pfs_scan: Cut simulated PFS data according to the pre-specified scan...

In phe9480/rgs: What the Package Does (One Line, Title Case)

Description

This function cuts the progression data according to the scan intervals. The observed PD date is determined by the immediate scan following the unobservable true PD date. The scan interval should be considered only for events after incorporating drop-off process. Data cutoff process, if needed, should follow the scan interval consideration. The event status should not be changed after considering the scan interval.

Usage

pfs_scan(
  time = data$survTimeCut,
  cnsr = data$cnsrCut,
  treatment = data$treatment,
  scanfreq0 = c(rep(6, 2), rep(9, 100)) * 7/(365.25/12),
  scanfreq1 = c(rep(6, 2), rep(9, 100)) * 7/(365.25/12)
)

Arguments

cnsr	Censor status variable (0: event; 1: censored)
treatment	Treatment arm code (0: control arm; 1: experimental arm)
scanfreq0	A vector of scan frequency for control arm, eg, every 6 weeks for 2 scans, then every 9 weeks thereafter. scanfreq = c(rep(6, 2), rep(9, 1000))*7
scanfreq1	A vector of scan frequency for experimental arm, eg, every 6 weeks for 2 scans, then every 9 weeks thereafter. scanfreq = c(rep(6, 2), rep(9, 1000))*7
PFS	Time to event variable

Examples

#Example: Simulate 1 sample from proportional hazards scenario. 
#Total 600 pts, 1:1 randomization, control median PFS 8 mo; 
#HR = 0.6

time0 = rexp(300, rate=log(2)/12)
time1 = rexp(300, rate=log(2)/12*0.6)
time = c(time0, time1)
cnsr = rep(0, 600)
treatment = c(rep("control", 300), rep("experiment", 300))
s = sample(1:600)
data = data.frame(cbind(time, cnsr))
data$treatment = treatment
data = data[s,]

scanfreq0 = scanfreq1 = c(rep(6, 2), rep(9, 100))*7/(365.25/12)
km<-survival::survfit(survival::Surv(time, 1-cnsr)~treatment, data=data)
plot(km,xlab="Month Since Randomization",ylab="Survival",
main="True unobservable PFS", lty=1:2,xlim=c(0,36))

data$PFS = pfs_scan(time = data$time, cnsr = data$cnsr, 
           treatment = data$treatment, scanfreq0 = scanfreq0, 
           scanfreq1=scanfreq1)

km2<-survival::survfit(survival::Surv(PFS, 1-cnsr)~treatment, data=data)
plot(km2,xlab="Month Since Randomization",ylab="Survival",
main="Observed PFS", lty=1:2,xlim=c(0,36))

phe9480/rgs documentation built on March 1, 2022, 12:26 a.m.