Monitoring Adap-ve Clinical Trials a Case Study

Monitoring Adap-ve Clinical Trials; a Case Study Zoran Antonijevic

AcknowledgementsCyrus Mehta Eric Silva ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 2

Presentation Contents• Study Design Overview • Opera-onal Challenges with Adap-ve Trials • Access Control Execu-on System (ACES) ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 3

The VALOR Trial for AML• Vosaroxin and Cytarabine combina-on evalua-ng Overall Survival in relapsed/refractory Acute MyeloidLeukemia (AML) • Phase 3, double-­‐blind, placebo-­‐controlled, mul-na-onal trial • Design for 90% power at 5% significance level ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 4

Sponsor’s Dilemma• Based on phase 2 data (N=69): – Assume 5/7 month median on Ctrl/Trtm (HR=0.71) – Require 375 events and 450 subjects @ 19/month • But phase 2 es-mates are subject to uncertainty: – What if 5/6.5 month median on Ctrl/Trtm (HR=0.77)? – HR = 0.77 is s-ll clinically meaningful – Require 616 events and 732 subjects @ 31/month – Not a feasible op-on for sponsor ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 5

Sponsor Adopts a Strategy ofStaged Investment• Power study to detect HR=0.71 up-­‐front • One interim analysis ager 50% informa-on – Stop early if overwhelming evidence of efficacy – Stop early for fu-lity if low condi-onal power – Increase number of events and sample size if results are promising ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 6

The Promising Zone Design• Par--on the interim outcome into three zones based on the interim es-mate of condi-onal power. – Favorable: CP ≥ Y%; no change to design – Promising: X% ≤ CP < Y%; increase resources – Unfavorable: CP ≤ X%; no change to design • Control type-­‐1 error by using Cui, Hung and Wang (1999) weighted sta-s-c modified for survival data ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 7

Adaptive Decision RuleInterim analysisat ~187 events:Interim AnalysisPlanned EndC effY%EfficacyFavorablePromising Zone (add 225 patients)X%C futUnfavorableFutilityCPCP = Conditional powerThe probability of success (statistical significance) at the end of the trial given current data trendInterim outcome par--oned into unfavorable, promising, and favorable zones ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 8

Data Monitoring Committee• DMC responsible for monitoring both, safety and efficacy • DMC may call for sample size increase only if interim result falls into Promising Zone ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 9

Avoidance of Operational Bias• Guidance documents by FDA and EMA for DMC and Adap-ve Trial Design: – Reference the importance of confiden-ality of interim results – Require that well-­‐trusted firewalls are established for trial conduct to provide assurance that opera-onal biases have not been introduced. – Requests an accurate recording of trial conduct and documenta-on – who saw what and when ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 10

Avoidance of Operational Bias• Must provide auditable evidence that GSD/SSR was strictly followed and based only on the pre-­‐specified decision rule • Ensure that firewalls were in place to protect unblinded analyses • Show evidence that Sponsor was not involved in ISC and DMC interac-ons and was not exposed to unblinded IA results ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 11

Access Control Execution System(ACES)• ACES is a secure, web-­‐based system used during the interim analysis to: – Centrally store interim analysis reports, mee-ng agendas and minutes, and DMC recommenda-ons – Assign team members to specific roles and grant explicit privileges to securely access data and informa-on ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 12

Traditional ProcessSponsor Create Documents (Protocol, SAP, DMC Charter) Store/Archive Documents Enroll Subjects & Collect Responses Send Response Data to ISC ISC Request additionalinformationSend Analysis to DMC Perform Analysis and Create Reports Create and Test Analysis Programs DMC Send Recommenda-on to Sponsor/ Steering Commitee Make Recommenda-on Steering Commitee Make Decision About Trial Ager decision… 1. DMC no-fied 2. Drug Supply no-fied 3. IVRS no-fied ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 13

Improving the ProcessEDC IRT CTMS DMC ISC Sponsor ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 14

Summary: Design• The adap-ve design mi-gates risk of ini-al over-­‐investment, and risk of failing to detect a relevant survival benefit • Sta-s-cal rigor: theore-cal and simula-on-­‐based guarantee that Type-­‐I Error is controlled • DMC may call for sample size increase only if interim result falls into Promising Zone • Study design prevents from back-­‐calcula-on of treatment effect • This design sa-sfies both sta-s-cal and opera-onal requirements s-pulated in FDA Drag Guidance and in EMA Reflec-on Paper on Adap-ve Design Clinical Trials ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 15

Summary: Process• Use of technology to control flow of informa-on between sponsor, independent sta-s-cal center and DMC – DMC portal for storage of sensi-ve documents – Interim analysis report generated and stored in DMC portal without possibility of external interven-on – Audit trail of access to all reports stored in the DMC portal to track ”who saw what and when” ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 16

Summary: Process• Integrated Systems can increase efficiency in the flow of data and informa-on, and how groups communicate • Opera-onal Bias is greatly reduced by using secure systems, and by limi-ng access to unblinded data • Trial Integrity is protected by documen-ng the conduct, logis-cs, and opera-on of the trial through security and audit trails • Regulatory Confidence in the consistent execu-on of adap-ve trials by implemen-ng systems that enforce ‘best prac-ce’ processes ©2012 Cytel Inc. JSM 2012; July 28 -­‐ August 2, 2012 17