|
Home:
Program Supplement
INTERVIEW WITH DR RICHARD PETO
DR NEIL LOVE: Before
we conclude our program with a number of other case discussions
by Dr Soloway and Dr Dawson, let me review the background to the
prostate cancer patient preferences study Mark and I are launching
this summer. My career in medical oncology since 1975 has, until
recently, focused on breast cancer. In producing a variety of nationally
distributed breast cancer education programs over the last 17 years,
I've had the unique opportunity to interface with many key national
and international investigators as breast cancer clinical research
has evolved. Over the last year, I noticed with great interest the
evolution of clinical research data on endocrine therapy of prostate
cancer - particularly the Early Prostate Cancer trials discussed
by Dr See. Last fall, I began chatting with my colleague, Mark Soloway,
about how these data compare to the early trial results in the 1980's
for tamoxifen in breast cancer. While the biology, natural history
and available clinical interventions for these two common diseases
are clearly not identical, Mark became quite interested in how the
thinking on breast cancer clinical research has evolved over time.
Specifically, in the late 1970s and early 80s, it was quite clear
to the medical oncology community that endocrine intervention for
breast cancer, including tamoxifen, when utilized for women with
clinical evidence of metastatic disease frequently delayed disease
progression. However, it did not have a major impact on survival,
and almost all of these women died from complications of metastases
unless another co-morbid condition interceded.
In 2002, it is still widely recognized that endocrine therapy for
women with metastatic disease has minimal effect on long-term survival.
But after a plethora of randomized trials of early so called adjuvant
therapy involving more than 100,000 women, it is also now very evident
that when endocrine therapy - specifically the anti-estrogen, tamoxifen
- is utilized at first diagnosis for women with localized breast
cancer, there is a major and profound impact both on the risk of
progression and survival. In the last decade, the age-adjusted mortality
of breast cancer in westernized areas of the world has decreased
more than 20 percent, probably to a great extent as a direct result
of the widespread use of adjuvant tamoxifen. The history of how
this evolved and the reaction of the oncology community to these
data is fascinating, and potentially relevant to physicians interpreting
the prostate cancer research database.
In retrospect, it's clear that for about 10 or 15 years, data was
available that was being applied to practice inconsistently. The
initial individual trials of the early use of tamoxifen demonstrated
a delay in recurrence, but no effect on survival. This led many
onocologists to conclude that there was no effect on mortality and
therefore, not of enough use to present as an option to patients.
However, in 1985, Oxford statistician Richard Peto pointed out that
there were two possible reasons that a mortality benefit for adjuvant
tamoxifen was not being observed. He believed either there was no
survival effect, or in fact tamoxifen did reduce mortality but not
enough events, specifically breast cancer deaths, had been observed
in the trials to detect the effect. Peto suggested that by combining
the available randomized trials into a meta-analysis, enough deaths
would be available to be able to determine if there were a survival
benefit, and his first public presentation of this concept was at
the 1985 NIH Consensus Conference on early breast cancer:
DR RICHARD PETO: I'm here to talk not on my own behalf,
but on behalf of a group of trialists who came together over the
weekend to see what, when all of their data were viewed together,
could be learned from them. These trialists came from many countries.
Many dozen trialists were involved and, together, they had information
on, I think, about 40,000 women who had participated in randomized
trials comparing one treatment with another treatment.
Now, the first reaction to this kind of massive enterprise is,
well, if you need to study so many women to find out whether there's
any difference, then the difference can't be worth knowing about.
I think this is quite wrong. I think there's danger in cancer research,
especially when we're talking about the major cancers, in over-optimism.
I mean, if you take, for example, what Secretary of State Hector
says, that the aim is to halve the risk of cancer by the year 2000.
Now, if you can halve the risk of cancer by the year 2000, you're
going to have to produce some absolutely vast changes in lots of
different types of cancer.
Then you say, how are we going to do this? What is going to contribute
substantially toward halving the risk of cancer by the year 2000?
Then really, only over-optimistic claims will seem relevant. The
danger of over-optimism is that, if you're trying to say how are
we going to knock 30 percent, 50 percent, 20 percent off total cancer
risks, then you're only going to be interested in things that are
going to knock five or ten percent off total cancer risks. There
are very, very few things that can do that. I mean, the only thing
I know is some substantial action against cigarette smoking. It's
the only thing I know that's up in the big league there. And if
you get so much interest in exaggerated hopes, if you're looking
only for big things, then the only things that'll contribute towards
your aim are exaggerated claims, over-optimistic claims, claims
by people who actually are not being realistic about what they can
achieve.
Of course, it may be that some vast breakthrough is coming. I mean,
we've seen vast breakthroughs in Hodgkin's disease, cancer of the
testes, treatment of leukemia. And since we've seen vast breakthroughs
in the treatment of rare diseases, maybe we're going to see a vast
breakthrough in the treatment of some of the common types of cancer.
There's no absolute reason why not. Let's hope that we are.
But if you look all the time at where are we going to get a vast
improvement here, where are we going to get a vast improvement there,
then, in general, you'll be concentrating your attention on unrealistic
and misleading hopes. And this may be to the detriment of more realistic
hopes. In fact, if you're looking all the time at claims that something
is going to save 10, 20, 30, 40, 50,000 lives a year, then realistic
claims that, by a lot of work, are actually going to save four or
5,000 lives a year by some particular therapy will start to seem
uninteresting.
Now, this is a pity, and it's particularly a pity in the context
of breast cancer, because, in breast cancer, moderate improvements
in prognosis are really worthwhile. They can be humanly worthwhile.
They are not large percentages, but they are large numbers of human
beings.
Every year in the U.S., there are 400,000 cancer deaths, roughly.
Of these, about 40,000 involve breast cancer. Now, realistically,
the kinds of changes that you're going to hear described, the kinds
of trials that you're going to hear described, might, if we're lucky,
involve avoidance of, let's say, 4,000 of those 40,000 deaths. Now,
avoidance of 4,000 deaths is humanly worthwhile. I mean, there's
nowhere near 4,000 people in this auditorium, and avoidance of the
instant deaths of all of us would certainly be worthwhile. So, 4,000
deaths is worth knowing about, but it's a small percentage. And
it's very difficult to pick out those kinds of small percentage.
So you've really got to get very accurate, very sensitive, large
randomized trials, the trials by the large groups. And even with
them, you've got to be very careful to avoid some particular pattern
in some particular trial, leading you to wild over-optimism or,
conversely, to unjustifiable pessimism about the possible value
of a particular treatment.
|
|
