Help for package HybridDesign

Type:

Package

Title:

Hybrid Design for Phase I Dose-Finding Studies

Version:

1.0

Maintainer:

Heng Zhou <heng.zhou@merck.com>

Description:

The Hybrid design is a combination of model-assisted design (e.g., the modified Toxicity Probability Interval design) with dose-toxicity model-based design for phase I dose-finding studies. The hybrid design controls the overdosing toxicity well and leads to a recommended dose closer to the true maximum tolerated dose (MTD) due to its ability to calibrate for an intermediate dose. More details can be found in Liao et al. 2022 <doi:10.1002/ijc.34203>.

Imports:

testit, ResourceSelection

Depends:

R (≥ 3.6)

License:

GPL-3

Encoding:

UTF-8

RoxygenNote:

7.1.2

NeedsCompilation:

Packaged:

2022-12-14 19:42:26 UTC; zhouheng

Author:

Heng Zhou [aut, cre], Feng Zhou [aut], Jason Liao [aut]

Repository:

CRAN

Date/Publication:

2022-12-15 09:10:02 UTC

Generate modified mTPI Design Decision Boundary

Description

Generate dose escalation and deescalation boundaries of modified Toxicity Probability Interval (mTPI) design with overdose control.

Usage

get_boundary_mtpi(
  target,
  ncohort,
  cohortsize,
  eps1 = 0.05,
  eps2 = 0.05,
  a = 1,
  b = 1,
  cutoff.eli = 0.95,
  tox.control = FALSE,
  cut.tox = 0.8,
  esc.control = FALSE,
  cut.esc = 0.5
)

Arguments

target

target toxicity rate

ncohort

the total number of cohorts

cohortsize

the cohort size

eps1

modified Toxicity Probability Interval (mTPI) design parameter epsilon1. Default: 0.05

eps2

modified Toxicity Probability Interval (mTPI) design parameter epsilon2. Default: 0.05

a

Beta prior shape parameter 1. Default: 1

b

Beta prior shape parameter 2. Default: 1

cutoff.eli

Posterior probability cutoff of eliminating dose due to unacceptable toxicity. Default: 0.95

tox.control

indicator of whether to perform toxicity control. If TRUE, change "stay" to "deescalation" if the posterior probability of DLT rate greater than target+eps2 is greater than the toxicity control cutoff cut.tox

cut.tox

toxicity control cutoff. Default: 0.8

esc.control

indicator of whether to perform escalation control. If TRUE, change decision of "escalation" to "stay" if the posterior probability of DLT rate less than target-eps1 is greater than the escalation control cutoff cut.esc

cut.esc

escalation control cutoff. Default: 0.5

Value

This function returns the table of escalation and deescalation boundaries.

Examples

get_boundary_mtpi(target=0.30, ncohort=10, cohortsize=3)

Generate operating characteristics for single-agent dose-finding studies using the Hybrid design

Description

Obtain the operating characteristics of the Hybrid design for single-agent dose-finding studies by simulation.

Usage

get_oc_hybrid(trueMTD, trueDLTvec, dose, target, ncohort, cohortsize,
                     eps1=0.05, eps2=0.05,a=1, b=1, cutoff.eli=0.95,
                     tox.control=TRUE, cut.tox=0.8, esc.control=FALSE, cut.esc=0.5,
                     ntrial, seednum=10000)

Arguments

trueMTD

the dosage of true maximum tolerated dose (MTD)

trueDLTvec

a vector of true dose-limiting toxicity (DLT) rates at each dose level

dose

a vector containing the numerical dosage of each dose level

target

target toxicity rate

ncohort

the total number of cohorts

cohortsize

the cohort size

eps1

mTPI design parameter epsilon1. Default: 0.05

eps2

mTPI design parameter epsilon2. Default: 0.05

a

Beta prior shape parameter 1. Default: 1

b

Beta prior shape parameter 2. Default: 1

cutoff.eli

Posterior probability cutoff of eliminating dose due to unacceptable toxicity. Default: 0.95

tox.control

cut.tox

toxicity control cutoff. Default: 0.8

esc.control

cut.esc

escalation control cutoff. Default: 0.5

ntrial

the total number of trials to be simulated

seednum

the random seed for simulation

Value

This function returns the operating characteristics of the Hybrid design as a list, including: (1) Percentage of correct selection of the true MTD in all simulated trials, (2) Percentage of selecting a dose above MTD in all simulated trials, (3) Percentage of selecting a dose below MTD in all simulated trials, (4) Average number of patients treated at MTD in all simulated trials.

Examples


get_oc_hybrid(trueMTD=12, trueDLTvec=c(0.15,0.20,0.25,0.30,0.35), dose=c(3, 6, 12, 18, 24),
              target=0.25, ncohort=10, cohortsize=3, eps1=0.05, eps2=0.05, a=1, b=1,
              cutoff.eli=0.95, tox.control=TRUE, cut.tox=0.8, esc.control=FALSE, cut.esc=0.5,
              ntrial=100, seednum=10000)

Implement Hybrid design with real data

Description

Obtain decision for the next dose level to be tested given current trial data.

Usage

hybrid(dose, nDLT, npts, currdose, nextdose=0, target, ncohort, cohortsize,
              eps1=0.05, eps2=0.05, a=1, b=1, cutoff.eli=0.95, tox.control=TRUE,
              cut.tox=0.8, esc.control=FALSE, cut.esc=0.5, regrule)

Arguments

dose

a vector containing the numerical dosage of each dose level

nDLT

a vector containing the number of patients who experienced dose-limiting toxicity at each dose level

npts

a vector containing the number of patients at each dose level

currdose

the dosage at the current dose level

nextdose

the dosage of next higher dose level; could be an intermediate dose

target

the target toxicity rate

ncohort

the total number of cohorts

cohortsize

the cohort size

eps1

modified Toxicity Probability Interval (mTPI) design parameter epsilon1. Default: 0.05

eps2

modified Toxicity Probability Interval (mTPI) design parameter epsilon2. Default: 0.05

a

Beta prior shape parameter 1. Default: 1

b

Beta prior shape parameter 2. Default: 1

cutoff.eli

Posterior probability cutoff of eliminating dose due to unacceptable toxicity. Default: 0.95

tox.control

cut.tox

toxicity control cutoff. Default: 0.8

esc.control

cut.esc

escalation control cutoff. Default: 0.5

regrule

indicator of whether to apply additional overdose control rule

Value

This function returns the decision of implementing the Hybrid design with real trial data as a list, including: (1) dose transition boundaries of modified mTPI design, (2) decision table of modified mTPI design, (3) the decision given current data, (4) the summary table of tested dose levels

Examples

hybrid(dose=c(2,4,8,16,22,28,40), nDLT=c(0,0,0,0,1,0,2), npts=c(3,3,4,6,9,5,16), currdose=40,
       nextdose=54, target=0.3, ncohort=10, cohortsize=3, eps1=0.05, eps2=0.05, a=1, b=1,
       cutoff.eli=0.95, tox.control=TRUE, cut.tox=0.8, regrule=TRUE)

Select the maximum tolerated dose (MTD) for single-agent dose-finding studies

Description

Select the maximum tolerated dose (MTD) when the single-agent dose-finding study is completed

Usage

hybrid_MTD_selection(target, dose, npts, nDLT, elimdose)

Arguments

target

the target toxicity rate

dose

a vector containing the numerical dosage of each dose level

npts

a vector containing the number of patients treated at each dose level

nDLT

a vector containing the number of patients who experienced dose-limiting toxicity at each dose level

elimdose

the dosage at the dose level which is excluded due to excessive toxicity

Details

hybrid.MTD.selection() selects the MTD based on isotonic estimates of toxicity probabilities. The isotonic estimates are obtained by the pooled-adjacent-violators algorithm (PAVA) (Barlow, 1972 <doi: 10.1080/01621459.1972.10481216>).

Value

The selected dosage as MTD

Note

The dose levels above elim are all excluded for MTD selection.

Examples

hybrid_MTD_selection(target=0.3, dose=c(2,4,8,16,22,28,40), npts=c(2,4,8,16,22,28,40),
                     nDLT=c(0,0,0,0,1,0,2), elimdose=28)