In cukarthik/CancerTreatmentCharacterization: What the package does (short line)

library(devtools)
library(dplyr)
library(ggplot2)
library(tidyr)
library(ggrepel)
library(lubridate)
library(forcats)
library(RColorBrewer)

outputFolder <- params$outputFolder
minCellCount <- params$minCellCount
cohortId <- params$cohortId
cohortName <- params$cohortName
databaseId <- params$databaseId
time_window_for_interventions <- 365
connection <- params$connection
cohortDatabaseSchema <- params$cohortDatabaseSchema

cancerCohortDataTable <- getCancerDataSet(cohortDatabaseSchema, cohortId, connection)

writeLines(paste("outputFolder = ", outputFolder, "\n",
      "minCellCount = ", minCellCount, "\n",
      "cohortId = ", cohortId, "\n",
      "cohortName = ", cohortName, "\n",
      "databaseId = ", databaseId, "\n",
      "time_window_for_interventions = ", time_window_for_interventions))

  cancerSpecificVectors <- getVectorsForSpecificCancer(cohortId)

  interventionsVector <- cancerSpecificVectors$interventions
  drugVector <- cancerSpecificVectors$drugs_vector
  timeWindowForInterventions <- time_window_for_interventions
  df1 <- cancerCohortDataTable %>%
    filter(intervention_type %in% interventionsVector | generic_drug_name %in% drugVector) %>%
    filter(difftime(cohort_start_date, intervention_date) <= timeWindowForInterventions) %>%
    distinct(person_id, dx_year, intervention_date, intervention_type, age_group) %>%
    arrange(dx_year, person_id, intervention_date, intervention_type) %>%
    group_by(person_id, intervention_date) %>% slice(1) %>% arrange(person_id, intervention_date) %>%
    ungroup(intervention_date) %>% select(person_id, dx_year, intervention_type)

  # interventionsPivotWide
  distinct_lung_interventions <- df1 %>% distinct(person_id, dx_year, intervention_type) %>%
    pivot_wider(names_from = intervention_type, values_from = intervention_type)

  ###Appending distinct patient intervention field to table###
  interventions_by_pt <- distinct_lung_interventions %>%
    unite(distinct_interventions, 3:ncol(distinct_lung_interventions), sep = ' + ', na.rm = TRUE)

  augmentedCancerDataSet <- cancerCohortDataTable %>%
    left_join(interventions_by_pt, by = c('person_id')) %>%
    select(-c(dx_year.x)) %>%
    rename(dx_year = dx_year.y)

  #clear out previous run data
  if (file.exists(outputFolder)) {
    unlink(outputFolder, recursive = TRUE)
  } else {
    dir.create(outputFolder, recursive = TRUE)
  }

  cancerSpecificVectors

#duplicate of generic intervention plot but does some addition filtering (i.e., 2000 and multiple interventions in a day)
df2 <- interventions_by_pt %>% arrange(person_id) %>% group_by(dx_year, distinct_interventions) %>% tally() %>%
      arrange(dx_year, desc(n)) %>% group_by(dx_year) %>% mutate(year_total = sum(n), pct = round(n*100/year_total)) %>% filter(dx_year >= 2000)

  file <- "percent_interventions_types_per_year_additional_filter"
  colourCount <-  length(unique(interventions_by_pt$distinct_interventions))
  getPalette <- colorRampPalette(brewer.pal(26, "Set3"))

#plot the data
  z <- ggplot(df2, aes(fill = distinct_interventions, x = dx_year, y = pct)) +
      geom_bar(position = 'fill', stat = 'identity') +
      geom_text(aes(label = n), position = position_fill(vjust = .5), size = 2.5) +
      labs(x = 'Year', y = 'percent', title = 'Percent and Count distributions of interventions, by year') +
      theme(legend.position = 'bottom', legend.text = element_text(size = 3), legend.key.size = unit(.25, 'cm'), legend.title = element_text(size = 6)) +
      scale_fill_manual(values = getPalette(colourCount))
# ggsave(file.path(Folder, 'Plots/Plot 0 - Percent distribution of intervention types, by year.pdf'))

saveAnalysis(x = z, data = df2, analysisFolder = outputFolder, fileName = file, cohortName, databaseId, minCellCount, fieldName = "n")
z

# Percent distribution of intervention types, by year
plot <- examineInterventionsPerYear(augmentedCancerDataSet %>% filter(dx_year >= 2000), cohortName, databaseId, outputFolder, minCellCount)
plot

#plot 1
#counting distinct diagnoses by year
plot <- examineDxPerYear(augmentedCancerDataSet, cohortName, databaseId, outputFolder, minCellCount)
plot

#plot 2
#average age at diagnosis by year
plot <- examineAvgAgeAtDx(augmentedCancerDataSet, cohortName, databaseId, outputFolder, minCellCount)
plot

#average drug classes by year
plot <- examineAvgNumDrugsByTreatmentClass(augmentedCancerDataSet, cohortName, databaseId, outputFolder, minCellCount)
plot

## All the below plots are based on the index date of each patient and the earliest drug intervention for irrespective of the year the drug (intervention) was taken.

#plot 11c
#first line chemotherapy in the adjuvant setting
adjuvant_chemo_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '0', generic_drug_name %in% cancerSpecificVectors$chemo_drugs) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)
plot <- examinePercentChemoForAdjuvantTherapy(adjuvant_chemo_records, cohortName, databaseId, outputFolder, minCellCount)
plot

#same for chemotherapy drugs for neoadjuvant setting instead of adjuvant setting
neoadjuvant_chemo_records <- cancerCohortDataTable %>%
    filter(neoadjuvant == '1', generic_drug_name %in% cancerSpecificVectors$chemo_drugs) %>%
    distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
    arrange(dx_year, person_id, intervention_date) %>%
    group_by(person_id) %>%
    slice(1)
plot <- examinePercentChemoForNeoAdjuvantTherapy(neoadjuvant_chemo_records, cohortName, databaseId, outputFolder, minCellCount)
plot

#plot 5 - adjuvant checkpoint inhibitors
adjuvant_checkpoint_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '0', generic_drug_name %in% cancerSpecificVectors$checkpoint_inhibitors) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)


#plot the data
title <- "Percent and count distributions of first line checkpoint therapy in the adjuvant setting, by year"
file <- "percent_of_distribution_first_line_checkpoint_therapy_adjuvant"
plot <- createPercentPlotForTherapy(adjuvant_checkpoint_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

#same for chemotherapy drugs for neoadjuvant setting instead of adjuvant setting
neoadjuvant_checkpoint_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '1', generic_drug_name %in% cancerSpecificVectors$checkpoint_inhibitors) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)

#plot the data
title <- "Percent and count distributions of first line checkpoint therapy in the neoadjuvant setting, by year"
file <- "percent_of_distribution_first_line_checkpoint_therapy_neoadjuvant"
plot <- createPercentPlotForTherapy(neoadjuvant_checkpoint_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

#adjuvant VEGF positive_drugs
adjuvant_vegf_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '0', generic_drug_name %in% cancerSpecificVectors$VEGF_positive_drugs) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)


#plot the data
title <- "Percent and count distributions of first line VEGF therapy in the adjuvant setting, by year"
file <- "percent_of_distribution_first_line_vegf_therapy_adjuvant"
plot <- createPercentPlotForTherapy(adjuvant_vegf_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

#neoadjuvant VEGF positive_drugs
neoadjuvant_vegf_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '1', generic_drug_name %in% cancerSpecificVectors$VEGF_positive_drugs) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)

#plot the data
title <- "Percent and count distributions of first line VEGF therapy in the neoadjuvant setting, by year"
file <- "percent_of_distribution_first_line_vegf_therapy_neoadjuvant"
plot <- createPercentPlotForTherapy(neoadjuvant_vegf_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

---

#adjuvant EGFR positive_drugs
adjuvant_egfr_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '0', generic_drug_name %in% cancerSpecificVectors$EGFR_positive_drugs) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)


#plot the data
title <- "Percent and count distributions of first line EGFR therapy in the adjuvant setting, by year"
file <- "percent_of_distribution_first_line_egfr_therapy_adjuvant"
plot <- createPercentPlotForTherapy(adjuvant_egfr_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

#neoadjuvant EGFR positive_drugs
neoadjuvant_vegf_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '1', generic_drug_name %in% cancerSpecificVectors$EGFR_positive_drugs) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)

#plot the data
title <- "Percent and count distributions of first line EGFR therapy in the neoadjuvant setting, by year"
file <- "percent_of_distribution_first_line_egfr_therapy_neoadjuvant"
plot <- createPercentPlotForTherapy(neoadjuvant_vegf_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

#adjuvant all systemic drugs for lung
adjuvant_all_antineoplastics_records <- cancerCohortDataTable %>%
  filter(neoadjuvant == '0', generic_drug_name %in% cancerSpecificVectors$drugs_vector) %>%
  distinct(person_id, dx_year, generic_drug_name, intervention_date) %>%
  arrange(dx_year, person_id, intervention_date) %>%
  group_by(person_id) %>%
  slice(1)

#plot the data
title <- "Percent and count distributions of first line all antineoplastics adjuvant setting, by year"
file <- "percent_of_distribution_first_line_all_lung_antineoplastics_adjuvant"
plot <- createPercentPlotForTherapy(adjuvant_all_antineoplastics_records, title, file, cohortName, databaseId, minCellCount, outputFolder)
plot

cukarthik/CancerTreatmentCharacterization documentation built on Dec. 19, 2021, 7:03 p.m.