SLIDES & TRANSCRIPTS
Friday, December 13, 2002

New Strategies for Prognostication and Treatment of Advanced RCC

Arie Belldegrun, M.D.

Thank you, Ian.

On this slide we tried to guess what Nick would put as a summary on one slide, everything that he is going to talk about in his talk. And then, having put everything that we can think of that has been tested or is being tested now in 2002 in new therapeutic approaches for kidney cancer, we asked the next question,

TOP

is anything promising, even if we stretched the indication? Is there anything promising?

And that's what is left, and even this is really stretching it quite significantly. Having said all of that, I think for now, we will stay with our treatment regimen that we started over 10 years ago, with the high dose interleukin-2 and IL-2 therapy, with some patients receiving interferon. And I will show you the data.

TOP

I don't see in the next five years or even longer period of time, to change dramatically this regimen, although clearly the allogeneic bone marrow transplant, as you have seen, is really encouraging, but it's not ready for prime time.

TOP

We started in 1989. At that time we didn't know if nephrectomy was the way to go or not, and we had a lot of debate, but we elected at that time to be consistent and to start with radical aggressive resection of all tumors, having had the opportunity to be at the NCI, and leave directly from here straight to UCLA.

So we started in 1989 with resection, followed by immunotherapy. So this is not a randomized, not prospective. It is highly biased, but these are the patients that we have accumulated. So after 10 years we didn't have at that time the database. We just went and reviewed the charts.

And we have come with our own data that patients who had nephrectomy and IL-2, had an advantage of about 11.9 months survival over patients who had nephrectomy, but did not receive any immunotherapy, again, with all the caveats I mentioned. So this was the original approach that showed that we were giving our patients about a year longer of life.

TOP

Now, having had the database now, we took the SWOG data that you just saw, and we added our own data with nephrectomy and IL-2, where all the patients were similar to the nephrectomy and interferon base as close as we could get, and you can see the difference that we have obtained. And again, I doubt that in the next at least five years, we will see any data maturing that would clarify whether this data is real, or is biased and selected.

TOP

Now, this is history. What we are looking at now is subgroup analyses, and how can we better select patients for therapy. And more importantly, can we select a group in whom we know that we tried the most aggressive approach, and then it subsequently failed. So maybe these patients should not be included in clinical trials, or more importantly, for any therapy that patients are coming for.

For example, this is a paper in press, but I was very surprised to see our own data, that if you have lung only metastases, or bone only metastases, your survival is the same. However, if you have multiple metastases, multi-organ metastases, this is your survival. Again, we need some more confirmation, but this is an extensive, exhaustive study that we have done, and it is against where are going (? What we are doing), and telling our own patients.

TOP

As far as lymph nodes, what do you do with patients with bulky lymph nodes? Here is our survival of patients, N0, M0 the work that Alan Pentack has put together, and there are two papers in press right now. If you look at patients who have either N1 disease with no metastases (N1M0), or N0 with metastases (N0M1), you can see survival is similar but with a completely different clinical pattern. That is, either nodes OR metastases. But if you have a combination of nodes AND metastases (N1M1), I believe it's a different breed of patient that have a worse prognosis than either of the prior groups.

TOP

So the survival of patients in whom we performed extensive lymph node dissection versus no lymph node dissection is shown here. This is highly statistically significant.

And again, these are selected patients, and it is not prospective, and not randomized.

TOP

So this is our take from lymph node dissection for our patients. If they have grossly negative nodes, a lymph node dissection provides limited staging value and no therapeutic value. However, if the nodes are grossly positive, we think that lymphadenectomy improves survival, with some added morbidity, and probably improves response rates. You don't take patients to immunotherapy, and do nephrectomy and leave the lymph nodes, something that we used to do years ago.

TOP

I'll move quickly to the staging and prognostication.

TOP

We have now the UCLA kidney cancer database. It is 1989-2001. We expanded it to 1,500 patients, 1,000 nephrectomies, 622 patients were placed in studies, and you can see high dose interleukin-2, 241 patients. Each patient has 263 variables. So there is a huge amount of information, and again, this is the kingdom of the fellows. They are coming each day with new ideas and new papers. And some of what I'm presenting here are essentially their ideas.

TOP

Based on that, we reached the conclusion after exhaustive analysis that the stage and grade and the ECOG in combination are more predictive than each of these alone.

TOP

So it's one 1, 2, and 3. So you can say you are a UCLA Integrated Staging System 1 (UISS1) and your overall survival is 94%. If you are 3, your overall survival is 39 percent. And more importantly if you are 5, even with the nephrectomy, immunotherapy, anything we have done, and survival is 0%. No patients survived.

TOP

So these patients shouldn't go in any clinical trial, and you can see this data.

TOP

In this issue of the Journal of Clinical Oncology, we have simplified and changed somewhat the UISS. And you can see exactly what Victor Reuter said, that this is a work in progress. It's not the one classification or the other. We keep changing our own classification as we go along and understand the disease better.

So in this issue of the JCO we have shown that not all localized diseases are the same, and not all metastatic diseases are the same. You have low, intermediate, and high risk patients, and therefore, any clinical trials in the future should take consider these patient groupings; the low metastatic patients probably will get the better responses as opposed to the sum of all the three groupings together. So you are comparing essentially apples and oranges in all the studies up to now.

TOP

We then went to Nijmegen in the Netherlands and looked at their database based on the UISS. Their data are not published yet. And you can see for localized disease, almost the same data.

TOP

MD Anderson has looked at their data, low, intermediate, and high.

TOP

So essentially, it's exactly the same in localized disease. There are no differences between the UCLA and any other centers that we have looked at, and it's about 1,700 patients.

TOP

If we look at the intermediate group of metastatics, there is no difference. So the main bulky group of patients, there is no difference in the results (between Nijmegen, UCLA, and Europe).

TOP

However, if we looked at the group at the Netherlands versus UCLA, you can see the patients in the UCLA study did somewhat better. I don't know why and what, but these are ways to show you that the ECOG and the SWOG data across the Atlantic are probably comparing again apples and oranges.

TOP

Can we predict response to immunotherapy? Probably we can predict response to immunotherapy as well.

TOP

TOP

In data that is submitted for publication predicting survival, you can see here on the work that has been done by Brad Leibovich -- he is now back at Mayo -- looking at nephrectomy and immunotherapy. Who can respond to immunotherapy? Who should we place on it, and who should we not? There is a low risk group, intermediate, and high-risk group patients that probably shouldn't be placed on any immunotherapy.

TOP

TOP

So with that, it's very clear that we need to change our data.

And add these molecular markers, which will improve the staging system that we are showing now.

TOP

The way we have done it is using tissue arrays.

Essentially, looking at the protein level.

TOP

TOP

And we have looked at a whole host of molecular markers. We are showing some of them here.

One of the most exciting tumor markers that we have is CA9/G250.

TOP

It is expressed in the majority, in 95 percent of renal cell carcinomas, clear cell carcinoma.

TOP

And when we looked at metastatic patients, here is the survival of patients who were CA9 positive, and here are the patients who lost one gene.

Look at the survival difference of one gene, whether one gene can make a difference as a prognostic marker. This is what Matt Bui is doing here. Survival was 16 months versus 8 months.

TOP

And this is seen across the high tumor grades, if we stratify to tumor grade, stage, and ECOG, and absence of nodal disease.

TOP

So CA9 is an independent predictor of survival, and loss of CA9 implies worse prognosis. And this paper is in publication in Clinical Cancer Research.

TOP

So if you take now the UISS4, and you look at what's happening with this UISS4, if it is CA9 positive, or CA9 negative, survival is completely different. So that's the first step in order to incorporate molecular markers with the new classification in prognostication markers.

TOP

And just as a closure, we looked at our patients who responded to immunotherapy. Nobody has this huge number of patients, but complete responders -- these are patients in whom all tumors completely disappeared, and are long-term survivors. All of these patients in our group were CA9 positive. We had no complete responder in any patient that had lost the CA9. And now, should you exclude these patients or not? Obviously, we are waiting for more data, and from other centers.

TOP

So in summary, for now IL-2 remains the mainstay of treatment, at least in our group.

And with aggressive surgery, we feel that this is the best therapy. Node dissection improves survival, if the nodes are grossly involved.

Prognostication, we need to focus on that, because I think that that's where the progression of the understanding of the disease, and improved treatment will come from. As well, molecular classification is a focus, to really find the prognostic factors.

TOP

Finally, the most important slide, this is the work that really has been done by all our five fellows here with Dr. Pantuck already moving to our faculty, and a lot of the residents, research fellows, and the other collaborators.
Thank you.

TOP