Sunteți pe pagina 1din 47

Adaptive Study Design Multi-Arm, Multi-Stage trials

Nicholas James Professor of Clinical Oncology University of Birmingham @Prof_Nick_James

Cancer research spending


Half of all drugs in trials are cancer drugs Global cancer drug market risen from $48bn (2008) to $75bn (est 2012) 25 drugs with sales >1bn pa Annual research spend:
pharmaceutical industry $6-8 bn NCI $3.6 bn EU public bodies Eur 1.4 bn

If we are to afford new drugs, we must make trials cheaper and quicker

Research environment
New treatments usually not better than current
About 30 to 40% are positive Both academia and industry

Trials require huge time, effort and cost

Must be a better way to select treatments for efficacy assessments

MAMS Trial Design


Multi-Stage: Are there reasons why we should continue investigating a treatment?
Require sufficiently encouraging activity to continue assessment

Focus away from insufficiently active regimens


Focus limited resources on regimens that may benefit patients

Add new treatments of interest

STAMPEDE Celecoxib -- ECCO 2011 [LBA20] N Clarke

Advantages of MAMS trials


1. Fewer patients 2. Less overall time
Concurrent assessment of agents Randomise from start One seamless trial One protocol Less bureaucracy
Multi-arm, Multi-stage

Traditional Approach T1 Phase II T2 T3 T4

C Phase II

T1

T2 T3 T4

C T1 Phase III C T3 C T4 Phase III

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Advantages of MAMS trials


3. Increased flexibility
Adapts to intermediate results Focus on more promising arms

Traditional Approach T1 Phase II T2 T3 T4

Multi-arm, Multi-stage

C Phase II

T1

T2 T3 T4

C T1 Phase III C T3 C T4 Phase III

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Advantages of MAMS trials


4. Reduced costs
Limited resources for trials Must use fairly and efficiently

Traditional Approach T1 Phase II T2 T3 T4

Multi-arm, Multi-stage

C Phase II

T1

T2 T3 T4

C T1 Phase III C T3 C T4 Phase III

STAMPEDE practicalities Heidelberg 2012 MR Sydes

MAMS Trial Design


Multi-Arm: Start randomising from the outset Assess many promising agents simultaneously
Compared to a common control arm

Multi-stage Incorporate interim activity assessments


Phase II type assessments of each research arm
STAMPEDE Celecoxib -- ECCO 2011 [LBA20] N Clarke

Clinical Setting
Men with metastatic or high-risk non-metastatic prostate cancer Long-term hormone therapy (HT) alone is the standard of care Investigating whether early use of additional therapies can improve overall survival Many interesting agents demand assessment
No clear reason to choose one particular regimen Do not want to choose arbitrarily Want to assess all interesting agents

Focussed on three distinct initial treatments


STAMPEDE Celecoxib -- ECCO 2011 [LBA20] N Clarke

STAMPEDE outcome measures


Outcome Measure Stage Primary Secondary

Pilot
Activity I-III (phase II)

Safety
Failure-free survival

Feasibility
Overall survival Toxicity / safety Skeletal events

Efficacy IV (Phase III)

Overall survival

Failure-free survival Toxicity / safety Skeletal events Quality of life

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Trial design
Stage Type 10 OM HRA Power Sig. Critical HR Control Events

1 2 3 4

Activity Activity Activity Efficacy

FFS FFS FFS OS

0.75 0.75 0.75 0.75

95% 95% 95% 90% 85%

0.500 0.250 0.100 0.025 0.013

1.00 0.92 0.89 -

114 215 334 400

Overall (per comparison)

FFS: failure-free survival OM: outcome measure OS: overall survival

Stop early for lack-of-benefit on intermediate primary OM Stop early for benefit only on definitive primary OM
STAMPEDE practicalities Heidelberg 2012 MR Sydes

FEASIBILITY ASSESSMENT

STAMPEDE trial
Launched 2005 Initial feasibility stage:
Would patients accept randomisation between 6 treatment arms? Would clinicians put the time into the study to make it work? Would the chosen treatments be safe?

STAMPEDE original design


A Deprivation
Therapy Androgen Control arm
Treatment detail Androgen Deprivation Therapy :: Standard hormones :: Given for >3 year

Man with high-risk prostate cancer starting long-term hormone therapy

RANDOMISATION

B ADT + Zoledronic Acid C ADT + Docetaxel D ADT + Celecoxib E ADT + ZA + Doc F ADT + ZA + Cel

Zoledronic Acid :: 3rd generation bisphosphonat :: IV for 2 years every 3 to 4 we

Docetaxel :: Taxane chemotherapy :: IV for 6 cycles over 18 weeks Celecoxib :: Cox-2 inhibitor :: Oral MRC for 1 PR08 year

CRUK/06/019 ISRCTN78818544

NCT00268476

Accrual: start
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib

Past accrual Possible future accrual Follow-up


Slides for Kim Chi and NCIC CTG: Oct-2011

Accrual: end of Pilot Phase


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib

Feasibility and safety confirmed


Past accrual Possible future accrual Follow-up

STOPPING ARMS WITH INSUFFICIENT ACTIVITY

Accrual: end of Activity Stage I


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib

Past accrual Possible future accrual Follow-up

All arms complete round 2 of their high jump!


Slides for Kim Chi and NCIC CTG: Oct-2011

Accrual: end of Activity Stage II


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib

Celecoxib fails round 3 of its contest


Past accrual Possible future accrual Follow-up
Slides for Kim Chi and NCIC CTG: Oct-2011

AS 2 analysis (original arms)


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Data frozen: IDMC meeting: Accrual: FFS events:

01-Feb-2011 31-Mar-2011 2043 patients total 209 on control arm (target 215)

IDMC recommended changes: : Stop accrual to 2 arms due to lack-of-benefit


Both celecoxib-containing arms (D and F)

: Remaining arms to continue accrual into their Activity Stage III


Arms B, C and E, plus control arm A
STAMPEDE practicalities Heidelberg 2012 MR Sydes

Failure-free Survival
Adjusted HR: 0.94 (95%CI 0.74-1.20)
(1-sided upper 75% upper CI 1.03)

Target HR: 0.924

Unadjusted HR = 0.98

2-year FFS 50% (As predicted)


STAMPEDE Celecoxib -- ECCO 2011 [LBA20] N Clarke

ADDING NEW AGENTS

Flexibility and extension


Design adapts to include further agents
Can add new research arms during trial

Might see as a new trial within STAMPEDE protocol Must be scientifically compelling case for inclusion
Proposed through Trial Management Group Survey of participants Peer review through original funder (Cancer Research UK)

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Principles
First priority is to ongoing research arms
Must not hamper accrual or maturity so must either: 1. Recruit better than predicted overall 2. Wait for arms to have dropped out

Accept that new arm will mature later than original research arms
Need to extend accrual to control arm Only compare patients on new arms to patients recruited contemporaneously to control arm
STAMPEDE practicalities Heidelberg 2012 MR Sydes

Criteria for new arms


Existing or pending licence in prostate cancer Deliverable across all (or most) trial centres (Pharmaceutical partner willing to support with drug and distribution costs) Model for introduction Draw up draft agreement with partner organisation(s) Needs clinical leader effectively the CI for the new substudy Apply to CRUK for extension of trial

Advantages?
1. Can start recruiting quicker than a new trial
Updated protocol = simple, substantial amendment Scientific review = amendment

2. Efficient use of volunteers


Patients contribute to more than one comparison Reduce competing trials Seamless accrual: no gaps between trials

3. Efficient use of resources


Much quicker start-up time: Start at full speed Much cheaper than separate trial Get answers more quickly STAMPEDE practicalities
Heidelberg 2012 MR Sydes

Disadvantages?
1. Original research arms could mature whilst new assessment ongoing
Treat like data emerging from an external trial TMG will reacts if needed Update backbone of therapy across arms if required Already done for some M0 patients in STAMPEDE: added RT to ADT following published data from SPCG-7 & MRC PR07/NCIC PR.3
STAMPEDE practicalities Heidelberg 2012 MR Sydes

STAMPEDE trial design


A

Composition of arms Abiraterone added Nov 2011

R A N D O M I S E
E C B

Hormone Androgen suppression Therapy (AS) (HT)

Control

Con trol a r m

HT + zoledronic acid

HT + docetaxel

HT + zoledronic acid + docetaxel

HT + G abiraterone

Accrual: from Nov-2011


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone

Past accrual Possible future accrual Follow-up


Slides for Kim Chi and NCIC CTG: Oct-2011

Local approvals of new comparison


Timely R&D approval for new protocol 1.0
Ready on activation Suspended until ready

0.8

New arm switched on for whole trial on set date Sites given ~4 wk notice to gain local approvals 80 sites ready to recruit on activation day!
Activation day

0.6

0.4

0.2

Accrual nearly seamless

0.0 0
Sites needing approval

4 5 6 7 8 9 10 11 12 13 14 15 Time (weeks) from notifying sites

104 (6) 98 (21) 77 (19) 58 (11) 47 (26) 21 (8) 13 (0) 13 (1) 12 (1) 11 (4) 7 (0) 7 (0) 7 (0) 7 (1) 6 (3) 3

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Aug 13

O ct 05 Jan 06

J ul 06

Jan 07

J ul 07

Jan 08

J ul 08

Jan 09

J ul 09

Jan 10

J ul 10

Jan 11

J ul 11

Jan 12

J ul 12

Jan 13

Z:\Prostate\Stampede\Data\Accrual\rand_planned.wmf

Accrual to new arms


D ate of randomisation

Accrual to abiraterone comparison


1500

Patients in abir ater one compar ison

1200

900

600

Closure of original arms

O ct 13

Apr 13

300

0
J an 12 J ul 12 J an 13 J ul 13 J an 14 J ul 14 J an 15

D ate of r andomisation

Z:\Prostate\Stampede\Data\Accrual\abi_comparison_accrual.wmf

J ul 13

Cost
Duration of trial Trial 1 2 arms

Trial 2

2 arms Duration of trial

MAMS 2+1 arms trial

Faster set up Lower cost as shared between arms New funder needs to provide red area costs, no cost to original funder for extra trial

Accrual: from autumn 2012


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H
Past accrual Possible future accrual Follow-up
Slides for Kim Chi and NCIC CTG: Oct-2011

ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib

Docetaxel and Zoledronic Acid pass their final high jump round

ADT + abiraterone ADT + RT

ADDING FURTHER NEW AGENTS

Adding a randomisation to a subgroup


Prostate radiotherapy standard of care in locally advanced cases Of unproven benefit but of interest in M1 cases
Patients eligible for STAMPEDE NEWLY DIAGNOSED M1 PATIENTS1 RANDOMISATION A G H ADT ADT + abiraterone ADT + RT to prostate A G ALL OTHER PATIENTS2 RANDOMISATION ADT ADT + abiraterone

WHATS THE CONTROL ARM?

Analysis: ZA comparison
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Fully powered: accrued through AS4

Analysis: Docetaxel comparison


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Fully powered: accrued through AS4

Analysis: Doc + ZA comparison


2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Fully powered: accrued through AS4

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

Analysis: Celecoxib comparisons

ADT-alone ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Less power: stopped accrual in AS2

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

Analysis: Abiraterone comparison


ADT-alone

ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Recruitment in Efficacy Phase 2

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 A B C D E F G H M1 only
In analysis Not in analysis

Analysis: M1/RT comparison


ADT-alone (M1-only)

ADT + zoledronic acid ADT + docetaxel ADT + celecoxib

ADT + zoledronic acid + docetaxel ADT + zoledronic acid + celecoxib


ADT + abiraterone ADT + RT (M1-only)

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Recruitment in Feasibility Phase

STAMPEDE trial
One question answered celecoxib probably not of value in prostate cancer Docetaxel and zoledronic acid completed intermediate efficacy tests FFS and OS data are awaited Trial expanded twice to add abiraterone, RT in M1 disease Potentially can add arms indefinitely and use as a programmatic platform

Conclusions
MAMS trials speed evaluation of new treatments by:
1. Testing many treatments simultaneously 2. Using LOB analyses to focus research efforts

Insufficiently active arms can successfully be stopped seamlessly in a MAMS trial Adding new research arms to an ongoing trial is achievable and desirable Addition of arms to subgroups feasible and effective e.g.
Stage Pattern of metastasis Biomarkers none yet tested in Stampede but possible in theory

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Refs: MAMS trials


Royston P, Parmar MKB, Qian W
Novel Designs for Multi-Arm Clinical Trials with Survival Outcomes, with an Application in Ovarian Cancer. Statistic Med 2003; 22: 22392256

Barthel FMS, Royston P, Parmar MKB


A menu-driven facility for sample size calculation in multi-arm, multi-stage randomised controlled trials with a survival-time outcome. The Stata Journal 2009; 9 (4): 505-523

Parmar MKB, Barthel F, Sydes MR et al


Speeding up the Evaluation of New Agents in Cancer. J Natl Cancer Inst 2008; 100 (17):1204-1214

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Refs: STAMPEDE methods & data


Sydes MR, Parmar MKB, James ND et al Issues in applying multi-arm multi-stage (MAMS) methodology to a clinical trial in prostate cancer: the MRC STAMPEDE trial. Trials 2009; 10 (39) James ND, Sydes MR, Parmar MKB et al Celecoxib plus hormone therapy versus hormone therapy alone for hormone-sensitive prostate cancer: first results from the STAMPEDE multiarm, multistage, randomised controlled trial. The Lancet Oncology 13(5): 549-558 Parker, C et al Prostate radiotherapy for men with metastatic disease: a new comparison in the STAMPEDE trial. BJU Int. 2013; 111(5):697-9 and Clin Onc editorial 2013; 25(5): 318-320

STAMPEDE practicalities Heidelberg 2012 MR Sydes

Acknowledgements
Matt Sydes and Max Parmar MRC CTU Noel Clark and Malcolm Mason Study Vice Chairs Funding:
Cancer Research UK (CRUK/016/09) Novartis free drug & educational grant Sanofi-Aventis discounted drug & educational grant Pfizer free drug & educational grant Janssen free drug & educational grant MRC core funding

All clinicians and hospital staff who have supported and continue to support the trial All patients who joined the trial and their families
STAMPEDE practicalities Heidelberg 2012 MR Sydes

S-ar putea să vă placă și