EWOC.Comb (version 1.0)

ewoc2simu: Generic EWOC2 simulation

Description

Generic function for simulating EWOC trials for 2 drugs combination

Usage

ewoc2simu(ntrials, nsamples, type, trho00, trho01, trho10, teta, nx, ny, tp, 
Min.Dose.A, Max.Dose.A, Min.Dose.B, Max.Dose.B, alpha, theta, vai, a01, 
b01, a10, b10, a00, b00, a, b, delta1x, delta1y, burn, mm, delta1, seed)
# S3 method for default
ewoc2simu(ntrials, nsamples, type, trho00, trho01, trho10, teta, nx, ny, tp, 
Min.Dose.A, Max.Dose.A, Min.Dose.B, Max.Dose.B, alpha, theta, vai, a01,
b01, a10, b10, a00, b00, a, b, delta1x, delta1y, burn=4000, mm=2000, delta1=0.05, seed)

Value

type

same as input parameter type

parameters

list of input parameters

priors

list of prior parameters

Dose.A

a matrix ntrials x nsamples containing the doses of drug A assigned for each patient in a trial and each trial in the simulation

Dose.B

a matrix ntrials x nsamples containing the doses of drug B assigned for each patient in a trial and each trial in the simulation

Resp

a matrix ntrials x nsamples containing ones and zeros indicating the occurance of DLT (1) and the absence of DLT (0) for each patient in the trial and each trial in the simulation

rho00

a numeric vector ntrials x 1 containing the estimated rho00 parameter for each trial in the simulation

rho01

a numeric vector ntrials x 1 containing the estimated rho01 parameter for each trial in the simulation

rho10

a numeric vector ntrials x 1 containing the estimated rho10 parameter for each trial in the simulation

eta

a numeric vector ntrials x 1 containing the estimated eta parameter for each trial in the simulation

postlow

a matrix ntrials x nsamples/2 containing posterior probability of DLT at lower doses (both 0 for durg A and B) at each step in a trial and each trial in the simulation

postdlts

a matrix (nx x ny x ntrials) x 4 containing posterior probability of DLT at each dose combination sets in each trial in the simulation. This is used to test whether or not a discrete set of MTDs was selected from a continous MTD curve is kept or dropped. It's avaiable only when type = 'discrete'

Arguments

ntrials

a number indicating the number of trials to be simulated

nsamples

a number indicating the number of patients enrolled for each clinical trial

type

a character indicating the type of design, could be 'continous' or 'discrete' or their initials

trho00

a numeric value indicating the true value of the parameter rho00, the probability of DLT when the levels of drugs A and B are both 0

trho01

a numeric value indicating the true value of the parameter rho01, the probability of DLT when the levels of drugs A and B are 0 and 1, respectively

trho10

a numeric value indicating the true value of the parameter rho10, the probability of DLT when the levels of drugs A and B are 1 and 0, respectively

teta

a numeric value indicating the true value of the eta, the interaction parameter

nx

a numeric value indicating the number of dose levels for drug A. It's only necessary if type = 'discrete'

ny

a numeric value indicating the number of dose levels for drug B. It's only necessary if type = 'discrete'

tp

a numerical vector indicating the true probabilities of DLT at each dose combinations, the order is by Drug B first, only necessary if type = 'discrete'

Min.Dose.A

a numeric value defining the lower bound of the support of the MTD for drug A

Max.Dose.A

a numeric value defining the upper bound of the support of the MTD for drug A

Min.Dose.B

a numeric value defining the lower bound of the support of the MTD for drug B

Max.Dose.B

a numeric value defining the upper bound of the support of the MTD for drug B

alpha

a numerical value defining the probability that dose selected by EWOC is higher than the MTD.

theta

a numeric value defining the proportion of expectd patients to experience a medically unacceptable, dose-limiting toxicity (DLT) if administered the MTD.

vai

a numeric value indicating variable alpha increment for each new cohort

a01

a numeric value for beta prior distribution associated with parameter rho01

b01

a numeric value for beta prior distribution associated with parameter rho01

a10

a numeric value for beta prior distribution associated with parameter rho10

b10

a numeric value for beta prior distribution associated with parameter rho10

a00

a numeric value for beta prior distribution associated with parameter rho00

b00

a numeric value for beta prior distribution associated with parameter rho00

a

a numeric value for gamma prior distribution associated with parameter eta

b

a numeric value for gamma prior distribution associated with parameter eta

delta1x

Maximum dose escalation at each step for drug A, the default is 0.2*(Max.Dose.A-Min.Dose.A if not assigned)

delta1y

Maximum dose escalation at each step for drug B, the default is 0.2*(Max.Dose.B-Min.Dose.B if not assigned)

burn

Number of iterations for adaption, see n.adapt in jags.model for detail

mm

Number of iterations to monitor, see n.iter in code.samples for detail

delta1

Threshold for toxicity

seed

a numeric value used in random number generation

References

Tighiouart M, Li Q and Rogatko A. A Bayesian adaptive design for estimating the maximuym tolerated dose curve using drug combinations in cancer phase I clinical trials. Statistics in Medicine. 2017, 36: 280-290.

Examples

# \dontshow{


# continous 
test1 = ewoc2simu(ntrials=1, nsamples=4, type="c", trho00=0.01,trho01=0.2, trho10=0.9,teta=20, 
Min.Dose.A=0, Max.Dose.A=1, Min.Dose.B=0, Max.Dose.B=1, alpha=0.25, theta=0.20, a01=1,b01=1,
a10=1,b10=1, a00=1,b00=1,a=0.8,b=0.0384)

print(test1)
plot(test1, type="MTD")
plot(test1, type="bias")
plot(test1, type="percent")

# discrete
tp = c(0.03,0.05,0.08,0.05,0.08,0.13,0.08,0.13,0.2,0.13,0.2,0.29,0.2,0.29,0.4,0.29,0.4,0.53)
test2 = ewoc2simu(ntrials=1, nsamples=4, type="d", nx=6, ny=3, tp=tp, 
Min.Dose.A=0, Max.Dose.A=1, Min.Dose.B=0, Max.Dose.B=1, alpha=0.25, theta=0.20, 
a01=1,b01=1,a10=1,b10=1,a00=1,b00=1,a=0.8,b=0.0384)

print(test2)
plot(test2, type="MTD")
plot(test2, type="percent")


# }
# \donttest{
# continous 
test1 = ewoc2simu(ntrials=10, nsamples=40, type="c", trho00=0.01,trho01=0.2, trho10=0.9,teta=20, 
Min.Dose.A=0, Max.Dose.A=1, Min.Dose.B=0, Max.Dose.B=1, alpha=0.25, theta=0.20, a01=1,b01=1,
a10=1,b10=1, a00=1,b00=1,a=0.8,b=0.0384)

print(test1)
plot(test1, type="MTD")
plot(test1, type="bias")
plot(test1, type="percent")

# discrete
tp = c(0.03,0.05,0.08,0.05,0.08,0.13,0.08,0.13,0.2,0.13,0.2,0.29,0.2,0.29,0.4,0.29,0.4,0.53)
test2 = ewoc2simu(ntrials=10, nsamples=40, type="d", nx=6, ny=3, tp=tp, 
Min.Dose.A=0, Max.Dose.A=1, Min.Dose.B=0, Max.Dose.B=1, alpha=0.25, theta=0.20, 
a01=1,b01=1,a10=1,b10=1,a00=1,b00=1,a=0.8,b=0.0384)

print(test2)
plot(test2, type="MTD")
plot(test2, type="percent")
# }