Help for package SAME

Type:

Package

Title:

Seamless Adaptive Multi-Arm Multi-Stage Enrichment

Version:

0.1.0

Maintainer:

Chengxue Zhong <czhong9106@gmail.com>

Description:

Design a Bayesian seamless multi-arm biomarker-enriched phase II/III design with the survival endpoint with allowing sample size re-estimation. James M S Wason, Jean E Abraham, Richard D Baird, Ioannis Gournaris, Anne-Laure Vallier, James D Brenton, Helena M Earl, Adrian P Mander (2015) <doi:10.1038/bjc.2015.278>. Guosheng Yin, Nan Chen, J. Jack Lee (2018) <doi:10.1007/s12561-017-9199-7>. Ying Yuan, Beibei Guo, Mark Munsell, Karen Lu, Amir Jazaeri (2016) <doi:10.1002/sim.6971>.

License:

GPL-2

Encoding:

UTF-8

LazyData:

true

RoxygenNote:

7.2.1

Imports:

boot, rjags, coda, extraDistr, survival, ggplot2, expint

Suggests:

testthat, mockery, knitr, rmarkdown

Depends:

R (≥ 3.3.0)

NeedsCompilation:

Packaged:

2023-01-13 14:24:41 UTC; alexz

Author:

Chengxue Zhong [aut, cre], Haitao Pan [aut], Hongyu Miao [aut]

Repository:

CRAN

Date/Publication:

2023-01-16 10:10:08 UTC

Function to identify the most promising treatment-biomarker-linked subgroup

Description

This function is used to estimate the effect size of each subgroup and to select the most promising subgroup.

Usage

conduct.phase2(formula, surv, event, data)

Arguments

formula

a formula object, with the combinations of treatment and biomarker term, e.g., formula = "T1:B1+T1:B2+T2:B1+T2:B2"

surv

survival time

event

the status indicator, 0=alive, 1=dead

data

a data.frame in which to interpret the variables named in the formula

Value

conduct.phase2() select the most effective subgroup and returns the estimated hazard ratio.

Examples

conduct.phase2(formula = "T1:B1+T1:B2+T2:B1+T2:B2", surv = "surv",
event = "death", data = "example.1")

Function to estimate the hazard ratios and other statistics of the selected subgroup

Description

This function is used to estimate the effect size of the selected subgroup.

Usage

conduct.phase3(data, eta, theta)

Arguments

data

a data.frame in which to interpret the variables named in the formula

eta

a cutoff probability for the strength of evidence for decision-making

theta

a clinically meaningful treatment effect size defined by clinicians

Value

conduct.phase3()

Examples

conduct.phase3(example.2,eta=0.8, theta=0.95)

A Time-to-event dataset containing the time and other attributes of 643 patients.

Description

A Time-to-event dataset containing the time and other attributes of 643 patients.

Usage

example.1

Format

A data frame with 643 rows and 6 variables:

T1: binary variable, receive treatment 1=1, not receive treatment 1=0
T2: binary variable, receive treatment 2=1, not receive treatment 2=0
B1: binary variable, biomarker 1 positive=1, biomarker 1 negative=0
B2: binary variable, biomarker 2 positive=1, biomarker 2 negative=0
death: the status indicator, alive=0, dead=1
surv: survival time or follow up time

...

A Time-to-event dataset containing the time and other attributes of 643 patients.

Description

A Time-to-event dataset containing the time and other attributes of 643 patients.

Usage

example.2

Format

A data frame with 643 rows and 6 variables:

T1: binary variable, receive treatment 1=1, not receive treatment 1=0
T2: binary variable, receive treatment 2=1, not receive treatment 2=0
B1: binary variable, biomarker 1 positive=1, biomarker 1 negative=0
B2: binary variable, biomarker 2 positive=1, biomarker 2 negative=0
death: the status indicator, alive=0, dead=1
surv: survival time or follow up time
survtime: survival time or follow up time
treatments: categorical vairable, indicating treatments received

...

Function to calibrate the cutoff points under null hypothesis

Description

This function is used to calibrate the cutoff points under null hypothesis using a multi-arm multi-stage biomarker-enriched design with time-to-event endpoints.

Usage

find.cutoffs(
  median.c,
  K,
  L,
  lfu,
  alpha,
  power,
  accrate,
  theta,
  bio.preva,
  FAtime.phase3,
  N.iter
)

Arguments

median.c

The median survival time for control group

K

Number of biomarkers

L

Information fraction in terms of the accumulative events in phase II stage, e.g., K = c(1/4,1/2,1)

lfu

Follow-up time

alpha

One-sided familywise error rate

power

Power

accrate

Accrual rate

theta

A clinically meaningful treatment effect size defined by clinicians

bio.preva

Prevalence of biomarker(s)

FAtime.phase3

the study ending time of phase III

N.iter

Number of iterations

Value

find.cutoffs() returns the calibrated cutoff points that can control the type I error rate.

Examples


find.cutoffs(median.c=12,K=2,L=c(1/4,1/2,1),lfu=0,alpha=0.05,power=0.9,
             accrate=15,theta=log(1.25),bio.preva=c(0.4,0.6),FAtime.phase3=48,
             N.iter=3)

Function to simulate Bayesian seamless multi-arm biomarker-enriched phase II/III designs

Description

This function finds the required number of events
using a multi-arm multi-stage biomarker-enriched design with time-to-event
endpoints.

Usage

sim.trial(
  median.c,
  hr,
  K,
  L,
  lfu,
  alpha,
  power,
  accrate,
  theta,
  bio.preva,
  FAtime.phase3,
  N.iter
)

Arguments

median.c

The median survival time for control group

hr

Alternative hazard ratio

K

Number of biomarkers

L

Information fraction in terms of the accumulative events in phase II stage, e.g., K = c(1/4,1/2,1)

lfu

Follow-up time

alpha

One-sided familywise error rate

power

Power

accrate

Accrual rate

theta

A clinically meaningful treatment effect size defined by clinicians

bio.preva

Prevalence of biomarker(s)

FAtime.phase3

the study ending time of phase III

N.iter

Number of iterations

Value

sim_trial() returns the nominal type I error rate and calibrated cutoff points, nominal power under user-defined hypothesis, empirical power under user-defined number of simulations, the duration of trial(time), the number of events (num_evs), the number of patients (num_pts) from different stages. The function can also display the number of events and patients under the selected subgroup, the distribution of decision zones and the estimated hazard ratio for the final analysis.

Examples


sim.trial(median.c=12,hr=c(1,1,1,0.6),K=2,L=c(1/4,1/2,1),lfu=0,
          alpha=0.05,power=0.9,accrate=15,theta=log(1.25),
          bio.preva=c(0.4,0.6),FAtime.phase3=48,N.iter=5)

Function to simulate Bayesian seamless multi-arm biomarker-enriched phase II/III designs with user-defined cutoff points

Description

This function finds the required number of events using a multi-arm multi-stage biomarker-enriched design with time-to-event endpoints with the user-defined cutoff points.

Usage

sim.trial.2(
  median.c,
  hr,
  K,
  L,
  lfu,
  alpha,
  power,
  accrate,
  theta,
  bio.preva,
  FAtime.phase3,
  eta,
  futility,
  superiority,
  N.iter
)

Arguments

median.c

The median survival time for control group

hr

Alternative hazard ratio

K

Number of biomarkers

L

Information fraction in terms of the accumulative events in phase II stage, e.g., K = c(1/4,1/2,1)

lfu

Follow-up time

alpha

One-sided family-wise error rate

power

Power

accrate

Accrual rate

theta

A clinically meaningful treatment effect size defined by clinicians

bio.preva

Prevalence of biomarker(s)

FAtime.phase3

the study ending time of phase III

eta

A cutoff probability for the strength of evidence for decision-making and defined by user.

futility

cutoff point for futility termination

superiority

cutoff point for superiority termination

N.iter

Number of iterations

Value

sim.trial.2() returns the nominal type I error rate, nominal power under user-defined hypothesis, empirical power under user-defined number of simulations, the duration of trial(time), the number of events (num_evs), the number of patients (num_pts) from different stages. The function can also display the number of events and patients under the selected subgroup, the distribution of decision zones and the estimated hazard ratio for the final analysis.

Examples


 sim.trial.2(median.c=12,hr=c(1,1,1,0.6),K=2,L=c(1/4,1/2,1),lfu=0,alpha=0.05,
             power=0.9,accrate=15,theta=log(1.25),bio.preva=c(0.4,0.6),
             FAtime.phase3=48,eta=0.2,futility=0.1,superiority=0.9,
             N.iter=3)