update

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: SurvivalPath
 Type: Package
 Title: Survival Path models for time-series survival data
-Version: 1.1.0
+Version: 1.1.1
 Author: Shen Lujun and Zhang Tao
 Maintainer: Zhang Tao <[email protected]>
 Description: Build the survival path of patients with a single disease.

NAMESPACE

@@ -7,3 +7,8 @@ export(matchsubgroup)
 export(plotKM)
 export(survivalpath)
 export(timedivision)
+import(dplyr)
+import(survival)
+import(survivalROC)
+import(survminer)
+import(treeio)

R/EvolutionAfterTreatment.R

@@ -1,7 +1,7 @@
 
-#'@title  Calculate the number of people (proportion) assigned to different branches after a certain
+#'@title  Calculate the number of participants (proportion) assigned to different sub-nodes after a certain
 #'node is treated by a specified plan
-#'@description Calculate the number of people (proportion) assigned to different branches after a
+#'@description Calculate the number of people (proportion) assigned to different sub-nodes after a
 #'certain node is treated by a specified plan
 #'@usage EvolutionAfterTreatment(
 #'df,
@@ -15,14 +15,14 @@
 #'@param  mytree  "mytree" in the return result of the survivalpath function
 #'@param source Raw data
 #'@param treatment Choose treatment
-#'@details By specifying the node, using mytree to calculate the relevant branch variables and subsequent nodes,
-#'and using the original data to calculate, filter out the group of subjects who branch to the next node after
-#'the node adopts different treatment plans. Calculate the number of people assigned to different nodes according
+#'@details By specifying the node, using mytree to calculate the relevant bifurcation variables and subsequent nodes,
+#'and using the original data to calculate the number or proportion of subjects who branch to each sub-nodes after
+#'the participants in parent-node adopts different treatment plans. Calculate the number of people assigned to different sub-nodes according
 #'to different treatment methods (proportion).
-#'@return A dataframe whose rows and columns are the next node and treatment plan
+#'@return A dataframe whose rows and columns are the next node and treatment plan, respectively
 #'@export
 #'@examples data("dataset")
-#'dataset = timedivision(X2021data,"ID","Date",period = 90,left_interval = 0.5,right_interval=1.5)
+#'dataset = timedivision(X2021data,"ID","Date",period = 90,left_interval = 0.5,right_interval=0.5)
 #'
 #'time <- list()
 #'status <- list()
@@ -32,11 +32,11 @@
 #'treatment <- list()
 #'for (i in 1:10){
 #'
-#'  data <- dataset[dataset['timenode']==i,]
+#'  data <- dataset[dataset['time_slice']==i,]
 #'
 #'  time <- c(time,list(data['OStime_new']))
 #'
-#'  status <- c(status,list(data['Status_new']))
+#'  status <- c(status,list(data['Status_of_death']))
 #'
 #'  tsid <- c(tsid,list(data['ID']))
 #'
@@ -46,14 +46,14 @@
 #'
 #'  tsdata <- c(tsdata,list(c_data))
 #'
-#'  c_treatment <- subset(data, select = c("Treatment2"))
+#'  c_treatment <- subset(data, select = c("Resection"))
 #'
 #'  treatment <- c(treatment,list(c_treatment))
 #'}
 #'
-#'tsdata <- classifydata(time,status,tsdata,tsid,cutoff=365*1)
+#'tsdata <- classifydata(time,status,tsdata,tsid,predict.time=365*1)
 #'
-#'result <- survivalpath(time,status,tsdata[[1]],tsid,time_slices = 10,treatments = treatment,p.value=0.05,degreeofcorrelation=0.7)
+#'result <- survivalpath(time,status,tsdata[[1]],tsid,time_slices = 8,treatments = treatment,p.value=0.05,degreeofcorrelation=0.7)
 #'
 #'mytree <- result$tree
 #'ggtree(mytree, color="black",linetype=1,size=1.2,ladderize = T, )+
@@ -71,15 +71,15 @@
 #'  theme(legend.title=element_blank(),legend.position = c(0.1,0.9))
 #'
 #'#plot KM curve
-#'treepoints = c(16,23)
+#'treepoints = c(32,55)
 #'plotKM(result$data, treepoints,mytree,risk.table=T)
 #'#Comparing the efficacy of treatment methods by drawing survival curves
-#'treepoints = c(17,22)
-#'compareTreatmentPlans(result$data, treepoints,mytree,dataset,"Treatment")
+#'treepoints = c(32,55)
+#'compareTreatmentPlans(result$data, treepoints,mytree,dataset,"Resection")
 #'
-#'treepoint=16
-#'A = EvolutionAfterTreatment(result$data, treepoint,mytree,dataset,"Treatment")
-#'mytable <- xtabs(~ `Treatment2`+treepoint, data=A)
+#'treepoint=32
+#'A = EvolutionAfterTreatment(result$data, treepoint,mytree,dataset,"Resection")
+#'mytable <- xtabs(~ `Resection`+treepoint, data=A)
 #'prop.table(mytable,1)
 #'
 
@@ -179,7 +179,7 @@ getnodePatients <- function(df,treepoint,mytree,source,treatment){
 
   result <- newdf[,1:3]
 
-  sourcedata <- source[which(source$timenode==time_slice-1),]
+  sourcedata <- source[which(source$time_slice==time_slice-1),]
   #print(dim(sourcedata))
   sourcedata <- sourcedata[sourcedata$ID %in% result$ID,]
 

R/classifydata.R

@@ -1,21 +1,21 @@
-#'@title  The same classification of different risk factors at each time node
+#'@title  Reclassify variables into binary variables based on survivalROC
 #'@description Based on the survivalROC package, using the survival time
-#'and survival state, calculate the optimal threshold for the binary classification
-#'of each variable of the initial node. According to the threshold, the variables of
+#'and survival state, calculate the optimal cutoff for the binary classification
+#'of each variable using data at the first time slice. According to the cutoff, the variables of
 #' different nodes are classified into two categories.
 #'@usage classifydata(
 #'time,
 #'status,
 #'timeslicedata,
 #'tspatientid,
-#'cutoff,
+#'predict.time,
 #'isfill=T
 #')
 #'@param time Event time or censoring time for subjects
 #'@param status Indicator of status, 1 if death or event, 0 otherwise
-#'@param  timeslicedata list object; in turn include the risk factors of subjects at different time nodes
-#'@param tspatientid list object; The unique identification number corresponding to the subject at each time node
-#'@param  predict.time Time point of the ROC curve
+#'@param  timeslicedata list object; each dataframe of the list represent data at individual time slices
+#'@param tspatientid list object; The unique identification number corresponding to the subject at each time slice
+#'@param  predict.time Time point to predict during drawing the ROC curve
 #'@param isfill  Logical value, used to confirm whether to fill in missing data. If it is True, then fill
 #'@details According to the first time node, the survival data is used to calculate the optimal classification threshold
 #'of different risk factors. The risk factor threshold calculated at the first time node is used to calculate
@@ -26,6 +26,7 @@
 #'@return cutoff  Optimal classification threshold
 #'@seealso survivalROC
 #'@export
+#'@import survivalROC
 #'@examples
 #'data("dataset")
 #'
@@ -43,11 +44,11 @@
 #'
 #'for (i in 1:10){
 #'
-#'  data <- dataset[dataset['timenode']==i,]
+#'  data <- dataset[dataset['time_slice']==i,]
 #'
 #'  time <- c(time,list(data['OStime_new']))
 #'
-#'  status <- c(status,list(data['Status_new']))
+#'  status <- c(status,list(data['Status_of_death']))
 #'
 #'  tsid <- c(tsid,list(data['ID']))
 #'
@@ -57,14 +58,14 @@
 #'
 #'  tsdata <- c(tsdata,list(c_data))
 #'
-#'  c_treatment <- subset(data, select = c("Treatment2"))
+#'  c_treatment <- subset(data, select = c("Resection"))
 #'
 #'  treatment <- c(treatment,list(c_treatment))
 #'}
 #'
 #'
-#' # cutoff
-#'tsdata <- classifydata(time,status,tsdata,tsid,cutoff=365*1)
+#' # predict.time
+#'tsdata <- classifydata(time,status,tsdata,tsid,predict.time=365*1)
 #'
 
 
@@ -75,7 +76,7 @@ classifydata <- function(time,status,timeslicedata,tspatientid,predict.time,isfi
   variables <- timeslicedata[[1]]
 
   result <- list()
-  predict.time <- list()
+  cutoff <- list()
   cut.variable <- list()
 
   for (v in 1:length(varnames)){
@@ -113,16 +114,17 @@ classifydata <- function(time,status,timeslicedata,tspatientid,predict.time,isfi
         timeslicedata[[i]][[v]] <- ifelse(timeslicedata[[i]][[v]] <= thresholds,0,1)
       }
 
-      predict.time <- c(predict.time,thresholds)
+      cutoff <- c(cutoff,thresholds)
+      #print(cutoff)
       cut.variable <- c(cut.variable,varn)
     }
 
-    names(predict.time)  <- cut.variable
-    predict.time <- data.frame(predict.time)
+    cutoff <- data.frame(cutoff)
+    names(cutoff)  <- cut.variable
 
   }
 
   (result$timeslicedata <- timeslicedata)
-  (result$predict.time <- predict.time)
+  (result$cutoff <- cutoff)
   result
 }

R/compareTreatmentPlans.R

@@ -5,9 +5,9 @@
 #   Test Package:              'Ctrl + Shift + T'
 
 
-#'@title  Draw the KM curve of different nodes after the specified treatment plan
+#'@title  Compare and Draw the KM curve of different nodes based on specified treatment plan
 #'@description Based on the survival tree, specify nodes and specify treatment methods,
-#'draw survival curves to compare prognostic effects.
+#'draw survival curves to compare treatments
 #'@usage compareTreatmentPlans(
 #'df,
 #'treepoints,
@@ -16,15 +16,15 @@
 #'treatment
 #')
 #'@param df "data" in the return result of the survivalpath function
-#'@param treepoints list object;Specify the node for drawing the KM curve, which is in the survival tree
+#'@param treepoints list object;Specify the node for drawing the KM curve, which is displayed in the survival path graphs
 #'@param  mytree  "mytree" in the return result of the survivalpath function
 #'@param source Raw data
-#'@param treatment Choose treatment
+#'@param treatment Choose treatment variable
 #'@details By specifying the node, using mytree to calculate the previous branch variables,
 #'and calculating with the original data, the group of subjects with different treatment plans
 #'at different nodes can be screened out. Draw the survival curve of patients based on the selected groups
 #'@seealso survminer
-#'
+#'@import survival
 #'@export
 #'@examples
 #'data("dataset")
@@ -38,11 +38,11 @@
 #'treatment <- list()
 #'for (i in 1:10){
 #'
-#'  data <- dataset[dataset['timenode']==i,]
+#'  data <- dataset[dataset['time_slice']==i,]
 #'
 #'  time <- c(time,list(data['OStime_new']))
 #'
-#'  status <- c(status,list(data['Status_new']))
+#'  status <- c(status,list(data['Status_of_death']))
 #'
 #'  tsid <- c(tsid,list(data['ID']))
 #'
@@ -52,12 +52,12 @@
 #'
 #'  tsdata <- c(tsdata,list(c_data))
 #'
-#'  c_treatment <- subset(data, select = c("Treatment2"))
+#'  c_treatment <- subset(data, select = c("Resection"))
 #'
 #'  treatment <- c(treatment,list(c_treatment))
 #'}
 #'
-#'tsdata <- classifydata(time,status,tsdata,tsid,cutoff=365*1)
+#'tsdata <- classifydata(time,status,tsdata,tsid,predict.time=365*1)
 #'
 #'result <- survivalpath(time,status,tsdata[[1]],tsid,time_slices = 10,treatments = treatment,p.value=0.05,degreeofcorrelation=0.7)
 #'
@@ -81,7 +81,7 @@
 #'plotKM(result$data, treepoints,mytree,risk.table=T)
 #'#Comparing the efficacy of treatment methods by drawing survival curves
 #'treepoints = c(17,22)
-#'compareTreatmentPlans(result$data, treepoints,mytree,dataset,"Treatment")
+#'compareTreatmentPlans(result$data, treepoints,mytree,dataset,"Resection")
 #'
 
 compareTreatmentPlans <- function(df,treepoints,mytree,source,treatment){
@@ -91,9 +91,13 @@ compareTreatmentPlans <- function(df,treepoints,mytree,source,treatment){
     newdf$treepoint <- d
     data.df <- rbind(data.df,newdf)
   }
-  #print(table(data.df[,3:4]))
-  name <- names(data.df)
-  fit <- survfit(Surv( time,status)~Treatment2+treepoint, data = data.df)
+
+  formula <- paste("Surv(time,status)~",paste(names(data.df)[3:4],collapse ="+") )
+
+  formula <- as.formula(formula)
+
+  fit <- surv_fit(formula,data=data.df)
+
   ggsurvplot(fit,data = data.df,
              pval = TRUE,
              #conf.int = TRUE,
@@ -136,7 +140,7 @@ getPatients <- function(df,treepoint,mytree,source,treatment){
   }
   #get  treepoint survival periods
   time_slice <- length(path)
-  print(path)
+  #print(path)
 
   for (i in time_slice:1){
 
@@ -175,14 +179,14 @@ getPatients <- function(df,treepoint,mytree,source,treatment){
   }
 
   result <- newdf[,1:3]
-  print(dim(result))
+  #print(dim(result))
 
-  sourcedata <- source[which(source$timenode==time_slice-1),]
-  print(dim(sourcedata))
+  sourcedata <- source[which(source$time_slice==time_slice-1),]
+  #print(dim(sourcedata))
 
   sourcedata <- sourcedata[sourcedata$ID %in% result$ID,]
 
-  result <-  subset(sourcedata,select=c("OStime_new","Status_new"))
+  result <-  subset(sourcedata,select=c("OStime_new","Status_of_death"))
   names(result) <- c("time","status")
   result <- cbind(result,sourcedata[,treatment])