study.deployment {espadon}

Deployment of DICOM files from multiple patients

Description

The study.deployment function duplicates DICOM data from multiple patients, so that it becomes data independent of the original data. This function simplifies the analysis of multi-center or multi-expert studies in dosimetry challenges, contouring consensus searches, etc.

Usage

study.deployment(
  pats.dir,
  deploy.dir,
  design.matrix = matrix(TRUE, nrow = length(dir(pats.dir)), ncol = 1, dimnames =
    list(basename(dir(pats.dir)), "expert_1")),
  pid.prefix = "",
  white.list = c("instance", "reference"),
  black.list = c("frame of reference", "class"),
  tag.dictionary = dicom.tag.dictionary()
)

Arguments

pats.dir

Name of the directory in which all patient directories are stored, each containing the DICOM files to be duplicated.

deploy.dir

Name of the directory where all patient files will be duplicated.

design.matrix

Boolean matrix. See Details.

pid.prefix

string vector of length 1 or string vector of length ncol(design.matrix), representing the prefix added to the new unique identifier of the deployed patient (tag (0010,0020)).

white.list

Names vector, representing a part of the DICOM tag name UI value representation, other than those defined by the DICOM standard, which will be modified. By default, the UID name containing 'instance' or 'reference' will be modified.

black.list

Names vector, representing a part of the DICOM tag name UI value representation, other than those defined by the DICOM standard, which will not be modified. By default, the frame of reference UID will not be modified.

tag.dictionary

Dataframe, by default equal to dicom.tag.dictionary, whose structure it must keep. This dataframe is used to parse DICOM files.

Details

The design.matrix argument defines how patients DICOM files will be deployed. The names of the lines must match the names of the directories contained in pats.dir. The names of the columns are for example the different experts or hospitals who will study the patient files. These experts will only review the patients files defined by rownames(design.matrix)[design.matrix[,"expert"]].

Value

Creates the deploy.dir directory, containing the expert directories defined by the design.matrix column names. Each expert directory contains as many patient directories as defined by the design.matrix row names. All patients will be independent of each other. The new created patients have the pats.dir as name, and expert name as first name, and an independent patient ID, with prefix pid.prefix.

Examples

# First, save toy.dicom.raw () raw data to a temporary file/pats.dir/toy_PM 
# for testing.
toy_PM.dir <- file.path (tempdir(), "pats.dir","toy_PM") 
dir.create (toy_PM.dir, recursive = TRUE) 
dcm.filename <- tempfile (pattern = "toyrtplan", tmpdir = toy_PM.dir,
                          fileext = ".dcm")
zz <- file (dcm.filename, "wb")
writeBin (toy.dicom.raw (), zz, size = 1)
close (zz)

# function test:
pats.dir <- dirname (toy_PM.dir)
deploy.dir <- file.path (tempdir(), "deploy.dir") 
design.matrix <- matrix(TRUE, nrow = length (dir (pats.dir)), ncol=3,
                        dimnames = list (basename (dir (pats.dir)),
                                         c("Dr Quinn","Dr Who","Dr House")))
design.matrix
study.deployment (pats.dir, deploy.dir, design.matrix, 
                 pid.prefix = c("zz_", "yy_", "xx_"))

# check result
list.files(deploy.dir, recursive = TRUE)
load.patient.from.dicom(deploy.dir)$patient
# Cleaning  temporary directory
unlink (pats.dir, recursive = TRUE)
unlink (deploy.dir, recursive = TRUE)