure and subsequent probability for response with these agents, as we saw with our cytotoxics.

Currently we are writing it to exclude it, in this particular version of it, but we have -- Mark and I have discussed that, in the future, if accrual looks like it is going to drag, we may need to re-look at that question again. I am just trying to make a plea that maybe there is a way that we can study somehow this particular question.

DR. BOOKMAN: One of the nice things about 213 is that it is a randomized trial. So, you can certainly include patients and stratify according to whether or not they have had prior biologic exposures.

I agree, we don't know the answers to those questions, and it is going to vary from study to study and from agent to agent, because some places are going to be more explicit and more strict than others. Some others will be more liberal.

DR. SWANSON: I was going to suggest that one way to include more patients in clinical trials is, when people get their initial surgery, to test them for VEGF and some of these other biomarkers, because I don't think it is routinely done. I notice that some clinical trials require that information.

DR. BOOKMAN: Overexpression of VEGF, for example, is very common in epithelial ovarian cancer. At this point we don't know enough in terms of biology to recommend treatment based on a marker like that.

So, we wouldn't want to inappropriately take patients out without validating data first. So, getting more information from tumor biopsies is great, and that is why we have the tissue bank and correlative endpoints in our studies.

I don't know for sure that what you are saying would actually enhance accrual, because our accrual is not generally selective.

You know, a lot of the studies I showed you, only a handful were really restricted to certain criteria like expression of HER-2-NU or expression of EGFR. The vast majority took all patients who were interested in participating.

DR. CIBULL: At the risk of really pissing people off, to get back to the original sort of big picture, it seems like the SPOREs are well-designed to do translational research and generate ideas for phase I studies, but poorly positioned to accrue enough patients to do phase II studies, whereas groups like the GOG were designed to do clinical trials, and asked to do translational research without sufficient funding.

They have the patients to do phase II trials relatively quickly, and certainly phase III trials. It is obvious that the reason that we are all here is to find some way for the SPOREs and the group like the GOG to get together and combine their resources.

One way to do that might be if we are running multiple phase IIs in ovarian cancer, is to have one of those tracks be set aside for SPORE-generated ideas, and that the SPOREs would fight it out among themselves which idea should be put at the top of the priority list with the consultation with the GOG. I am just throwing that out for discussion.

DR. DANCEY: I think that is what I was trying to say, that I think the SPOREs can do a good job of getting the idea, or potentially doing the phase I portion, but then, to get a phase II trial done at a SPORE site, it can't be done.

If we work together and fed those into the GOG, they could have potentially a queue where they weren't eligible for any of the other trials, and there are still lots of women in good shape that could go on a new study.

I think, yes, how you would prioritize them there, it might be fighting it out, but it is going to work out some so that the trial is ready, the drug is available, the study is funded. I mean, I think it could work.

DR. CHRISTIAN: Could I follow up on that just for a moment? The GOG clearly has the greatest accrual potential among the resources that are available nationally.

Yet, as Janet Dancey pointed out, even GOG studies take three years to complete. So, I think we need some strategies to speed up the identification of important new phase II agents.

I, frankly, don't care whether they come from SPOREs, from industry, from wherever. I think we need a prioritization method that selects the most promising agents, and then evaluates them very quickly.

So, what are we going to do about that? Is it the time we spend between stages and then doing analysis that slows us down? What is it that prevents a national group from getting these studies done faster?

DR. BOOKMAN: Well, speaking as a representative of a national group, we are very sensitive to this issue. In fact, every single LOI that we have submitted in the last few months for investigational agents has been disapproved by the IDB, in part because our projected start dates were too far into the future.

Now, we are in the awkward position of wanting to be honest about our projected start dates, but at the same time, we want access to these interesting agents.

So, it has prompted a series in discussions, culminating in our meeting yesterday, to try to move some of these things along.

The three-year time frame does not reflect the -- it is not a simple thing. First of all, we may get LOI approval a year and a half before we have an opening in the queue to open the study.

So, we manage our time and effort accordingly, knowing that that endpoint, that milestone, is coming along. So, we are not going to write the protocol within 30 days if it is not going to open for a year and a half.

Now, it is probably a proven LOI that is going to open a year and a half later, and it is up to us, I think, to try to figure out how to work more efficiently within the group to get studies to open more quickly.

This has become a more acute issue recently, and I think it is something we can beat, and we did talk about some ideas yesterday in terms of better ways to divide resources in the group to promote rapid completion of studies.

As I said, one study we opened and closed within 35 weeks, but on your calendar it shows two years, because you are going from the time of LOI approval, not from when we actually opened the study.

DR. CIBULL: One of the factors is, of course, reimbursement for phase II trials. Particularly if the NCI wished to foster that kind of cooperation between SPOREs and the GOG, it might consider applying more resources to phase II trials using novel agents created by the SPOREs and tested by the GOG.

DR. OZA: One of the other opportunities to perhaps consider is cooperation with phase II consortia, and also through the GCIG.

I wear two different hats. One is as an NCI phase II consortium hat, where we have had, in fact, pretty good accrual to studies, but perhaps a more formalized collaborative effort with the GOG or some of the SPOREs might be very useful.

Secondly, through the GCIG, I think that there is a potential for collaboration, particularly with SPORE-initiated ideas to be taken into phase II settings.

DR. MONK: I was just going to say that these sorts of conversations are very instructive, because we have everyone at the table.

I want to support Dr. Bookman's comments, that we are trying to broaden eligibility. We are trying to broaden evaluation of response and progression free survival. I guess the second point is cooperation is key.

So, I would also hope that the NCI could, as they have been, continue to shorten the timeframe that it takes for an LOI to be reviewed, shorten the time with the CIRB.

You know, all of these things add up. Certainly we, as clinicians, and the people that do these trials, are working hard and so is the NCI.

I think together, each of us at every step, should cooperate to quicken the timing, broaden the eligibility, alter our evaluation criteria. I think we can get these things done in a very timely fashion, as Dr. Bookman has proposed.

DR. BIRRER: I think that Dr. Bookman was onto something in an important discussion related to narrowing the focus of specific GOG institutions, and maybe, if the GOG infrastructure will allow a larger number of simultaneous trials done by focused groups.

For example, we have in development an interferon gamma bevacizumab trial that is being done by GOG institutions who also are participants in an immunobiology cooperative group.

So, the immunobiology cooperative group is doing the basic science work for that trial, and it will involve 10 institutions.

That way, some of the IRB concerns will be allayed. The protocol will be open longer in that institution, but they will focus on getting that trial done.

DR. SEIDEN: I think it was Michael, or maybe it was David, had a list of what is it that slows down the trials, and it had a long list of possible delays.

Our institution, going through the IRB is about a 100-day process. Activating after it goes through the IRB is about another 100-day process.

An analysis of why it took 200 days showed that every single one of those things on that list were problems. It isn't an issue of what exactly is the problem. It is a 100 sort of different point problem.

Everyone who needs to speed things up actually isn't in this room. In the Harvard system, to open a trial across the consortium requires now about 45 signatures, from three different nursing committees, three different radiology committees, three different pharmacy committees, a number of lawyers.

I think, realistically, if you think about even the trial process 10 years ago, the number of people who now say, I better get a look at this and sign off of this from whatever kingdom I belong to is getting complicated.

One way that possibly government and regulatory agencies can help is the burden of reporting everything to IRBs really does slow things down.

Michael had that interesting slide of how many reports go to the IRB. If I see one more erbitux rash safety report for 076DD, I am just going to kill myself. I mean, it just is nuts.

My point is, not only are there all those safety reports, but just the amount of reporting and paperwork just clogs the whole thing down. Can we get some sanity into this system?

DR. CHRISTIAN: One of the things we put on one of the slides was the risk of causing digression, which I don't actually want to do, but following up on your point about IRB approvals and safety reports, you know, one way to approach it might be actually to use the CTSU and the central IRB, which would obviate the need to distribute individual safety reports to every IRB, and might allow people actually broadly to open up these trials quickly.

So, it is something that we actually ought to seriously think about or talk about, perhaps after this session, in terms of whether that could at least help with some of these problems.

DR. SEIDEN: Just to close on that, as we add these new layers of time saving processes, it is very important to make sure that the old layers also goes away.

What ends up, in our institution, is the institution says, okay, if there are nine other central layers, that is fine with us, but we are not getting rid of our layers.

DR. OZOLS: LOIs from GOG get pretty extensive review. I guess they get sent to CTEP. Why should CTEP review them and turn them down? I am just curious, when something goes through GOG and goes to CTEP, what basis do you turn them down on?

DR. CHRISTIAN: I am not going to try to give a lengthy answer to that. I will give you a couple of examples of reasons.

One would be it is duplicative, because the trial is being done somewhere else, and there are a few other people doing GYN phase II studies, because they require smaller sample sizes.

Another reason would be, as I think Mike just mentioned, that the projected start date is two years down the road, and we don't usually place studies two years down the road.

So, we might disapprove it for that reason and either eventually place it somewhere else, or approve it at a time in closer proximity

DR. SPRIGGS: I would like to touch on another topic that Dr. Wolf brought up, and that is the issue of how to hold INDs for investigational agents that may be created within a single institution.

I know that the GOG has struggled with this and has really insufficient infrastructure to really do that. I am wondering whether or not there are ways that IDB can help to facilitate getting those institutionally created projects out to a larger group of investigators.

DR. CHRISTIAN: Well, we actually have an extensive program for working with investigators on agents that they have developed, either through the RAID program, which you heard a little bit about, or through our traditional drug development group, where NCI agrees to actually sponsor trials and assume the regulatory responsibilities for the agent.

So, there are actually a number of mechanisms by which individual investigators in academic settings, who are developing promising new agents, can get NCI assistance.

Again, there is a lot of information on our website about that, or you should really speak with one of here today, if that is an issue.

DR. BIRRER: I wouldn't want to supersede Ted, but I did have one comment about the phase II trials from a translational science standpoint, because Mike's committee and my committee at GOG frequently discuss this, and usually end up screaming at each other.

I think that a number of issues have come up which are important to crystalize. First of all -- and I think we have a bit of a consensus on this within GOG - that there are lots of markers, there are lots of phase IIs, but for any given phase II trial, even one that is active, even where the agent is active, the total number of responses are going to be fairly small.

The cost of doing biopsies, the complications, the ethics involved, make it a big deal. So, we have evolved, over a period of time, to be more selective in what we conduct.

Guidance from the CTEP on this is helpful, because the general flavor we get when the LOIs come out is that institutions that do translational research, or proposals to have translational research, will be favored.

I guess I understand that, but it does put us in a difficult situation. I don't think that translational endpoints on all these trials will be helpful.

In addition, I would ask whether people considered the flexibility of phase II trial design where, where, if you get an active agent -- for instance, Avastin is active, but even in Bob Burger's phase II trial, you are only talking about a small number of responses.

The chance that we are going to correlate a marker to that response, I would guess, as a scientist, is going to be pretty small.

Has anybody thought about the possibility that, in selected cases, you would actually take your phase II trial out to do further accrual, so you could actually get enough responses to evaluate carefully a given set of biomarkers, so you would have that data going into the phase III trials?

For instance, now we are going into 218 using Avastin, and frankly, we don't know, really, any validated biomarker.

DR. CHRISTIAN: Let me just comment on that, too, because I think we are all grappling with these same issues, of what is the best kind of clinical trial design in which to address some of these translational endpoints.

I think that we agree that simply doing exploratory studies across a wide range of clinical trials isn't the right way to do it, that in fact in order to correlate things you need an active agent, and so we maybe should be focusing on those agents where we have activity, and perhaps accruing larger sample sizes in order to address those.

I guess the difficulty that we deal with is that, the larger the sample size, the longer the trial takes, and we are then delaying what we think is an active agent forward.

Certainly one can make strong arguments for trying to understand mechanistically how they are working so as to optimize it, but it is a very complicated decision-making process, I think, that we all need to spend more time thinking about, what is the optimal clinical trial design as we move forward?

I think we didn't have much time to talk about that in this session. I think speed of accrual, getting these patients accrued quickly and in larger numbers so that we can address some of these issues is important.

I agree with you, and I think CTEP agrees with you generally, that we really do need to think very differently about the way we incorporate correlative laboratory studies.

DR. BIRRER: One way we have gotten around that is, like with the cetuximab trial we are doing, the trial is powered on response rate for a certain number of patients, but there is a provision that, if there is enough activity to get a signal, accrual would be expanded only if the primary endpoint is met.

So, you can build it in as a trip mechanism without having to commit to a much larger study, and that may have no signal at all.

DR. WOLF: Let me comment, if this is a single agent study, it becomes a third stage in a phase II evaluation.

So, your first stage is exposure if the agent is inactive. The second stage is to more fully evaluate the level of activity and, if it meets that threshold, you could then expand to a third stage.

At that point, you would think accrual might actually be quite good, patients and investigators would rally around an agent that looks promising.

In that third stage, you would be accruing sufficient numbers of patients to address a specific predictive marker question, for example.

The reason to do that or, if you like, the prerequisites for doing that is that you have some idea of what that predictive marker is, you have a robust assay that you are prepared to bring forward and, when I say robust, it is all of the criteria that I think everybody here who has been involved in translational research, has an understanding of.

I would not recommend that if it is an exploratory, well, we are not really sure if this is the best target kind of scenario, but I think it is certainly a reasonable design and one that I would welcome further discussion around, in an area that we might develop a predictive marker that could be incorporated in a phase III trial in an efficient way.

DR. WEISENTHAL: Just briefly, it seems like there is an increasing need for resources for translational studies. There is an increasing need for resources for phase II studies.

There is a finite limited amount of total resources, and the only suggestion that I heard, as far as getting things done, is to put more money in the system, which isn't really going to happen.

A suggestion that I would have has to do with raising the bar for doing a phase III study. I don't want to say anything too inflammatory about ovarian cancer, but just to give you briefly an example of what I mean, if you take the situation of metastatic breast cancer, there have been literally scores of prospective randomized phase III studies done in metastatic breast cancer involving literally thousands of patients.

There was an interesting study published last spring where they asked oncologists, what do you use to treat metastatic breast cancer.

In the case of an academic medical center, off-protocol, what 85 percent of the medical oncologists would use is Xeloda or 5-FU.

When they asked the same question of community oncologists, 85 percent of them use intravenous combination chemotherapy with a whole variety of agents.

The point is that, after 25 years of doing prospective randomized trials, there is no clarity that has been brought to it at all.

I think that with only slightly toning it down, you could probably make the same sort of criticism about ovarian cancer.

The question is, you could maybe take some of the resources away from phase III trials. In other words don't do a phase III trial until it is really obvious that it is something worth testing.

Agenda Item: Wrap-Up Session.

DR. TRIMBLE: Why don't we move on to the wrap-up session. Before I forget, we moved up the time for dinner to 6:00 o'clock, for our informal barbecue at the hotel, so that local people can get home a little earlier, and their out of town guests can get to bed or go out to the nightclubs of Washington a little bit earlier.

I worked through the session with Mike and with Jane and Drs. Michaele Christian and Jeff Abrams, to come up with some of the ideas that we saw and heard as potential action items.

I just kind of wanted to go through those briefly with you all to see if there were other ideas or additions as to who needed to be involved in those.

The first we heard a lot about through all the sessions was using CA 125 as an eligibility criterion for enrollment of clinical trials in patients who did not have measurable disease, as well as using CA 125 as a response criteria.

Certainly, as I said, there was a great deal of interest in that. It would seem reasonable that we might want to consider a working group including representatives from GOG and the Gynecologic Cancer Intergroup and CTEP to explore that as a possibility. Does that sound as a reasonable action item?

As has been mentioned, Dr. Pazdur at the FDA plans a workshop within the next several months on endpoints in ovarian cancer trials.

He has asked for potential speakers and potential panelists. Both NCI and the Gynecologic Cancer Intergroup and GOG have provided him with names.

I understand he also plans to ask ASCO and AACR for potential names, but that the FDA will determine the final agenda for that conference and then present, at some point, their results, or the recommendations from that conference as conceived by FDA, will be presented to the ODAC, but I don't really have a schedule for that.

The next action item that I heard this morning, and it seemed to get consensus, is that we needed to separate out patients with clear cell and mucin epithelial ovarian cancer, that they were not the same, did not have the same molecular biology as other subtypes, and it was appropriate to put them on separate trials. Does that sound reasonable?

Obviously, we need to work with our fellow members in the Gynecologic Cancer Intergroup because these are such rare histologies that we probably need to work together.

DR. MANNEL: For point of information, the GOG has already voted through the protocol committee to do just that. So, mucinous tumors and clear cell tumors are being looked at as part of the rare study committee for that reason.

DR. TRIMBLE: Good, okay. Then there was some suggestion of whether we could use some novel statistical approaches for phase II trials.

Here, too, I think we need to have a working group of statisticians from GOG and CTEP and possibly the Gynecologic Cancer Intergroup to look at some alternatives to our current strategies. Does that sound reasonable? Are there other groups that need to be involved in that?

Now, certainly we heard a lot of discussion about how the SPOREs and other institutions, including GOG, could better work together to facilitate innovative work coming out of the SPORE and other cancer centers.

We did have a brief breakfast meeting this morning with representatives of GOG and SPOREs to discuss this. So, I think it seems apparent that we need to have a working group that would again include the representatives of the SPOREs, the GOG and NCI, to see how we can help get these trials done. Any other groups that we need to involve in that working group?

COMMENT: What about the NO1 holders?

DR. TRIMBLE: I think you are right. We need to include those contract holders. I mean, you all are contract holders. So, that would bring your group in.

One issue that we heard the example of what has been done in breast cancer using paraffin-fixed tissue resources to evaluate a panel of gene markers.

We might want to consider something like that in ovarian cancer, certainly between the resources that we have available in our various tissue banks, between what David Bowtell has put together in Australia, thanks to the DOD grant, the GOG tumor bank, the various SPORE banks, as well as the banks in some of our larger cancer centers.

We might be able to use that frozen material to look at, or to validate a microarray analysis, and then use that, once we have identified the appropriate genes, to see if we could work to develop the RT-PCR for use on paraffin.

So, this has been done, I think reasonably successfully in breast cancer, using the resources of the NSABP and industry. Does this sound like something we should explore?

DR. BAST: I think Mike Birrer also suggested that it might be helpful to have an ongoing working group look at genomic and proteomic in general, because there is a tremendous amount of information already out there that we could make use of.

Given the paradigm of how things have gone in breast, to select markers that we already know about from individual marker studies, and to use the genome, both to look at stage one disease, and also for prognosis. That might really well be worth doing, but have people who think regularly and get together to talk about this on a regular basis would be helpful.

DR. BOYD: Just very briefly, there are at least as many data, if not more, currently available in ovary cancer than there were for breast cancer at the time that the particular company began developing its assay.

The key, of course, was focus, dedication and money from -- in this case a corporate entity teamed with a Cooperative Group, that individual PIs or even groups of PIs just simply don't have the time and the motivation to pursue.

I think at a consortium group level, we could certainly be as successful in ovary cancer, or have the same prospect for success in ovary cancer, as did the group that has developed what up to now would appear to be a successful breast assay.

DR. HOWELL: I completely agree. We need a project team focused on this that needs to marry high-level informatics into this as well, the people who are really good at data mining and turning things out of this.

It really is a shame that we have not done with ovarian cancer what has been done in breast. It is such an obvious thing to do, and it is going to take a lot of coordination and, therefore, a czar to make this happen across the groups.

DR. CHRISTIAN: I think it is actually important to define the process by which we will make that happen. I think, as has been pointed out, there probably exist significant amounts of data right now in ovarian cancer, and what we need to do is to have a systematic approach to reviewing that, and perhaps engaging an industry partner as was done in breast cancer, or a contractor or some entity that will actually do that careful analysis.

So, there are statistical resources that could be brought to bear as well. I think if we feel this is an important objective coming out of this State-of the-Science meeting, an important opportunity, then we should define how we will move forward.

Lisa, I don't know if you had any other comments about this, but I think, again, if we think this is a high priority, then we need to define a process and get on with it.

DR. MCSHANE: I think it is doable.

DR. ABRAMS: The point that I would like to add is, you do need to define your question. So, really, what do you want. Do you want a prognostic marker in early-stage ovarian cancer that is going to tell you who to treat with chemotherapy and who not, or do you want predictive assays for later stages, who is going to benefit from specific chemotherapy.

I think it would help the breast cancer effort move forward if they defined the question and they went after that question and they made sure they had large sets of valid tests and validation sets in the Cooperative Groups, and it turned out to be NSABP patients that they tested it on.

As you plan this, that is what you have got to look for, is both what is your question and what is your stored paraffin samples that will then answer what you come up with.

DR. BAST: I think you have outlined at least a couple of the things that would be worth doing. Certainly one in early-stage disease, to look at who needs chemotherapy, as tough as that is going to be, at least with frozen tissue.

I think certainly the second issue is really, who needs taxane therapy and who doesn't, and the third is, who is going to do so abysmally with conventional carboplatinum Taxol that you need to do something at least in addition, maybe something completely different.

Those are three different problems, but I think at least some of the data are out there to deal with all three of them.

DR. LIVELY: I think another reason to emphasize focusing early on, on perhaps one high priority question is, again, looking back at GHI's experience, one of the keys to their success was a huge amount of effort and time and money that they invested in their laboratory processes, devising a set of primers for their RT-PCR assays that worked on paraffin, worked on old paraffin and new paraffin, and were completely documented, that they had this reliable, robust assay.

That was an industrial scale development effort. I am not saying it is not worth doing, but it is going to be something that we shouldn't underestimate. It is going to be necessary if this test ever actually is a success.

DR. SEIDEN: I wouldn't underestimate how hard figuring out what the right question is. I think everyone would agree it would be wonderful to know who, with early-stage disease, didn't need treatment.

For that study, you would need hundreds of samples from people who were not treated and did well. That repository doesn't exist in the United States .

DR. BIRRER: I appreciate the last few comments because, as I was sitting here, I felt I was going to be in the position of being the pessimist among us.

I think the general direction is the right one. My view, as someone who is, I think, reasonably involved in this field, is that the timeline I see is slightly different.

I favor more of a working group of scientists who are involved in this to review the data with statisticians, and then decide, really, what the next step is.

The reason I say that is because, yes, there are lists out there. There are lists for responsive and sensitive tumors. Very few of the genes overlap and they are completely different from the lists that are generated for survival.

Some of the lists are generated from micro-dissected specimens. Many of them are not. It is extremely complex and I would argue more complex than what was available in the breast data.

The early ovarian cancer question is very interesting. That data has not been published. I don't know of any gene lists that predict recurrence from early ovarian cancer. So, we are in the discovery mode on that.

I think -- I don't want to be too pessimistic. I think it is the right direction and we ought to put a timeline on it, but we should be rationale about it. The only thing worse than not doing it, is doing it badly.

DR. BOWTELL: I would just like to say that I support Mike's comments on that. I think that the breast cancer is a terrific model of that, identifying the best of the best and using that to go forward, and that reduces the problem of multiple testing a lot.

We are a bit behind where they are with breast cancer, at the stage when the GHI test was being developed. So, I think probably these things need to proceed in parallel.

I think we have got to look critically at the data sets that are out there at the moment, and I suspect we still need some more data on larger sets, given that virtually all the sets are less than 100 at the moment.

I think there probably is material there, and there are genes that could be taken, and the GHI list was not just derived from microarray data. It was derived from other sorts of data that could be systematically analyzed. I think if there can be a consortium effort come out of this, I think it would be very positive.

DR. KOHN: I think one additional point to consider is, even though we may not have the power that we need to have data sets, and they might have a lot of useful information, we ought to look at some techniques to try to find a way to bring all the data sets that are published together, so they can be modeled across each other.

DR. TRIMBLE: Okay, I think we have got some direction on that issue. Turning now to IP therapy, I heard that there was a feeling that we needed to build on our experience and our achievements in IP, that we needed to work to reduce the toxicity associated with IP, to investigate to see if we could find out what it is about the IP approach that is helping patients to live significantly longer, and that, in general -- and that as we design phase III trials for women with optimally debulked ovarian cancer, that we need to consider some of the therapy, or giving some of the therapy by an IP approach.

Certainly we also heard the need for education through our professional societies, through advocacy groups and through NCI, about the benefits and risks associated with IP therapy.

DR. BOOKMAN: I was thinking about the fact that, when the NCI puts forth this release, it is kind of open-ended in a lot of ways.

It would be appealing if there was some way to get something positive out of it besides just circulating the information.

I don't know the answer to this question I am posing but, for example, what is going to happen perhaps is that many physicians and patients are going to be embarking on that approach, but with a variety of methods that they choose on their own to deal with things like processes, dose schedules, number of cycles and so forth.

Might there be value to, for example, create a registry where physicians could indicate how they have treated their patients and provide very limited data on extent of residual disease, viable, and outcomes, without details, reporting.

At least, then until there is a more established standard, we could collect some information that might help inform us about the problems and benefits of doing this in an IP setting.

DR. TRIMBLE: We have had discussions with the NCI's Division of Cancer Control and Population Sciences, which is responsible for tracking outcomes and working on diffusion of innovations.

We have had some thoughts about how to track uptake of the recommendations for IP chemotherapy. They have a cancer research network that encompasses several large HMOs, so we can work with them to gather data prospectively.

Long term, look at the SEER Medicare files, which has its disadvantages because its patients are older than age 65.

We also will be discussing, with the principal investigators of an NCI-sponsored project called CanCORS, which is focused, I think, on lung and colon cancer, but they have been following several large cohorts from diagnosis through survivorship, and asking them whether they might consider, or be receptive, to the idea of adding ovarian cancer to their list. That would obviously require some additional money.

DR. BOOKMAN: Any of those things would greatly enhance the value of -- if we put out the release now with all the uncertainties that exist -- [off microphone].

DR. TRIMBLE: The other two entities that we also hope to work with are the National Comprehensive Cancer Network which has asked NCI for potential projects and we think this would be a very good project to work with, and the American College of Surgeons commission on the National Cancer Data Bank, which has also been asking for projects. They do track 80 percent of cancer patients in the United States , as far as initial treatment. So, we plan to work with as many organizations as we can to evaluate uptake of these recommendations, as well as on education on how to deliver that therapy.

DR. BOOKMAN: It would be nice to find some mechanism of collecting some of that data.

DR. TRIMBLE: We will explore that. It is expensive to put together a cohort study.

DR. OZOLS: It looks like GOG will not be doing an immediate randomized trial of IP therapy. You know, I think all that stuff, in one sense, may be moot.

Supposing Taxol, carbo and Avastin is really good, and maybe it is even better on patients with small volume disease.

I think we should reconsider, while you are doing these phase II studies, to see if you can decrease the toxicity of this regimen and give more data about it, why not include the small volume disease patients in those centers who want to participate in a randomized trial.

Supposing, you know, two years from now that IP carbo/Taxol has its biggest impact in a patient with small volume disease and Avastin may not be a good drug for IP.

You may have a whole changing baseline and we are missing an opportunity to see how good Taxol/carbo/Avastin is in small volume disease, all given intravenously, while we have this window here.

DR. TRIMBLE: That is a tough issue. I will have to think about that.

DR. MANNEL: I think one of the things that we are fortunate in, in ovarian cancer, is that the GCIG last year stated that, for clinical trials, that IV Taxol/carbo should be considered the standard arm.

I think we are looking at modifying that to include some IP therapy. I am not sure exactly what your final draft if going to show, but I do think that it would be important that the alert would say that it is IP cisplatinum.

I share the concern that if we just say IP therapy is the key, we are going to start getting a lot of IP carboplatinum, IP Taxol, whatever, which has very, very little data to support it, and we may be doing a disservice to our ovarian cancer constituents.

So, I think we need to make clear that the data supports IP cisplatinum. I know, at least in some of the reiterations I have seen from you, you have made that clear, and I don't know what the final one is going to say, but I do think that is important.

DR. TRIMBLE: We aren't at the final point yet, but all the versions have said that the three most recent trials have used platinum, given IP.

DR. BOOKMAN: I wanted to address the point that Bob raised about the eligibility on the phase III front line trial with bevacizumab.

That was a compromise, because GOG is running a trial of maintenance therapy and we felt that, because the front line phase III with bevacizumab included maintenance or placebo for extended duration, those patients would not be eligible for the other maintenance trial, and this was an effort to basically allow the group to run both studies at the same time.

Now, should the maintenance trial, for some reason, suffer an untimely demise, then of course it would be a very good idea to roll those patients immediately into the trial.

DR. OZOLS: Don't you think there is a distinct possibility that the Avastin may work perhaps best in the patients who have the smallest amount of disease? We don't know that, of course, but why not take this opportunity to put all those patients on that study?

DR. BOOKMAN: I think that is a very important point, but a year ago we were faced with not doing the Avastin trial at all, and it was only because of some focused discussions on how to do things together that we were able to move forward in that direction.

I think you are raising a good point, and I will certainly support it, but again, it is a complex issue, the navigation of the phase III process. I am not justifying it. I am just citing my recollection of the historical record, so to speak.

DR. TRIMBLE: Why don't we move on. The last two items I have, the first is for some discussion that we had yesterday between GOG and IDB, is to see if we could work to develop a kind of mini-consortia for phase II trials within GOG, similar to the phase I working group within GOG that would help GOG speed up its queue.

We know that, obviously, from CTEP's perspective we need to open our trials sooner rather than later, particularly when we approach the pharmaceutical industry asking them to collaborate with NCI and our groups.

We have to be able to guarantee timely opening and accrual to our clinical trials. So, I think it is important that GOG work with IDB to see what we can do to improve or to speed up the queue within GOG. Certainly we are delighted that Dr. Dancey has volunteered to become the new IDB representative to GOG.

Finally, we discussed exploring whether some combination of the CTSU and central IRB might help to reduce the paperwork associated with safety reporting for clinical trials, but that is something that CTEP will have to work on and is a long-term issue.

DR. CHRISTIAN: I just wanted to clarify that a little bit. The reason to suggest that actually was because I think there is potential to increase and speed accrual.

I think another possible effect, however, would be to reduce the need for safety reporting. So, that wasn't the primary thing. The primary reason for suggesting that really was because I think a broader pool of potential people interested in enrolling on these important phase II studies could be accessed using that mechanism. So, hopefully we could finish these trials faster, too.

DR. FLEMING: I know it has been discussed in many venues, but certainly I have numerous meetings with our chair of IRB and hospital oversight review committees.

I think if OHRP came out with some sort of strongly worded statement supporting this, and explaining why they think it is okay which, to my understanding, hasn't occurred from them, maybe it would be helpful.

DR. ABRAMS: This is an interesting thing, because OHRP is never going to say that you shouldn't use your local IRB, you should just use the central IRB.

They clearly have said -- and on the central IRB website it said -- that the central IRB is a perfectly acceptable and good mechanism for multi-center trials. There is a letter from the director of OHRP on the website.

I think what we are running up against is the deans in charge of this area of research in the different universities have some reluctance, for many different reasons.

We have tried to overcome this by showing them that they would be much better off on the witness stand saying that the government did it, it is their fault, rather than my university.

We are making headway, but I think we have gotten out of ORHP about as much support as they are going to give. It is a perfectly acceptable mechanism, it is within their guidelines, and they think it is a good idea for multi-center studies, and they have written that in to us.

DR. ROSE: I was just going to say that certainly, talking to some different institutions recently, looking at why one study wasn't included in GOG, the cost of putting some of the studies through the IRBs at each institution, at the major institutions in the United States, the relatively small numbers of patients that are going to be accrued to each one.

Speaking to Michael, probably 40 percent of GOG institutions put through their phase II studies. So, we have to address that mechanism if we want to increase our pool.

If it is going to cost you $5,000 to put a study through your IRB that you are going to accrue two patients, or you may have an income of $4,000 from, it is just a money-losing thing, and people aren't doing it because they see it is fruitless in that sense.

We are working at 40 percent capacity of what we have. Whether we would come up with funds that would overcome that initial IRB cost, whether the national government can legislate something saying that this is a nationally funded study, that your IRB cannot charge you, or you go back to the central IRB and saying that it is not a necessary review, because we are losing 60 percent of our phase II accrual.

DR. BOOKMAN: Just to clarify one of the things that Ted mentioned quickly, one of the approaches we would like to take is to target the participation on the trials.

In other words, if a phase II goes out, what you really need is 10 institutions to commit to it. You don't need 50 to put it through their IRB and only 10 actually accrue.

So, if we could find a way of doing that, and keep it within the resources available for the group, that might be a more effective solution. Then you are only putting it through your IRB if you want to actively participate in the trial.

DR. ROSE: We can go back to GOG Stat office and see who puts what patients on protocol, but I would imagine it is the same.

DR. ARMSTRONG: I think all of us area aware of how many hurdles that we have now that we didn't have 10 or 15 years ago for clinical trials, but I think we probably should step back and again recognize what has been the fact all along, which is the single biggest reason that we don't have accrual to clinical trials, is because most people don't get treated on clinical trials.

We have members of the advocacy community here. I certainly think one of the take home messages -- you guys have asked us, what is the priority?

I think the patient perception that the clinical trials are only for when there is nothing else available for you, and your doctor is ready to give up on you. It is probably the single biggest pervasive thing out there that is a hindrance to going on a clinical trial.

If there is maybe one message that the advocacy community could get at, it is that actually the opposite is true, that you would get better care, better medical care on a clinical trial than any other way.

You talk about patients not knowing about phase II trials. You have heard how much work it is. If we had phase II trials available, we are definitely going to talk to patients about them.

The issue is that most patients are not being treated at places where they have access to them. So, if you guys can get the word out to patients, about who is eligible, where the trials are, we could get these done in a much faster period of time and we could get the information more quickly.

You guys have asked for some marching orders, and that would be what I would say is the thing for you guys, is to get the message out, about clinical trials not being a last ditch thing, that clinical trials are really how we move the science forward.

DR. TRIMBLE: That is the last thing on my list of action items. If there are projects that you think we should be doing, or that we discussed this morning and I didn't pick up, please write them down, give them to Jane, Michael, or myself, or e-mail them to us, or bring them to our wrap-up session tomorrow.

We thank you for your valiant efforts today. I know it has been a long day and many of you traveled long distances to get here. So, I want to thank you for that.

TOP