Clinical Trial Simulation To Assist In COPD Trial Planning And Design With A Biomarker-Based Diagnostic: When To Pull The Trigger?

Figures & data

Diagram 1.  CTOORA: (1) Clinical characteristics simulated from a COPD population, with dependencies between exacerbation history and disease progression, (2) Two patient molecular types are simulated given assumed prevalence, (3) Simulated patients are registered according to a mock protocol, and randomized to treatment arms, (4) Biomarker (panel) Dx test is performed, to predict patient molecular type (2), (5) Simulated targeted treatment effect differential by patient molecular type (2), (6) Simulated clinical outcomes given patient histories, treatment, and molecular-type. (7) Biomarker segmentation (assay) merged to clinical outcomes.

Diagram 2. Simulation Plan: Phase IIb trial simulated for study size, RRE, biomarker Dx cutoffs, and COPD phenotype prevalences. At the end of Phase IIb the Go/No Go decision made to pursue Phase III without a biomarker, with a biomarker (enrichment), or discontinue. The overall probability of Phase III success is weighted by the respective end of Phase IIb decision probabilities. Decision and success probabilities estimated from 5,000 simulated trial sequences.

Diagram 3.  Selected Simulation Results, Phase IIb Decision and Phase III Success Probabilities (a) Phase IIb simulated with N = 110 subjects per arm and biomarker Dx test sensitivity of 70%. There was a 46% probability of enriching with a biomarker in Phase III, and a 59% overall probability of Phase III success. (b) Increasing Phase IIb study size to N = 190 subjects per arm increased probability of enriching with biomarker in Phase III to 56%, and overall probability of Phase III success to 73%. (c) As in (b) although with increased Dx test sensitivity of 90%. Probability of enriching with biomarker in Phase III increased to 67%, while the overall probability of Phase III success of 75% was negligible from (b) at 73%.

Table 1.   Simulated Phase IIb Decision and Phase III Success Probabilities with a Biomarker, RRE = 25% to Targeted Treatment in the All Comers Population

Phenotype A		Biomarker Dx Test+					Ph IIb Decision Probability++			Ph III Success Probability†
N/Arm	Prevalence*	Sensitivity	Specificity	Test Pos.%	PPV	NPV	All Comers	Enrichment	Stop	All Comers	Enrichment	Overall‡
110	30%	70%	90%	28%	75%	88%	24%	48%	28%	54%	99%	60%
110	30%	80%	80%	38%	63%	90%	21%	49%	30%	54%	99%	60%
110	30%	90%	70%	48%	56%	94%	18%	50%	32%	55%	98%	59%
110	40%	70%	90%	34%	82%	82%	24%	46%	30%	57%	99%	59%
110	40%	80%	80%	44%	73%	86%	20%	48%	32%	56%	99%	59%
110	40%	90%	70%	54%	67%	91%	16%	51%	33%	55%	97%	58%
110	50%	70%	90%	40%	88%	75%	23%	44%	33%	57%	99%	57%
110	50%	80%	80%	50%	80%	80%	19%	48%	33%	56%	96%	57%
110	50%	90%	70%	60%	75%	88%	14%	50%	36%	58%	94%	55%
140	30%	70%	90%	28%	75%	88%	27%	53%	20%	54%	99%	67%
140	30%	80%	80%	38%	63%	90%	24%	53%	23%	54%	99%	65%
140	30%	90%	70%	48%	56%	94%	18%	58%	24%	55%	98%	67%
140	40%	70%	90%	34%	82%	82%	26%	51%	23%	57%	99%	65%
140	40%	80%	80%	44%	73%	86%	23%	54%	23%	56%	99%	66%
140	40%	90%	70%	54%	67%	91%	17%	59%	24%	55%	97%	67%
140	50%	70%	90%	40%	88%	75%	26%	49%	25%	57%	99%	63%
140	50%	80%	80%	50%	80%	80%	20%	54%	26%	56%	96%	63%
140	50%	90%	70%	60%	75%	88%	16%	58%	26%	58%	94%	64%
190	30%	70%	90%	28%	75%	88%	31%	57%	12%	54%	99%	73%
190	30%	80%	80%	38%	63%	90%	27%	59%	14%	54%	99%	73%
190	30%	90%	70%	48%	56%	94%	21%	64%	15%	55%	98%	74%
190	40%	70%	90%	34%	82%	82%	31%	56%	13%	57%	99%	73%
190	40%	80%	80%	44%	73%	86%	24%	61%	15%	56%	99%	74%
190	40%	90%	70%	54%	67%	91%	18%	67%	15%	55%	97%	75%
190	50%	70%	90%	40%	88%	75%	30%	55%	15%	57%	99%	72%
190	50%	80%	80%	50%	80%	80%	24%	61%	15%	56%	96%	72%
190	50%	90%	70%	60%	75%	88%	17%	68%	15%	58%	94%	74%
285	30%	70%	90%	28%	75%	88%	40%	56%	4%	54%	99%	77%
285	30%	80%	80%	38%	63%	90%	31%	64%	5%	54%	99%	80%
285	30%	90%	70%	48%	56%	94%	23%	71%	6%	55%	98%	82%
285	40%	70%	90%	34%	82%	82%	37%	59%	5%	57%	99%	79%
285	40%	80%	80%	44%	73%	86%	29%	66%	6%	56%	99%	81%
285	40%	90%	70%	54%	67%	91%	20%	73%	7%	55%	97%	82%
285	50%	70%	90%	40%	88%	75%	37%	57%	6%	57%	99%	78%
285	50%	80%	80%	50%	80%	80%	27%	67%	6%	56%	96%	79%
285	50%	90%	70%	60%	75%	88%	18%	76%	6%	58%	94%	82%

*Frequency of phenotype A in simulated COPD population

⁺Summary measures of biomarker based Dx test performance to predict Phenotype A, see Supplment

⁺⁺Probability of end of Phase IIb decision for Phase III design, estimated over 5,000 simulations

^†Respective probabilities of Phase III success, estimated over 5,000 simulations

^‡Weighted average of Phase III all comers or enrichment success probabilities by respective Phase IIb decision probabilities⁺⁺,

see Supplement and Diagram 2

Table 2.   Simulated Phase IIb Decision and Phase III Success Probabilities with a Biomarker, RRE = 20% to Targeted Treatment in the All Comers Population

Phenotype A		Biomarker Dx Test+					Ph IIb Decision Probability++			Ph III Success Probability†
N/Arm	Prevalence*	Sensitivity	Specificity	Test Pos.%	PPV	NPV	All Comers	Enrichment	Stop	All Comers	Enrichment	Overall‡
110	30%	70%	90%	28%	75%	88%	20%	37%	43%	30%	98%	42%
110	30%	80%	80%	38%	63%	90%	17%	37%	46%	29%	94%	40%
110	30%	90%	70%	48%	56%	94%	15%	39%	46%	31%	86%	38%
110	40%	70%	90%	34%	82%	82%	19%	36%	45%	32%	94%	40%
110	40%	80%	80%	44%	73%	86%	17%	38%	45%	31%	88%	39%
110	40%	90%	70%	54%	67%	91%	14%	39%	47%	32%	79%	35%
110	50%	70%	90%	40%	88%	75%	19%	33%	48%	31%	88%	35%
110	50%	80%	80%	50%	80%	80%	17%	37%	46%	32%	79%	35%
110	50%	90%	70%	60%	75%	88%	13%	39%	48%	32%	71%	32%
140	30%	70%	90%	28%	75%	88%	23%	41%	36%	30%	98%	47%
140	30%	80%	80%	38%	63%	90%	20%	44%	36%	29%	94%	47%
140	30%	90%	70%	48%	56%	94%	16%	45%	39%	31%	86%	44%
140	40%	70%	90%	34%	82%	82%	20%	40%	40%	32%	94%	44%
140	40%	80%	80%	44%	73%	86%	18%	44%	38%	31%	88%	44%
140	40%	90%	70%	54%	67%	91%	14%	45%	41%	32%	79%	40%
140	50%	70%	90%	40%	88%	75%	21%	37%	42%	31%	88%	39%
140	50%	80%	80%	50%	80%	80%	17%	42%	41%	32%	79%	39%
140	50%	90%	70%	60%	75%	88%	13%	44%	43%	32%	71%	35%
190	30%	70%	90%	28%	75%	88%	25%	50%	25%	30%	98%	57%
190	30%	80%	80%	38%	63%	90%	21%	51%	28%	29%	94%	54%
190	30%	90%	70%	48%	56%	94%	18%	53%	29%	31%	86%	51%
190	40%	70%	90%	34%	82%	82%	25%	48%	27%	32%	94%	53%
190	40%	80%	80%	44%	73%	86%	21%	51%	28%	31%	88%	51%
190	40%	90%	70%	54%	67%	91%	16%	54%	30%	32%	79%	48%
190	50%	70%	90%	40%	88%	75%	26%	44%	30%	31%	88%	47%
190	50%	80%	80%	50%	80%	80%	21%	49%	30%	32%	79%	45%
190	50%	90%	70%	60%	75%	88%	15%	53%	32%	32%	71%	42%
285	30%	70%	90%	28%	75%	88%	31%	56%	13%	30%	98%	64%
285	30%	80%	80%	38%	63%	90%	24%	59%	17%	29%	94%	62%
285	30%	90%	70%	48%	56%	94%	20%	63%	17%	31%	86%	61%
285	40%	70%	90%	34%	82%	82%	30%	54%	16%	32%	94%	60%
285	40%	80%	80%	44%	73%	86%	24%	60%	17%	31%	88%	60%
285	40%	90%	70%	54%	67%	91%	19%	65%	17%	32%	79%	57%
285	50%	70%	90%	40%	88%	75%	29%	53%	18%	31%	88%	56%
285	50%	80%	80%	50%	80%	80%	23%	60%	17%	32%	79%	55%
285	50%	90%	70%	60%	75%	88%	16%	65%	19%	32%	71%	51%

*Frequency of phenotype A in simulated COPD population

+Summary measures of biomarker based Dx test performance to predict Phenotype A, see Supplment

++Probability of end of Phase IIb decision for Phase III design, estimated over 5,000 simulations

†Respective probabilities of Phase III success, estimated over 5,000 simulations

‡Weighted average of Phase III all comers or enrichment success probabilities by respective Phase IIb decision probabilities⁺⁺,

see Supplement and Diagram 2

Table 3.   Simulated Phase IIb Decision and Phase III Success Probabilities with a Biomarker, RRE = 25% to Targeted Treatment in the All Comers Population with Reduced Sensitivity and Specificity

Phenotype A		Biomarker Dx Test^+					Ph IIb Decision Probability^++			Ph III Success Probability^†
N/Arm	Prevalence*	Sensitivity	Specificity	Test Pos.%	PPV	NPV	All Comers	Enrichment	Stop	All Comers	Enrichment	Overall^‡
110	30%	60%	80%	32%	56%	82%	29%	35%	36%	54%	98%	50%
110	30%	70%	70%	42%	50%	84%	25%	38%	37%	54%	95%	50%
110	30%	80%	60%	52%	46%	88%	22%	42%	36%	55%	91%	50%
110	40%	60%	80%	36%	67%	75%	29%	35%	36%	57%	97%	50%
110	40%	70%	70%	46%	61%	78%	24%	39%	37%	56%	93%	50%
110	40%	80%	60%	56%	57%	82%	21%	41%	38%	55%	89%	48%
110	50%	60%	80%	40%	75%	67%	28%	34%	38%	57%	94%	48%
110	50%	70%	70%	50%	70%	70%	24%	38%	38%	56%	90%	48%
110	50%	80%	60%	60%	67%	75%	20%	42%	38%	58%	86%	48%
140	30%	60%	80%	32%	56%	82%	32%	39%	29%	54%	98%	46%
140	30%	70%	70%	42%	50%	84%	28%	43%	29%	54%	95%	43%
140	30%	80%	60%	52%	46%	88%	25%	46%	29%	55%	91%	40%
140	40%	60%	80%	36%	67%	75%	33%	39%	28%	57%	97%	46%
140	40%	70%	70%	46%	61%	78%	27%	42%	31%	56%	93%	44%
140	40%	80%	60%	56%	57%	82%	23%	48%	29%	55%	89%	39%
140	50%	60%	80%	40%	75%	67%	33%	38%	29%	57%	94%	46%
140	50%	70%	70%	50%	70%	70%	27%	42%	31%	56%	90%	43%
140	50%	80%	60%	60%	67%	75%	21%	48%	31%	58%	86%	39%
190	30%	60%	80%	32%	56%	82%	39%	42%	19%	54%	98%	62%
190	30%	70%	70%	42%	50%	84%	34%	47%	19%	54%	95%	63%
190	30%	80%	60%	52%	46%	88%	26%	53%	21%	55%	91%	62%
190	40%	60%	80%	36%	67%	75%	39%	42%	19%	57%	97%	63%
190	40%	70%	70%	46%	61%	78%	31%	50%	19%	56%	93%	64%
190	40%	80%	60%	56%	57%	82%	25%	55%	20%	55%	89%	63%
190	50%	60%	80%	40%	75%	67%	39%	42%	19%	57%	94%	62%
190	50%	70%	70%	50%	70%	70%	32%	50%	18%	56%	90%	63%
190	50%	80%	60%	60%	67%	75%	25%	55%	20%	58%	86%	62%
285	30%	60%	80%	32%	56%	82%	49%	43%	8%	54%	98%	69%
285	30%	70%	70%	42%	50%	84%	41%	50%	9%	54%	95%	70%
285	30%	80%	60%	52%	46%	88%	31%	59%	9%	55%	91%	71%
285	40%	60%	80%	36%	67%	75%	48%	44%	9%	57%	97%	69%
285	40%	70%	70%	46%	61%	78%	40%	52%	9%	56%	93%	70%
285	40%	80%	60%	56%	57%	82%	30%	61%	9%	55%	89%	71%
285	50%	60%	80%	40%	75%	67%	48%	43%	9%	57%	94%	67%
285	50%	70%	70%	50%	70%	70%	39%	51%	10%	56%	90%	68%
285	50%	80%	60%	60%	67%	75%	30%	61%	9%	58%	86%	70%

*Frequency of phenotype A in simulated COPD population

⁺Summary measures of biomarker based Dx test performance to predict Phenotype A, see Supplment

⁺⁺Probability of end of Phase IIb decision for Phase III design, estimated over 5,000 simulations

^†Respective probabilities of Phase III success, estimated over 5,000 simulations

^‡Weighted average of Phase III all comers or enrichment success probabilities by respective Phase IIb decision probabilities⁺⁺,

see Supplement and Diagram 2

Barnes PJ, Celli BR. Systemic manifestations and comorbidities of COPD. Eur Respir J 2009; 33(5):1165–1185.

 PubMed Web of Science ®Google Scholar

Allen-Ramey FC, Gupta S, DiBonaventura MD, Patient characteristics, treatment patterns, and health outcomes among COPD phenotypes. Int J Chron Obstruct Pulmon Dis 2012; 7:779–787.

 PubMed Web of Science ®Google Scholar

Toy EL, Gallagher KF, Stanley EL, Swensen AR, Duh MS. The economic impact of exacerbations of chronic obstructive pulmonary disease and exacerbation definition: a review. COPD 2010; 7(3):214–228.

 PubMedGoogle Scholar

Mannino DM, Buist AS. Global burden of COPD: risk factors, prevalence, and future trends. Lancet 2007; 370(9589):765–773.

 PubMed Web of Science ®Google Scholar

Lopez AD, Shibuya K, Rao C, Mathers CD, Hansell AL, Held LS, Schmid V, Buist S. Chronic obstructive pulmonary disease: current burden and future projections. Eur Respir J 2006; 27:397–412.

 PubMed Web of Science ®Google Scholar

Richard Russell R, Anzueto A, Weisman I. Optimizing management of chronic obstructive pulmonary disease in the upcoming decade. Int J Chron Obstruct Pulmon Dis 2011; 6:47–61.

 PubMedGoogle Scholar

Vestbo J, Hurd SS, Agusti AG, Jones PW, Vogelmeier C, Anzueto A, Global Strategy for the Diagnosis, Management and Prevention of Chronic Obstructive Pulmonary Disease, GOLD Executive Summary. Am J Respir Crit Care Med 2012; Epub ahead of print.

 Web of Science ®Google Scholar