Clinical characterization of cancer treatment using the Oncology CDM

Analytics use case(s): Characterization
Study type: Clinical Application
Tags: OHDSI-Korea, FEEDER-NET, Oncology WG
Study lead: Hokyun Jeon, Seng Chan You
Study lead forums tag: Hokyun Jeon
Study start date: Februrary 25, 2020
Study end date: -
Protocol: -
Publications: -
Results explorer: -

This study aims to charaterize the cancer treatment based on the Oncology CDM

CancerTxPathway

Introduction

Tool for extracting chemotherapy cycle records from single medication records in CDM database. Then, visualize the treatment pathway using extracted chemotherapy and treatments. As a proof of the study, the chemotherapy-induced neutropenia onset timing is suggested.

Technology

CancerTxPathway is an R package codes for whole process of the study.

Dependencies

install.packages("DatabaseConnector")

install.packages("collapsibleTree")

install.packages("data.table")

install.packages("dplyr")

install.packages("ggplot2")

install.packages("ggthemes")

install.packages("reshape2")

install.packages("scales")

install.packages("highcharter")

install.packages("gridExtra")

install.packages("viridis")

install.packages("tidyverse")

install.packages("hrbrthemes")

install.packages("plotly")

install.packages("SqlRender")

install.packages("listviewer")

install.packages("tidyr")

install.packages("networkD3")

install.packages("ggbeeswarm")

install.packages("flexdashboard")

Getting started

In R, use the following commands to download and install:

install.packages("devtools")

devtools::install_github("ohdsi-studies/CancerTxPathway")

library(CancerTxPathway)

library(flexdashboard)

How to run

Parameter setting for algorithm :

Database parameters :

# Details for connecting to the server:
connectionDetails <- DatabaseConnector::createConnectionDetails(dbms='pdw',
                                                                server=Sys.getenv("PDW_SERVER"),
                                                                schema='cdmDatabaseSchema',
                                                                user=NULL,
                                                                password=NULL,
                                                                port='port')

oracleTempSchema <- NULL
cdmDatabaseSchema <- "cdm_database_schema.dbo"
cohortDatabaseSchema <- "cohort_database_schema.dbo"
vocaDatabaseSchema <- "voca_database_schema.dbo"
oncologyDatabaseSchema <- "oncology_database_schema.dbo" # Schema for Episode table and Episode_eventtable, default = cdmDatabaseSchema


createCohortTable <- FALSE # Create cohort table for your cohort table
createEpisodeAndEventTable <- TRUE # warning: existing table might be erased

episodeTable <- "episode_table"
episodeEventTable <- "episode_event_table"
cohortTable <- "cohort"

maxCores <- 4

Then run the following :

Generate episode and episode event table :

## Episode table and episode Event generation
executeExtraction(connectionDetails,
                  oracleTempSchema = NULL,
                  cdmDatabaseSchema,
                  cohortDatabaseSchema,
                  vocaDatabaseSchema = cdmDatabaseSchema,
                  oncologyDatabaseSchema = cdmDatabaseSchema,
                  createCohortTable = FALSE,
                  createEpisodeAndEventTable = FALSE,
                  createTargetCohort = FALSE,
                  episodeTable,
                  episodeEventTable,
                  cohortTable,
                  maxCores = 4)

Parameter setting for visualization :

Before the setting parameters for visualization, insert your cohort information in inst/csv/cohortDescription.csv file. Then, set the parameters :

outputFolder <- 'output folder path'
outputFileTitle <- 'output file title'
targetCohortIds <- c(4:11)
episodeCohortCreate <- TRUE
minSubject <- 0 # under 0 patients are removed from plot

# Usage Pattern graph
fromYear <- 1998
toYear <- 2018

# Iteration Heatmap
identicalSeriesCriteria <- 60 # Regard as a same treatment when gap dates between each cycle less than 60 days
maximumCycleNumber <- 18 # Ignore patients who received regimen more than 18 iteration

# Treatment Pathway
collapseDates <- 0
conditionCohortIds <- 1 # restrict target patients with certain condition_occurrence
treatmentLine <- 3 # Treatment line number for visualize in graph
minimumRegimenChange <- 1 # Target patients for at least 1 regimen change

# Cohort for surgery and event
surgeryCohortIds <- 42 # Colectomy
eventCohortIds <- 45 # Neutropenia

# ignore the event in range of +- treatmentEffectDates
treatmentEffectDates <- 2

Visualization results export

Excute after the parameter settings and cohort description in csv file :

plots <- CancerTxPatterns(connectionDetails,
                          oracleTempSchema,
                          cdmDatabaseSchema,
                          cohortDatabaseSchema,
                          oncologyDatabaseSchema,
                          vocaDatabaseSchema,
                          cohortTable,
                          episodeTable,
                          outputFolder,
                          outputFileTitle,
                          targetCohortIds,
                          episodeCohortCreate = FALSE,
                          createEpisodeCohortTable,
                          fromYear = 1998,
                          toYear = 2018,
                          identicalSeriesCriteria = 60,
                          maximumCycleNumber = 18,
                          minSubject = 0,
                          collapseDates = 0,
                          conditionCohortIds,
                          treatmentLine = 3,
                          minimumRegimenChange = 1,
                          surgeryCohortIds,
                          eventCohortIds,
                          treatmentEffectDates = 2)

License

CancerTxPathway is licensed under Apache License 2.0

Development

CancerTxPathway is being developed in R Studio.

GaengSu/CancerTxPathway