I, too, want to thank all the speakers and the session chairs for an excellent session. As was said, I think when we set out to do this meeting, it wasn't easy to figure out who was working on gastric cancer. It certainly was an education for me and the other members of the planning committee.
I certainly learned a lot today. I have to warn you, this is the first time I have given this particular talk. What I would like to do is to summarize briefly what we have heard over the past day and a half.

To start off, in the present status session, Dr. Correa remarked on the geographic variation as well as the changes in gastric cancer incidence seen in the west, which clearly demonstrate an important environmental influence in this disease. Based on that, as well as data in animals and in humans, he has demonstrated, among others, a model of gastric carcinogenesis, which involves the progression from normal epithelium to metaplasia, dysplasia, to cancer. Unfortunately, unlike colon cancer, the molecular events associated with gastric carcinogenesis are not so well described. As we have heard, this is a very heterogeneous disease where, in fact, probably this is not the only pathway. Clearly, what is needed in future studies is for us to better understand the events that occur in this and the other pathways, not only because it will allow us to understand the biology, but almost certainly it will give us the next series of targets for novel therapies that will be specific, I hope, for gastric cancer.

Dr. Powell also very nicely summarized our understanding of the pathology and molecular markers. He pointed out, again, that there is significant histologic and biologic heterogeneity in gastric cancer, with far too many classification systems that really inadequately define the varieties of gastric cancer. However, we know now that molecular alterations such as E-cadherin, which is more common in diffuse-type cancers, trefoil factor one, which appears more common in intestinal type, MSI, more common in intestinal among others, may, in part, confer useful prognostic information, but also will likely improve our understanding of gastric cancer subtypes in biology, and I will come back to that in a subsequent review.

Dr. Karpeh gave an excellent presentation, summarizing the status of staging and surgical resection. He described the value of laparoscopy and/or peritoneal cytology.
What is not available from this data is the cost effectiveness of laparoscopy as well. We realize that, with the enhancement of PET imaging and other things, well, maybe that may some day even overtake laparoscopy. Clearly, for distal cancers, a subtotal gastrectomy is adequate, but we must always remind ourselves to be aware of the margin. He pointed out that interoperative margin assessment is important.
He also discussed, as did other speakers, the importance of extent of lymph node dissection. Based on the current data, D-1 is probably the recommended procedure, but there are far too many D-zero procedures going on. I think some approach to standardization is required. He also pointed out that splenectomy and pancreatectomy increase the morbidity and mortality of these procedures, and probably did confound, in part, those D-1, D-2 trials. I will be the third medical oncologist to mention the sentinel node technology. It seems like such an attractive notion, but I guess our surgical colleagues inform us that that is just not a reality.

Dr. Macdonald summarized some very important data. Obviously, in terms of adjuvant chemotherapy, at least 19 studies have been null and not born fruit. Clearly, 0116 clearly indicates a survival benefit for adjuvant chemoradiation which is now the standard. In the treatment of advance therapy, there is clearly no standard. As we did hear later in the day, and as we are all aware, the introduction of novel targeted agents is needed. In the meantime, I think we have to rely on multicenter phase III studies to define what we think is the best available therapy for advanced disease.
Future opportunities were reviewed by Dr. Macdonald including defining a follow-up study to the adjuvant study, and I will actually touch upon these other points later on.

I actually want to thank Dr. Moss, who gave an excellent presentation and also agreed to speak here only five days ago, because of some changes in our schedule. It was rather short notice. He gave an excellent presentation, and very excellent on h. pylori. Clearly, this is a major etiologic factor in distal gastric cancer. Ironically, it may in fact be protective for proximal lesions. As such, the decrease in h. pylori prevalence may explain why distal lesions are decreasing and proximal increasing in this country. H. pylori infection is ubiquitous in the world. Clearly, many -- most don't get cancer. As he described, the susceptibility to cancer among infected individuals is related to strain variation, topographical variation, such as the age you are infected, as well as genetically determined host factors just recently described, including IL1 germ line polymorphisms. The mechanisms of h. pylori are not entirely clear, but include direct effects, cytokines, growth factors, as well as, curiously, initial increase in apoptosis, as we were told, but maybe subsequent apoptotic resistance. Obviously, h. pylori doesn't directly affect the proposals of the intergroup per se, but it is clear that the elucidation of h. pylori pathogenesis could result in defining novel treatments, in part because of the example that was given, which is that h. pylori does stimulate IL-8. Later in the day we heard that IL-8 is linked strongly to angiogenesis. If multiple myeloma and inhibition of IL-6 proves to be therapeutic, well, maybe in gastric cancer, if we can get a novel target to inhibit IL-8, that, too, might be a reasonable approach.

Dr. Fenoglio-Peiser, I want to thank her, because she was the only remaining member of that cohort for that session after the first design of that agenda. I was glad that she stayed. As did many other speakers, she described the fact that gastric cancer arises from a variety of precursor lesions, again demonstrating the remarkable heterogeneity of these tumors. We realize that there are a lot of small studies looking at molecular features of gastric cancer, but all conducted around the world suffer from the fact of retrospective design plus the other things that we are all familiar with, including multivariate analysis and underpowered studies. What is needed and what she is obviously working on is utilizing tumor samples from the intergroup study to confirm the prognostic value of these markers but, no less importantly, to develop a new classification system to hopefully overtake our inadequate systems right now, to use both histologic as well as molecular data, as we really define the subtypes of gastric cancer.

Ultimately, when that data becomes available, we will need to validate that in another large prospective cohort or clinical trial. To do that, we obviously need to make tissue submissions as close as possible to be a mandatory component in all future gastric cancer studies. We need to integrate DNA and protenomic arrays in that. We need to, in addition, I think to further our understanding, to develop blood and fresh tissue collection as much as possible in these studies. If we do that, I think we will allow us a new classification of gastric cancer. It will help us to target therapies to the logical subtype of gastric cancer. It will also help us refine our eligibility criteria or, more likely, the stratification criteria in terms of what we do with GE junction versus proximal versus distal, as we better classify these tumors. Finally, with this resource, we can better examine future hypotheses as they emerge.

We heard from the delegation from Texas, both in person and by short wave, about the fascinating new targets in cancer, some of which occasionally involve gastric cancer, all of which were very interesting. These targets, as well as their many related pathways clearly are going to be the next generation of studies that we do.
It included such things as the inhibition of angiogenesis, with Dr. Ellis describing the fact that VEGF and PDECGF correlate with vessel count and stage, in intestinal, although not diffuse, cancers. As well, anti-VEGF receptor, as well as soluble FLT-1 did inhibit growth and increase survival in mouse models. Similarly, we heard about signal transduction, both in terms of growth factor receptor inhibition and affecting downstream signalling.

As well, we heard about apoptosis, in terms of the fact that inflammation, which appears to be important in gastric cancer pathogenesis also may drive apoptotic resistance. As well, it looks like death receptors as well as pro-apoptotic BCL-2 family members may be inactivated in gastric cancer models. In fact, maybe the NSAIDS will have a role stimulating apoptosis. Nonetheless, we realize that mouse models don't necessarily correlate with human models. As well, we will have to figure out how do we define what is the best of all of these targets for our future studies, or we just do them all as the targets become available. How do we integrate these biologics into our trials? Do we use them alone? Do we combine them with radiation? Do we combine them with chemotherapy or all of the above?
What population do we look at? We heard that angiogenesis is probably more important in intestinal versus diffuse. Also, do we look at it in adjuvant versus advanced studies? Clearly, again, we need to have the resources to look at these things. Most important, we have to have the availability of these agents to our cooperative groups from the manufacturers in order to do these elegant studies.
As well, we heard about how you develop a clinical trial using these new agents.

Clearly, the goal would be to develop reliable surrogate markers to optimize the dose using optimal biologic response versus MTD. As well, explore the mechanisms of action and resistance of these biologic agents within these trials. However, the problems as we heard very nicely summarized was, we don't have validated surrogate tissues available. We need to link the changes in the targets with a clear clinical response which hasn't been done yet. We need to define the optimal timing of assessment as well as biopsies, to know what these drugs are doing. As well, we heard at length from several of the speakers about the great heterogeneity of the results in terms of within tumor, within biopsy, within patients, and that has to be addressed.

We also heard about novel imaging techniques which, in part, will allow for the rapid assessment of drug efficacy sooner than we would otherwise learn from a CAT scan, perhaps. More important, I think more relevant down the road would be non-invasive alternatives to assess receptor expression, enzyme inhibition, as well as other downstream events from these biologics. This is clearly what we need if we are going to study these drugs biologically in large populations. There are a lot of challenges in terms of defining these outcomes because it clearly is a different paradigm.

Dr. Schilsky very nice, I think, offered one potential paradigm in which we look at resectable cancer, we do a biopsy, we give them a new agent. At the time of resection, we assess them for response. We obtain more adequate tissue following the treatment. As well, we decide after that whether we give them adjuvant therapy plus or minus the new agent. This is, of course, one model to look at the intergroup setting. Several others could exist. The question he asked yesterday that we probably didn't have time to answer was that the study of these novel agents is going on right now without really doing some of these biological assessments.
The question, of course is, is it worth going through all this trouble that was described. I think there is no clear yes or no, but I think in part, yes, because I think it would be helpful to understand the biology and resistance related to these drugs.
I think it would allow us to screen some of these agents more rapidly, and certainly important to industry in that regard. Also, by doing it, I think it would allow us to move these drugs more quickly into the adjuvant setting, rather than waiting for large metastatic studies before we do something like that.

Last night we heard a very nice discussion about end points in clinical trials by Rick Pazdur. As he pointed out, survival remains the gold standard, but other surrogate end points, such as time to tumor progression, response rate, quality of life, must be "reasonably likely to predict a clinical benefit," which even he couldn't entirely define for us, and I guess we will wait to hear further what that absolutely means. Surrogate end points are often used, as you know, in an accelerate approval process with the promise of phase IV studies that must follow. The point that he made which was important is that, when you do this, you can't do a randomized study for the indication you have subsequently approved. So, it is a complex process when accelerated approval occurs.

Other issues that he pointed out included the issue of creative oncology. That is, sometimes designing studies, not because of the important scientific question, but because it is the safest and quickest and cheapest approach to get your drug approved. Finally, a point that was made during the discussion was, we have to work on an effective partnership with industry, since most of these important agents are coming from industry.

We heard this morning about how do we best achieve loco-regional control. Clearly, chemoradiation based on intergroup 0116 is the new standard. As occurred, about 15 years or so ago, with the first positive study with rectal cancer chemoradiation, there clearly is going to be a steep learning curve of the application for this to the general community, and it will certainly be, to a large part, under the guidance of Drs. Gunderson, Tepper and Smalley, I think, to educate all of us as to how to do this safely, both within the community practice, and in our subsequent trials.
Clearly, future studies following 0116 are required which, as Dr. Gunderson pointed out, could be including other multi-agent chemotherapies or neoadjuvant therapy.
In terms of surgery, again, we heard that D-2 is not clearly superior to D-1 dissections.

There may be a role for a computer-guided lymph node dissection, as demonstrated by the Maruyama index. I think the caveat, of course, is that the data that we have gotten in the past from Asia has not always been applicable to our patient population. Clearly, the approach now to look at it prospectively, and validate it is obviously the logical approach to look at this rigorously. At the same time, given the issues with lymph node dissection and lymph node assessment, I think educate is maybe a little patronizing, but we need to at least work with the surgeons and pathologists in this country in the general community to try to get standards of surgery as well as pathologic assessments of these gastrectomies.

In terms of intraperitoneal therapy, clearly the natural history of gastric cancer naturally lends itself to intraperitoneal therapy. Although initial of interest, my take -- particularly from the editorials recently in the JCO -- is that in ovarian cancer, the enthusiasm for intraperitoneal chemotherapy seems to be waning. Despite the results of individual centers of excellence, it is not clear that intraperitoneal chemotherapy is a potentially viable alternative in large cooperative group studies, but I guess we will wait to see more data.

Lastly, we heard about why is gastric cancer so resistant and how do we overcome this resistance. Clearly, there are multiple events related to resistance, and we know only a few of them. We heard very nicely about several of them this morning, including cell cycle regulation, in part related to apoptotic resistance, and potentially by inhibiting CDKs in conjunction with either chemotherapy or radiation.
We can overcome that, as well as some of the novel agents, by inducing growth arrest. It may also overcome that resistance. We heard about molecular profiling of tumors, something we have heard a lot about in the context of colon cancer, in that we could look at the changes in the tumor to protect 5 FU resistance, cisplatin resistance, and now more recently taxol resistance as well. I think more interesting is to look at the germ line polymorphisms. That is, what inherited mutations do we have that make us more or less resistant and potentially more subsequent to toxicity to either 5 FU or platinum or probably many other drugs.

Finally, we heard about the fact that p53 is a very complex player within cell cycle mechanisms, apparently regulating apoptosis, cell cycle regulation, DNA repair.
In some circumstances, it may increase the sensitivity to chemotherapy. In others, it may decrease sensitivity to therapy.

Obviously, with all of these things, we have to first better know what all the methods or mechanisms of resistance are. In addition, most relevant to us, we have to assess these factors in large, prospective clinical trials.

There was a lot covered in the last 36 hours or so, I think of great interest, and hopefully of some value.

What do I think the challenges are from all of this? Clearly, we don't have any studies right now in gastric cancer in the intergroup. So, we have to design more intergroup studies, either in the adjuvant or the advanced setting. Some of that discussion occurred this morning and, as well, will occur later today. We have to clearly integrate tumor and blood banking into our trials. We have to validate the biological end points as well as the molecular markers within our large cohorts. We have to develop methodology to examine these new agents, as was discussed as length yesterday afternoon. We have to learn more about how to modify the multiple mechanisms of resistance that we know about now, and almost certainly will know more next year. We have to standardize radiation and educate the community, as well as us all, so we can design safe and effective future trials. We have to improve surgery and pathology in the community throughout this country, so that we can get past this concern about D-zero resections and heterogeneity within surgical and pathologic assessments.

Finally, I think a point that was made several times is that gastric cancer is quite heterogeneous. Hopefully over the next five years we will learn how to refine our classification of this disease. We will develop trials that focus on these subclasses. As well, we will target our next generation of therapies to these individual subtypes of gastric cancer, to achieve the greatest clinical benefit for our patients. Thanks again.

TOP