This python script facilitates standardized conversion of DICOM studies into nifti format by doing the following:
- Near autonomous conversion of DICOM images to medical image processing-friendly nifti format
- Deidentification
- Assignment of a standardized naming convention relevant to a given annotation routine, for example, labling liver lesions on T1 and ADC sequences before and after treatment (easily customized for other uses)
Currently can be used for conversion of pre- and post-treatment MRI abdomen studies AND CT angiography studies by using the appropriate config file
SimpleITK
Numpy
PyDicom
NiPype
Slugify
Script assumes a nested folder eg tree study-->series-->DICOM (standard when exportingw with Osirix or Horos). This should easily by adapted for other folder structures if working with exported images from other DICOM viewers/PACS servers.
There are two script calls needed:.
-
Grab Metadata: create a .csv with relevant metadata in table for tagging of desired series by domain expert (or algorithm in future)
-
Convert From Table: conversion, deidentification and standardized naming of converted sequences using the annotated tag column from 1) Metadata table.
python osirix_dicom_to_nifti.py --grabMetadata --tablePath ./path/to/metadata/table.csv --dicomDir ./path/to/folder/containing/study/folders
This will generate a .csv file with columns of metadata for each series of an MRI study meant to allow inference of which series' are desired for conversion.
Heads up: this can be a lot of rows as modern MRI studies generate a lot of series!
An empty tag column will be generated and with some domain knowledge the user must fill-in with the appropriate tag number.
Example is as follows (take note of the number inputted in the tag column):
| MRN | ACC | Machine | Series Path | Acq Time | Series Number | Series Desc | Tag(0=pre,1=ea,2=ea_sub,3=la,4=la_sub,5=pv,6=pv_sub,7=ev,8=ev_sub,9=adc) |
|---|---|---|---|---|---|---|---|
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/trufisp-loc-1 | 110028.625 | 1 | Trufisp_Loc | |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/cor-haste-2 | 110142.875 | 2 | COR HASTE | |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/3point-dixon-t2-star-031511-fa10-modified-6611-opp-6 | 110544.7325 | 6 | 3-point Dixon_T2 star_03-15-11_FA10_modified 6/6/11_opp | |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/3point-dixon-t2-star-031511-fa10-modified-6611-in-5 | 110544.735 | 5 | 3-point Dixon_T2 star_03-15-11_FA10_modified 6/6/11_in | |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-pre-62612-13 | 110828.9625 | 13 | AX VIBE PRE (6/26/12) | 0 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-14 | 111014.0575 | 14 | AX VIBE POST 2ART,1MIN,3MIN EQU | 1 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-sub-15 | 111014.0575 | 15 | AX VIBE POST 2ART,1MIN,3MIN EQU_SUB | 2 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-16 | 111035.2275 | 16 | AX VIBE POST 2ART,1MIN,3MIN EQU | 3 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-sub-17 | 111035.2275 | 17 | AX VIBE POST 2ART,1MIN,3MIN EQU_SUB | 4 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-18 | 111114.47 | 18 | AX VIBE POST 2ART,1MIN,3MIN EQU | 5 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-sub-19 | 111114.47 | 19 | AX VIBE POST 2ART,1MIN,3MIN EQU_SUB | 6 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-500 | 111131.799 | 500 | AX VIBE POST 2ART,1MIN,3MIN EQU | |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-20 | 111316.04 | 20 | AX VIBE POST 2ART,1MIN,3MIN EQU | 7 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/ax-vibe-post-2art-1min-3min-equ-sub-21 | 111316.04 | 21 | AX VIBE POST 2ART,1MIN,3MIN EQU_SUB | 8 |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/cor-vibe-post-3-min-22 | 111407.52 | 22 | 000001 | Avanto |
| 9999999 | 000001 | Avanto | ./Y90_Seg/name/body-mt-sinai-protocols-abdomen/pace-diffusion-50400800-10113-adc-29 | 112810.835 | 29 | PACE Diffusion 50-400-800 10-1-13_ADC | 9 |
| 9999999 |
Once the table above is properly annotated with the correct tags for the desired series, the second step is conversion.
python osirix_dicom_to_nifti.py --convertFromTable --tablePath ./path/to/metadata/table.csv --dicomDir ./path/to/folder/containing/study/folders --niftiDir ./path/to/nifti/output/folder
Study folders within designated nifti output folder will be named according to accession number for study, so conversion of these folder names to random strings using a secure look-up table is encouraged for thorough de-identification.