index rss mastodon twitter github linkedin email
Álvaro Ramírez
sponsor

Álvaro Ramírez

19 December 2015 View DICOM files from your X-ray

Got a CD with my chest X-ray from the hospital. Was expecting a pdf or an image of sorts, but the CD content was rather different. For starters, it was targeted at Windows users (AUTORUN.INF, MediaViewerLauncher.EXE and a bunch of DLLs):

$ find . -exec file --mime-type '{}' \;

./AUTORUN.INF: text/plain
./DICOMDIR: application/dicom
./MediaViewerLauncher.EXE: application/octet-stream
...
./Libraries/BASEPRINTER.DLL: application/octet-stream
./Libraries/CDDATABURNER.DLL: application/octet-stream
./Libraries/COM.DLL: application/octet-stream
...
./Libraries/ACE.DLL: application/octet-stream
./Libraries/ACE_SSL.DLL: application/octet-stream
./Libraries/ATL90.DLL: application/octet-stream
...
./DICOM/PAT_0000: application/x-directory
./DICOM/PAT_0000/STD_0000/SER_0000/OBJ_0001/IM_0001: application/dicom
./DICOM/PAT_0000/STD_0000/SER_0001/OBJ_0001/ED_0001: application/dicom
./DICOM/PAT_0000/STD_0000/SER_0002/OBJ_0001/ED_0001: application/dicom
./Worklist/ClinicalInfo/067eccde-b299-e511-9114-005056ad3afe.mht: text/html
./Worklist/Report/067eccde-b299-e511-9114-005056ad3afe.mht: text/html
./Worklist/Worklist.wl: application/octet-stream

I'm on a Mac, so most of these files were not useful to me. The more interesting files were IM_0001 and ED_0001 with "application/dicom" MIME type. DICOM files stand for Digital Imaging and Communications in Medicine. How to view these on a Mac? OsiriX viewer is an option. OsiriX, though on the heavy side (100.7MB download), it rendered the X-ray successfully.

2015-12-03-osirix-x-ray-small.png

Unsurprisingly, ImageMagick's convert utility also handles DICOM files. Converting to PNG worked well.

$ convert ./DICOM/PAT_0000/STD_0000/SER_0001/OBJ_0001/ED_0001 ED_0001.png

2015-12-03-imagemagick-x-ray-small.png

DICOM files also hold patient's metadata and optional reports. The file format is well known. OsiriX gives you access to it, but a few lines of python can also extract it for you. First install the pydicom package:

$ sudo pip install pydicom

Running the python interpreter is enough to peak at the metadata:

>>> import dicom
>>> ds = dicom.read_file("./DICOM/PAT_0000/STD_0000/SER_0000/OBJ_0001/IM_0001")
>>> ds
(0008, 0000) Group Length                        UL: 400
(0008, 0005) Specific Character Set              CS: 'ISO_IR 100'
(0008, 0016) SOP Class UID                       UI: Computed Radiography Image Storage
(0008, 0020) Study Date                          DA: '20151203'
(0008, 0021) Series Date                         DA: '20151203'
(0008, 0023) Content Date                        DA: '20151203'
(0008, 0030) Study Time                          TM: '120519.000000'
(0008, 0031) Series Time                         TM: '120520.000000'
(0008, 0033) Content Time                        TM: '120643.000000'
(0008, 0060) Modality                            CS: 'CR'
(0008, 0070) Manufacturer                        LO: 'Canon Inc.'
...

There were other DICOM files with a report:

>>> import dicom
>>> ds = dicom.read_file("./DICOM/PAT_0000/STD_0000/SER_0001/OBJ_0001/ED_0001")
>>> ds
(0008, 0005) Specific Character Set              CS: 'ISO_IR 100'
(0008, 0016) SOP Class UID                       UI: Encapsulated PDF Storage
...
(0042, 0012) MIME Type of Encapsulated Document  LO: 'application/pdf'

DCMTK is another alternative tool to extract DICOM metadata. The source is available and can be built:

$ tar xf dcmtk-3.6.0.tar.gz
$ cd dcmtk-3.6.0
$ cmake .
$ make

Or installed via homebrew:

$ brew install dcmtk

DCMTK includes dcmdump. You can use it to dump DICOM files:

$ dcmdata/apps/dcmdump DICOM/PAT_0000/STD_0000/SER_0000/OBJ_0001/IM_0001
# Dicom-File-Format

# Dicom-Meta-Information-Header
# Used TransferSyntax: Little Endian Explicit
(0002,0000) UL 192                                      #   4, 1 FileMetaInformationGroupLength
(0002,0001) OB 01\00                                    #   2, 1 FileMetaInformationVersion
(0002,0002) UI =ComputedRadiographyImageStorage         #  26, 1 MediaStorageSOPClassUID
(0002,0003) UI [1.2.392.200046.100.2.1.1.42667.20151203120519.1.1.1] #  52, 1 MediaStorageSOPInstanceUID
(0002,0010) UI =LittleEndianExplicit                    #  20, 1 TransferSyntaxUID
(0002,0012) UI [1.3.46.670589.42.1.4.4.5]               #  24, 1 ImplementationClassUID
(0002,0013) SH [PhilipsISPACS445]                       #  16, 1 ImplementationVersionName
...

Of interest, David Clunie's Medical Image Format Site.