DR. BORDEN: George, while people are coming up: with the exon 11 mutation, we have completely stopped the car. Yet, we do not have complete responses. Do you or Allan have speculations as to why the car has not come to a complete, regressed stop?

DR. DEMETRI: Well, we have some thoughts about that, about what is keeping a subset of these cells alive. You know, we've thrown them into a cell cycle arrest. We don't really know.

I would actually encourage Jonathan to speculate. I am real good about speculation and Chris keeps me honest about that, but I try to be visionary without being, again, delusional.

I would like Jonathan to comment on that or Mike or anyone else. I have my own thoughts but the molecular biologists are more skilled at commenting on that. What do you think, Jonathan?

DR. FLETCHER: Unfortunately, I don't have an answer, but it was pretty much going to be my question for the group as well. I think it is fascinating that in some of the patients you are seeing that 60 percent of the tumor melts away fairly quickly and then you are left with this nidus. In some cases, it is dead stuff; but in other cases, you can say you have brought the cells to a standstill.

In culture, we do see that with some of the models, that you can basically select for a group of cells in which the pathways are affected by the inhibitor. They are no longer the normal GIST signaling profiles, but the cells are nonetheless knocked out.
So, I think we will have some clues to that from the in vitro models and the mouse models. It certainly brings up the point as to why there aren't more biopsies being done or considered in patients who have residual disease, not just at the time of progression, but at the time that they have been stable for a number of months to get at this issue more routinely, as to what is still on and what needs to be taken care of.

DR. DEMETRI: I think it also brings up the fact that we are all going in the same direction, that likely combinations are going to be necessary, that this drug is not so much a killing drug as removing an anti-apoptotic block. You remove that and a large number of the cells can go to apoptosis, but it is not encouraging the others to go over that precipice. So, we would probably need other things to help that along.

DR. BAKER: I would like to extend that same question in a slightly different way. We have known about absence of complete remission for some time, and we all hold ourselves out at least in this room, or outside of the room, as experts in sarcomas. What would be typically done would be to surgically extricate the disease that remains. Do you think that is commonly going on now and, if not, why not?

DR. DEMETRI: We are commonly doing it at our site. What I worry about is how common it is out in the community. I think there is still a lot more education that needs to go on. I remember when we were at the NCI talking to them about designing the phase III study, one of the first things Bob Benjamin said was, I hope you are not going to prevent us from taking out residual disease after it shrinks. We want to take it out. In our study, as well as the European study, they were specifically written to allow that to happen. I think there are a lot of people who still are worried that if they take out the tumor, somehow the patient will not be eligible to continue on the drug.

DR. BAKER: That is what I am trying to get at. Now we have several ongoing trials. How often is this occurring?

DR. DEMETRI: That people are having residual disease resected?

DR. VERWEIJ: For the EORTC studies, it is a minority of patients. The number of patients where we have succeeded to take everything out, n equals one -- one patient only. For all the rest of the patients, there was more tumor left than we thought there was by doing CT scans or MRIs or whatever.

It is always very small deposits. I am not a surgeon but it seems to me that it is very technically difficult to take away all the small nodules that are anywhere in the abdomen.

DR. VAN OOSTEROM: There are other problems, of course. The European problem is that in Europe imantinib is not yet registered. So, we can only treat patients in studies. The second is that the best chance, of course, is in a neoadjuvant situation with, for example, local extensive tumor and no liver metastasis.

Most patients have liver metastasis, so that it is difficult to remove all the remaining tumors. If you have only primary tumors, then most patients don't come to centers at the beginning. They come for gastric bleedings or bleedings in the colon or somewhere, and local surgeons see it. Only after the operation they know it was a GIST, but up to that moment we are not involved. I think at this moment, since no one at least in Europe is educated for sending bleeding tumors directly to a center, and the fact that the drug is not recompensated at this moment by the Social Security, it is a difficult problem.

DR. O'SULLIVAN: We had several patients that had localized disease, had it partially resected on the outside and have been treated with neoadjuvant therapy, either because they didn't want another operation right away or to assess response to the Gleevec drug. In about five cases, we have been able to selectively resect those patients who didn't have multifocal disease and have put them back on drug.
Others who had extensive multifocal disease, because of severe symptoms, we have debulked and then had one patient who has remained free of disease after that large debulking, despite having peritoneal studding on the surface. Another patient who debulked is now going on the next Gleevec in that trial, but because of demonstrated resistance to Gleevec, is still progressing on therapy; I think the benefit of therapy in these patients needs to be further defined.

DR. DEMETRI: I think that is an important thing. We took a unique group of patients into these phase III studies -- frankly the phase II and the phase III globally -- and we were dealing with almost a pent-up demand of people with massive metastasis. We wrote the studies to let them go on if they just had one foot and four toes in the grave, that they could still get onto these studies. So, the number of people who went on with an enormous tumor burden, even if Gleevec got rid of 90 percent of it, it would still not make sense to go back in and resect because they are still left with 30 liver mets and all sorts of other disease.

It would be very different to take the new study that Burt Eisenberg is now chairing, this neoadjuvant approach where like Alan was saying the aim is to now shrink disease. That can also be recurrent disease.

DR. EISENBERG: That explained it very well. The indication for the study was born out of an interest of what is the surgical role to play in this specific target and with good response rates. Clearly, patients who have bulk disease will tend to have residual disease. Whether it is alive or dead, we don't know for sure.

We also saw mixed responses -- patients who had good response in one place and then progressed in another place. The question is, should we take out the spot that is progressing and leave the spot that is not progressing? These questions are outstanding and the clinical trial right now is not only to do a biological study, because there are a lot of things that we need to know, as George said, about molecular mechanism of response and resistance with good tumor, but also to address surgical. manipulations that come into play, and, specifically, debulking. I mean, the ovarian cancer model here may be similar. Can we get rid of enough disease to make minimal disease for drug?

DR. BORDEN: Allan, several months ago, when I asked you, you told me the index patient had been repeatedly biopsied and continued to have viable tumor. Can you speak to that? Patient zero had been repeatedly biopsied, and I wondered if there was just an update in terms of that.

DR. DEMETRI: Yes, the patient has required an increase in the dose, but is still maintaining a response now with an increase in the dose. So, there is some interesting biology seemingly going on within that patient. She never lost every last tumor cell. I think that is the other important thing that we learned. We got a little bit full of ourselves as medical oncologists feeling that, yes, this is great, we can rid of surgery altogether. We actually got taken down a few pegs by the fact that there are no really convincing frequency of complete responses.

I am sure there are a few cases here and there, but I think generally this drug is still, by itself, going to be part of a multidisciplinary management plan, which is what Larry was alluding to -- that resecting residual disease before it can have a chance to generate the next generation mutations and activate other pathways, to allow Jonathan and others to figure out what they are, is probably the best management for the patient right now.

DR. FLETCHER: George, this must be one of your moments of honesty, which we all enjoyed. It is true. Actually, I would say 90 percent of the patients who have been biopsied, having responded, still have clearly viable tumor.

In fact, actually seeing necrotic tumor is unbelievably infrequent. I would say less than 10 percent, possibly even less than 5 percent. All it is doing is that the proliferative rate seems to have dropped dramatically, but morphologically, they are as alive as alive can be. So, we need to bear that in mind.

specific question for you. You can call it speculation or delusion, take your pick. You said that we might learn from the people who become kit unresponsive -- the kit resistant group -- that we may be able to learn things from that that we can extrapolate to other tumor types. What concerns me is that that is in danger of becoming wild speculation again. That is the whole game of, hey, we have got a drug that does something, so let's chuck it at something else. The wonderful thing about the kit story is that it was mechanistically so perfect. That is why everybody loves to talk to it -- because it has been about the first rational piece of medical oncology in 25 years, because it is really -- we all know that is the case. It is a wondrous story. The danger is we mustn't then go off in crazy tangents. They need to be thought through rather than just being an excuse to chuck a drug at another tumor type.

DR. DEMETRI: I agree. I think the point is, though, that if we see that a MAP kinase pathway or an AKT pathway is activated after Gleevec fails, and then it perhaps offers a nice proof-of-concept about inhibition of that pathway is then useful.
It may only have one of those 10 or 15 percent contribution rates. That is why I was going through all that sensitivity testing of the different things.

It might be very hard to show in a complex disease like breast cancer, but a disease like resistant GIST, where you have got the markers and you can tie them up to the hard endpoints, might still be a good way of proving the point for other pathways.

DR. FLETCHER: George, say that you discover that, of the Gleevec resistant patients, they have all done something funny to their MAP kinases. Then you come running to the pathologists saying, I want you to screen all the synovial sarcomas to see if they have got MAP kinase switched on, because then I can give them Gleevec?

In other words, how would that actually happen in real life, that if you find the second line biochemical event, the downstream event that has made them Gleevec insensitive, how is that going to be extrapolated to other tumor types?

DR. DEMETRI: I think it may well be that way, just the same way that the Gleevec story played into the second-generation trials. For those of you who don't know, the second-generation trials of Gleevec to look for this exploratory activity took a different approach. So, let's forget what the histology is. Let's just see if they have an activated target that Gleevec could shut down --in other words, activated abl, activated PDGFR, activated kit. If you have got one of those, you can get into a study.

That was the theory anyway. The problem is, as Chris is alluding to, finding the activation dose is a much bigger screening task than was set up in an infrastructure globally, and certainly in Jonathan's lab it was more than we could do. So, the screen did not become functional activation. The screen simply became, is there a protein there that reacts with the antibody? As we have just found out, just because you have got a hood ornament doesn't mean that you are really a Jaguar. You could put a hood ornament for that car on a Ford, and it is still a Ford.

The hood ornament approach with the immunohistochemistry may be misleading us a little bit in that study, but that is what we are trying to figure out. If we look at AKT, that might be something we would want to screen for.

DR. BRENNAN: Just one comment. We do have another extraordinary opportunity with such a cytostatic drug, because a cytostatic drug with limited side effects is the perfect drug for an adjuvant trial. It is the perfect question to ask biologically and clinically in a prospective randomized trial already supported now by the NCI. Placebo control, because you will learn a great deal about those that either never recur or those that do recur that can then go on to the drug.

Remember, essentially, all people with high-risk growth, the 10-centimeter central necrosis will ultimately recur and die.

I just remind everybody that the opportunity in this randomized trial, using such a drug as an adjuvant has enormous both meaningful clinical outcomes and meaningful biological questions. That would be something, I think, for this group to support.

DR. MELTZER: I just wonder, with regard to residual viable tumor on therapy, whether anyone yet can say anything about kit expression, mutation status, and phosphorylation status.

DR. JONATHAN FLETCHER: You know, what I can say is that, as I alluded to in an earlier question, we really haven't done much with residual tumor that is just maintaining its own but not progressing, but it is stable. With Mike Heinrich, we have been evaluating progressing disease, which is generally not just all of the GIST in the patient, but some nidus that breaks loose and begins to grow while the rest is still under control.

Our sense is that it will be quite possible to assign resistance categories, if you will, into bins, where some of this may have to do with a target no longer being effectively suppressed by the drug. In some cases, it may be that the target is different but, by a combination of genomic and protein-level analysis, that there will be at least three or four general bins that you will assign the resistant tumor to, that will either involve a modulation where the target is still active; kit is still active but for one reason or another -- invariably, there will be several such reasons -- it is no longer being effectively shut down by the drug; in other cases, where that target may no longer be relevant, at least to that aspect of the patient's GIST.

So, those studies are very preliminary. I don't know Mike if you want to comment more, but I think they will be quite revealing.

DR. HEINRICH: Nothing to add really. I totally agree with -- maybe Chris Fletcher has a better idea, but I have never really seen any of the well-responding lesions that have been biopsied or resected. Certainly not aware of any molecular studies on them.

DR. JONATHAN FLETCHER: It wasn't the stable disease. It was the residual disease.

DR. SORENSON: I was asking about residual viable disease.

DR. CHRISTOPHER FLETCHER: Responders with viable tumors.

DR. JONATHAN FLETCHER: I took it to mean that a patient who had been treated for some time and then still has some disease.

The analysis that we have done -- mostly by the good graces of the Fox Chase group, who have provided us with some very nice core biopsies that were about a week or two into therapy -- in patients who are just seeing their initial imatinib, we have a very good sense as to, biochemically, what happens to those patients within the space of a week, in terms of shutting down kit and shutting down the pathways.
Extending that to the residual disease, which is still present six months later, you have had this gradual response and now you are left with something. We haven't done anything with those such tumors and I think that would be tremendously useful.

DR. DEMETRI: I think we have only recently started resecting them, Jonathan. Really, only in the last two months have we been sending you those stable ones, as opposed to selecting for the progressors.

DR. JONATHAN FLETCHER: Just to echo some of the comments that were made by George, I do see this as a tremendous opportunity to develop paradigms that will be useful generally for targeted therapy.

The idea of a specific molecule may or may not be important, but the general idea of resistance, how does that occur, and what do you need to do either upfront or as a back-up plan to have contingencies and to take care of that resistance, does it involve several ways of going after the primary target, or are there ways of eluding the primary target such that you absolutely have to have downstream targets or ultimate targets in mind?

I think that we need information from the residual tumor that is left after six months to tell us what we need to do better to knock that thing on its back, but we also need studies, which are clearly ongoing, on the progressive lesions to tell us what we need to do as combination therapy, either upfront or as back-up plans to deal with the inevitable resistance that will emerge in these patients.

DR. DE MATTEO: Just two comments, one to just add to what Dr. Brennan said. Part of the rationale for the adjuvant trial is that is also that it may be more than cytostatic in the presence of minimal residual disease. It was difficult to determine what the dose and length of treatment would be for the adjuvant trials and we, with a lot of input, agreed to 400 milligrams a day for a year. Nobody knows what the right answer to that should be, though.

My second point is that I have noticed that a lot of the clinicians will give up or will be much more resistant to recommend some of the patients with liver metastases for surgery, even patients with limited liver metastases.

You know, we are able to remove tumor in some of these patients, to either debulk them upfront or to remove stable disease. I would keep that in mind. It seems overall that people are more willing to refer somebody with a few peritoneal modules than a few liver tumors.

DR. BENJAMIN: Robert Benjamin, Houston. I am confused by the discussion on the pathology on the treated patients. On the tumors that have come out, are they characterized first based on clinical response and then given to both Chris and Jonathan, or only selected ones to Jonathan, so that Chris is seeing the global answer to the clinical question and Jonathan is seeing a highly selected subset?
The corollary question; when the tumor comes out, Chris told us earlier today that he doesn't know how to quantitatively assess the response based on tumor necrosis, because he doesn't know how to determine spontaneous from induced necrosis. Is there an attempt made to quantify the necrosis within the tumor, so that some sort of prospective data can be ascertained?

DR. CHRISTOPHER FLETCHER: I can't answer the first question you had about whether Jonathan and I are getting different things, because you know how dysfunctional we all are and we only speak to each other at meetings like this. Of course, I am being facetious.

We can't tell one sort of necrosis from another. You know that, and we have known that for years. I think it is interesting that actually most GISTs, when they are excited -- usually if they haven't been treated they tend to go cystic in the middle, but they are not necessarily spontaneously necrotic tumors. They are either very viable with cystic holes, or they are just one big piece of meat. The occasional ones that we have seen with response, I have guessed it is treatment response because you get this sort of great acre of dead looking tumor that is unusual in appearance for a GIST generally.

No, we are not doing quantitative things because I don't know how to quantitate it and we don't sample the whole thing, bottom line.

DR. BENJAMIN: So, the answer is, what you said -- when we look at the tumor, it looks just like untreated tumor. You are talking about the part that you have already selected out that you know hasn't responded, rather than the global tumor, which is largely necrotic.

DR. CHRISTOPHER FLETCHER: No, Bob, it is not that. It is that George can send all these poor patients to get biopsied constantly. They only have to walk through the door of the Farber to get a needle stuck in them. So, we get all these protocol biopsies on them being followed up. The response or no response to treatment -- the tumors usually look completely unchanged. I cannot tell. If George gives me a biopsy after three months, I can't tell, by looking at it, whether it has been a Gleevec responder or is Gleevec resistant.

DR. BENJAMIN: So, you are not getting surgical specimens on these tumors?

DR. CHRIS FLETCHER: Only on a small proportion.

DR. SAXMAN: I will give Dr. van Oosterom the last word, and then we will continue this discussion during the breakout session.

DR. VAN OOSTEROM: I think the message that comes out -- and George agrees -- is that the detection that Gleevec works in kit positive, mutated GIST tells you only that we have a drug, just as 40 or 50 years ago with methotrexate, 5-FU etc. You have to realize that, for Digoxin, we still don't know exactly how it works.

For this drug, we have the impression that we now how it works against a special type of cancer in part of the patients.

I am a little bit older than George. The second group of tumors that I always treat are testicular cancer, in which we, in the beginning, also thought under the push of Larry Einhorn that we could do on our own. We now know that we needed surgeons.
Here in sarcomas, we need the surgeons also in the GIST as it were. I think we should, every time that we see these patients discuss with them and see if it is the appropriate time to operate them, even if they have massive liver metastasis and they are responders.

The ultimate time is at this moment not yet known. For me -- and I stated it earlier in my presentation and George did as well -- I think it is only the beginning of our understanding how it works.

I hope that, with all the investigations that we do now, we will discover also in GIST other pathways downstream.

Maybe we have that drug already. I think in the future we will do just as we did in classical chemotherapy. First, we had one drug, then we had a second drug which was addressing another target.

We will end up by giving two or three drugs in GIST patients in the future. Still, we will need to discuss the patients with the surgeons to take the remaining tumors out, to confirm sometimes complete responses or partial responses. I think that will happen in GIST and I hope also in other sarcoma types as time goes by, that we will see the targets. We will also do multiple targeted therapies. That is my hope for the future.

DR. SAXMAN: Thank you.

TOP