package_tutorial

library(AdjKMCIF)
library(tidyr)
library(DT)
library(data.table)
library(KMsurv)
library(ggplot2)

Covariate-Adjusted KM curves and probability

Example data

data(bmt)
bmt$arm <- bmt$group
bmt$arm = factor(as.character(bmt$arm),levels=c("2","1","3"))
#bmt$z10 = as.character(bmt$z10)
bmt$z3 = as.character(bmt$z3)
bmt$t2 = bmt$t2*12/365.25

Adjusted KM

Coxph model

result1 = adjusted_KM (data = bmt,time = "t2", status = "d3",group = "arm",
                       covlist = c("z1","z3"),stratified_cox = "No", reference_group = NULL)
table_res1 = spread(result1, class, prob)
head(table_res1)
#>         time         1         2         3
#> 1 0.03285421 0.9924879 0.9958479 0.9891992
#> 2 0.06570842 0.9849745 0.9916806 0.9784332
#> 3 0.32854209 0.9774536 0.9874946 0.9676936
#> 4 0.52566735 0.9699628 0.9833105 0.9570340
#> 5 1.05133470 0.9624386 0.9790928 0.9463641
#> 6 1.14989730 0.9548973 0.9748504 0.9357076

adjKM_plot(result1,data=bmt)

Alternative: use ggplot to finalize figure

Users can also use ggplot to customize figures as needed (e.g.change x ticks to every 6 months, change lab names)

ggplot(result1,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

save results

write.csv(table_res1,"cox_KM_res.csv")

Stratified Cox model: Gail & Byar’s approach

result2 = adjusted_KM (data = bmt,time = "t2", status = "d3",group = "arm",
                       covlist = c("z1","z3"),stratified_cox = "Yes", reference_group = "G&B")

table_res2 = spread(result2, class, prob)
head(table_res2)
#>         time         1         2         3
#> 1 0.03285421 0.9740336 0.9816549 0.9780277
#> 2 0.06570842 0.9740336 0.9816549 0.9780277
#> 3 0.32854209 0.9740336 0.9816549 0.9780277
#> 4 0.52566735 0.9740336 0.9816549 0.9561560
#> 5 1.05133470 0.9740336 0.9816549 0.9342545
#> 6 1.14989730 0.9740336 0.9633376 0.9342545

adjKM_plot(result2,data=bmt)

Alternative: use ggplot to finalize figure

ggplot(result2,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

save results

write.csv(table_res2,"GB_KM_res.csv")

Stratified Cox model: Storer’s approach

result3 = adjusted_KM (data = bmt,time = "t2", status = "d3",group = "arm", 
                       covlist = c("z1","z3"),stratified_cox = "Yes", reference_group = "arm:2")

table_res3 = spread(result3, class, prob)
head(table_res3)
#>         time         1         2         3
#> 1 0.03285421 0.9720931 0.9816549 0.9783523
#> 2 0.06570842 0.9720931 0.9816549 0.9783523
#> 3 0.32854209 0.9720931 0.9816549 0.9783523
#> 4 0.52566735 0.9720931 0.9816549 0.9567957
#> 5 1.05133470 0.9720931 0.9816549 0.9352011
#> 6 1.14989730 0.9720931 0.9633376 0.9352011

adjKM_plot(result3,data=bmt)

Alternative: use ggplot to finalize figure

ggplot(result3,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

save results

write.csv(table_res3,"storer_KM_res.csv")

Bootstrap CI

Cox PH model

result1_1 = boot_ci_adj_km(boot_n=100,ci_cut=c(0.025,0.975),data = bmt, time = "t2", status = "d3",group = "arm",
                         covlist = c("z1","z3"), stratified_cox = "No", reference_group = NULL)

Figure

adjKM_CI_plot(result1_1,bmt)

result1 = do.call(rbind,result1_1)

ggplot(data.frame(result1),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_km = do.call(rbind,result1_1)
write.csv(res_km,"boots_km_cox.csv")

Stratified Cox: G&B for event = relapse

result1_2 =  boot_ci_adj_km(boot_n=100,ci_cut=c(0.025,0.975),data = bmt, time = "t2", status = "d3",group = "arm",
                         covlist = c("z1","z3"), stratified_cox = "Yes", reference_group = "G&B")

adjKM_CI_plot(result1_2,bmt)

result2 = do.call(rbind,result1_2)

ggplot(data.frame(result2),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_km = do.call(rbind,result1_2)
write.csv(res_km,"boots_km_cox.csv")

Stratified Cox: Storer for event = relapse

result1_3 =  boot_ci_adj_km(boot_n=100,ci_cut=c(0.025,0.975),data = bmt, time = "t2", status = "d3",group = "arm",
                         covlist = c("z1","z3"), stratified_cox = "Yes", reference_group = "arm:2")

Figure

adjKM_CI_plot(result1_3,bmt)

result3 = do.call(rbind,result1_3)

ggplot(data.frame(result3),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_km = do.call(rbind,result1_3)
write.csv(res_km,"boots_km_cox.csv")

Covariate-Adjusted CIF curves and probability

Example data

data(bmt)
bmt$arm <- bmt$group
bmt$arm = factor(as.character(bmt$arm),levels=c("2","1","3"))
bmt$z3 = as.character(bmt$z3)

bmt$CenCI <- 0
for(ii in 1:137) {
  if (bmt$d3[ii] == 0){
    bmt$CenCI[ii] <- 0
    }else {
    if (bmt$d2[ii] == 1){
      bmt$CenCI[ii] <- 1
  }else{
      bmt$CenCI[ii] <- 2}}
}
bmt$t2 = bmt$t2*12/365.25

Adjusted CIF

FG model for event = relapse

result1 =adjusted_CIF(data = bmt,time = "t2", status = "CenCI",group = "arm", covlist = c("z1","z3"),event_code=1,stratified = "No", reference_group = NULL)
table_res1 = spread(result1, class, prob)
head(table_res1)
#>         time           1           2          3
#> 1 0.03285421 0.007337385 0.003216491 0.01206072
#> 2 0.06570842 0.007337385 0.003216491 0.01206072
#> 3 0.32854209 0.007337385 0.003216491 0.01206072
#> 4 0.52566735 0.007337385 0.003216491 0.01206072
#> 5 1.05133470 0.007337385 0.003216491 0.01206072
#> 6 1.14989730 0.007337385 0.003216491 0.01206072

adjCIF_plot(result1,data=bmt)

Use ggplot to finalize figure

Users can also use ggplot to customize figures as needed (e.g.change x ticks to every 6 months, change lab names)

ggplot(result1,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

save results

write.csv(table_res1,"cox_CIF_res.csv")

Stratified FG: G&B approach for event = relapse

result2 = adjusted_CIF(data = bmt,time = "t2", status = "CenCI",group = "arm",covlist = c("z1","z3"),event_code=1,stratified = "Yes", reference_group = "G&B")
table_res2 = spread(result2, class, prob)
head(table_res2)
#>         time          1          2          3
#> 1 0.03285421 0.02588104 0.01835945 0.02200836
#> 2 0.06570842 0.02588104 0.01835945 0.02200836
#> 3 0.32854209 0.02588104 0.01835945 0.02200836
#> 4 0.52566735 0.02588104 0.01835945 0.02200836
#> 5 1.05133470 0.02588104 0.01835945 0.02200836
#> 6 1.14989730 0.02588104 0.01835945 0.02200836

adjCIF_plot(result2,data=bmt)

save results

write.csv(table_res2,"GB_CIF_res.csv")

Use ggplot to finalize figure

ggplot(result2,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

Stratified FG: Storer’s approach for event = relapse

result3 = adjusted_CIF (data = bmt,time = "t2", status = "CenCI",group = "arm",covlist = c("z1","z3"),event_code=1,stratified = "Yes", reference_group = "arm:2")

table_res3 = spread(result3, class, prob)
datatable(table_res3)%>%
  formatRound(columns=c("1","2","3"),digits = 3)

adjCIF_plot(result3,data=bmt)

Alternative: use ggplot to finalize figure

ggplot(result3,aes(x=time,y = prob, group =class))+
      geom_step(aes(linetype=class,color = class),size=1.5)+
      theme_classic()+
      ylim(c(0,1))+
      scale_x_continuous(name="Time (Months)",breaks = seq(0,84,6))+
      ylab("Probability")

save results

write.csv(table_res3,"storer_CIF_res.csv")

Bootstrap CI

FG model for event = relapse

result1_1 = boot_ci_adj_cif(boot_n=100,ci_cut=c(0.025,0.975),data = bmt,time = "t2", status = "CenCI",group = "arm",covlist = c("z1","z3"),event_code=1,"No",NULL)

Figure

adjCIF_CI_plot(result1_1,bmt)

result1 = do.call(rbind,result1_1)

ggplot(data.frame(result1),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_boot = do.call(rbind,result1_1)
write.csv(res_boot,"boots_cif_cox.csv")

Stratified FG: G&B for event = relapse

result1_2 = boot_ci_adj_cif(boot_n=100,ci_cut=c(0.025,0.975),data = bmt,time = "t2", status = "CenCI",group = "arm",covlist = c("z1","z3"),event_code=1,"Yes","G&B")

Figure

adjCIF_CI_plot(result1_2,bmt)

result2 = do.call(rbind,result1_2)

ggplot(data.frame(result2),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_boot = do.call(rbind,result1_2)
write.csv(res_boot,"boots_cif_cox.csv")

Stratified FG: Storer for event = relapse

# unstratified Fine-Gray regression model 
result1_3 = boot_ci_adj_cif(boot_n=100,ci_cut=c(0.025,0.975),data = bmt,time = "t2", status = "CenCI",group = "arm",covlist = c("z1","z3"),event_code=1,"Yes","arm:2")

Figure

adjCIF_CI_plot(result1_3,bmt)

result3 = do.call(rbind,result1_3)

ggplot(data.frame(result3),aes(x=time,y = t_sub_mean, group =class ))+
  geom_step(aes(linetype=class,color = class),size=1.2)+
  geom_ribbon(aes(ymin=lower,ymax=upper,fill=class),alpha=0.3)+
  ylim(c(0,1))+
  theme_classic()+
  scale_x_continuous(name = "Time (Months)",
    breaks = seq(0, 84, 6)) +
    ylab("Probability")

save bootstrap results

res_boot = do.call(rbind,result1_3)
write.csv(res_boot,"boots_cif_cox.csv")