AJR F and L Medical Products: Radiation Protection & More
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     


This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yam, C.-S.
Right arrow Articles by Larson, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yam, C.-S.
Right arrow Articles by Larson, M.
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Hotlight (NEW!)
Right arrow
What's Hotlight?
AJR 2004; 183:529-533
© American Roentgen Ray Society


Computers in Radiology

A Simple Method for Extracting DICOM Images from a Magnetooptic Disk

Chun-Shan Yam1, Arkadiusz Sitek, Vassilios Raptopoulos and Michael Larson

1 All authors: Department of Radiology, Beth Israel Deaconess Medical Center, 1 Deaconess Rd., WCC, Rm. 306, Boston, MA 02215.

Received October 9, 2003; accepted after revision December 7, 2003.

 
Address correspondence to C.-S. Yam (csyam{at}caregroup.harvard.edu).


Abstract
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
OBJECTIVE. Our objective was to develop a simple and easy-to-use method to extract DICOM images from magnetooptic (MO) disks to the computer desktop for research purposes.

CONCLUSION. The method we developed allows users to extract DICOM images directly from MO disks to a PC desktop. The hardware component that we used is commercially available and is plug-and-play. The system is lower in cost than a clinical workstation. Users do not need to have special computer skills to use our method. DICOM images can be transferred directly from the MO disks to computer desktop folders using drag-and-drop. In our implementation, we store the DICOM files in a shared folder in our hospital network, so users can access the data from their office or research computers.


Introduction
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
Magnetooptic (MO) recording is a common industry practice for long-term data storage and external backup. Most medical scanner manufacturers also use MO technology as their standard archiving medium for medical images [1, 2]. One of the major advantages of the MO disk is that data can be rewritten repeatedly without significant damage to the disk. Most of the industrial-grade MO disks can be rewritten millions of times before noticeable failure is observed. The capacity of MO disks is large, allowing the storage of multiple patient studies in a single cartridge. For example, a typical MO disk with a storage capacity of 5.2 GB can store up to 10,000 CT or 40,000 MR images. Because of this unique advantage, most PACS vendors also use MO disks as their primary long-term archive. This kind of storage is called a "jukebox" and may contain hundreds of MO disks of patient data. A typical PACS jukebox can store several terabytes of data.

The use of MO disks for data storage is already standard practice in radiology for both clinical and research endeavors. However, this use raises compatibility and accessibility issues for the research community. First, MO disks can be read only by the manufacturer's proprietary software and hardware. Users must have the same type of archiving system (such as a clinical workstation) to read the data. In most cases, the users would need to find and get permission to share a compatible clinical workstation or purchase their own systems. The typical cost of a basic clinical workstation with MO disks ranges from $50,000 to $100,000. Although a clinical workstation may be available to read the images, it is difficult to transfer the data to a stand-alone desktop computer for research purposes. Transfer requires the installation of a data transfer mechanism such as File Transfer Protocol (FTP) on the clinical workstation [3, 4]. However, the use of FTP on a clinical workstation is not recommended and is usually prohibited by the vendor.

Many researchers in radiology face the problem of downloading data from the MO disks, but this problem is acute for researchers participating in multicenter clinical trials. Our goal was to develop a low-cost and simple-to-use method of extracting the DICOM images from MO disks.


System Installation
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
The schematic of our MO data extraction system is shown in Figure 1A. A standard desktop PC (GX110, Dell) running Windows 2000 Professional operating system (Microsoft) and an external 5.2-GB MO storage drive (RMO-S551/SD, Sony Electronics) serve as the primary workstation to extract the MO data. A connection was established between the PC and the MO drive with a commercially available peripheral component interconnect (PCI) small computer system interface (SCSI) card (model 2906, Adaptec). The total cost of this system, including the PC and MO drive, is usually less than $1,500.



View larger version (17K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1A. —System schematic and setup. Diagram of overall system for reading magnetooptic (MO) media to shared files. Drive used is 5.2 GB rewritable drive (RMO-S551/SD, Sony Electronics) with built-in firmware for displaying MS DOS-based (Microsoft) file system that is currently being used in most MO media. SCSI = small computer system interface.

 

The system installation is a two-step plug-and-play process. Step 1 installs the PCI interface card; for most of the newer PC systems, installing a PCI card is simple and does not require any tools. Step 2 connects the external MO drive to the host PC via the PCI interface card. We used a 6-ft (180-cm) SCSI cable to connect the drive to the PCI card. Most SCSI devices have two SCSI connectors to connect ("daisy-chain") multiple devices to the same PC. For a successful SCSI connection loop, the last SCSI device in the chain must be terminated. Because the MO drive is the only SCSI device used in our system, we used the built-in terminator to end the connection loop. The system setup with the MO drive connected to the host PC is shown in Figure 1B.



View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1B. —System schematic and setup. Photograph of completed system setup shows MO drive connected to host PC with external SCSI cable. System installation is plug-and-play.

 

Installation is now complete, and the system is ready for data extraction. Because the MO drive is a SCSI device, it must be powered on before the host PC is turned on. This rule is important for using external SCSI devices. Once the system is turned on, the MO drive will be automatically mounted to the Windows operating system. For our system setup, the MO drive appears as a removable drive icon "Drive G:\" on the desktop. We did not need to install any driver software for the SCSI card or the MO drive because we are using the Windows 2000 Professional operating system. Windows NT users will need to install drivers, but the installation process is intuitive and easy to follow. Several system reboots may be needed. Users do not need special computer skills to perform the hardware installation.

The ability to read MO disks in our system is mainly due to the platform-independent industry standards of SCSI and PCI connectivity. Hardware running on Mac OS, Linux, and Unix operating systems can also use this system schematic.


Implementation
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
We have tested our system with some commonly used rewritable MO disks from Sony (Sony Electronics) and Maxoptix (GE Healthcare) with different storage capacities: 600 MB and 1.2, 2.3, and 5.2 GB. Figures 2A, 2B and 3A, 3B are screen-capture images showing the file contents of a GE disk and a Siemens MO disk, respectively. In both cases, we used 2.3-GB disks. Both the GE and Siemens MO disks are compliant with the current DICOM 3.0 guidelines [1]. DICOM specifies protocols and formats for the exchange of images, time-based waveforms, reports, and associated information for medical applications. In most DICOM storage media, a set of DICOM information is described by an index file, DICOMDIR, which accompanies the files that it references. As seen in Figures 2A and 3A, the DICOMDIR index file appears in the root directory (or top folder) for both GE and Siemens MO disks. Individual DICOM files are stored in subfolders with their names in alphanumeric order. Figures 2B and 3B are screen-capture images of some representative DICOM files from the subfolders in GE and Siemens disks, respectively.



View larger version (51K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 2A. —Screen-capture images of system directories of magnetooptic (MO) disk from GE Healthcare. Screen-capture image shows DICOMDIR index file in root directory.

 


View larger version (66K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 2B. —Screen-capture images of system directories of magnetooptic (MO) disk from GE Healthcare. Screen-capture image of subdirectory of MO disk shows DICOM files named in alphanumeric index. Note that files are compressed in this particular case.

 


View larger version (67K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 3A. —Screen-capture images of system directories of magnetooptic (MO) disk from Siemens. Screen-capture image shows DICOMDIR index file in root directory.

 


View larger version (60K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 3B. —Screen-capture images of system directories of magnetooptic (MO) disk from Siemens. Screen-capture image of subdirectory of MO disk shows DICOM files named in numeric index. Note that files are uncompressed in this particular case.

 

Because the names of the folders and files in the MO disk follow an internal index and not the actual demographics such as patient name or scanning date, it is difficult to locate specific DICOM files. To correlate the folders and files to demographic information, we extracted this information from the DICOMDIR file. The DICOMDIR file is basically an ASCII file containing information on how images are stored in the MO media. In fact, DICOMDIR is a collection of DICOM objects written in sequential order [1]. Each object contains three individual entities: tag, length, and data. The tag is a 4-byte segment indicating the identity of the object, such as patient name and study date. The length is another 4-byte segment representing the type and size of the information. The data are the actual information that creates the image.

Following this schematic, we wrote a Visual Basic for Applications (VBA) macro in MS Word (Microsoft) to decode the patient name, study date, and file location for each individual image. Table 1 shows some of the decoded DICOM objects of a GE DICOMDIR file from a HiSpeed helical CT scanner (GE Healthcare). With this table, users can easily identify their DICOM files and then drag-and-drop to copy the selected files directly from the MO disk. In our implementation, we store all the DICOM files in a shared folder in our hospital network, so that users can access their data from their office computers. The ability to identify desired images derives from the nature of the DICOM standard rather than the choice of VBA in MS Word. Any system that can read SCSI MO disks can be programmed to read the DICOMDIR file and extract the desired study filenames.


View this table:
[in this window]
[in a new window]

 
TABLE 1 Decoded DICOM Objects from the DICOMDIR File

 


Data Analysis
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
Although it is not in the scope of this article to discuss the selection or application of image processing software in radiology research, we tested some of the extracted DICOM images using ezDICOM software, which is an open-source freeware program for radiology education and research [5]. To load an image in the application, drag and drop the DICOM files into the application window or its desktop icon. In our implementation, we selected the DICOM file from the MO disk and dragged it to the application icon. The DICOM images are displayed at the original window and level settings. Figure 4A is the screen-capture image of a CT scan loaded in ezDICOM. Basic image manipulation and pixel analysis can be archived through simple mouse actions. The software has a builtin color map for functional imaging data analysis. Also, the detailed DICOM information can be displayed by pressing the F3 key on the keyboard. Figure 4B is a screen-capture image of the detailed DICOM information extracted from the CT image shown in Figure 4A.



View larger version (83K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 4A. —Screen-capture CT loaded in ezDICOM software. CT displayed in abdomen window shows content of subcutaneous and intraabdominal fat for this subject.

 


View larger version (81K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 4B. —Screen-capture CT loaded in ezDICOM software. Screen-capture image of detailed DICOM information extracted using ezDICOM software shows detailed DICOM header information for CT image (A) that lists in ezDICOM software when F3 key is pressed. List contains all header information for that image and includes patient name or identification number; study date, protocol, and time; series description; image position; kVp and mA; slice thickness; and more.

 


Performance
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
Our system has been in use in our department for more than 1 year. We have used it to extract DICOM files mainly for research studies. We installed the system in our media laboratory as a central location and saved all the extracted data in a secured network folder on our hospital network. Users are required to have a valid username and password to access this shared folder. Because all research studies conducted in our institution are compliant with the Health Insurance Portability and Accountability Act (HIPAA), no patient-identifying information has been used. However, for studies performed before the HIPAA standard took effect, patient names may be used to label scans. In such cases, we suggest researchers use only the patient's initials to name the extracted files.


Limitation
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
Although our system has proven to be compatible with DICOM-compliant MO disks from GE and Siemens, other non-DICOM archiving media such as tapes and write-once-read-many cartridges, many of which require obsolete hardware, will not be compatible. In our experience, such old archiving media are also incompatible even with new hardware supplied by the same manufacturer. However, with the increasing demand for digital images in radiology, a greater number of medical scanner manufacturers are becoming DICOM-compliant. More information on current DICOM development, vendor and imaging conformance, and software resources is available at the Radiological Society of North America Internet site [6].


Discussion
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 
We developed a simple method for extracting DICOM images from MO disks using a commercially available PC and an external MO drive. The system is low in cost ($1,500) compared with a typical clinical workstation ($50,000-100,000), and the hardware installation is basically plug-and-play. Users do not need special computer skills to use our method. Images can be transferred directly from the MO disk to a PC desktop folder or a shared folder in a network drive. We have tested our system with MO disks from GE and Siemens. We also showed the use of ezDICOM software for basic image processing. The success of this solution mainly depends on the industry standards of SCSI, PCI, and DICOM. Although our system was implemented on a Windows platform, other systems such as Mac OS, Linux, and Unix can also be used.


References
Top
Abstract
Introduction
System Installation
Implementation
Data Analysis
Performance
Limitation
Discussion
References
 

  1. American College of Radiology, National Electrical Manufacturers Association. ACR-NEMA Digital imaging and communications standard. Washington, DC: National Electrical Manufacturers Association, 1985. Publication 300-1985
  2. Bidgood WD Jr, Horii SC. Introduction to the ACR-NEMA. RadioGraphics1992; 12:345 -355[Abstract]
  3. Gale DR, Gale ME, Schwartz RK, Muse VV, Walker RE. An automated PACS workstation interface: a timesaving enhancement. AJR 2000;174:33 -36[Abstract/Free Full Text]
  4. Maldjian JA, Listerud J. Automated teaching file and slide database for digital images. AJR2000; 175:1249 -1251[Abstract/Free Full Text]
  5. ezDICOM Web site. Available at www.psychology.nottingham.ac.uk/staff/cr1/ezdicom.html. Accessed May 28, 2004
  6. Radiological Society of North America Web site. Available at rsna.org/practice/dicom/dicom.html. Accessed May 28, 2004

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
Am. J. Roentgenol.Home page
C.-S. Yam
Freeware Program for Converting Image Sequences to Flash Movie Loops
Am. J. Roentgenol., January 1, 2008; 190(1): W68 - W72.
[Abstract] [Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
C.-S. Yam
A solution for using dynamic data sets in electronic presentations.
Am. J. Roentgenol., August 1, 2006; 187(2): W218 - W226.
[Abstract] [Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
C.-S. Yam
Preparation of digital movie clips for online journal publication.
Am. J. Roentgenol., July 1, 2006; 187(1): W93 - 101.
[Abstract] [Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
C.-S. Yam and M. Larson
Solutions for Extracting Images from Non-DICOM Proprietary Storage Media
Am. J. Roentgenol., March 1, 2005; 184(3): 1027 - 1027.
[Full Text] [PDF]


This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yam, C.-S.
Right arrow Articles by Larson, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yam, C.-S.
Right arrow Articles by Larson, M.
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Hotlight (NEW!)
Right arrow
What's Hotlight?


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS