|
|
|
|
|
|
|
|
|
|
Drug Development for Chronic Cancers: Time to Think Differently?
Drug Development for Chronic Cancers: Time to Think Differently?
+ Author Affiliations
- Corresponding author: David R. Spriggs, MD, Memorial Sloan-Kettering Cancer Center, Howard 901, 1275 York Ave, New York, NY 10021; e-mail: spriggsd@mskcc.org.
As the treatments for epithelial cancers have
inched forward, the chronic management of patients with high
performance status
has become a common problem in the practice of
oncology. This conundrum is clearly illustrated by the excellent, yet
negative,
randomized trial by Colombo et al1
that accompanies this editorial. This large, well-designed clinical
trial tested another modestly active agent, patupilone,
in comparison with pegylated liposomal doxorubicin
(PLD) in patients with platinum-resistant ovarian cancer. Sadly, this
study
did not meet its prespecified superiority end point,
namely an overall survival improvement compared with the use of PLD.
What drug development lessons can be learned
from yet another trial that fails to meet its designed superiority
outcome?
There are several potential points for consideration by solid tumor oncologists and gynecologic oncologists as we design strategies to incrementally improve the lives of patients with cancer until we find the rare breakthroughs that will irrefutably extend patient survival.
There are several potential points for consideration by solid tumor oncologists and gynecologic oncologists as we design strategies to incrementally improve the lives of patients with cancer until we find the rare breakthroughs that will irrefutably extend patient survival.
First, we need to acknowledge that the
barriers for achieving major advances are formidable. Previously treated
solid tumors
have been selected in a devilishly Darwinian crucible
of genetic instability, heterogeneity, and mutagenic primary therapy.
Drug-resistant cancers tend to be broadly refractory
to additional agents, regardless of mechanism. Platinum-resistant tumor
response rates to topotecan, PLD, and trabectedin are
all half of what would be expected in the untreated or
platinum-sensitive
patient groups. Without a predictive biomarker to
enrich for patients who are likely to benefit, many nonresponders will
be
present in both arms of a randomized trial in
platinum-resistant ovarian cancer, making it difficult to detect small
but potentially
important signs of clinical benefit. This low response
rate and the brief duration of improvement combine to dictate very
large and costly sample sizes. Yet even adequately
sized superiority trials like that of Colombo et al1 are likely to fail. Table 1 details the recent phase III randomized trials in platinum-refractory ovarian cancer.1–9
From these trials, performed over more than a decade, it is apparent
that classic chemotherapy development has hit a glass
ceiling in platinum-resistant ovarian cancer. One must
conclude that inferiority is a real possibility in these trials
(because
the approved drugs have reproducible, modest effects),
therapeutic equivalence is an optimistic outcome, and superiority to
current treatment is a distant, unlikely achievement.
New superiority trials with classic chemotherapy must be regarded like
third Hollywood marriages: a “triumph of hope over
experience.”10
View this table:
Table 1.
Recent Phase III Randomized Trials in Platinum-Refractory Ovarian Cancer
Second, our models for predicting success are
unreliable. The phase II experience that led to this phase III trial is
not
yet published but is described as a single-arm,
nonrandomized study. These modestly sized, nonrandomized trials are
treacherous
guides and have increasingly been criticized for their
lack of predictive power.11,12
The reliability of single-arm, phase II trials is undermined by
undefined patient selection factors and the wide variability
that is caused by the statistics of small numbers. The
sponsor often finds that the data set is incomplete, but the ticking
of the patent clock forces a choice between a risky
phase III trial and abandoning an asset of potential value. Although
predictive
biomarkers are widely advocated, the number of
responses is frequently low and the biomarkers are not validated until
long
after the phase III decisions are forced.13
Finally, the overall survival end point,
demanded by purists and regulatory authorities, is seriously weakened by
the growing
number of modestly successful treatments that are
offered to patients who need treatment after their trial participation
ends.
Although this is not an issue in the patupilone trial
by Colombo et al,1 the analysis by Broglio and Berry14
is particularly sobering for other trials with an overall survival end
point. The simulations by Broglio and Berry suggest
that extended survival after completion of trial
participation is likely to lead to a growing dissociation between
progression-free
survival (PFS) and overall survival end points. Long
post-trial survival, coupled with the use of subsequent (albeit
marginal)
treatment options for platinum-resistant ovarian
cancer, make this point very salient to our assessment of new drugs in
the
randomized trial setting. Topotecan, weekly
paclitaxel, gemcitabine, and bevacizumab all have measurable antitumor
activity
in the population that was studied by Colombo et al,
but the effect of epothilone treatment on subsequent response to these
and other agents is unknown. The uncertainty is
compounded further by predictable PLD crossover in the patupilone arm.
Solutions to this conundrum are complicated,
but several approaches may be considered. In the absence of a robust
selection
marker, perhaps therapeutic equivalence is a more
realistic goal, despite the well-known statistical penalties that are
intrinsic
to equivalency trials, not the least of which is the
relatively large number of patients required to power such studies
appropriately.
The history delineated in Table 1
confirms that inferiority is a risk, but superiority is exceedingly
rare. PFS outcomes that are similar to the best current
therapy might not be such a bad outcome, assuming that
the new drug has a superior toxicity profile. In the last year, PLD
has been rarely available to patients as a result of
problems with production. An alternative drug like patupilone would have
been welcome to patients and doctors alike during that
hiatus. We may also want to reconsider the value of a strictly defined
PFS end point when expected post-trial survival
exceeds 6 months, as it often does in well-managed patients with solid
tumors.
And perhaps a more realistic economic valuation is in
order for therapies with modest incremental benefits. A lower cost might
decrease society's reluctance to accept a new drug's
approval for limited survival benefits. This is an appropriate time to
re-evaluate all of our traditional preconceptions
about cancer drug development. The present system seems to be serving
patients,
companies, regulators, and oncologists poorly.
AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
0 comments :
Post a Comment
Your comments?
Note: Only a member of this blog may post a comment.