DR. WANG: We have about ten minutes for discussion, and then we can move on to the next session.

PARTICIPANT: Tony, a technical question on the PSA response side of the equation. We know that PSA after initiation of hormonal therapy has a rapid drop as PSA is down regulated, then a slower drop. How do you actually calculate the doubling time on that side of the equation?

DR. D'AMICO: I didn't have time to put all the details in. We require for that very reason that each PSA measurement, both in the pre and post hormonal setting be separated by three months, and that the absolute value of the difference, particularly for the rising phase, is to be at least 0.2. So we have a nine month sampling interval, I guess is the answer, at a minimum, and the median is actually closer to a year, getting three points in the post hormonal therapy setting.

You're right, you will see this. But there is one thing that I don't think is well appreciated. It is not that the initial studies were right or wrong, but if you have somebody who is falling fairly slowly, meaning they are probably going to get down to one and then start to rise, there is a small contribution even in that falling part of the curve that makes it less steep, if you will. So by capturing information over the course of about a year, nine months to a year afterwards, you can get that.

There is one other question that someone might ask, and let me address it now. As the value of the doubling time and the pre hormonal therapy piece goes up, 18 months, 24 months, its accuracy goes down. The most accurate values of the doubling time, the ones with the smallest confidence interval or standard deviations, are the smallest values. They are the ones you usually have more points in general, because the patient is more concerned, because the numbers are going up quickly, in order to calculate. But once you get about 12 months, a 12 month doubling time and a 24 month doubling time are almost indistinguishable mathematically.

PARTICIPANT: This is a question for Dr. D'Amico. Looking at your patients who received radiation therapy, was that broken down into patients who received brachytherapy versus external beam therapy? Did you look at that at all, in terms of differences? Also, were there any other clinical factors in terms of grade of tumor, Gleason grade that were important at all?

DR. D'AMICO: The question is, in terms of radiation, whether people had interstitial therapy or brachytherapy. All the data I showed you was external beam radiation therapy. We have not yet acquired a brachytherapy database with long enough follow-up and large enough sample size to do these analyses based on death in the PSA area.

In terms of other prognostic factors, that question was asked at the CapCure meeting. We did go back and we did look. It is interesting, that once you know the doubling time, the prostatectomy Gleason score as well as seminal vesicle invasion or capsular penetration are not prognostic on a multi-variable analysis.

Dr. Partin, who couldn't be here, actually has similar data. I have seen the update of the Pound study, and they show the same thing.

PARTICIPANT: Question for Dr. D'Amico. I appreciate all the work that you are doing with PSA doubling time less than three months, and trying to do something for these patients, who you have shown will invariably fail.

I think a larger group of patients is those group of patients higher up on the curve, between three months and say 12 months. As you just pointed out, anything above 12 months is indistinguishable. That group of patients in between there is patients that could really go either way.

Maybe a nice surrogate end point for you to look at for possibly inclusion in FDA trials is a PSA doubling time say greater than 12 months, and using medication such as proscar or COX-2 inhibitors that will allow us to take a patient's PSA doubling time, just as you showed with their hormonal therapy, from something between six and nine months, and bring it up to 12 months, and turn this into a chronic biochemical disease or occurrence, something that won't cause clinical disease.

DR. D'AMICO: There are two points to be made. In the short time, I couldn't show all of that. But we have set up some guidelines for the doubling times that are long. The clinical trial that we would think about doing is nothing in one arm versus a novel therapy that has low toxicity, Celebrex or something of that sort.

For the ones who are less than three months, that is where the hormonal therapy can be justified up front. Three months is very strict, and it is based on the fact that the median follow-up in the study is seven years. As the median follow-up becomes ten or 12 years, six months or eight months may actually be more reasonable.

It is just that death is a function of time, and so right now, we are seeing three months as a surrogate. It may be in another five years maybe six months as a surrogate. So you might be able to loosen that strap a little bit as well.

PARTICIPANT: Thanks very much for the presentation. Two quick things, if I may. I'd like to point out, as somebody who writes software for a living, that there are lots of ways of dealing with the details, as you undoubtedly know. I'd like to propose that some standard algorithm for handling doubling time calculation that could be put up on the SUO website might be a very great public service that would get everybody on the same page with respect to how these numbers are generated.

Secondly, since Ken advertised his trial, I'd like to say we have finished a 300 patient randomized trial in this exact design, which will be presented at the AUA. It unfortunately showed absolutely no difference in any clinical outcome so far with an almost seven year follow-up.

So personally, I'm a little skeptical that this is going to be helpful in the very earliest stages, even though I am quite enthusiastic about the topic in other settings.

DR. PIENTA: That is somewhat disappointing to hear. What chemotherapy did you use?

PARTICIPANT: Just a couple of questions. First, is Nadir PSA following hormonal therapy, did you find that that actually was or was not a predictor that might be a much easier one than the entry and exit ratio?

DR. D'AMICO: We looked at that. The answer is, on a univariate analysis, it is. On a multi variable with doubling time, it is not.

The reason why is, a lot of people become undetectable, and subsequently rise. The issue is, I wonder if you used an ultra sensitive scale where you actually could measure lower values, if the Nadir would have a chance of being as important as or more important than the doubling time. But at least with using 0.1 as the lower limit, it wasn't. It is only important univariable, not multi variable.

PARTICIPANT: A second question is, in terms of entry of patients into neo-adjuvant or even adjuvant trials, is there any information on PSA doubling time prior to the initial therapy in terms of additional stratification of patients for clinical trials in that setting?

DR. D'AMICO: I would think that doubling time could be used as a patient selection factor and a stratifier, but there clearly are going to be other things that you are going to want to balance for, because other things have been mentioned -- surgery trial, prostatectomy grade, pathologic state, perhaps margin status, although that is debatable.

I think that you want to collect that information or at least have it available, to show that the concepts are balanced. Doubling time is having a very easy opportunity here to make its way in, because the clinical parameters we have currently are limited. As you pointed out on many occasions, once we have molecular type markers that are actually a little bit better at finding the subsets within this heterogeneous disease of prostate cancer that do poorly, perhaps when you put that into a multivariate analysis with doubling time, it will become equally important or an independent predictor.

We are still dealing basically with volume and grade. So you've got volume and grade, and all you've done with doubling time is that time response. There are very primitive techniques that we have known for decades, just quantifying it a little bit better.

DR. WANG: Ken, since it doesn't look like PSA is predictive of outcome in a hormone refractory group, what do you propose we use as end points for novel agents in that group? Do we have agents that are powerful enough to detect survival at a detectable level? Or do we need some other end point?

DR. PIENTA: I think in the hormone refractory setting, patients still die at a quick enough rate that we can use survival without question as the end point. I think as you know, there are a lot of agents in the pipe that look very promising, and we are all waiting for this year's ASCO to look at the SWOG study in 9916 of estramustine versus Mitozantrum prednisone, as well as the tax 327 study, which was Taxotere prednisone versus Mitozantrum prednisone alone.

I really don't feel pushed in the hormone refractory setting to have a surrogate marker, so to speak, until we really start helping people live a lot longer.

DR. WANG: In a clinical trial, what would you think is a significant prolongation of survival that would make you excited about using an agent as a routine or building with another agent? How many months is enough?

DR. PIENTA: As you know, it is a moving target of how much disease a person starts with. I'm not sure months are the way to actually measure it. I would like to see in randomized trials a 25 percent to 33 percent increase in survival in months, as compared to whatever control arm happens to be.

I think already, people are discussing the fact that there may be some differences in the control arm time to survival in the phase three studies that have been currently done being analyzed. That is going to be a problem for us.

So I don't know the months, but I would like to see at least that sort of increase in time.

DR. WANG: Thanks very much. We will move now to the second late-breaking session.

TOP