BAYESIAN ADAPTIVE DESIGN

1
BAYESIAN ADAPTIVE DESIGN INTERIM ANALYSIS

• Donald A. Berry
• dberry_at_mdanderson.org

2
Some references

• Berry DA (2003). Statistical Innovations in
  Cancer Research. In Cancer Medicine e6. Ch 33. BC
  Decker. (Ed Holland J, Frei T et al.)
• Berry DA (2004). Bayesian statistics and the
  efficiency and ethics of clinical trials.
  Statistical Science.

3
Benefits

• Adapting examples
• Stop early (or late!)
• Change doses
• Add arms
• Drop arms
• Final analysis
• Greater precision (even full follow-up)
• Earlier conclusions

4
Goals

• Learn faster More efficient trials
• More efficient drug/device development
• Better treatment of patients in clinical trials

5
OUTLINE EXAMPLES

• Extraim analysis
• Modeling early endpoints
• Seamless Phase II/III trial
• Adaptive randomization
• Phase II trial in AML
• Phase II drug screening process
• Phase III trial

6
EXTRAIM ANALYSES

• Endpoint CR (detect 0.42 vs 0.32)
• 80 power N 800
• Two extraim analyses, one at 800
• Another after up to 300 added pts
• Maximum n 1400 (only rarely)
• Accrual 70/month
• Delay in assessing response

Modeling due to Scott Berry ltscott_at_berryconsultan
ts.comgt
7

• After 800 pts accrued, have response info on 450
  pts
• Find pred prob of stat sig when full info on 800
  pts available
• Also when full info on 1400
• Continue if . . .
• Stop if . . .
• If continue, n via pred prob
• Repeat at 2nd extraim analysis

8
vs 0.80
9
MODELING EARLY ENDPOINTS LONGITUDINAL MARKERS

• Example CA125 in ovarian cancer
• Use available data from trial ( outside of
  trial) to model relationship over time with
  survival, depending on Rx
• Predictive distributions
• Use covariates
• Seamless phases II III

10
CA125 data predictive distributions of survival
for two of many patients gt
Modeling due to Scott Berry ltscott_at_berryconsultan
ts.comgt
11
Patient 1
Treatment
Days
12
Patient 1
13
Patient 2
Days
14
Patient 2
15
Methods

• Analytical
• Multiple imputation

16
SEAMLESS PHASES II/III

• Early endpoint (tumor response, biomarker) may
  predict survival?
• May depend on treatment
• Should model the possibilities
• Primary endpoint survival
• But observe relationships

Inoue, et al (2002 Biometrics)
17
Conventional drug development
Survivaladvantage
Market
Good resp
No survivaladvantage
Not
No resp
Stop
Phase 3
Phase 2
gt 2 yrs
6 mos
9-12 mos
Seamless phase 2/3
lt 2 yrs (usually)
18
Seamless phases

• Phase 2 1 or 2 centers 10 pts/mo, randomize E
  vs C
• If pred probs look good, expand to Phase 3
  Many centers 50 pts/mo (Initial centers continue
  accrual)
• Max n 900
• Single trial survival data combined in final
  analysis

19
Early stopping

• Use pred probs of stat sig
• Frequent analyses (total of 18) using pred probs
  to
• Switch to Phase 3
• Stop accrual for
• Futility
• Efficacy
• Submit NDA

20
Comparisons

• Conventional Phase 3 designs Conv4 Conv18,
  max N 900
• (same power as adaptive design)

21
Expected N under H0
22
Expected N under H1
23
Benefits

• Duration of drug development is greatly shortened
  under adaptive design
• Fewer patients in trial
• No hiatus for setting up phase 3
• All patients used for
• Phase 3 endpoint
• Relation between response survival

24
Possibility of large N

• N seldom near 900
• When it is, its necessary!
• This possibility gives Bayesian design its edge
• Other reason for edge is modeling
  response/survival

25
ADAPTIVE RANDOMIZATIONGiles, et al JCO (2003)

• Troxacitabine (T) in acute myeloid leukemia (AML)
  combined with cytarabine (A) or idarubicin (I)
• Adaptive randomization to IA vs TA vs TI
• Max n 75
• End point Time to CR (lt 50 days)

26
Adaptive Randomization

• Assign 1/3 to IA (standard) throughout (until
  only 2 arms)
• Adaptive to TA and TI based on current results
• Results ?

27
(No Transcript)
28
Drop TI
Compare n 75
29
Summary of results

• CR lt 50 days
• IA 10/18 56
• TA 3/11 27
• TI 0/5 0
• Criticisms . . .

30
SCREENING PHASE II DRUGS

• Many drugs
• Tumor response
• Goals
• Treat effectively
• Learn quickly

31
Standard designs

• One drug (or dose) at a time no drug/dose
  comparisons
• Typical comparison by null hypothesis RR 20
• Progress hopelessly slow!

32
Standard 2-stage design

• First stage 20 patients
• Stop if 4 or 9 responses
• Else second set of 20

33
An adaptive allocation

• When assigning next patient, find r P(rate
  20data) for each drug
• Assign drugs in proportion to r
• Add drugs as become available
• Drop drugs that have small r
• Drugs with large r ? phase 3

34
Suppose 10 drugs, 200 patients
Identify nugget With probability In average n

• 9 drugs
• have mix
• of RRs
• 20 40,
• 1 has 60
• (nugget)

gt99
110
lt70
Adaptive
Standard
50
Standard
Adaptive
Adaptive also better at finding 40, sooner
35
Suppose 100 drugs, 2000 patients
Identify nugget With probability In average n

• 99 drugs
• have mix
• of RRs
• 20 40,
• 1 has 60
• (nugget)

gt99
1100
lt70
Adaptive
Standard
500
Standard
Adaptive
Adaptive also better at finding 40, sooner
36
Consequences

• Treat pts in trial effectively
• Learn quickly
• Attractive to patients, in and out of the trial
• Better drugs identified sooner move through
  faster

37
PHASE III TRIAL

• Dichotomous endpoint
• Q P(pE gt pSdata)
• Min n 150 Max n 600
• After n 50, assign to arm E with probability Q
• Except that 0.2 P(assign E) 0.8
• (Not optimal, but )

38
Recommendation to DSMB to

• Stop for superiority if Q 0.99
• Stop accrual for futility if P(pE pS lt
  0.10data) gt PF
• PF depends on current n . . .

39
PF
40
Common prior density for pE pS

• Independent
• Reasonably non-informative
• Mean 0.30
• SD 0.20

41
(No Transcript)
42
Updating

• After 20 patients on each arm
• 8/20 responses on arm 1
• 12/20 responses on arm 2

43
(No Transcript)
44
Assumptions

• Accrual 10/month
• 50-day delay to assess response

45
Need to stratify. But how?

• Suppose probability assign to experimental arm
  is 30, with these data . . .

46
(No Transcript)
47
One simulation pS 0.30, pE 0.45
Superiority boundary
Final Std 12/38 19/60
20/65 Exp 38/83 82/167 87/178
48
One simulation pE pS 0.30
Futility boundary
9 mos. End Final Std 8/39
15/57 18/68 Exp 11/42 32/81 22/87
49
Operating characteristics
50
FDA Why do this? Whats the advantage?

• Enthusiasm of PIs
• Comparison with standard design . . .

51
Adaptive vs tailored balanced design w/same
false-positive rate power (Mean number patients
by arm)
ORR Arm pS 0.20 pE 0.35 pS 0.20 pE 0.35 pS 0.30 pE 0.45 pS 0.30 pE 0.45 pS 0.40 pE 0.55 pS 0.40 pE 0.55
ORR Arm Std Exp Std Exp Std Exp
Adaptive 68 168 79 178 74 180
Balanced 171 171 203 203 216 216
Savings 103 3 124 25 142 36
52
Consequences of Bayesian Adaptive Approach

• Fundamental change in way we do medical research
• More rapid progress
• Well get the dose right!
• Better treatment of patients
• . . . at less cost

53
OUTLINE EXAMPLES

• Extraim analysis
• Modeling early endpoints
• Seamless Phase II/III trial
• Adaptive randomization
• Phase II trial in AML
• Phase II drug screening process
• Phase III trial