Quantifying the Usefulness of PD Biomarkers in Phase 2 Screening Trials of Oncology Drugs

1
Quantifying the Usefulness of PD Biomarkers in
Phase 2 Screening Trials of Oncology Drugs

• Eric Holmgren
• Genentech Inc.
• 1 DNA Way
• South San Francisco, CA

2
Outline of Talk

• Phase 2 Screening Trials
• Model of Screening Trials in Drug Development
• Efficiency
• Without accounting for discovery costs
• With accounting for discovery costs

3
Outline of Talk

• Surrogate Endpoint / PD Biomarker
• Power of a PD Biomarker as a function of
  treatment benefit measured by Phase 3 end point.
• Efficiency of Screening Trials based on PD
  Biomarker
• Example

4
Clinical Development of New Molecules

• Phase 1 Dose escalation with 3-6 subjects per
  dose cohort. Determine Maximum tolerable Dose
• Phase 2 Can be a Randomized controlled trial, 30
  100 patients per arm. Establish evidence of
  activity. Under Powered for Phase 3 endpoint
  (Survival)
• Phase 3 Randomized Controlled Trial with power
  greater than or equal to 80 percent to establish
  Clinical Benefit (Survival)

5
What do we hope to learn from a Phase 2 Trial?

• Is the drug Safe?
• Of concern throughout a Drug Development program
• Is it worthwhile to Conduct Phase 3?
• Based on Phase 3 Endpoint?
• Based on PD Biomarkers?
• What dose should be used in Phase 3?

6
Is it worthwhile to conduct a phase 3 Trial?

• Phase 2 studies in Oncology are often used as
  screens to determine whether a new molecule is
  sufficiently active to test in a Phase 3 clinical
  trial (Korn et al, Gray et al).
• Unfortunately, these Phase 2 trials are
  substantially underpowered for making definitive
  assessments of activity.

7
Is it worthwhile to conduct a phase 3 Trial?

• From 1991 2000, only 1 in 20 oncology compounds
  developed by the 10 largest pharmaceutical
  companies received regulatory approval by
  European or US regulatory authorities.

8
Table 3 Out of Pocket Costs by Phase of Drug
Development in millions of 2000 Dollars (DiMassi
et al),,

• Mean denotes the average cost in the collection
  of drug development programs analysed
• Exp denotes the expected cost per molecule
  accounting for the probability of entering that
  phase of drug development.

9
Is it worthwhile to conduct a phase 3 Trial?

• The hope is that compounds that will ultimately
  fail can be identified with a Phase 2 screening
  trial before Phase 3 trials begin.

10
Reasons to be Cautious

• A bad screening trial will not detect a large
  number of active molecules

11
Model of the Screening Trials
12
Model of Screening Trials
We model the drug development problem as having
two states of nature
13
A number of important quantities
14
A number of important quantities
15
Efficiency
16
(No Transcript)
17
(No Transcript)
18

• There are three things associated with a Phase
  2/3 approach to note in the two figures presented
  above.
• the reduction in the probability that a specific
  molecule will have a positive Phase 3 study.
• the large decrease in the expected sample size
  and
• the gain in efficiency for hazard ratios less
  than 0.80.

19
Relative Efficiency
We can factor efficiency as follows
20
Relative Efficiency
21
Admissible Phase 2/3 Programs
alpha_20.20

• Phase 2 trials with fewer events as a percentage
  of Phase 3 are admissible
• Represents an upper bound on events that can be
  efficiently observed in Phase 2

22
Admissibility after considering costs of drug
discovery.

• The costs of developing a drug are not solely
  made up of the costs of enrolling patients into
  Phase 2 and Phase 3 trials. Substantial costs are
  also incurred identifying molecules that would be
  good candidates to enter clinical trials.

23
Table 3 Out of Pocket and Capitalized Costs by
Phase of Drug Development (DiMassi et al), in
millions of 2000 Dollars,

• Mean denotes the average cost in the collection
  of drug development programs analysed
• Exp denotes the expected cost per molecule
  accounting for the probability of entering that
  phase of drug development.

24
Admissibility after considering costs of drug
discovery

• With 71 percent of molecules that enter phase I
  going ahead to Phase 2, the RD cost per molecule
  entering Phase 2 is 100 million.
• Adding in the costs of running Phase 1 trials we
  end up with a total pre Phase 2 cost of 130.5
  million. This is roughly equal to the estimated
  cost of a Phase 3 trial of 119.2 million.

25
Relative efficiency including costs of drug
discovery
We can factor the expression for relative
efficiency as follows
,

26
Relative efficiency including costs of drug
discovery

27
Relative efficiency
Without counting Costs of discovery
With counting Costs of Discovery
28
C0 (no costs prior to Phase 2)
C1 (costs prior to Phase 2 cost of Phase 3).
29
alpha
Efficiency
Power
Numberof Subjects
30
alpha
Number of Subjects
Efficiency
Power
31
Phase 2,3 Strategies
32
Phase 2/ 3 Adaptive
33
Phase 2 Screen/Phase 3 Adaptive Design
34
Preliminary Conclusions

• When considering only the efficiency of a
  clinical trials program, Phase 2 screening trials
  can substantially increase the efficiency with
  which drugs that provide clinical benefit are
  identified.
• If there are substantial costs in identifying
  drug candidates to test in clinical trials, then
  Phase 2 screening trials as a prelude to Phase 3
  trials may reduce the efficiency of the drug
  development process as a whole.

35
PD Biomarkers

• To assess the value of PD biomarkers we will
  evaluate the impact of their use as a primary
  endpoint in Phase 2 on relative efficiency

36
Examples of PD Biomarkers

• DCE-MRI (dynamic contrast enhanced magnetic
  resonance imaging) for the detection of changes
  in the permeability of tumor vasculature,
• FDG PET (fluorodeoxyglucose positron emission
  tomography) for the detection of changes in
  Glucose metabolism and
• FLT PET (fluoro-L-thymidine positron emission
  tomography) for the detection of changes in DNA
  synthesis.

37
Surrogacy

• If a PD biomarker can be shown to be a surrogate
  of survival Prentice then the justification for
  its use to assess a drugs activity in Phase 2 is
  clear.
• However, it is relatively uncommon that a marker
  meets all the requirements to be considered a
  surrogate Fleming, Demets.

38
Surrogacy

• The proportion of treatment effect explained,
  and are some of the measures that have
  been proposed to quantify how strong the
  surrogacy is. This literature is reviewed in Weir
  and Walley

39
Surrogacy

• How strong does surrogacy have to be to be
  meaningful?
• We will address this question by looking at the
  implications for a drug development program when
  a biomarker is used as a primary endpoint in
  Phase 2

40
Surrogates

• Need to express the power of a PD biomarker in a
  Phase 2 trial in terms of the Hazard ratio for
  Survival

41
We model the relationship between Survival and
the pd biomarker as follows
where
42
Power of Phase 2 Screen in terms of hazard ratio
for survival.

,
43
(No Transcript)
44
Relative Efficiency for a marker based Phase 2
Screening study.
.
45
Relative Efficiency as a function of Marker and
Rand. Effect Variance
46

• The parameters for the model and their variances
  can be estimated in a meta analysis as suggested
  by Buyse et al
• The variability of relative efficiency is
  obtained using the delta method.

47
Estimation
.
48
The likelihood for a group of N studies is
49
(No Transcript)
50
(No Transcript)
51

• Power calculations show that 20 to 30 Phase 2
  trials of different molecules will have
  reasonable power to determine the usefulness of a
  PD Biomarker as described above
• A large biotech or a consortium of companies
  could assemble such a data set over several years.

52
References

• Prentice RL. Surrogate endpoints in clinical
  trials definition and operational criteria.
  Statistics in Medicine 1989 8431-440
• Fleming TR, Demets DL. Surrogate end points in
  clinical trials are we being misled? Annals of
  Internal Medicine 1998 125605-613
• Weir CJ, Walley RJ. Statistical evaluation of
  biomarkers as surrogate endpoints a literature
  review. Statistics in Medicine 25183-203
• Buyse M et al. Validation of surrogate endpoints
  in meta-analyses of randomized experiments.
  Biostatistics 2000 149-67
• Holmgren, Are Phase 2 screening trials in
  oncology obsolete, Statistics in Medicine, 2007
• Holmgren, Quantifying the usefulness of PD
  biomarkers in Phase 2 screening trials of
  Oncology Drugs, Statistics in Medicine, 2008

53
References

• Korn E.L., et al. Clinical Trial Designs for
  Cytostatic Agents Are New Approaches Needed. J
  Clin Oncol 19265-272.
• Gray R. et al. Phase II Clinical Trial Design
  Methods in Translational Rsearch from the
  Genintourinary Committee at the Eastern
  Cooperative Oncology Group. Clin Cancer Res
  200612(7) April 1, 2006.
• Roberts Jr TG, Lynch Jr TJ, Chabner BA. The Phase
  III Trial in the Era of Targeted Therapy
  Unraveling the "Go or No Go" Decision. Journal of
  Clinical Oncology, 21(19), 2003, pg 3683-3695
• Von Hoff, D . There are no bad anticancer
  agents, only bad clinical trial
  designs-Twenty-first Richard and Hinda Rosenthal
  Foundation Award Lecture. Clin Cancer Res 1998 4
  1079-1986.
• DiMasi J. et al. The price of innovation new
  estimates of drug development costs. Journal of
  Health Economics 22 (2003) 151-185.
• Raghavan An Essay on Rearranging the Deck Chairs
  Whats Wrong with the Cancer Trials System Clin
  Cancer Res 200612(7) April 1, 2006.
• .