Execute the CRM

Description

applied_crm is used to execute the continual reassessment method with specified design options to determine the dose for the next subject.

Usage

applied_crm(prior, target, tox, level, no_skip_esc = TRUE,
  no_skip_deesc = TRUE, global_coherent_esc = TRUE, stop_func = NULL,
  ...)

Arguments

Details

For maximum likelihood estimation, the variance of the estimate of beta (post.var) is approximated by the posterior variance of beta with a dispersed normal prior.

The empiric model is specified as F(d, beta) = d^exp(beta). The logistic model is specified as logit (F(d,beta)) = intcpt + exp(beta) * d. For method="bayes", the prior on beta is normal with mean 0. Exponentiation of beta ensures an increasing dose-toxicity function.

This function is largely a wrapper for the dfcrm function crm. It provides functionality for additional design choices for the CRM including global coherency and stopping for excess toxicity and stopping when sufficient number of subjects are dosed at MTD.

Value

An object of class "mtd" is returned as per package "dfcrm", additional information is provided if a stopping function is used.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)
applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
            global_coherent_esc = TRUE, stop_func = NULL)

Simulate CRM trials using specified design options

Description

applied_crm_sim is used to simulate trials using the continual reassessment method with specified design options to determine the operating characteristics.

Usage

applied_crm_sim(true_tox, prior, target, max_sample_size, first_dose,
    num_sims, cohort_size = 1, dose_func = applied_crm, ...)

Arguments

Value

A list containing two further lists. The first of these lists contains the operating charateristics of the design, the second contains the underlying data for each of the simulation iterations.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

# It may take quite long for large num_sims
prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
true_tox <- c(0.15, 0.25, 0.45)
first_dose <- 1
num_sims <- 5  # recommend doing 5000 simulations for the final design

applied_crm_sim(true_tox, prior, target, max_sample_size = 30, first_dose,
                num_sims, cohort_size = 1, dose_func = applied_crm)

Execute the TITE-CRM

Description

applied_titecrm is used to execute the time-to-event continual reassessment method with specified design options to determine the dose for the next subject.

Usage

applied_titecrm(prior, target, tox, level, followup, obswin,
    no_skip_esc = TRUE, no_skip_deesc = TRUE, global_coherent_esc = TRUE,
    stop_func = NULL, ...)

Arguments

Details

The adaptive weighting scheme is given in Cheung and Chappell (2000) given in the reference list.

Value

An object of class "mtd" is returned as per package "dfcrm", additional information is provided if a stopping function is used.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Cheung, Y. K. and Chappell, R. (2000). Sequential designs for phase I clinical trials with late-onset toxicities. Biometrics 56:1177-1182.

Examples

prior    <- c(0.1, 0.3, 0.5)
target   <- 0.2
tox      <- c(0, 0, 1, 0, 1, 1)
level    <- c(1, 1, 1, 2, 2, 2)
followup <- c(96, 82, 77, 60, 51, 44)
obswin   <- 80

applied_titecrm(prior = prior, target = target, tox = tox,level = level,
                followup = followup, obswin = obswin)

Simulate TITE-CRM trials using specified design options

Description

applied_titecrm_sim is used to simulate trials using the time-to-event continual reassessment method with specified design options to determine the operating characteristics.

Usage

applied_titecrm_sim(true_tox, prior, target, max_sample_size,
    first_dose, num_sims, cohort_size = 1, obswin, minfu, recrate, dose_func
    = applied_titecrm, ...)

Arguments

Value

A list containg two further lists. The first of these lists contains the operating charateristics of the design, the second contains the underlying data for each of the simulation iterations.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

# It may take quite long for large num_sims
prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
true_tox <- c(0.05, 0.2, 0.35)
first_dose <- 1
num_sims <- 5  # recommend doing 5000 simulations for the final design
obswin = 80

applied_titecrm_sim(true_tox = true_tox, prior = prior, target = target,
                              max_sample_size = 21, first_dose = first_dose,
                              num_sims = num_sims, cohort_size = 3,
                              obswin = obswin, minfu = 20, recrate = 3,
                              dose_func = applied_titecrm)
                              

Simulate TITE-CRM trials using specified design options

Description

applied_titecrmts_sim is used to simulate trials using the two-stage time-to-event continual reassessment method with specified design options to determine the operating characteristics.

Usage

applied_titecrmts_sim(true_tox, prior, target, max_sample_size,
    num_sims, cohort_size = 1, obswin, minfu, recrate, initdes,
    dose_func = applied_titecrm, ...)

Arguments

Value

A list containg two further lists. The first of these lists contains the operating charateristics of the design, the second contains the underlying data for each of the simulation iterations.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

# It may take quite long for large num_sims
prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
true_tox <- c(0.05, 0.2, 0.35)
first_dose <- 1
num_sims <- 5  # recommend doing 5000 simulations for the final design
obswin = 80

applied_titecrmts_sim(true_tox = true_tox, prior = prior, target = target,
                      max_sample_size = 21, num_sims = num_sims,
                      cohort_size = 3, obswin = obswin, minfu = 20,
                      recrate = 3, initdes = c(rep(1, 3), rep(2, 3), rep(3, 15)),
                      dose_func = applied_titecrm)
                      

Produce the Dose Transition Pathways

Description

calculate_dtps is used to produce the dose transition pathways for the continual reassessment method with specified design options. These pathways present the possible model recommendations based on all permumations of trial outcomes.

Usage

calculate_dtps(next_dose, cohort_sizes, prev_tox = c(), prev_dose =
    c(), dose_func = applied_crm, ...)

Arguments

Value

Produces a dataframe containing all possible permutations of outcomes for each cohort based on cohort_sizes and the recommended doses for such permutations.

Examples

prior  <- c(0.1, 0.2, 0.5)
target <- 0.15
prev_tox <- c(0, 0, 0)
prev_dose <- c(2, 2, 2)
cohort_sizes <- c(2, 3)

next_dose = applied_crm(prior = prior, target = target,
                        tox = prev_tox, level = prev_dose)$mtd

dose_func <- applied_crm

DTP = calculate_dtps(next_dose, cohort_sizes, prev_tox = prev_tox,
                      prev_dose = prev_dose, dose_func = applied_crm,
                      prior = prior, target = target)

Produce DTP flow diagram

Description

dtpflow will produce a flow diagram of the possible paths for the next three cohorts of subjects.

Usage

dtpflow(dtptable, cohort.labels = c('C1', 'C2', 'C3'))

Arguments

Details

The function will produce a visual flow diagram for the first three cohorts of the provided dataframe.

Examples

prior  <- c(0.1, 0.2, 0.5)
target <- 0.15
prev_tox <- c(0, 0, 0)
prev_dose <- c(2, 2, 2)
cohort_sizes <- c(2, 3, 3)

next_dose = applied_crm(prior = prior, target = target,
                        tox = prev_tox, level = prev_dose)$mtd

dose_func <- applied_crm

DTP = calculate_dtps(next_dose, cohort_sizes, prev_tox = prev_tox,
                      prev_dose = prev_dose, dose_func = applied_crm,
                      prior = prior, target = target)

dtpflow(dtptable = DTP, cohort.labels = c('C1', 'C2', 'C3'))

Plot of posterior estimates from the CRM

Description

Provides functionality for plotting the posterior estimates of probabilities of toxicity at each dose level for both the most recent update and for past cohort updates if specified.

Usage

plot_crm(crm, dose_labels, cohort_sizes = NULL, file = NULL,
                height = 600, width = 750, dose_func = NULL, ...,
                ylim = c(0, 1), lwd = 1, cex.axis = 1, cex.lab = 1,
                cex = 1, cohort.last = F)

Arguments

Details

Produces a plot of current dose-toxicity estimates including the priors and outputs a .png of plot to current directory if 'file' is provided. Potential for history of estimates by cohort if cohort.sizes is provided; dose_func is required to do this.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

crm <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                   global_coherent_esc = TRUE, stop_func = NULL)

plot_crm(crm, dose_labels = c("1", "2", "3"))

Stopping for consensus

Description

This is a function for use with applied_crm for the stop_func argument. The rule will suggest stopping in the scenario that a particular number of patients has already been treated at the current recommended MTD.

Usage

stop_for_consensus_reached(x, req_at_mtd)

Arguments

Details

This function is an example of a possible stopping function to be used with applied_crm, it will modifiy the 'mtd' class object produced by applied_crm to include a logical value under the name 'stop' indicting whether or not the trial should stop. The package dtpcrm contains a few of these functions for possible use with applied_crm.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

stop_rule <- function(x){
  x <- stop_for_consensus_reached(x, req_at_mtd = 6)
}

crm <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                   global_coherent_esc = TRUE, stop_func = stop_rule)

Stopping for excess toxicity - Empiric method

Description

This is a function for use with applied_crm for the stop_func argument. The rule will suggest stopping in the scenario that the probability of toxicity being greater than a specifed value at a defined dose is greater than some further specified certainty value.

Usage

stop_for_excess_toxicity_empiric(x, tox_lim, prob_cert, dose = 1,
    nsamps = 10^6, suppress_dose = TRUE)

Arguments

Details

This function is an example of a possible stopping function to be used with applied_crm, it will modifiy the 'mtd' class object produced by applied_crm to include a logical value under the name 'stop' indicating whether or not the trial should stop. The package dtpcrm contains a few of these functions for possible use with applied_crm.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

stop_rule <- function(x){
  x <- stop_for_excess_toxicity_empiric(x, tox_lim = 0.25, prob_cert = 0.85)
}

crm <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                   global_coherent_esc = TRUE, stop_func = stop_rule)

Stopping for excess toxicity - Logistic method

Description

This is a function for use with applied_crm for the stop_func argument. The rule will suggest stopping in the scenario that the probability of toxicity being greater than a specifed value at a defined dose is greater than some further specified certainty value.

Usage

stop_for_excess_toxicity_logistic(x, tox_lim, prob_cert, dose = 1,
    nsamps = 10^6, suppress_dose = TRUE)

Arguments

Details

This function is an example of a possible stopping function to be used with applied_crm, it will modifiy the 'mtd' class object produced by applied_crm to include a logical value under the name 'stop' indicating whether or not the trial should stop. The package dtpcrm contains a few of these functions for possible use with applied_crm.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

stop_rule <- function(x){
  x <- stop_for_excess_toxicity_logistic(x, tox_lim = 0.25, prob_cert = 0.85)
}

crm <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                   global_coherent_esc = TRUE, stop_func = stop_rule)

Stopping for sample size reached

Description

This is a function for use with applied_crm for the stop_func argument. The rule will suggest stopping in the scenario that a maximum number of subjects has been recruited.

Usage

stop_for_sample_size(x, max_sample_size)

Arguments

Details

This function is an example of a possible stopping function to be used with applied_crm, it will modifiy the 'mtd' class object produced by applied_crm to include a logical value under the name 'stop' indicting whether or not the trial should stop. The package dtpcrm contains a few of these functions for possible use with applied_crm.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

stop_rule <- function(x){
  x <- stop_for_sample_size(x, max_sample_size = 20)
}

crm <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                   global_coherent_esc = TRUE, stop_func = stop_rule)

Provide a summary of applied_crm output

Description

summary_crm is used to return a dataframe of the summary of the output from applied_crm.

Usage

summary_crm(x)

Arguments

Details

This function takes an object of class "mtd" and produces a dataframe containing a summary of information within the object. Specifically it shows the dose levels, prior probabilities, number of evaluable patients, number of DLTs and the posterior probability estimates along with confidence/probability intervals if estimated in the underlying object.

Value

Dataframe of the summary of the output from applied_crm.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

prior  <- c(0.1, 0.3, 0.5)
target <- 0.2
tox    <- c(0, 0, 1, 0, 1, 1)
level  <- c(1, 1, 1, 2, 2, 2)

crm_obj <- applied_crm(prior, target, tox, level, no_skip_esc = TRUE, no_skip_deesc = TRUE,
                       global_coherent_esc = TRUE, stop_func = NULL)

summary_crm(crm_obj)