SLIDES & TRANSCRIPTS
Wednesday, February 2, 2000

Report from Working Group A
Charles Schiffer, MD

DR. SCHIFFER: In keeping with the model of drugs as low tech, we turn from antibodies. We confronted a much wider range of issues and potentially a much wider range of possible agents for study. While there was some discussion, maybe at times too much about the specifics of different agents, what we tried to focus on eventually, and I think succeeded in doing, is strategies about how to evaluate new agents and also new laboratory principles as efficiently as possible.

TOP

In the area of compelling and remaining Phase III issues, there was a great deal of discussion about recent trials incorporating multidrug resistance reversing agents such as PSC-833 and why they haven't worked.

There is an question there about what to do with all the recent data that has accumulated about reversal of multidrug resistance and specifically P glycoprotein. There was a clear-cut consensus that there was a positive clue, a strong positive clue out of the recent positive SWOG study which utilized cyclosporine as an MDR reversing agent and that this should be pursued in subsequent trials. In fact, it is being pursued. There was much less consensus, with fairly strong opinions on both sides of the coin, about whether PSC-833 should be continued to be explored with the same zeal.

There are lots of reasons why the past PSC-833 experimental trials may not have been ideal. In retrospective review, many people felt that recently designed studies perhaps excluded some important issues and perspective. One of the big criticisms of the PSC studies was that it is necessary to reduce the dose of chemotherapy in the treatment group receiving PSC which might be to the disadvantage of at least one-third of patients whose pharmacokinetics were not effected.

There are new agents coming along that, putatively at least, do not affect PK but are actually more potent inhibitors of PGP than PSC. Without having discussed this, I know that there is strong interest amongst the people who are still interested in MDR and believe that these new compounds should indeed be tested.

I think aside from this, maybe the other message is that, from all of the other possibilities that one might imagine, nothing else was really discussed. Importantly, nothing else is rising to the top in terms of doing a Phase III trial immediately.

If there is further evaluation of PSC, and maybe even cyclosporine, it is apparent that are a number of variables that have to be addressed. Some of these relate to whether the anthracycline is given by continuous infusion or bolus, some relate to which choice of modulator (PSC or cyclosporine). Looking at the required number of patients that will be needed to address these questions, these are not trials that just a single cooperative group could do. It indicates that there is going to have to be some closer planning at the cooperative group chairs meetings about how to prioritize the issues that were discussed at great length and come up with either one or two studies that are complementary and not duplicative.

TOP

With regard to duplicative studies, we discussed why we all did duplicative studies. This is because the same idea would come across at the same period of time, whether it be high-dose Ara-C, growth factors, transplant, or MDR-modulation, and we all pursued these. These are the four big things that have been done in the last 15 years or so, and I think it is going to be a challenge in the future to avoid duplicative studies as new things come along. So there has to be planning so at least the studies are studying the same manipulation and that studies are complementary, if you will, rather than parallel.

Given the absence at the moment of compelling questions other than the MDR question, there was a long discussion about Phase III studies which address clinically and maybe economically important questions. How many courses of high dose ARA-C post remission? Should these kinds of studies be done; do you need three rather than one? The economic impact of that is obvious and I would say that opinion was somewhat split on this. I think we probably would all agree that if there are a host of Phase III questions that look like they could advance the overall survival of patients, they would be infinitely preferred to questions that might not advance therapy or ultimate outcome, but rather might make therapy cheaper or a little more tolerable.

I guess this really might address the issue, Bruce, about whether you can have two tiers of studies? That is, that some studies that might be directed to community physicians to address questions like this, and there are a few such questions that are relevant to be answered, and other studies which take advantage of totally new technologies or new scientific approaches.

I don't know whether we have enough patients to go around for both. The reason that the cooperative groups have not done these studies in the past is because they have been budgetarily limited and these questions simply haven't risen to the top of the pile. But it is something that generated a lot of discussion for which there is no clear-cut answer.

Another problem, of course, with embarking upon a large study of that type is, if a new more compelling idea comes along in a year and one-half, you are stuck in the middle and would have to delay addressing this important question until the other study was finished.

TOP

In terms of important Phase II questions, the discussion was a little bit easier. We saw a number of slides from a number of investigators that had all sorts of ideas and some of them are probably even good and worth doing. What we wanted to focus on was not which might be more interesting than the others, but in fact, how to start moving this stuff along. There was, I think, general agreement that the cooperative groups should be able to do this and should be able to do this expeditiously.

One problem with opening a Phase II study that has one arm is by the time you get it through your IRB, it is closed if it is a 30-or-40-patient study in a common disease. One simple strategy was to use the so-called "randomized Phase II," which is not randomized as much for statistical purposes of comparing the arms but for the practical purposes of having a study that is open for a longer period of time for a larger number of institutions. ECOG has been doing this.

There are clearly lots of things to study, and it may very well be that certain groups may focus on particular themes that may correlate with particular correlative laboratory interests of the particular group.

There was a nice discussion this morning of more novel designs of studies of new agents in AML, with two points. One is to use new agents up front in people who are at very high risk of failure with conventional therapy. In particular, there was a focus on people with, for example, myelodysplasia who are rumbling along and may not need immediate treatment for their apparent cytopenia but who may be candidates for the use of selective agents if that can be done as initial therapy prior to cytotoxic therapy.

Another model was suggested. One of the problems with using totally unknown agents in even poor risk patients is that there is a finite CR rate in almost all patients. You could deprive such patients of the possibility of achieving even short CRs, which otherwise certainly does add to their quality of life. Thus, an alternative proposal was made that you could study (again some agents may be more appropriate for this than others) certain agents as post-remission therapy in such patients with the end point being two-fold, perhaps. One obviously is prolongation of remission, and the second is serial evaluations of minimal residual disease.

It was recognized that this can be an interesting but potentially inefficient way of doing things because if you started with 100 poor-risk patients and then you have to get them into remission, you might have 30 at the other end. Not all patients will be suitable for further therapy. It will take a different type of statistical analysis than we have been accustomed to using to decide whether this is in fact a good and better model of introducing new agents than the traditional way we have done it in the past.

TOP

We had a bit of a discussion about what is a ipositivei result, the implication of a positive result of course being that you want to use it as initial therapy either in induction or post-remission consolidation in newly diagnosed patients. It is frankly very difficult to arrive at a consensus of what turns you on in terms of the balance between response rate, toxicity, and cost, but it is my bias that there is a crying need for systematic looks at the results of Phase II trials to see what evidence you glean from Phase II that subsequently is verified in Phase III trials that makes it a real drug rather than something that gets discarded. The problem with that approach is that you are not going to find that many real drugs.

There was, also, a lot of discussion of using quantitative MRD detection for minimal residual disease by a variety of methods as a surrogate means of assessing benefit from new agents. It is clear that you can add a surrogate that needs correlation with the outcome of clinical trials to approximate truth and to help make decisions about whether something is a true positive.

The issue here is not just doing it in the laboratory, because that is probably the easiest thing to do now. The issue here is actually getting the samples in a reliable fashion, and this will be discussed at some length, I think, by Marty Tallman. But there are real problems in establishing important endpoints if you are missing key samples in a relatively small number of patients.

TOP

In the area of science and priorities, there was a lot of really good and fun discussion. We focused on the concept that what we are not doing in the majority of patients who are not cured is eradicating or sufficiently suppressing the leukemia stem cell. The obvious question is what is the unique biology of the leukemia stem cell and does it differ from the biology that has been so well defined when studying the large overall population of leukemia cells? There was a lot of discussion about the best ways of doing this -- microarrays, all sorts of things which we don't have to get into right now. But what was clear was that this should be an important focus with regard to drug resistance, because it is obvious that it is the drug resistance characteristics of these stem cells that wind up killing the majority of our patients. There was overwhelming consensus about the interest and importance of obtaining and comparing diagnostic samples versus relapse samples, but what actually may be more informative samples is collecting samples in the middle during therapy to assess minimal residual disease.

This again brings up the practical issues of getting such samples reliably. There were some very good points made about the fact that the nature and size of this conference and the people who were involved precluded a higher level scientific discussion of different ways of approaching this question. We believe it is very likely that the group would benefit from a smaller conference with individuals with unique expertise in this particular laboratory area.

TOP

Some of the issues that came up from this discussion is that this is a new ball game scientifically and in terms of biostatistics, particularly with respect to all the microarray technologies, distinguishing between the subtleties of 100 different shades of orange and yellow and green and blue. The point was made that this new approach would require a totally different biostatistical emphasis than is traditional for group biostatisticians. If we are going to potentially do this type of science on large numbers of samples, we cannot do it without the new biostatistics and genome informatics. One suggestion that was made was perhaps this is something that can be provided or coordinated by the NCI as a resource for all of the groups and all of the tissue banks because this is going to be generic to leukemia and all other forms of cancer.

TOP

Our last and overriding issue which I think was thematic for the entire meeting is the recognition of the fact that leukemia is really an orphan disease. Both the pharmaceutical companies, as you heard last night in spades, and, sorry but to some extent, also the NCI seems to have this view. Now that you are able to do science on solid tumors we have lost our monopoly. It is clearly important for us to be convincing to the NCI about how good the state of the science is in leukemia, how unique it is, and how it serves as a model often for the science in other areas.

It is clear, for example, that there are going to be other things analogous to STI-571 coming down the pike, but these are going to be for disorders that are even less common than CML. They are going to be for AML cases with t(8;21) and maybe even only a subset of t(8;21) or something like that. The groups are the only resources in the world that can allow expeditious testing of such compounds. The groups must figure out how to position themselves so that they can respond competitively and almost instantaneously when these opportunities come along.

DR. LARSON: Questions or additions?

If not, we will go on to Marty's presentation.

TOP