Transcript PROMISE
PROspective Multicenter Imaging Study
for Evaluation of Chest Pain
Udo Hoffmann MD
MGH
ACRIN CardioVascular Committee
October 2nd, 2010
What is PROMISE??
A pragmatic randomized controlled trial
Comparing noninvasive testing strategies for
patients with suspected CAD
10,000 subjects at >200 sites
Funded by NHLBI
Study Timeline
Grant Awarded October 2009
Targeted First Subject Enrolled June 2010
First Investigator Meeting Summer 2010
Last Subject Enrolled Summer 2012
Database Lock Fall 2014
Background and Rationale
Evaluation of Chest Pain Syndrome is the most
common clinical cardiology problem
Large and growing costs ($14.1 billion for imaging)
Differing ACC/AHA guideline recommendations
Lack of trial data on effect of imaging care
Calls for studies showing improved health outcomes by
patients / physicians / insurers / policy makers
Imaging for CAD: Wow vs Value?
Imaging: Improved assessment of cardiac function,
anatomy, and pathology
Does this translate into improved diagnostic
accuracy or assessment of risk?
How about improved outcomes? Lower cost?
Critical Questions for NI Testing
What is the population being tested today?
Do current tests perform well for
Diagnosis (yield)?
Prognosis (events)?
What about new technology like CTA?
What is the right way to evaluate NI testing
for CAD diagnosis?
Current Use of Stress Testing in Stable CP:
Low Test Yield and Few Clinical Events
UHC Claims Data:
84,656 pts w/o CAD; M 45-64 yo; W 50-64 yo
CP visit + stress test w/in 30 days
Kaplan Meier plots: 1 year test yield and event rates
Obstructive CAD at Cath: NCDR 2005-2007
376,430 pts without CAD/MI or prior PCI/CABG
Undergoing diagnostic cath to R/O CAD
59% of patients with positive stress tests had no obstructive CAD
on invasive angio (False positive)
NEJM. 2010 362:886-95
What is the Population Currently Being
Tested for Suspected CAD?
Very large numbers of pts being tested
Most are low risk for CAD and for MACE
Bayesian principles preclude high accuracy
Multiple testing choices
Exercise ECG, Stress Echo, Stress Nuclear
All provide functional assessments
Large numbers of pts undergoing cath
Most do not have obstructive CAD
Current practice: Imperfect NI testing strategies and
clinical diagnostic/prognostic assessments
PROMISE Will Address 3 Fundamental Questions
Which is the right noninvasive test for a patient with
new CAD symptoms?
What is the correct role of CTA in the evaluation of
stable chest pain?
Should new imaging tests be required to prove that they
improve outcomes?
The PROMISE Hypothesis
Question: Is functional or anatomic testing the best initial
testing strategy for diagnosis of stable symptoms
concerning for CAD?
Hypothesis: Information derived from an initial anatomic
strategy (CTA) will drive superior health outcomes
compared to a functional strategy (ischemia testing)
Why a CTA-Superior Hypothesis?
Anticipate population with low disease prevalence
15% CAD, 40% Non-Obstructive CAD, 45% Normal
Superior test performance
false negative test results/untreated CAD
coronary events (death, MI, unstable angina)
false positive test results/unnecessary caths
invasive procedures with complications
Better detection of non-obstructive CAD
Improved preventive treatment and adherence
More confidence in CTA results over functional test results
Longer ‘warranty’ period with fewer repeat tests
hospitalizations during follow up
Trial Design Philosophy
General principles of a pragmatic trial
Effectiveness, not efficacy
Large, simple study with real world care
Maximize generalizability
New paradigm for imaging research
Prospective and randomized
Clinical endpoints
Goal: Change care; require demo of clinical superiority
Balancing efficacy and effectiveness
Site certification and testing quality control - Dx Testing Core
Optimal medical therapy - 1 and 2 prevention sheets
Assure ‘Best practices; Usual care’
PROMISE Trial Design
Symptoms suspicious for significant CAD,
Requiring non-emergent noninvasive testing
Randomization
Anatomic strategy
64+ slice CTA
Functional strategy
Pharmacologic
Stress imaging
Exercise ECG or
Exercise Imaging
Clinical results immediately available to care team
Subsequent testing/mgmt per care team + guideline care
Average f/u 30 months
1º = 30 mo death, MI, Complications, UA hosp
2º = MACE components, Costs, QOL
Safety: Radiation exposure
PROMISE Trial – Inclusion Criteria
Stable symptoms suspicious for significant CAD,
Requiring non-emergent noninvasive testing
No prior W/U for this episode of CP
Planned non-invasive evaluation
Men: > 55 years
Women: > 65 years
Men: 45- 54 years
Women: 50- 64 years
+ One risk factor
Randomize
Exclusion Criteria
Diagnosed or suspected ACS; Unstable
Known CAD, recent CV eval or known heart disease
MI, PCI, CABG or CAD ≥50% lesion
Cath or NI CV test for CAD <12 months
Other causes of sxs: HCM, heart failure, etc
Contraindication to radiation exposure, beta
blockers or contrast agents
Unable to participate in long term follow up
Patient Flow and Follow Up
Screening, enrollment, randomization
Blood biomarker and Omics repository
Randomized test performed w/in 30 days
Images, ECGs and cath films repository
Subsequent care per site MD
Site f/u visit or phone - 60 days
DCRI F/U mail and phone – q 6 mos for 2-4y
Assessments of symptoms, interval events, IF
f/u, medications, CV risk Rx, QOL, costs
An Imaging Research Paradigm Shift
New paradigm for imaging research
Prospective and randomized
Clinical endpoints
Goal: Change care Requires demo of clinical superiority
Balancing efficacy and effectiveness
Diagnostic Testing quality control
Optimal medical therapy - 1 and 2 prevention sheets
Assure ‘Best practices; Usual care’
Qualification of Testing Sites
Equipment, protocol and report template review
Upload 1- 2 test images to ACRIN; reports to DCRI
Meet 100% completeness and quality
Test case review
Functional testing – COCATS II or equivalent
Cardiac CT – COCATS III or review test series
Ongoing QC
100% technical for completeness and quality
20% MD over-read for interpretation
Primary Endpoint
Time to first event in major cardiovascular events including:
Death
Myocardial infarction (MI)
Unstable angina requiring hospitalization
Major complications from CV procedures & testing:
Stroke
Major bleeding
Anaphylaxis – requiring circulatory support
Renal failure - defined as requiring dialysis
Secondary Endpoints
Death or MI or unstable angina hospitalization
Death or MI
Major complications from CV procedures & testing (stroke,
bleeding, renal failure)
Medical costs, resource use, and incremental cost
effectiveness
Health related quality of life
Secondary Safety Endpoint
Cumulative radiation exposure
Statistical Analysis Plan
10,000 subjects
Usual care arm: Estimated rate of death / MI /
USA Hospitalization/ Major procedural
complication over 30 mo: 9.0%
CTA arm: Estimated relative reduction of 20%,
or rate of 7.2%
Primary analysis is CTA superiority
Power > 90% even if event rate to 8%
Power = 87% if effect magnitude to 17.5%
Threshold for superiority at p=0.05 level is an
effect magnitude of 13.5%
Statistical Analysis Plan: Non-Inferiority
Non-inferiority: The results will be evaluated to test
the hypothesis that CTA is not worse than
standard of care by a clinically meaningful amount
Additional pre-specified analyses
Non-inferiority analysis if superiority not met;
Power > 80% for margin 1.10 (HR)
Precision of risk/benefit estimates
Test performance characteristics: dx, px
Why a Secondary Non-Inferiority Hypothesis?
Without prior outcome trials in NI testing, we do not know:
The true effect of standard of care
The acceptable non-inferiority margin
The margin needed to inform clinical care
The margin needed for reimbursement
A primary hypothesis of non inferiority (margin of 47 vs 53%)
would require >15,000 patients
If one test is NOT found to be clinically superior, then calculating
cost effectiveness is impossible:
effectiveness / cost
Clinical choice would be based on cost and safety only…
Costs for cardiovascular procedures are changing rapidly, so
cant calculate an enduring true effect
Substudies
Radiation exposure
Incidental findings
Site vs Core lab test interpretation
Test diagnostic and prognostic accuracy
Performance vs Cath and Event prediction
All modalities
Blood biomarker repository
CV Risk- lipids hsCRP, etc
Myocardial injury- hsTn
‘Omics repository: RNA, DNA, Proteomics
PROMISE Sites
212 overall
164 cardiology, 39 primary care, 6 radiology, 3
ER, 1 anesthesia
Summary
Large, pragmatic RCT evaluating diagnostic
strategies in stable CAD symptoms
10,000 patients; >200 US sites; Up to 4 year FU
Functional (usual care) vs anatomic (CTA) testing
Subsequent usual dx and tx care up to local MD
Uses 1 clinical and 2 economic outcomes
Studies real world effectiveness of testing and
medical care, in multiple specialty settings
Highly experienced investigative team and advisors
You are a part of it!
https://www.promisetrial.org