DOI: 10.1148/rg.242035131
RadioGraphics 2004;24:343-356
© RSNA, 2004
CT with 3D Rendering of the Tendons of the Foot and Ankle: Technique, Normal Anatomy, and Disease1
Robert H. Choplin, MD,
Kenneth A. Buckwalter, MD,
Jonas Rydberg, MD and
Joshua M. Farber, MD
1 From the Department of Radiology, 0279, Indiana University School of Medicine, 550 N University Blvd, Indianapolis, IN 46202-5253. Presented as an education exhibit at the 2001 RSNA scientific assembly. Received May 12, 2003; revision requested June 12 and received July 24; accepted July 25. Address correspondence to R.H.C. (e-mail: rchoplin@iupui.edu).
 |
Abstract
|
|---|
Three-dimensional rendering of computed tomographic data with volume rendering (VR), shaded surface display (SSD), and maximum intensity projection has been performed for over 20 years. In the foot and ankle, no one image reformatting technique is satisfactory for displaying every anatomic relationship or disease process. Two-dimensional multiplanar reformatted (MPR) images are the basic images used for diagnosis. MPR images are especially useful for identifying small fractures. VR is useful for demonstrating the relationships between ankle tendons and the underlying osseous structures, and SSD is useful when fractures extend to the articular cortex and a disarticulated view is desired. Three-dimensional images are helpful in patients with congenital deformities, arthritis, and trauma.
© RSNA, 2004
Index Terms: Ankle, anatomy, 46.92 • Foot, anatomy, 46.92 • Foot, CT, 46.12117 • Images, display, 46.12117 • Images, three-dimensional, 46.12117 • Tendons, 46.12117, 46.92
|
Introduction
|
|---|
Methods used to acquire and display x-ray images of the ankle include radiography, computed tomography (CT) performed in a predefined plane, two-dimensional (2D) multiplanar reformation (MPR) of the CT data, and three-dimensional (3D) projections of the CT data. It is important to select the proper technique for optimal display of the anatomy and disease processes.
Axial CT is, by design, a 2D cross-sectional imaging method. In regions of complex anatomy such as the ankle and hindfoot, axial CT scans can be challenging to interpret. For example, the precise location of the articular facets of the subtalar joints may be difficult to appreciate on routine axial scans. MPR images created from the CT data provide an additional perspective that can improve depiction of the subtalar joint anatomy (Fig 1). However, because MPR images are cross-sectional, they cannot provide a global depiction of the relationship of the bones to the surrounding tendons and ligaments.
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1d. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1e. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1f. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1g. Articular facets of the subtalar joint. (a) On an axial MPR image obtained at the level of the talocalcaneal articulations, the posterior calcaneal facet (black arrow) is well depicted, but the middle (arrowhead) and anterior (curved white arrow) facets are not. There is a fracture at the medial end of the navicular bone (straight white arrow). (b-d) Coronal MPR images (b is the most posterior, d is the most anterior) show the posterior facet (arrow in b), middle facet (arrowhead in c), and anterior facet (curved arrow in d). The fracture at the medial end of the navicular bone is again noted (straight arrow in d). (e-g) Sagittal MPR images show the posterior facet (arrow in e), middle facet (arrowhead in f), and anterior facet (arrow in g).
|
|
Three-dimensional images enable radiologists to highlight and display the individual osseous and soft-tissue components of complex anatomy.
Three-dimensional rendering is the process of creating 2D images that convey the 3D relationships of an object or objects. The various rendering techniques make use of different amounts of information from the CT data set to display the desired structure or structures. The large data sets obtained with spiral and multi–detector row CT have improved the potential for creating superior 3D images, which is accomplished with software programs running on powerful workstations. A variety of software programs are available; they may be vendor specific or created by third parties.
In this article, we review the types of 3D rendering techniques available: volume rendering (VR), shaded surface display (SSD), and maximum intensity projection (MIP), noting the strengths and weaknesses of each technique and comparing these 3D techniques with MPR, a 2D technique. We also discuss CT technique and the applications of VR in demonstrating the normal anatomy and disease processes of the bones and tendons in the foot and ankle.
|
Three-dimensional Rendering Techniques
|
|---|
The primary methods for rendering a volume of CT data to create 3D images are VR (Fig 2), SSD (Fig 3), and MIP (Fig 4) (1).
View larger version (77K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. VR imaging of the foot and ankle (same patient as in Fig 1). (a) VR image of the bones depicts the osseous structures of the foot and ankle. (b) VR image of the bones and tendons demonstrates the relationship between the two types of tissue. The plantar fascia and some of the plantar musculature have an attenuation similar to that of tendons.
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. VR imaging of the foot and ankle (same patient as in Fig 1). (a) VR image of the bones depicts the osseous structures of the foot and ankle. (b) VR image of the bones and tendons demonstrates the relationship between the two types of tissue. The plantar fascia and some of the plantar musculature have an attenuation similar to that of tendons.
|
|
Volume Rendering
As a result of improvements in computer software and hardware over the past 10 years, VR imaging has become a common 3D display technique (2,3). The steps required to create a VR image are illustrated in Figure 5. Virtual light rays are cast from a viewing point through the tissue or tissues of interest. These light rays are attenuated according to their interaction with the tissues they encounter. Tissues are defined by assigning a specific tissue type (eg, muscle) to a specific range of Hounsfield units. Each tissue is assigned a color, brightness, and degree of opacity depending on its range of Hounsfield units. Usually, the tissue of most interest is rendered as the most opaque (eg, 100%), whereas other tissues are rendered as less opaque (eg, 50%) or clear (0%). This process is semiautomatic and may be performed by choosing or modifying a manufacturer-supplied protocol or by creating a custom protocol. Either the image or the histogram of the CT data set may be used to refine the tissue definitions. With use of the histogram, the height and shape of the brightness and opacity curves may be changed interactively to optimize tissue display. The rendered image data provide almost no depth cues, but the assignment of the various tissue opacities results in the appearance of depth. Finally, the virtual rays are cast from an essentially infinite number of positions on a virtual sphere surrounding the CT data set. This last feature enables one to rotate the image on a workstation to view the image from any desired perspective.
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Drawings illustrate the steps needed to create a VR image. (a) Step 1: Choose a viewing point. Step 2: Create a histogram of the CT data set to determine the total number of voxels and the total range of Hounsfield units. (b) Step 3: From the histogram, determine the range of Hounsfield units for the tissues of interest. (c) Step 4: Assign opacity values to the tissues of interest. Opacity values range from 0% (transparent) to 100% (fully opaque). The tissue of interest in this example is bone. (d) Step 5: Cast rays from the viewing point to each point in the volume. Step 6: Determine the opacity value of all the voxels along the path of each ray. Step 7: Place the sum of all the opacity values for each ray in the final image. Because the opacity values are summed, there is no depth information (unlike with SSD).
|
|
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Drawings illustrate the steps needed to create a VR image. (a) Step 1: Choose a viewing point. Step 2: Create a histogram of the CT data set to determine the total number of voxels and the total range of Hounsfield units. (b) Step 3: From the histogram, determine the range of Hounsfield units for the tissues of interest. (c) Step 4: Assign opacity values to the tissues of interest. Opacity values range from 0% (transparent) to 100% (fully opaque). The tissue of interest in this example is bone. (d) Step 5: Cast rays from the viewing point to each point in the volume. Step 6: Determine the opacity value of all the voxels along the path of each ray. Step 7: Place the sum of all the opacity values for each ray in the final image. Because the opacity values are summed, there is no depth information (unlike with SSD).
|
|
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. Drawings illustrate the steps needed to create a VR image. (a) Step 1: Choose a viewing point. Step 2: Create a histogram of the CT data set to determine the total number of voxels and the total range of Hounsfield units. (b) Step 3: From the histogram, determine the range of Hounsfield units for the tissues of interest. (c) Step 4: Assign opacity values to the tissues of interest. Opacity values range from 0% (transparent) to 100% (fully opaque). The tissue of interest in this example is bone. (d) Step 5: Cast rays from the viewing point to each point in the volume. Step 6: Determine the opacity value of all the voxels along the path of each ray. Step 7: Place the sum of all the opacity values for each ray in the final image. Because the opacity values are summed, there is no depth information (unlike with SSD).
|
|
View larger version (44K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5d. Drawings illustrate the steps needed to create a VR image. (a) Step 1: Choose a viewing point. Step 2: Create a histogram of the CT data set to determine the total number of voxels and the total range of Hounsfield units. (b) Step 3: From the histogram, determine the range of Hounsfield units for the tissues of interest. (c) Step 4: Assign opacity values to the tissues of interest. Opacity values range from 0% (transparent) to 100% (fully opaque). The tissue of interest in this example is bone. (d) Step 5: Cast rays from the viewing point to each point in the volume. Step 6: Determine the opacity value of all the voxels along the path of each ray. Step 7: Place the sum of all the opacity values for each ray in the final image. Because the opacity values are summed, there is no depth information (unlike with SSD).
|
|
The major advantage of VR is its capacity to display multiple tissues and show their relationships to one another. VR programs retain all of the information within the CT data set and display one or more tissues depending on the chosen protocol. This technique is helpful in musculoskeletal applications, where it may be particularly important to display the relationships of bone to soft tissue.
Preparatory editing of the image data may be required. For example, if a display of the articular surface of the bone is desired, the unwanted bone must be "removed" manually to accomplish joint disarticulation. This modification may be performed on a section-by-section basis or by editing the displayed volume interactively.
To show separate tissues in different colors and with different attenuation values, VR programs retain and manipulate the entire CT data set. The larger data set requires a more powerful workstation with more memory than is needed for either SSD or MIP. However, this limitation has been offset by the evolution of powerful and inexpensive workstations and the decreased cost of electronic storage. Because the creation of VR images involves the manipulation of separate tissues having different colors and attenuation values, the learning curve for navigating VR software is longer than for SSD or MIP.
Shaded Surface Display
Historically, SSD has been the most commonly used 3D technique (4). SSD became popular initially because it requires a smaller amount of data than VR and can therefore be implemented on less powerful computers. In addition, SSD images require less electronic storage space. The steps required to create an SSD image are shown in Figure 6. Tissues are defined by assigning a specific tissue type (eg, bone) to a specific threshold of Hounsfield units. Voxels below this threshold are discarded (ie, made transparent), and voxels above this threshold are assigned to the tissue. The set of retained voxels represents the surface of the tissue of interest. Virtual light rays are cast toward the object until they encounter a voxel at or above the threshold level. The SSD image takes on a 3D quality when a gray value is assigned that is proportional to the distance between the vantage point and the surface of the tissue of interest. Afterward, the object may be rotated in space and viewed from different vantage points, recalculating the surface display by determining the distance between the new vantage point and the surface of the tissue of interest. SSD is usually chosen when one is interested in the 3D shape of only one tissue type.
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. Drawings illustrate the steps required to create an SSD image. (a) Step 1: Choose a viewing point. Step 2: Select a threshold value (in Hounsfield units) for the tissue of interest (eg, bone >150 HU). (b) Step 3: Cast rays from the viewing point to each point in the volume. Step 4: Identify the first point along each ray that is at the selected threshold value. Step 5: Measure the distance from the viewing point to the threshold point and place a gray pixel in the final image; discard all other points along the ray. Short distances produce bright pixels and longer distances produce dark pixels. This coding provides depth information by means of shading.
|
|
View larger version (50K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. Drawings illustrate the steps required to create an SSD image. (a) Step 1: Choose a viewing point. Step 2: Select a threshold value (in Hounsfield units) for the tissue of interest (eg, bone >150 HU). (b) Step 3: Cast rays from the viewing point to each point in the volume. Step 4: Identify the first point along each ray that is at the selected threshold value. Step 5: Measure the distance from the viewing point to the threshold point and place a gray pixel in the final image; discard all other points along the ray. Short distances produce bright pixels and longer distances produce dark pixels. This coding provides depth information by means of shading.
|
|
Because of the thresholding function, preparatory editing of the images may be required to remove unwanted structures or tissues that may obstruct the view. For example, contrast material–filled vessels and the skeleton have a similar attenuation, and the former must be edited out of the data set before the skeleton may be displayed. With older SSD software, the editing was tedious and sometimes required manual editing of each individual CT section. The time required to create an SSD image was a major drawback. With modern programs, improved editing tools allow easier removal of unwanted portions of an image. The combination of editing and thresholding results in only 10% of the CT data being used.
Tissues may be shown as either completely or partially opaque. It is possible to display multiple objects (tissues), but they must be rendered separately and then "fused" into the final display. The tissues retain their 3D spatial relationships (Fig 7). Although it is possible to render the bones and tendons of the ankle with SSD, the number and course of tendons makes this process tedious. If disarticulation of a joint is needed with SSD, the individual bones may be rendered separately and one or more of the bones removed from the image.
The SSD technique has several disadvantages. First, creating the displays may be time consuming because of preprocessing editing, although the advent of more powerful workstations and improved rendering programs has significantly ameliorated this disadvantage. Second, the size of objects may be inaccurately displayed if an inappropriate threshold is chosen. Because of the binary nature of SSD, voxels that volume average two different tissues may or may not be rendered in the display depending on the chosen threshold. As a result, the rendered image may contain stairstep artifacts and the size of objects in the image may be affected. Finally, if there are wide gaps between similar tissues (eg, distracted fracture fragments), the program may render only one of the fragments. To render the entire bone, the fragments must be rendered separately and then fused into the final image.
Maximum Intensity Projection
The skeleton may also be displayed with MIP (1). The steps needed to create an MIP image are shown in Figure 8. With this technique, a virtual light ray is cast through the CT data set to locate the voxel with the maximum number of Hounsfield units. These maximum-attenuation voxels are retained and all other voxels discarded. As with SSD, these voxels can represent up to 90% of the information in the CT data set. MIP does not display the soft tissues in a useful way because they do not represent the maximum-attenuation voxels throughout the volume when bone is present. In addition, because there is no shading, MIP images do not appear to have depth (Fig 9), and multiple views must be acquired and rotated to demonstrate spatial relationships. MIP images are most useful when metal is present because of the decreased artifacts associated with this rendering technique (Fig 9). In nonmusculoskeletal applications, MIP is most frequently chosen to display blood vessels that demonstrate either contrast material filling at CT or high signal intensity at magnetic resonance (MR) imaging.
View larger version (51K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8. Drawings illustrate the steps needed to create an MIP image. Step 1: Choose a viewing point. Step 2: Cast rays from the viewing point to each point in the volume. Step 3: Determine the maximum pixel value of all the pixels along the path of each ray. Step 4: Place the maximum pixel value of each ray in the final image.
|
|
View larger version (80K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. Three-dimensional rendering with metal present. (a) On a multitissue VR image, artifacts from metal (arrows) interfere with tendon visualization. (b) MIP image shows the bone and metal but not the soft tissues and tendons.
|
|
View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. Three-dimensional rendering with metal present. (a) On a multitissue VR image, artifacts from metal (arrows) interfere with tendon visualization. (b) MIP image shows the bone and metal but not the soft tissues and tendons.
|
|
|
Technical Limitations of 3D Rendering
|
|---|
Fine detail can be lost with 3D rendering. With SSD and MIP, large portions of the original data set are discarded. With VR, most of the voxel information present in the original data set is retained, so that detailed structures can sometimes be identified. Unfortunately, unless image editing is performed, overlapping structures can obscure some of the detail in the region of interest, underscoring the need to review VR images from multiple perspectives. Consequently, primary diagnosis of bone abnormalities (eg, fractures) is best accomplished with the source CT scans or MPR images (Fig 10). Three-dimensional images are useful in providing a global depiction of the bones and other tissues when complex fractures or anatomic relationships are present.
View larger version (165K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Limitations of 3D rendering technique. (a) Sagittal MPR image obtained in a 23-year-old man who sustained an acute twisting injury and presented with pain and swelling of the left ankle shows a small fracture of the anterior process of the calcaneus (arrow). (b, c) Neither a VR image of the bones (b) nor an SSD image (c) demonstrates the anterior process fracture.
|
|
View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Limitations of 3D rendering technique. (a) Sagittal MPR image obtained in a 23-year-old man who sustained an acute twisting injury and presented with pain and swelling of the left ankle shows a small fracture of the anterior process of the calcaneus (arrow). (b, c) Neither a VR image of the bones (b) nor an SSD image (c) demonstrates the anterior process fracture.
|
|
View larger version (113K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10c. Limitations of 3D rendering technique. (a) Sagittal MPR image obtained in a 23-year-old man who sustained an acute twisting injury and presented with pain and swelling of the left ankle shows a small fracture of the anterior process of the calcaneus (arrow). (b, c) Neither a VR image of the bones (b) nor an SSD image (c) demonstrates the anterior process fracture.
|
|
In patients with fractures involving a joint, disarticulation with removal of the bone on one side of the joint allows visualization of the fracture lines at the joint surface (Fig 11). Such views may be created with either VR or SSD.
View larger version (101K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Global assessment of fractures involving joints. (a) SSD image demonstrates a pilon fracture of the distal tibia (white) and fibula (light blue). Arrows indicate fracture, dark blue area indicates the tarsal bones. (b) Disarticulated SSD image of the joint surface shows the fracture components in the tibia (white) and fibula (blue).
|
|
View larger version (83K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Global assessment of fractures involving joints. (a) SSD image demonstrates a pilon fracture of the distal tibia (white) and fibula (light blue). Arrows indicate fracture, dark blue area indicates the tarsal bones. (b) Disarticulated SSD image of the joint surface shows the fracture components in the tibia (white) and fibula (blue).
|
|
|
CT Technique
|
|---|
CT technique is very important because it determines the quality of the data set used for subsequent 3D rendering. We use a four-channel CT scanner (Philips MX8000; Philips Medical Systems, Cleveland, Ohio) with the scanning and reconstruction parameters listed in Table 1. The section thickness and reconstruction interval are selected such that our data have nearly isotropic voxels.
A medium sharp or bone algorithm will improve the quality of MPR images. A softer algorithm will improve the quality of subsequent SSD images of bone surfaces or VR images of soft-tissue structures like tendons by reducing the apparent noise of the CT data set. If both MPR and 3D images are desired, initial reconstruction with a bone algorithm followed by a second reconstruction of the raw data with a smoother algorithm may be desirable. If the original raw data cannot be reconstructed, it may be possible to negatively enhance (smooth) the original reconstructed images and use these smoother images for any subsequent 3D imaging.
Although overlapping sections are preferable for creating MPR images, the large data set may result in prolonged computation times for some VR programs. Selecting contiguous images (ie, every other image of a 50% overlapped data set) will halve the computation time requirement and speed further processing.
|
VR Applications in the Ankle and Foot
|
|---|
With VR, it is possible to demonstrate both the tendons and bones of the ankle and hindfoot. This depiction is possible because the tendons are surrounded by fat and have an attenuation of approximately 90 HU, which is intermediate between the attenuation of bone and muscle. In the setting of acute trauma, surrounding edema and hemorrhage may hamper the ability to visualize tendons.
Normal Anatomy
The normal VR appearance of the bones and tendons of the ankle and hindfoot are shown in Figure 12. The origins, insertions, and actions of the tendons are shown in Table 2.
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12a. Normal tendinous foot and ankle anatomy. (a) VR image (posterior medial view) shows the tibialis posterior tendon (1), flexor digitorum tendon (2), flexor hallucis longus tendon (3), and plantar aponeurosis (4). (b) VR image (lateral view) demonstrates the Achilles tendon (5), peroneus brevis tendon (6), and peroneus longus tendon (7). (c) VR image of the dorsum of the midfoot shows the tibialis anterior tendon (8), extensor hallucis longus tendon (9), and extensor digitorum tendon (10). The origins, insertions, and actions of these tendons are shown in Table 2.
|
|
View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12b. Normal tendinous foot and ankle anatomy. (a) VR image (posterior medial view) shows the tibialis posterior tendon (1), flexor digitorum tendon (2), flexor hallucis longus tendon (3), and plantar aponeurosis (4). (b) VR image (lateral view) demonstrates the Achilles tendon (5), peroneus brevis tendon (6), and peroneus longus tendon (7). (c) VR image of the dorsum of the midfoot shows the tibialis anterior tendon (8), extensor hallucis longus tendon (9), and extensor digitorum tendon (10). The origins, insertions, and actions of these tendons are shown in Table 2.
|
|
View larger version (86K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12c. Normal tendinous foot and ankle anatomy. (a) VR image (posterior medial view) shows the tibialis posterior tendon (1), flexor digitorum tendon (2), flexor hallucis longus tendon (3), and plantar aponeurosis (4). (b) VR image (lateral view) demonstrates the Achilles tendon (5), peroneus brevis tendon (6), and peroneus longus tendon (7). (c) VR image of the dorsum of the midfoot shows the tibialis anterior tendon (8), extensor hallucis longus tendon (9), and extensor digitorum tendon (10). The origins, insertions, and actions of these tendons are shown in Table 2.
|
|
Disease Processes
We have found VR images to be helpful as adjunct images in patients with foot deformities, arthritis, and trauma. The following disease processes are illustrated in this article: hypertrophic peroneal tubercle (Fig 13), rheumatoid arthritis with a thin peroneal brevis tendon (Fig 14), rheumatoid arthritis with severe erosion (Fig 15), degenerative joint disease with tendon thinning (Fig 16), old calcaneal fracture with displaced peroneal tendons (Fig 17), and acute calcaneal fracture with tendon entrapment (Fig 18).
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13a. Hypertrophic peroneal tubercle in a 22-year-old man who presented with a hard, tender mass on the lateral aspect of the foot. (a) SSD image of the skin shows a protuberance just distal to the lateral malleolus (arrow). (b) Lateral VR image of the bones and tendons shows that a hypertrophic peroneal tubercle of the calcaneus is responsible for the protuberance. The peroneal tendons are splayed around the tubercle (arrow). (c) Lateral SSD image shows enlargement of the peroneal tubercle (arrow). Blue area indicates the fibula.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13b. Hypertrophic peroneal tubercle in a 22-year-old man who presented with a hard, tender mass on the lateral aspect of the foot. (a) SSD image of the skin shows a protuberance just distal to the lateral malleolus (arrow). (b) Lateral VR image of the bones and tendons shows that a hypertrophic peroneal tubercle of the calcaneus is responsible for the protuberance. The peroneal tendons are splayed around the tubercle (arrow). (c) Lateral SSD image shows enlargement of the peroneal tubercle (arrow). Blue area indicates the fibula.
|
|
View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13c. Hypertrophic peroneal tubercle in a 22-year-old man who presented with a hard, tender mass on the lateral aspect of the foot. (a) SSD image of the skin shows a protuberance just distal to the lateral malleolus (arrow). (b) Lateral VR image of the bones and tendons shows that a hypertrophic peroneal tubercle of the calcaneus is responsible for the protuberance. The peroneal tendons are splayed around the tubercle (arrow). (c) Lateral SSD image shows enlargement of the peroneal tubercle (arrow). Blue area indicates the fibula.
|
|
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14. Rheumatoid arthritis in a 43-year-old woman with severe erosive joint destruction. Lateral VR image shows marked thinning of the peroneus brevis tendon (black arrow). White arrow indicates the peroneus longus tendon.
|
|
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15. Rheumatoid arthritis in a 43-year-old woman with severe erosive joint destruction. Medial multitissue VR image emphasizes the tendons. Marked synovial proliferation is seen (arrow) and appears similar to muscles and tendons because of similar attenuation.
|
|
View larger version (83K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16a. Degenerative joint disease with tendon thinning in a 51-year-old man with ankle pain. (a) VR image of the bones shows osteophytes and bone proliferation at the ankle joint (arrows). (b-d) Medial VR image (b) and axial MPR images (c, d) show thinning of the posterior tibial tendon (arrow).
|
|
View larger version (99K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16b. Degenerative joint disease with tendon thinning in a 51-year-old man with ankle pain. (a) VR image of the bones shows osteophytes and bone proliferation at the ankle joint (arrows). (b-d) Medial VR image (b) and axial MPR images (c, d) show thinning of the posterior tibial tendon (arrow).
|
|
View larger version (107K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16c. Degenerative joint disease with tendon thinning in a 51-year-old man with ankle pain. (a) VR image of the bones shows osteophytes and bone proliferation at the ankle joint (arrows). (b-d) Medial VR image (b) and axial MPR images (c, d) show thinning of the posterior tibial tendon (arrow).
|
|
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16d. Degenerative joint disease with tendon thinning in a 51-year-old man with ankle pain. (a) VR image of the bones shows osteophytes and bone proliferation at the ankle joint (arrows). (b-d) Medial VR image (b) and axial MPR images (c, d) show thinning of the posterior tibial tendon (arrow).
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17a. Healed calcaneal fracture with displaced peroneal tendons in a 43-year-old man. (a) Coronal MPR image shows calcaneal deformity with lateral displacement of a large caudal fragment. The peroneal tendons are also laterally displaced (arrow). In addition, there is lateral fibular displacement and disruption of the normal distal tibiofibular articulation. (b) Anterolateral VR image of the ankle and foot again demonstrates lateral displacement of the peroneal tendons (arrow). Note also the proximal fibular osteophyte. (c) Lateral VR image shows that the peroneal tendons are impinged upon by the fibular osteophyte (arrow).
|
|
View larger version (94K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17b. Healed calcaneal fracture with displaced peroneal tendons in a 43-year-old man. (a) Coronal MPR image shows calcaneal deformity with lateral displacement of a large caudal fragment. The peroneal tendons are also laterally displaced (arrow). In addition, there is lateral fibular displacement and disruption of the normal distal tibiofibular articulation. (b) Anterolateral VR image of the ankle and foot again demonstrates lateral displacement of the peroneal tendons (arrow). Note also the proximal fibular osteophyte. (c) Lateral VR image shows that the peroneal tendons are impinged upon by the fibular osteophyte (arrow).
|
|
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17c. Healed calcaneal fracture with displaced peroneal tendons in a 43-year-old man. (a) Coronal MPR image shows calcaneal deformity with lateral displacement of a large caudal fragment. The peroneal tendons are also laterally displaced (arrow). In addition, there is lateral fibular displacement and disruption of the normal distal tibiofibular articulation. (b) Anterolateral VR image of the ankle and foot again demonstrates lateral displacement of the peroneal tendons (arrow). Note also the proximal fibular osteophyte. (c) Lateral VR image shows that the peroneal tendons are impinged upon by the fibular osteophyte (arrow).
|
|
View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18a. Acute calcaneal fracture with tendon entrapment in a 25-year-old man. (a) Sagittal MPR image shows fracture of the calcaneus with depression of the Boehler angle. (b) Coronal MPR image shows lateral and superior displacement of the calcaneal fragment, which results in entrapment of the peroneal tendons (arrow). (c) Posterolateral VR image shows entrapment of the peroneal tendons between the fibula and calcaneus (arrow). (d) Posterolateral VR image of the bones shows the proximity of the fibula to the calcaneus.
|
|
View larger version (149K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18b. Acute calcaneal fracture with tendon entrapment in a 25-year-old man. (a) Sagittal MPR image shows fracture of the calcaneus with depression of the Boehler angle. (b) Coronal MPR image shows lateral and superior displacement of the calcaneal fragment, which results in entrapment of the peroneal tendons (arrow). (c) Posterolateral VR image shows entrapment of the peroneal tendons between the fibula and calcaneus (arrow). (d) Posterolateral VR image of the bones shows the proximity of the fibula to the calcaneus.
|
|
View larger version (94K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18c. Acute calcaneal fracture with tendon entrapment in a 25-year-old man. (a) Sagittal MPR image shows fracture of the calcaneus with depression of the Boehler angle. (b) Coronal MPR image shows lateral and superior displacement of the calcaneal fragment, which results in entrapment of the peroneal tendons (arrow). (c) Posterolateral VR image shows entrapment of the peroneal tendons between the fibula and calcaneus (arrow). (d) Posterolateral VR image of the bones shows the proximity of the fibula to the calcaneus.
|
|
View larger version (101K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18d. Acute calcaneal fracture with tendon entrapment in a 25-year-old man. (a) Sagittal MPR image shows fracture of the calcaneus with depression of the Boehler angle. (b) Coronal MPR image shows lateral and superior displacement of the calcaneal fragment, which results in entrapment of the peroneal tendons (arrow). (c) Posterolateral VR image shows entrapment of the peroneal tendons between the fibula and calcaneus (arrow). (d) Posterolateral VR image of the bones shows the proximity of the fibula to the calcaneus.
|
|
|
Discussion
|
|---|
Three-dimensional rendering of CT scans was developed by the late 1970s (5). Early programs and computers were difficult and time consuming to use, which limited the usefulness of the technique. By the late 1980s, MPR images had become easy to create and were being widely used for the evaluation of ankle and foot fractures.
Three-dimensional imaging was first applied to the skeleton in 1984 by Vannier et al (6) in the evaluation of craniofacial abnormalities. In 1987, Scott et al (7) used VR for display of acetabular fractures. Three years later, Magid et al (8) used 2D and 3D displays to study 15 patients with ankle fractures and found that the displays altered management in five cases. In 1991, Vannier et al (9) evaluated the usefulness of 3D images in 28 patients with acetabular fractures. They found that the 3D images were often equivalent to conventional radiographs and CT scans in helping identify the fractures. Also in 1991, Allon and Mears (10) described the use of 3D images in the evaluation of and preoperative planning for 22 patients with 30 calcaneal fractures. They developed a standardized protocol for demonstrating these fractures and found this protocol to be helpful in displaying the anatomy.
Not all authors have found 3D images to be helpful, however. Tanyu et al (11) used 3D images to evaluate 23 patients with calcaneal fractures and found no improved diagnostic accuracy. Instead, the authors discovered that the use of 3D rendering resulted in artifacts and that the quality of the rendered images was highly dependent on which thresholds were chosen for rendering. Nevertheless, they noted that the referring surgeons appreciated the 3D images. Freund et al (12) evaluated 45 patients with 47 calcaneal fractures. The patients underwent conventional radiography as well as CT with MPR and SSD images. Orthopedic surgeons and radiologists evaluated images before and after electronic disarticulation. Articular facet involvement was quantified using the axial images as the standard of reference. According to the axial images, there were 90 fractures of facets. Three-dimensional images with electronic disarticulation showed 82 of 90 fractures (91%), sagittal MPR images showed 63 of 90 (70%), and 3D images without disarticulation showed five of 90 (5.5%). The number of fragments was best seen on sagittal MPR images obtained in patients with up to four fragments and was best seen on the axial images if there were more than five fragments. Three-dimensional images showed relatively few fragments. Height reduction of the calcaneus was best seen on sagittal and 3D images. The authors concluded that 3D images should not be obtained in isolation, even if the joints are disarticulated for 3D display. Moreover, they concluded that 3D images without disarticulation were superfluous.
Initial reports of 3D rendered images of the ankle concentrated on the osseous structures. As VR programs have become more widespread, interest in evaluating the tendons and soft tissues has increased (13). To our knowledge, there are no published series of patients with VR evaluation of the ankle tendons. Our patients represent a nonconsecutive retrospective group for whom 3D images have been created. Although these images have been presented to the referring clinicians, there has been no rigorous evaluation of their impact. Further studies are needed to clarify the effectiveness of these images. It is likely that the images will not increase diagnostic accuracy so much as they provide a global display of disease for managing physicians and their patients.
Although 3D rendering programs and workstations have become significantly easier to use over the past 10 years, further improvement is needed. A partial listing of available 3D software packages can be accessed at http://biocomp.stanford.edu/3dreconstruction/software/index.html. The amount of time required for postprocessing of the data to create 3D images should be noted. These images can be created by a skilled operator within 5–10 minutes after the CT data become available. Unfortunately, there is no standardized way of creating the images with consistent results. The final images frequently require input from the radiologist to ensure appropriate display of disease.
Because of its high soft-tissue contrast, MR imaging is the technique of choice for demonstrating muscle and tendon disease. However, MR imaging is a sectional technique that is most frequently performed with gaps between sections. Although MR imaging can be performed with 3D sequences and contiguous sections, these scans take longer and are usually not performed. It is possible to enter MR imaging data into 3D rendering programs, but MR images are more difficult to render than CT scans because CT-based rendering programs are optimized to use Hounsfield units with predictable ranges for different tissues. Unlike CT scans, MR images consist of pixel values with unpredictable ranges. In addition, there are patients who may be unable to undergo MR imaging, and CT with 3D rendering provides an alternative method of evaluation. Ultrasonography is useful in selected cases but is incapable of providing a global display of anatomy.
|
Conclusions
|
|---|
Top
Abstract
Introduction
