Summary

• Assignment description

• Setting the enviroment

• Descriptive Analysis

• Survival Analysis

• Deeper Analysis

Dataset Description

The response variable, SurvTime, is the survival time in days of a lung cancer patient.

The covariates are:

•Cell (type of cancer cell),

•Therapy (type of therapy: standard or test),

•Prior (prior therapy: 0=no, 10=yes),

•Age (age in years),

•DiagTime(time in months from diagnosis to entry into the trial)

•Kps(performance status).

A censoring indicator variable Censor is created from the data, with the value 1 indicating a censored time and the value 0 indicating an event time. Since there are only two types of therapy, an indicator variable, Treatment, is constructed for therapy type, with: value 0 for standard therapy and value 1 for test therapy.

Exercises

1. what was the maximum survival time for the cell type adeno?

2. what is the average age of subjects in this study?

3. which cell type appeared the most during this study?

4. Calculate descriptive statistics for all numeric variables within this dataset?

5. Perform a survival analysis to assess the survival time (variable SurvTime)? based on the cancerous cells (var Cell)? Consider applying survival functions/kaplan meier quartiles/cumulative incidence function/cox regression etc.

6. Perform an appropriate multivariable analysis to analyze the effect of independent variables age on the hazard ratio between the different cancerous cells (var Cell)?

Load Libraries

pacman::p_load(
  tidyverse,  # A set of many useful libraries
  readxl,     # To import the dataset from Excel
  here,       # To avoid problems with file directories
  janitor,    # To clean data in a fast way
  gt,         # Output tables
  gtsummary,  # Output tables for models and survival data
  survival,   # To manage survival data
  ggsurvfit,  # To plot survival analysis
  tidycmprsk  # To fit survival models
)

Import Data

data_imported <- read_xlsx(
  here("Akkodis interview", "Oncology_dataset_for_R.xlsx")
  )

data_cleaned <- data_imported |> 
  clean_names() |> 
  remove_empty() |> 
  remove_constant() 

data_cleaned |> 
  slice_sample(n = 1, by = c(therapy, cell)) |> 
  gt() |> 
  tab_header("Sample of the dataset cleaned",
             "The sample was stratified by the variables therapy and cell")

what was the maximum survival time for the cell type adeno?

data_cleaned |> 
  filter(
    surv_time == max(surv_time),
    .by = cell
    ) |> 
  select(obs, cell, surv_time, kps, age) |> 
  gt() |> 
  tab_header("Maximum Survival Time by Cell", 
             "Adeno's row is bolded") |> 
  tab_style(
    style = cell_text(weight = "bold"),
    locations = list(
      cells_body(rows = cell == "adeno")
      )
  )

what is the average age of subjects in this study?

data_cleaned |>  
  summarise(
    across(age, list(
      mean = mean,
      median = median,
      sd = ~ round(sd(.)),
      Q1 = ~ quantile(., .25),
      Q3 = ~ quantile(., .75),
      min = min,
      max = max
    ),
    .names = "{fn}"
    )
  ) |> 
  pivot_longer(everything(), names_to = "statistics") |> 
  gt() |>  
  tab_header(
    "Age", 
    "Mean and other position and variance indicators"
    ) |> 
  tab_style(
    style = cell_text(weight = "bold"),
    locations = list(
      cells_body(rows = statistics == "mean")
      )
  )

which cell type appeared the most during this study?

data_cleaned |> 
  tabyl(cell) |> 
  adorn_pct_formatting(digits = 0) |> 
  gt() |> 
  tab_header("Cell's frequency") |> 
  tab_style(
    style = cell_text(weight = "bold"),
    locations = list(
      cells_body(rows = n == max(n))
      )
  )

Calculate descriptive statistics for all numeric variables within this dataset?

data_cleaned |>
  select(-obs) |> 
  summarise(
    across(
      where(is.numeric),
           list(
      mean = mean,
      median = median,
      sd = ~ round(sd(.)),
      Q1 = ~ quantile(., .25),
      Q3 = ~ quantile(., .75),
      min = min,
      max = max
    ),
    .names = "{col}-{fn}"
    )
  ) |> 
  pivot_longer(everything(), names_to = "statistics") |> 
  separate(col = statistics, sep = "-", into = c("column", "statistics")) |> 
  pivot_wider(names_from = statistics, values_from = value) |> 
  gt() |>  
  tab_header("Descriptive Statistics", "All numeric variables") 

A Focus on Survival Time

• We expect surv_time to follow an exponential distribution

ggstatsplot::gghistostats(data_cleaned, surv_time, binwidth = 70, title = "Histogram of surv_time")