R: Generate operating characteristics for drug combination...

get.oc.comb {BOIN}

R Documentation

Generate operating characteristics for drug combination trials

Description

Obtain the operating characteristics of the BOIN design or waterfall design for drug combination trials. The BOIN design is to find a MTD, and the waterfall design is to find the MTD contour (i.e., multple MTDs in the dose matrix)

Usage

get.oc.comb(target, p.true, ncohort, cohortsize, n.earlystop=NULL, startdose=c(1, 1),
                   titration=FALSE,p.saf=0.6*target, p.tox=1.4*target, cutoff.eli=0.95,
                   extrasafe=FALSE,offset=0.05, ntrial=1000, mtd.contour=FALSE,
                   boundMTD=FALSE, seed=6)

Arguments

`target`	the target DLT rate
`p.true`	a `J*K` matrix `(J<=K)` containing the true toxicity probabilities of combinations with `J` dose levels of agent A and `K` dose levels of agent B
`ncohort`	a `1*J` vector specifying the number of cohorts for each of `J` subtrials if `mtd.contour=TRUE`; Otherwise, a scalar specifying the total number of cohorts for the trial.
`cohortsize`	the cohort size
`n.earlystop`	the early stopping parameter. If the number of patients treated at the current dose reaches `n.earlystop`, stop the trial or subtrial and select the MTD based on the observed data. When the waterfall design is used to find the MTD contour, `n.earlystop=12` by default.
`startdose`	the starting dose combination level for drug combination trial
`titration`	set `titration=TRUE` to perform dose escalation with cohort size = 1 to accelerate dose escalation at the begining of the trial.
`p.saf`	the highest toxicity probability that is deemed subtherapeutic (i.e. below the MTD) such that dose escalation should be undertaken. The default value is `p.saf=0.6*target`.
`p.tox`	the lowest toxicity probability that is deemed overly toxic such that deescalation is required. The default value is `p.tox=1.4*target`.
`cutoff.eli`	the cutoff to eliminate an overly toxic dose for safety. We recommend the default value of (`cutoff.eli=0.95`) for general use.
`extrasafe`	set `extrasafe=TRUE` to impose a more stringent stopping rule
`offset`	a small positive number (between 0 and 0.5) to control how strict the stopping rule is when `extrasafe=TRUE`. A larger value leads to a more strict stopping rule. The default value `offset=0.05` generally works well.
`ntrial`	the total number of trials to be simulated
`mtd.contour`	set `mtd.contour=TRUE` to select the MTD contour (claiming multiple MTDs). Otherwise, BOIN design is used to search for a single MTD.
`boundMTD`	set `boundMTD=TRUE` to impose the condition: the isotonic estimate of toxicity probability for the selected MTD must be less than de-escalation boundary.
`seed`	the random seed for simulation

Details

The operating characteristics of the BOIN design or waterfall design are generated by simulating trials under the prespecified true toxicity probabilities of the investigational dose combinations. If titration=TRUE, we perform dose escalation with cohort size = 1 at the begining of the trial: starting from startdose, if no toxicity is observed, we escalate the dose; otherwise, the titration is completed and we switch to cohort size = cohortsize. Titration accelerates the dose escalation and is useful when low doses are believed to be safe.

The BOIN and waterfall designs have two built-in stopping rules: (1) stop the trial/subtrial if the lowest dose is eliminated due to toxicity, and no dose should be selected as the MTD; and (2) stop the trial/subtrial and select the MTD if the number of patients treated at the current dose reaches n.earlystop. The first stopping rule is a safety rule to protect patients from the case in which all doses are overly toxic. The rationale for the second stopping rule is that when there is a large number (i.e., n.earlystop) of patients assigned to a dose, it means that the dose-finding algorithm has approximately converged. Thus, we can stop the trial/subtrial early and select the MTD to save sample size and reduce the trial duration.

For some applications, investigators may prefer a more strict safety stopping rule than rule (1) for extra safety when the lowest dose is overly toxic. This can be achieved by setting extrasafe=TRUE, which imposes the following more strict safety stopping rule: stop the trial if (i) the number of patients treated at the lowest dose >=3, and (ii) Pr(toxicity\ rate\ of\ the\ lowest\ dose > \code{target} | data) > \code{cutoff.eli}-\code{offset}. As a tradeoff, the strong stopping rule will decrease the MTD selection percentage when the lowest dose actually is the MTD.

Value

get.oc.comb() returns the operating characteristics of the BOIN combination or waterfall design as a list. For the BOIN combination design, including: (1) true toxicity probability at each dose level ($p.true), (2) selection percentage at each dose level ($selpercent), (3) the number of patients treated at each dose level ($npatients) (4) the number of toxicities observed at each dose level ($ntox), (5) the total number of patients in the trial ($totaln), (6) the total number of toxicities observed for the trial ($totaltox) (7) the pecentage of correct selection ($pcs), (8) the total percentage of patients treated at the MTD ($npercent). (9) the percentage of early stopping without selecting the MTD ($percentstop) For the the waterfall design, including: (1) true toxicity probability at each dose level ($p.true), (2) selection percentage of dose combinations ($selpercent), (3) the number of patients treated at each dose combination ($npatients) (4) the number of toxicities observed at each dose combination ($ntox), (5) the total number of patients in the trial ($totaln), (6) the total number of toxicities observed for the trial ($totaltox) (7) the total percentage of correct selection at the MTD contour ($pcs.contour), (8) the total percentage of patients treated at MTD contour ($npercent.contour) (9) the total percentage of patients treated above MTD contour ($npercent.above.contour) (10) the total percentage of patients treated below MTD contour ($npercent.below.contour)

Note

We should avoid setting the values of p.saf and p.tox very close to the target. This is because the small sample sizes of typical phase I trials prevent us from differentiating the target DLT rate from the rates close to it. The default values provided by get.oc() are strongly recommended, and generally yield excellent operating characteristics.

Author(s)

Suyu Liu, Liangcai Zhang, Yanhong Zhou, and Ying Yuan

References

Liu S. and Yuan, Y. (2015). Bayesian Optimal Interval Designs for Phase I Clinical Trials, Journal of the Royal Statistical Society: Series C, 64, 507-523.

Lin R. and Yin, G. (2017). Bayesian Optimal Interval Designs for Dose Finding in Drug-combination Trials, Statistical Methods in Medical Research, 26, 2155-2167.

Yan, F., Zhang, L., Zhou, Y., Pan, H., Liu, S. and Yuan, Y. (2020).BOIN: An R Package for Designing Single-Agent and Drug-Combination Dose-Finding Trials Using Bayesian Optimal Interval Designs. Journal of Statistical Software, 94(13),1-32.<doi:10.18637/jss.v094.i13>.

Zhang L. and Yuan, Y. (2016). A Simple Bayesian Design to Identify the Maximum Tolerated Dose Contour for Drug Combination Trials, Statistics in Medicine, 35, 4924-4936.

Examples


###### drug-combination trial ######

##### combination trial to find a single MTD  ######

## get the operating characteristics for BOIN design
p.true <- matrix(c(0.01,0.03,0.10,0.20,0.30,
                 0.03,0.05,0.15,0.30,0.60,
                 0.08,0.10,0.30,0.60,0.75), byrow=TRUE, ncol=5)

oc.comb <- get.oc.comb(target=0.3, p.true, ncohort=20, cohortsize=3,
   n.earlystop=12, startdose=c(1,1), ntrial=100)

summary(oc.comb)
plot(oc.comb)


## get the operating characteristics with titration for BOIN design
oc.comb <- get.oc.comb(target=0.3, p.true, ncohort=20, cohortsize=3,
   n.earlystop=12, startdose=c(1,1), titration=TRUE, ntrial=100)
summary(oc.comb)
plot(oc.comb)


##### combination trial to find the MTD contour ######

## find the MTD contour using waterfall design
oc.comb <- get.oc.comb(target=0.3, p.true, ncohort=c(10,5,5), cohortsize=3,
   n.earlystop=12, startdose=c(1,1), ntrial=100, mtd.contour=TRUE)

summary(oc.comb)
plot(oc.comb)

[Package BOIN version 2.7.2 Index]