Surrogate Endpoints: Surrogate Endpoints: A Regulatory View A Regulatory View

Greg Campbell, Ph.D.Greg Campbell, Ph.D.Director, Division of BiostatisticsDirector, Division of BiostatisticsCenter for Devices and Center for Devices and Radiological Radiological HealthHealthFood and Drug AdministrationFood and Drug Administration

22

FDA’s Critical Path FDA’s Critical Path Opportunities ListOpportunities List #1 Biomarker Qualification#1 Biomarker Qualification

– One of five questions is “What types and One of five questions is “What types and levels of evidence are needed to accept a levels of evidence are needed to accept a biomarker as a surrogate endpoint for biomarker as a surrogate endpoint for product efficacy?”product efficacy?”

#6 Surrogates Outcomes for #6 Surrogates Outcomes for Cardiovascular Drug Eluting StentsCardiovascular Drug Eluting Stents

#23 Imaging Biomarkers in #23 Imaging Biomarkers in Cardiovascular DiseaseCardiovascular Disease

http://www.fda.gov/oc/iniatitives/criticalpath/reports/opp_list.pdfhttp://www.fda.gov/oc/iniatitives/criticalpath/reports/opp_list.pdf

33

Prentice’s Criteria for Prentice’s Criteria for Validation of a Validation of a Surrogate: Mathematical Surrogate: Mathematical FormulationFormulation

For surrogate S, true endpoint T and For surrogate S, true endpoint T and treatment Ztreatment Z

1.1. f(S|Z) is not f(S)f(S|Z) is not f(S)2.2. f(T|Z) is not f(T)f(T|Z) is not f(T)3.3. f(T|S) is not f(T)f(T|S) is not f(T)4.4. f(T|S,Z) = f(T|S)f(T|S,Z) = f(T|S)

Prentice (1989) Prentice (1989) Stat in MedStat in Med as in Burzykowski, Molenberghs, as in Burzykowski, Molenberghs, Buyse (2005). Buyse (2005). The Evaluation of Surrogate Endpoints The Evaluation of Surrogate Endpoints

44

Prentice CriteriaPrentice Criteria

The four elements of the Prentice The four elements of the Prentice criteria are difficult to achieve criteria are difficult to achieve simultaneously. The fourth one in simultaneously. The fourth one in particular implies that the entire particular implies that the entire treatment effect on T is captured by S treatment effect on T is captured by S (100% explained). (100% explained).

There is nothing in the criteria that There is nothing in the criteria that prevents one from going from a binary prevents one from going from a binary true endpoint to a continuous surrogate.true endpoint to a continuous surrogate.

55

Surrogates in DrugsSurrogates in Drugs

Fleming, T. and DeMets, D. Fleming, T. and DeMets, D. (1996). Surrogate end points in (1996). Surrogate end points in clinical trials: Are we being clinical trials: Are we being misled? misled? Ann. Int. Med.Ann. Int. Med. 125125:605-:605-613.613.

““A correlate does not a surrogate A correlate does not a surrogate make”make”

66

Surrogates in Medical Surrogates in Medical DevicesDevices DeMets, D. (2000). The role of DeMets, D. (2000). The role of

surrogate outcome measures in surrogate outcome measures in evaluating medical devices. evaluating medical devices. Surgery Surgery 128128:379-385. :379-385.

77

Accelerated Approval Accelerated Approval for Drugsfor Drugs

21 CFR (314 and 601) Accelerated Approval 21 CFR (314 and 601) Accelerated Approval Rule Rule – for serious or life-threatening illness for serious or life-threatening illness – it allows the use of surrogate or non-ultimate it allows the use of surrogate or non-ultimate

clinical endpoints when the effect on a surrogate clinical endpoints when the effect on a surrogate end point is “reasonably likely” to predict clinical end point is “reasonably likely” to predict clinical benefitbenefit

– post-market data is required “to verify and post-market data is required “to verify and describe the drug’s clinical benefit and to resolve describe the drug’s clinical benefit and to resolve the remaining uncertainty as to the relation of the remaining uncertainty as to the relation of the surrogate endpoint up on which approval was the surrogate endpoint up on which approval was based to clinical benefit, or the observed clinical based to clinical benefit, or the observed clinical benefit to ultimate outcomes.”benefit to ultimate outcomes.”

88

Device LawDevice Law

The FDA has no corresponding The FDA has no corresponding rule for devices.rule for devices.

99

Validation vs. Validation vs. Qualification of Qualification of SurrogatesSurrogates Prentice criteria are most difficult to Prentice criteria are most difficult to

achieve.achieve. Is the surrogate “reasonably likely to Is the surrogate “reasonably likely to

predict clinical benefit”?predict clinical benefit”? How much does knowledge of the How much does knowledge of the

surrogate contribute to the prediction surrogate contribute to the prediction of the primary endpoint?of the primary endpoint?

Assoc. Comm. Janet Woodcock refers Assoc. Comm. Janet Woodcock refers to the “qualification” of surrogatesto the “qualification” of surrogates

1010

ExamplesExamples

Progression-Free Survival (PFS) at Progression-Free Survival (PFS) at 3 years as a surrogate for Overall 3 years as a surrogate for Overall Survival (OS) at 5 years for Survival (OS) at 5 years for adjuvant colorectal cancer adjuvant colorectal cancer (Sargent et al)(Sargent et al)

CD4 count and HIV-1 RNA for CD4 count and HIV-1 RNA for AIDS. AIDS.

1111

What is a Drug-Eluting What is a Drug-Eluting Stent?Stent?Example: Cordis’ Cypher™ Sirolimus-Eluting Coronary Stent

Stent Platform Stent Platform & Delivery & Delivery SystemSystem

Carrier(s)Carrier(s) DrugDrug

Components

1212

Drug-Eluting Coronary Drug-Eluting Coronary StentsStents

Drug-eluting stents have dramatically Drug-eluting stents have dramatically reduced the restenosis rate compared to BM reduced the restenosis rate compared to BM stents.stents.

Target Lesion Revascularization (TLR) is often Target Lesion Revascularization (TLR) is often the ultimate (true) endpoint of interest at 9 the ultimate (true) endpoint of interest at 9 months. TLR is any repeat percutaneous months. TLR is any repeat percutaneous intervention of the target lesion or bypass intervention of the target lesion or bypass surgery of the target lesion.surgery of the target lesion.

Surrogate candidate: Late luminal loss is the Surrogate candidate: Late luminal loss is the difference in millimeters between the difference in millimeters between the diameter of a stented segment post-diameter of a stented segment post-procedure compared with the follow-up procedure compared with the follow-up angiogram at 6 or 9 months, a continuous angiogram at 6 or 9 months, a continuous measure.measure.

1313

Types of Late Loss in Types of Late Loss in DESDES

Late Loss in-stent—late loss within the Late Loss in-stent—late loss within the length of the stentlength of the stent

Late Loss in-segment—late loss within Late Loss in-segment—late loss within the stent plus 5 mm on either sidethe stent plus 5 mm on either side

Late Loss can be measured either in Late Loss can be measured either in mm or as a percentage of the mm or as a percentage of the (expanded) blood vessel lumen (expanded) blood vessel lumen diameter immediately after a stent diameter immediately after a stent procedure. This latter is referred to as procedure. This latter is referred to as Percent Diameter Stenosis (%DS).Percent Diameter Stenosis (%DS).

1414

Drug-Eluting StentDrug-Eluting Stent

1515

Late LossLate Loss

This variable Late Loss (LL) does not save This variable Late Loss (LL) does not save time since the angiography is at virtually the time since the angiography is at virtually the same time as TLR.same time as TLR.

The interest in LL is related to sample size The interest in LL is related to sample size reduction associated with the use of a reduction associated with the use of a continuous as opposed to a binary outcome.continuous as opposed to a binary outcome.

There is a possible concern about the There is a possible concern about the measurement error since LL relies on two measurement error since LL relies on two angiographies at two time points and the angiographies at two time points and the associated diameter measurements.associated diameter measurements.

There have now been a number of There have now been a number of randomized trials involving drug-eluting randomized trials involving drug-eluting stents.stents.

1616

Why LL Why LL MayMay be be Plausible as a Plausible as a SurrogateSurrogate

The main reason to do TLR is that The main reason to do TLR is that there is evidence that there has been there is evidence that there has been narrowing, and this is confirmed with narrowing, and this is confirmed with imaging. So in a study in which every imaging. So in a study in which every patient undergoes angiography at 6 or patient undergoes angiography at 6 or 9 months, the result could be the 9 months, the result could be the decision to do TLR. It is unusual to do decision to do TLR. It is unusual to do such imaging in the real world without such imaging in the real world without some clinical symptoms.some clinical symptoms.

In short, it In short, it couldcould be directly in the be directly in the causal pathway.causal pathway.

1717

Late Loss and TLRLate Loss and TLR

When restenosis hits 50% or more When restenosis hits 50% or more most interventionalists agree to most interventionalists agree to reintervene.reintervene.

lLogistic regression and ROC methods lLogistic regression and ROC methods are beig used to investigate the are beig used to investigate the relationship of LL compared to TLR.relationship of LL compared to TLR.

At this point, FDA has not agreed to At this point, FDA has not agreed to the general acceptance of LL or %DS the general acceptance of LL or %DS as a surrogate for TLR.as a surrogate for TLR.

1818

Intermediate Temporal Intermediate Temporal EndpointsEndpoints An intermediate endpoints is identified An intermediate endpoints is identified

by Temple (JAMA, 1999) as a clinical by Temple (JAMA, 1999) as a clinical endpoint but not the true (ultimate) one. endpoint but not the true (ultimate) one.

Here an intermediate temporal endpoint Here an intermediate temporal endpoint is the true (ultimate) endpoint but at an is the true (ultimate) endpoint but at an earlier time point. earlier time point.

One example Age-Related Macular One example Age-Related Macular Degeneration (Buyse et al, 2000) where Degeneration (Buyse et al, 2000) where 6-month visual acuity is used as an 6-month visual acuity is used as an intermediate temporal endpoint for the intermediate temporal endpoint for the true endpoint namely one-year visual true endpoint namely one-year visual acuity.acuity.

1919

Orthopedic ExampleOrthopedic Example

Use 12-month success on a spinal fixation Use 12-month success on a spinal fixation device as the temporal intermediate endpoint device as the temporal intermediate endpoint of the ultimate (true) endpoint of 2-year of the ultimate (true) endpoint of 2-year success.success.

A patient can go from success to failure or A patient can go from success to failure or failure to success.failure to success.

Useful in adaptive designs (Bayesian or Useful in adaptive designs (Bayesian or frequentist)frequentist)

Such models could be used to investigate Such models could be used to investigate whether 12-month success is a reasonable whether 12-month success is a reasonable surrogate for 24-month success. However no surrogate for 24-month success. However no such surrogate has as yet been established.such surrogate has as yet been established.

2020

Prediction ModelsPrediction Models Use of prediction model within a study Use of prediction model within a study

prospectivelyprospectively in a Bayesian or frequentist manner in a Bayesian or frequentist manner Bayesian exampleBayesian example

– Predict the true endpoint from an intermediate Predict the true endpoint from an intermediate temporal one (2 years from 1 year) using some 2-year temporal one (2 years from 1 year) using some 2-year data is required to build (and all the 2-year data to data is required to build (and all the 2-year data to then validate) this model.then validate) this model.

– Use posterior probability to gauge the prediction.Use posterior probability to gauge the prediction.– Concern about the model assumptions.Concern about the model assumptions.– Predictive model for effectiveness may not address Predictive model for effectiveness may not address

long-term safety (this is true for DES example as well) long-term safety (this is true for DES example as well) since “there is no surrogate for safety” (Temple, 1999)since “there is no surrogate for safety” (Temple, 1999)

2121

GeneralizabilityGeneralizability

Can a surrogate be established based Can a surrogate be established based on a single study or only based on a on a single study or only based on a meta-analysis?meta-analysis?

Does the surrogate generalize Does the surrogate generalize – to other studies?to other studies?– to other populations?to other populations?– to other devices? to other devices? – to other companies?to other companies?– to other diseases?to other diseases?– For DES, to other drug coatings for a stent For DES, to other drug coatings for a stent

or to other stent designs for a drug or to other stent designs for a drug coating?coating?

2222

GeneralizabilityGeneralizability

All medical devices and all indicationsAll medical devices and all indications– There is general agreement that meta-There is general agreement that meta-

analysis of high-quality randomized clinical analysis of high-quality randomized clinical trials would be necessary.trials would be necessary.

One or more classes of medical One or more classes of medical devices for some indicationsdevices for some indications

One class of similar medical devices by One class of similar medical devices by one manufacturer for a single one manufacturer for a single indicationindication

2323

Total Product Life Cycle Total Product Life Cycle (TPLC) for Devices(TPLC) for Devices

“Ensuring the Health of the Public Throughout the Total Product “Ensuring the Health of the Public Throughout the Total Product Lifecycle Lifecycle . . . Its Everybody’s Business” . . . Its Everybody’s Business”

2424

CDRH’s Vision of the CDRH’s Vision of the PipelinePipeline

2525

2626

A Number of RCTsA Number of RCTs

Taxus IV, V, VI Taxus stentTaxus IV, V, VI Taxus stent Sirius Cypher stentSirius Cypher stent Ravel CypherRavel Cypher DeliverDeliver Reality CypherReality Cypher Endeavor II, III EndeavorEndeavor II, III Endeavor