Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2013;2013:169526.
doi: 10.1155/2013/169526. Epub 2013 Jul 18.

SIVIC: Open-Source, Standards-Based Software for DICOM MR Spectroscopy Workflows

Affiliations
Free PMC article

SIVIC: Open-Source, Standards-Based Software for DICOM MR Spectroscopy Workflows

Jason C Crane et al. Int J Biomed Imaging. .
Free PMC article

Abstract

Quantitative analysis of magnetic resonance spectroscopic imaging (MRSI) data provides maps of metabolic parameters that show promise for improving medical diagnosis and therapeutic monitoring. While anatomical images are routinely reconstructed on the scanner, formatted using the DICOM standard, and interpreted using PACS workstations, this is not the case for MRSI data. The evaluation of MRSI data is made more complex because files are typically encoded with vendor-specific file formats and there is a lack of standardized tools for reconstruction, processing, and visualization. SIVIC is a flexible open-source software framework and application suite that enables a complete scanner-to-PACS workflow for evaluation and interpretation of MRSI data. It supports conversion of vendor-specific formats into the DICOM MR spectroscopy (MRS) standard, provides modular and extensible reconstruction and analysis pipelines, and provides tools to support the unique visualization requirements associated with such data. Workflows are presented which demonstrate the routine use of SIVIC to support the acquisition, analysis, and delivery to PACS of clinical (1)H MRSI datasets at UCSF.

Figures

Figure 1
Figure 1
Multidimensional MRSI data visualization. (a) 4D brain MRSI data in SIVIC. Spectra from individual voxels are shown on the right. The left panel shows the spatial localization of each MRSI voxel on a reference anatomical image. The color overlay is a 3D metabolite map derived from spectral quantification of individual peaks. (b) 5D dynamic MRSI data. Metabolite peaks are derived from each point in a time series of 4D MRSI volumes. 3D dynamics of individual metabolites are represented by time curves in the bottom row for two different metabolites. The example at the bottom is from hyperpolarized  13C MRSI of a rat.
Figure 2
Figure 2
SIVIC software suite components. SIVIC applications (top) are built using the SIVIC Kit (svk) bottom. The svk is a C++ library representing a model, view, controller (MVC) design. View classes provide components that graphically display data and acquisition constructs represented by svkImageData objects (yellow). The controller layer utilizes svk IO (readers, writers) and svk algorithm classes to provide analysis functionality. The underlying svk model is represented by specific implementations of IO, algorithm, and data structure classes. Some specific examples of each class hierarchy are shown in the model (bottom box).
Figure 3
Figure 3
File formats supported by the svk IO layer are shown. SIVIC provides support for parsing raw data formats such as the GE P-file and Varian FID files, though interpreting the data may be sequence specific requiring customization to svk reader software classes. Version numbers indicate target SIVIC release to provide support.
Figure 4
Figure 4
Generalized DICOM MRSI workflow. MRS data is acquired and encoded in vendor specific formats (red, orange, and pink). SIVIC tools reconstruct data and/or convert to DICOM format (green) to send to PACS. DICOM data can be retrieved for visualization in the reading room or on a research workstation for processing and visualization using the SIVIC GUI or command line tools.
Figure 5
Figure 5
On-scanner MRSI workflow. SIVIC running on the scanner reads raw MRSI vendor data and anatomical DICOM MRI images. MRSI data is reconstructed and DICOM MRS, DICOM MRI metabolite maps, and DICOM secondary capture (SC) images are exported and sent to PACS. DICOM SC and DICOM MRI images are viewed in the reading room. DICOM MRI (anatomical and metabolite maps) and DICOM MRS images may be viewed on a research workstation running SIVIC or other DICOM applications. CPU intensive on-scanner reconstruction may require a computational cluster for real-time analysis during an exam.
Figure 6
Figure 6
SIVIC GUI running on a GE 7T scanner console. Raw data from a phantom acquisition is shown in the right SIVIC panel. The left SIVIC panel shows the MRSI voxel grid spatially referenced to the reference image. The yellow box represents the PRESS volume localization, and purple regions represent sat bands. The SIVIC GUI is configured to run from configurable menu buttons on the scanner's operator console (right side).
Figure 7
Figure 7
Phantom MRS data reconstructed and quantified using the SIVIC GUI on a 7T GE scanner console. The right panel shows spectra from the 16 selected voxels. The voxels are spatially referenced to the image in the left panel. The color overlay on the left is a metabolite map representing the choline peak height. The blue text above the spectra gives the exact value of the current overlay for each voxel.
Figure 8
Figure 8
DICOM MRSI exam in OsiriX PACS (a) and DCM4CHEE PACS (b): Raw Data Storage SOP class (1.2.840.10008.5.1.4.1.1.66, RAW), reconstructed MRSI, MR Spectroscopy SOP class (1.2.840.10008.5.1.4.1.1.4.2, MRS), Secondary Capture SOP class (1.2.840.10008.5.1.4.1.1.7, SC), metabolite maps (Enhanced MR Image Storage SOP class (1.2.840.10008.5.1.4.1.1.4.1, EMRI).
Figure 9
Figure 9
CNI metabolite maps (bottom color overlay) derived from MRSI data in SIVIC are exported as standard DICOM MR Image Storage SOP instances, which can be loaded into 3D DICOM image analysis software packages (shown here in 3D Slicer). Derived maps are thus amenable to multimodal analysis. The top panel shows ADC maps (color) on FLAIR images. The bottom panel shows the same anatomical locations on a T1 contrast enhanced image.
Figure 10
Figure 10
Off-scanner MRSI workflow. SIVIC command line tools running on the scanner convert vendor raw MRS data to DICOM Raw Data Storage SOP instances. Anatomical MRI and raw DICOM data is sent to PACS. A research workstation retrieves DICOM images from PACS. MRSI data is reconstructed and DICOM MRS, DICOM MRI metabolite maps and DICOM secondary capture (SC) images are exported and sent to PACS. DICOM SC and DICOM MRI data is viewed in the reading room. DICOM MRI (anatomical and metabolite maps) and DICOM MRS images may be viewed on a research workstation running SIVIC or other DICOM applications.
Figure 11
Figure 11
SIVIC generated DICOM secondary capture report for UCSF MRS exam. The series consists of 8 images shown here. The color overlay represents the choline to NAA index. Spatial referencing to T1 postcontrast image, volume localization (yellow), and sat bands (purple shading) are shown. The final two images are summary representations of the acquisition referenced to the anatomical images.
Figure 12
Figure 12
svk raw data readers handle acquisition-specific data reorganization. This includes vendor-specific header parsing and acquisition-specific data reordering and resampling. The output of an svkImageReader is always an svkImageData object, represented by a DICOM header and data sampled on a regular grid that is suitable for FFT-based reconstruction. This permits the use of a common set of independent downstream reconstruction and processing algorithms, independent of the acquisition sequence or vendor data format. An svkImageReaderFactory reads the raw data files to create the appropriate type of svkImageReader for the specific input format and acquisition type. Vendor- and acquisition-specific readers load data and associated mappers resample data to a regular grid.

Similar articles

See all similar articles

Cited by 32 articles

See all "Cited by" articles

References

    1. Okuda DT, Srinivasan R, Oksenberg JR, et al. Genotype-phenotype correlations in multiple sclerosis: HLA genes influence disease severity inferred by 1HMR spectroscopy and MRI measures. Brain. 2009;132(1):250–259. - PMC - PubMed
    1. Laprie A, Catalaa I, Cassol E, et al. Proton magnetic resonance spectroscopic imaging in newly diagnosed glioblastoma: predictive value for the site of postradiotherapy relapse in a prospective longitudinal study. International Journal of Radiation Oncology Biology Physics. 2008;70(3):773–781. - PubMed
    1. Chang J, Thakur S, Perera G, et al. Image-fusion of MR spectroscopic images for treatment planning of gliomas. Medical Physics. 2006;33(1):32–40. - PubMed
    1. Westman E, Wahlund L-O, Foy C, et al. Magnetic resonance imaging and magnetic resonance spectroscopy for detection of early Alzheimers Disease. Journal of Alzheimer’s Disease. 2011;26(3):307–319. - PubMed
    1. Josan S, Spielman D, Yen Y-F, Hurd R, Pfefferbaum A, Mayer D. Fast volumetric imaging of ethanol metabolism in rat liver with hyperpolarized [1-13C]pyruvate. NMR in Biomedicine. 2012;25(8):993–999. - PMC - PubMed

LinkOut - more resources

Feedback