DOI: 10.1148/rg.261055066
RadioGraphics 2006;26:299-308
© RSNA, 2006
Informatics in Radiology (infoRAD)
Navigating the Fifth Dimension: Innovative Interface for Multidimensional Multimodality Image Navigation1
Antoine Rosset, MD,
Luca Spadola, MD,
Lance Pysher, MD and
Osman Ratib, MD, PhD
1 From the Department of Radiology, Geneva University Hospital, Rue Micheli-du-Crest 24, 1211 Geneva 14, Switzerland (A.R., L.S.); Department of Radiology, St Thomas More Hospital, Canon City, Colo (L.P.); and Department of Radiology, David Geffen School of Medicine at UCLA, Los Angeles, Calif (O.R.). Recipient of a Cum Laude award for an infoRAD exhibit at the 2004 RSNA Annual Meeting. Received March 21, 2005; revision requested July 8 and received October 14; accepted October 26. All authors have no financial relationships to disclose.
Address correspondence to A.R. (e-mail: rossetantoine{at}bluewin.ch).
 |
Abstract
|
|---|
The display and interpretation of images obtained by combining three-dimensional data acquired with two different modalities (eg, positron emission tomography and computed tomography) in the same subject require complex software tools that allow the user to adjust the image parameters. With the current fast imaging systems, it is possible to acquire dynamic images of the beating heart, which add a fourth dimension of visual information—the temporal dimension. Moreover, images acquired at different points during the transit of a contrast agent or during different functional phases add a fifth dimension—functional data. To facilitate real-time image navigation in the resultant large multidimensional image data sets, the authors developed a Digital Imaging and Communications in Medicine–compliant software program. The open-source software, called OsiriX, allows the user to navigate through multidimensional image series while adjusting the blending of images from different modalities, image contrast and intensity, and the rate of cine display of dynamic images. The software is available for free download at http://homepage.mac.com/rossetantoine/osirix.
© RSNA, 2006
|
Introduction
|
|---|
During the past decade, three-dimensional (3D) imaging has rapidly become an essential tool for various diagnostic imaging examinations. With current multisection computed tomography (CT) and magnetic resonance (MR) imaging techniques, it is possible to acquire within a few seconds high-resolution volumetric data sets that have identical resolution in all three dimensions. The isotropic data sets provide exquisitely detailed anatomic information. With 3D volume rendering, it is possible to generate from these tomographic data high-quality images that provide a realistic anatomic view of the body and organs (1). Three-dimensional volume-rendered images offer an attractive alternative to axial source images and provide better depiction of morphologic features and the spatial relationships between them (2–4). However, 3D volume rendering and multiplanar reformatting are applicable in the clinical routine only if the technology is easy to use and allows radiologists to generate and manipulate images in real time. Radiologists must acquire specific skills to manage and navigate extremely large image series in multidimensional space. The standard picture archiving and communication system (PACS) workstations used for diagnostic imaging must be adapted to allow radiologists to navigate image series with ease in any direction and to generate different views of the image data in real time (ie, with downloading and processing times of less than 1 second). To perform these complex tasks, the software programs used for image display and manipulation also must be based on innovative techniques and navigation devices, including a convenient and efficient graphical user interface.
Recent technical innovations in radiologic modalities allow the acquisition of images with a temporal resolution that is sufficient for observation and analysis of dynamic function and physiologic mechanisms. With the latest generations of CT scanners and MR imagers, image series can be acquired that show the transit of an intravenous contrast agent through the body or through a given organ (5). With the improved capability for fast image acquisition, it is also possible to acquire dynamic image series of rapidly moving anatomic structures, such as a beating heart. These high-temporal-resolution images often are described as four-dimensional images.
Two fast-growing domains of medical imaging are functional and molecular imaging, which enable the in vivo characterization of biologic processes at the cellular and molecular levels, respectively. Positron emission tomography (PET) with a radiolabeled tracer is the most commonly used molecular imaging method, but other techniques are emerging that involve the use of various molecular markers with other imaging modalities (eg, MR imaging). These techniques offer new perspectives for functional imaging and a hope of increasing both the sensitivity and the specificity of imaging for the detection of various diseases. The new techniques are complementary to those of conventional anatomic CT and MR imaging. The combination of functional or molecular data with anatomic information adds a new dimension to images. The functional or molecular component of the image data is often referred to as a fifth dimension (Fig 1). Radiologists need new tools and new ways to interpret the images that result from multidimensional examinations. The new software platforms must offer simple and intuitive user interfaces and powerful processing capabilities that allow radiologists to concentrate their attention on the image content and to arrive at an accurate diagnostic interpretation.
View larger version (65K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Three-dimensional PET/CT fusion image series obtained with the use of OsiriX software shows five degrees of blending and superposition of images from the two modalities, on a progressive scale from CT alone (left) to PET alone (right).
|
|
With this in mind, we set out to develop a Digital Imaging and Communications in Medicine (DICOM)-compliant software program that would allow unlimited multidimensional navigation in a single interactive environment. We sought a completely new platform that would allow users to efficiently and conveniently navigate through large sets of multidimensional data without the need to purchase expensive hardware or software. Most important, we elected to develop our program in accordance with an open-source paradigm based on the integration of existing open-source software libraries and to allow other institutions and developers to contribute to our project.
|
Materials and Methods
|
|---|
Our goal was to create a single common platform that would facilitate the interpretation and management of the wide variety of image data obtained with dynamic, functional, and molecular imaging. It seemed reasonable to assume that, thanks to the current general adoption of the DICOM standard, all imaging systems of any modality would produce DICOM files and transmit those files by using DICOM network protocols (6). We therefore developed a software program that is fully DICOM compatible, a program that can read, write, send, and receive DICOM files from any imaging system and that is usable with different vendors’ DICOM profiles. We chose a Unix-based operating system (Mac OS X; Apple Computer, Cupertino, Calif) as our software platform for the following reasons: (a) It offers a strong multiprocessor and multithread environment with a well-designed and easy-to-use graphical user interface; (b) it runs on a fifth-generation high-performance microprocessor that is optimized for graphic applications (eg, PowerPC 970; IBM, New York, NY); and (c) it has an optimized and fully integrated OpenGL application program interface that provides fast implementation of interactive graphic capabilities (7). The 64-bit microprocessor, with a capacity of more than 4 GB of random access memory (RAM), enables the speedy processing of large data sets by the software program. Given the large data set acquired in a 64–detector-row CT examination, the 4-GB RAM available with 32-bit microprocessors (eg, G4 or Pentium) quickly becomes a limitation.
In order to minimize our development time and obtain a powerful and robust software program, we based our development framework on well-known open-source toolkits: Papyrus (available at http://www.expasy.ch/UIN/html1/projects/papyrus.html) for DICOM file management; Offis (available at http://dicom.offis.de) and PixelMed (available at http://www.pixelmed.com) for the DICOM network functions; OpenGL for fast two-dimensional (2D) and 3D display; VTK (Visualization Toolkit, available at http://public.kitware.com/VTK) for 3D rendering; and ITK (the National Library of Medicine Insight Segmentation and Registration Toolkit, available at http://itk.org) for segmentation and registration. These toolkits were developed primarily for use with Unix-based operating systems, and they can be used without adaptation on the OS X Unix kernel. The toolkits, which are written in C, C++, or Java language, are open source and available free of charge. Our software, OsiriX, is written in Objective C/C++, a language similar to C++ but with an easier syntax and a dynamic symbolic linking similar to that in Java. Objective C is the official language of choice for the OS X platform. We used the open-source and free GCC compiler (available at http://gcc.gnu.org). We integrated all the toolkits without major problems in OsiriX within about 9 months with a single computer programmer.
For easy processing and navigation of 3D temporal and functional image data, we needed to give special attention to the graphical user interface. We used all the features available in the OS X operating system (eg, drag and drop, multiple windows, thumbnails, and customizable tool-bars).
Because the OsiriX software is designed to facilitate the manipulation of image data from all time points of the examination in multidimensional space, the user has the ability to switch from the axial plane to any oblique plane while changing the section thickness, displaying a dynamic image sequence, and adjusting the blending of fused anatomic and functional images. The simultaneous adjustment of the various display parameters is facilitated by specially designed tools, and toolbars can be further customized by the user as needed. To further enhance the efficiency of display adjustment, several keyboard shortcuts are available. However, given the number of parameters that must be adjusted to optimize the display of a given image data set, it can be cumbersome to adjust all the different parameters by using the mouse and other traditional tools. Therefore, we evaluated various hardware devices for the simultaneous adjustment of display parameters (eg, contrast and intensity, zoom, rate of cine display, section thickness, transparency, blending of fused images). We selected a programmable jog wheel, a device that is commonly used in video editing, because it allows easy adjustment of multiple display parameters with a single click on a button, followed by fine tuning with the jog wheel (see Fig 2). This device offers a larger number of programmable buttons than does a classic multibutton mouse and provides speed and high precision for controlling the frame rate of dynamic image series such as four-dimensional (4D) cardiac CT scans or adjusting the percentage of blending in PET/CT fusion images.
View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. Photograph shows the use of a programmable multifunction jog wheel, which allows users to adjust image navigation and display parameters in OsiriX more easily and efficiently than is possible with a traditional mouse. WL = window level, WW = window width.
|
|
OsiriX was developed as an open-source software program and is distributed free of charge, in accordance with the GNU General Public License (8). Because we used open-source components to create OsiriX, we also distribute it as freeware so as to maintain conformity with the stipulations of the GNU General Public License. We believe that the open-source model provides benefits to users through the provision of free software and to developers by generating collaboration toward a shared but modifiable platform. The source code and compiled program may be downloaded free of charge from our Web site (http://homepage.mac.com/rossetantoine/OsiriX).
|
Results
|
|---|
We successfully deployed the OsiriX software within an existing PACS infrastructure. To our existing capabilities to receive, query, and retrieve image data from any DICOM-compliant device, OsiriX added the value of efficient management and manipulation of the large image data sets generated at 3D MR imaging and multidetector CT, functional and molecular image data sets obtained with hybrid PET/CT, and dynamic image series obtained with electrocardiographically gated cardiac CT and cardiac MR imaging. To determine the value of our multidimensional rendering and visualization techniques and to validate the adequacy and convenience of the graphical user interface, we tested the following applications: the rendering of PET/CT fusion images, the fusion of MR images obtained with different sequences and techniques, and the combination of dynamic contrast-enhanced images and functional images obtained at different time points in the same examination.
PET/CT Fusion Images
PET/CT images are a perfect example of fusion images, which are created by combining anatomic image data (in this case, from CT) with functional image data (in this case, from PET). Various color blending and transparency techniques may be used to achieve image fusion (9–11). The most commonly used technique is based on the combination of a gray-scale anatomic image (eg, from CT) with a color-coded functional image (eg, from PET). Investigators in a number of clinical studies have shown that the interpretation of combined PET/CT images results in an improvement of overall diagnostic accuracy beyond that achieved with the interpretation of PET images and CT images separately. Chin et al (12) reported their findings in a series of 30 patients examined with PET and CT for mediastinal lymph node involvement and concluded that combined information from PET and CT yielded the highest diagnostic accuracy (90%). Similarly, Weng et al (13) reported a higher diagnostic accuracy with combined PET and CT than with either PET or CT alone for lung cancer staging.
OsiriX software can be used to generate a fused image set from two series of images by simply dragging and dropping one image series over the other. After the fused image set is generated, the user can further adjust the degree of blending or transparency of one image over the other. The hybrid image series, like any single-modality image series, then can be manipulated by using the available software tools for processing and display. The user can choose to generate images in an oblique plane of adjustable thickness across the fused sets of images or to apply 3D postprocessing techniques such as volume rendering or maximum intensity projection. Figure 3 shows the use of OsiriX to display PET/CT fusion images as an axial tomographic section, a coronal reconstruction, and a 3D volume-rendered image.
View larger version (61K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Fusion of functional (PET) and anatomic (CT) image data by using OsiriX. Left: Two-dimensional axial fusion image. Middle: Multiplanar reformatted fusion image. Right: Three-dimensional volume-rendered fusion image.
|
|
Fusion of MR Images from Different Sequences
Whereas PET/CT is the most widely known hybrid modality that benefits from fused image display and rendering, little attention has been given to the application of fusion techniques to MR images acquired with different sequences in the same examination. Different MR imaging techniques can provide high-definition anatomic information as well as functional and biologic information. While T1-weighted sequences provide good anatomic detail and tissue differentiation, T2-weighted sequences can provide valuable information about edema and the fluid content of the tissues. Fusion MR images can be generated with the same technique used to obtain PET/CT images, by superimposing the color-coded functional T2-weighted image over the grayscale T1-weighted anatomic image. Figure 4 shows the fusion of T1- and T2-weighted short inversion time inversion recovery images of a scaphoid fracture.
View larger version (177K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. OsiriX-based fusion of MR images obtained with different sequences. (a) T1-weighted MR image demonstrates a linear area of signal hypointensity along the waist of the scaphoid (arrowhead). (b) Image obtained with a short inversion time inversion recovery sequence shows a corresponding area of edema (arrow). (c) Image obtained with fusion of a and b clearly demonstrates abnormal signal intensity along the scaphoid fracture line and fluid within the radioscaphoid articulation (arrow).
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. OsiriX-based fusion of MR images obtained with different sequences. (a) T1-weighted MR image demonstrates a linear area of signal hypointensity along the waist of the scaphoid (arrowhead). (b) Image obtained with a short inversion time inversion recovery sequence shows a corresponding area of edema (arrow). (c) Image obtained with fusion of a and b clearly demonstrates abnormal signal intensity along the scaphoid fracture line and fluid within the radioscaphoid articulation (arrow).
|
|
View larger version (181K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c. OsiriX-based fusion of MR images obtained with different sequences. (a) T1-weighted MR image demonstrates a linear area of signal hypointensity along the waist of the scaphoid (arrowhead). (b) Image obtained with a short inversion time inversion recovery sequence shows a corresponding area of edema (arrow). (c) Image obtained with fusion of a and b clearly demonstrates abnormal signal intensity along the scaphoid fracture line and fluid within the radioscaphoid articulation (arrow).
|
|
Combination of Contrast-enhanced Images
The addition of a contrast medium to the CT or MR imaging examination allows the radiologist to obtain functional information by observing the transit of the contrast medium through different organs and anatomic structures. With color coding and image fusion techniques, it is possible to generate functional images that contain different temporal data. To illustrate this concept, we used OsiriX to blend abdominal MR angiograms acquired during the arterial and portal phases (Fig 5).
View larger version (83K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5. OsiriX-based fusion of dynamic contrast-enhanced abdominal MR images acquired in the late portal phase (left) and the early arterial phase (right). The fusion image (center) clearly shows the relation between the veins and arteries.
|
|
Dynamic Cardiac CT Images
With the latest generation of multi–detector row CT scanners, it is possible to acquire 3D images of the heart in different phases of the cardiac cycle, with resultant data sets in four dimensions (14). OsiriX allows the display and manipulation of dynamic images of a beating heart in real time. It is possible, for example, to generate oblique sections with adjustable thickness across the heart to adequately analyze different segmental wall motion anomalies. An example of oblique reformatting of a 4D cardiac image data set is shown in Figure 6.
View larger version (103K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Screen capture shows use of the 4D viewer for multiplanar reconstruction of a cardiac CT image series (20 phases during a single cardiac cycle). The 4D viewer allows simultaneous navigation through 3D and 4D space with real-time selection of the plane of orientation during observation of the beating heart.
|
|
Functional Cardiac MR Images
Color-coded image fusion also can be used to combine dynamic cardiac MR images obtained with different MR imaging sequences. Images obtained with velocity-encoded sequences, for example, can be combined with images obtained with dynamic cine sequences to provide flow velocity information superimposed on anatomic information. The technique used is similar to the superimposition of color Doppler ultrasonographic (US) images on dynamic 2D echocardiograms. Figure 7 shows a 4D image series obtained from the fusion of velocity-encoded MR images with dynamic cine MR images. Velocity encoding and phase encoding were applied by using a lookup color table similar to that used in Doppler US, which resulted in images similar in appearance to Doppler-enhanced echocardiograms.
View larger version (89K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7. Cardiac MR image fusion with use of OsiriX. Fusion image (right), obtained by blending a velocity-encoded and phase-encoded image (left) with an MR angiographic image (not shown), shows an aortic coarctation with areas of high velocity and turbulence.
|
|
Most important, the software tools were designed to be easily accessible and simple to use. The OsiriX user interface is fully customizable; the user selects the tools to be displayed on the toolbar and eliminates those that are not used on a regular basis. The user can use any of the tools at any time and on any type of image. Furthermore, the different tools may be applied in different ways, by using sliding cursors, keyboard shortcuts, or programmable buttons on a multi-button mouse or a multifunction jog wheel. Major efforts were invested to make the navigation and manipulation of large sets of multidimensional images as simple as possible.
|
Discussion
|
|---|
Use of the current generation of PET/CT scanners for cardiac examinations may result in image data sets in all five dimensions. We therefore designed our software to allow users to interactively navigate through the five dimensions. The same image fusion technique and the same 4D viewer may be used also to combine images obtained with different MR imaging sequences. Radiologists traditionally analyze image data sets by mentally combining the information provided by the different acquisition sequences. This concept was referred to by Vogel et al as visual fusion (15). A better definition would probably be "the mental integration or fusion of data from different images displayed separately." The added value that fusion images provide to interpreting physicians as well as referring physicians has been well demonstrated in studies of PET/CT (12,13). Our preliminary observations also confirm that the ability to review and manipulate fused images increases the usability of the information derived from the images. The OsiriX software, which is cost free and easy to use, provides a value commensurate with that previously attainable only with expensive high-performance display workstations.
One of the current limitations of the OsiriX software is that it can be used to fuse only images that are already coregistered. The current version does not provide the tools needed for coregistration of images obtained at different examination times, but future versions will include such tools, which will be adapted from the open-source Insight Toolkit library. The current version of the software is capable of fusing images acquired with hybrid systems such as PET/CT scanners, because the images from both modalities are acquired on the same scanner and without any patient motion and therefore are inherently coregistered. Image fusion may be applied also to images obtained with different sequences during the same examination and without patient movement, as in MR imaging examinations that include several pulse sequences. The flexibility and ease of use of the software led some users to explore new applications of image fusion (eg, the combination of functional and anatomic images).
To further facilitate and improve the image navigation and manipulation functions in five dimensions, we explored innovative solutions by using advanced pointing devices and multidimensional navigation devices that can be used in conjunction with the standard mouse and keyboard. We found that the device that is most adaptable to such functions is a standard programmable jog wheel that is widely used by professional video editors. This device allows rapid navigation through multiple image data sets and multiple dimensions with a single hand. The addition of this low-cost pointing device increases the ability of the user to navigate rapidly and to change display functions in real time.
Because OsiriX software is open source, any user with programming skills will be able to adapt and modify the program for specific needs. The open-source approach also allows the community of users to share new developments that could emerge from different institutions using the OsiriX program. We believe that OsiriX can rapidly evolve to assist radiologists and physicians in interpreting new multidimensional examinations. It has greater potential for rapid adaptation to the emerging needs of the medical community than do commercial software solutions. The latter are typically less flexible because their evolution is driven by market rules; vendors are reluctant to adopt innovations that serve only a minority of users and that may lack sufficient commercial value.
|
Conclusions
|
|---|
New possibilities for display of radiologic images are evolving. Conventional sets of 2D tomographic sections are giving way to 3D volumetric image data sets, which in turn may be extended into a fourth or fifth dimension with the addition of temporal and functional data acquired with fast CT or MR imaging systems and with combined PET/CT scanners. To allow radiologists and clinicians to conveniently and efficiently interpret these large image data sets, traditional image viewers commonly available on PACS workstations have to be redesigned and tailored to a new paradigm of multidimensional image display, navigation, and manipulation. This is why we elected to develop a completely new image navigation environment. By making it available free of charge, in accordance with open-source standards, we expect to encourage the participation of other institutions that are willing to invest time and resources in the development and exploration of new means for complex image interpretation and diagnostic tasks. The OsiriX software also provides an attractive and cost-effective alternative to radiologists and health care providers who have increasing needs for multidimensional image processing and manipulation capabilities but cannot afford the high-priced workstations currently available. The availability of OsiriX also facilitates communication between radiologists and referring physicians by allowing them to share the same convenient platform for image display and navigation. We also believe that such tools are critical for patient care and patient management, given the increasing number of image-based therapeutic and surgical procedures. In these innovative procedures, it is critical for adequate patient care that the performing physician have full access to the same tools and image data that are available to the interpreting radiologists. Furthermore, we also believe that a wider availability of 3D visualization tools and multidimensional navigation tools will greatly enhance the quality of communication between physicians and patients. Physicians can use these tools to educate patients about their illness and to better explain the effect of various possible treatments or therapeutic procedures.
OsiriX also provides a completely new perspective on medical imaging in general by providing the necessary tools for exploring the fourth dimension (time) and the fifth dimension (organic function or molecular process) in cardiac CT and PET/CT examinations. With the rapid evolution of imaging modalities and the shift toward molecular imaging examinations, the new paradigm of multidimensional image navigation may become even more prevalent in the performance of general diagnostic tasks, in which the interpreting physician will have to integrate information from functional and/or molecular imaging into the diagnostic process. The addition of image fusion and multidimensional visualization to conventional image display and interpretation methods—for example, in the combination of T1- and T2-weighted MR images or of functional and dynamic MR images—also can facilitate image review and improve the quality of interpretation. A four-dimensional review that includes dynamic cardiac images or a combination of flow velocity–sensitive acquisitions with phase-encoded acquisitions is becoming a requirement for adequate interpretation of cardiac and cardiovascular imaging examinations.
The development of OsiriX as a new image navigation platform and its distribution as an open-source software product allowed us to demonstrate the potential value of 4D viewers and image fusion with functional or molecular data. We believe that such tools will be an essential part of the next generation of DICOM viewers and PACS workstations, just as 3D rendering and multiplanar reformatting have made their way into the mainstream and have become standard techniques of modern radiology.
|
TAKE-HOME POINTS
|
|---|
OsiriX is a free and open-source Digital Imaging and Communications in Medicine– compliant viewer.
OsiriX facilitates multimodality image fusion.
OsiriX allows manipulation of four-dimensional image data during interactive display.
|
Footnotes
|
|---|
Abbreviations: DICOM = Digital Imaging and Communications in Medicine, 4D = four-dimensional, PACS = picture archiving and communication system, 3D = three-dimensional, 2D = two-dimensional
|
References
|
|---|
- Salgado R, Mulkens T, Bellinck P, Termote JL. Volume rendering in clinical practice: a pictorial review. JBR-BTR 2003;86(4):215–220.
- Mahnken AH, Wildberger JE, Sinha AM, et al. Value of 3D-volume rendering in the assessment of coronary arteries with retrospectively ECG-gated multislice spiral CT. Acta Radiol 2003; 44(3):302–309.[CrossRef][Medline]
- Kamel IR, Fishman EK. Recent advances in CT imaging of liver metastases. Cancer J 2004;10(2): 104–120.[Medline]
- Inoue T, Tamaki Y, Hamada S. Usefulness of three-dimensional multidetector-row CT images for preoperative evaluation of tumor extension in primary breast cancer patients. Breast Cancer Res Treat 2005;89(2):119–125.[CrossRef][Medline]
- Flohr T, Ohnesorge B, Bruder H, et al. Image reconstruction and performance evaluation for ECG-gated spiral scanning with a 16-slice CT system. Med Phys 2003;30(10):2650–2662.[CrossRef][Medline]
- Wiggins RH 3rd, Davidson HC, Harnsberger HR, Lauman JR, Goede PA. Image file formats: past, present, and future. RadioGraphics 2001;21(3): 789–798.[Abstract/Free Full Text]
- OpenGL overview. OpenGL Web site. http://opengl.org. Updated January 2005. Accessed January 2005.
- GNU General Public License (GPL). GPL Web site. http://www.gnu.org/copyleft/gpl.html. Updated June 1991. Accessed February 10, 2005.
- Lemke AJ, Niehues SM, Hosten N, et al. Retrospective digital image fusion of multidetector CT and 18F-FDG PET: clinical value in pancreatic lesions—a prospective study with 104 patients. J Nucl Med 2004;45(8):1279–1286.[Abstract/Free Full Text]
- Shimizu S, Suzuki H, Maki H, et al. A novel image fusion visualizes the angioarchitecture of the perforating arteries in the brain. AJNR Am J Neuroradiol 2003;24(10):2011–2014.[Abstract/Free Full Text]
- Behrenbruch CP, Marias K, Armitage PA, et al. Fusion of contrast-enhanced breast MR and mammographic imaging data. Med Image Anal 2003; 7(3):311–340.[CrossRef][Medline]
- Chin R Jr, Ward R, Keyes JW, et al. Mediastinal staging of non-small-cell lung cancer with positron emission tomography. Am J Respir Crit Care Med 1995;152(6 pt 1):2090–2096.[Abstract]
- Weng E, Tran L, Rege S. Accuracy and clinical impact of mediastinal lymph node staging with FDG-PET imaging in potentially resectable lung cancer. Am J Clin Oncol 2000;23(1):47–52.[CrossRef][Medline]
- Saito K, Saito M, Komatu S, et al. Real-time four-dimensional imaging of the heart with multi-detector row CT. RadioGraphics 2003;23(1):e8. http://radiographics.rsnajnls.org/cgi/content/full/23/1/e8. Accessed February 10, 2005.[Abstract/Free Full Text]
- Vogel WV, Oyen WJ, Barentsz JO, Kaanders JH, Corstens FH. PET/CT: panacea, redundancy, or something in between? J Nucl Med 2004;45(suppl 1):15S–24S.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
R.K. Fulbright, C. Hoffmann, H. Lee, A. Pozamantir, J. Chapman, and I. Prohovnik
MR Imaging of Familial Creutzfeldt-Jakob Disease: A Blinded and Controlled Study
AJNR Am. J. Neuroradiol.,
October 1, 2008;
29(9):
1638 - 1643.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. L. Ehman, W. R. Hendee, M. J. Welch, N. R. Dunnick, L. B. Bresolin, R. L. Arenson, S. Baum, H. Hricak, and J. H. Thrall
Blueprint for Imaging in Biomedical Research
Radiology,
July 1, 2007;
244(1):
12 - 27.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
