Skip to main content
Advanced search
Skip main navigation
Other
Pictorial Essay
November 23, 2012

Musculoskeletal Imaging with Multislice CT

Authors: Kenneth A. Buckwalter, Jonas Rydberg, Kenyon K. Kopecky, Keith Crow, and Edward L. YangAuthor Info & Affiliations
Volume 176, Issue 4
https://doi.org/10.2214/ajr.176.4.1760979
PDF
Multislice CT has brought about major advances in bone and joint imaging. A volumetric image set with isotropic properties can be obtained in a single acquisition with a 0.5-mm slice width. Multislice CT allows extended anatomic coverage with thin slices; large patients and patients with metal hardware in their bodies now can be scanned without sacrificing diagnostic quality. To take full advantage of these capabilities, production of multiplanar reformatted (MPR) images has become an integral part of the examination. Different three-dimensional rendering techniques can be applied to reduce the large image sets into clear pictures for the referring physician and the patient. In our 8-year experience, the detailed MPR images obtained from volumetric data obviate reviewing the original slice data. Multislice helical CT was introduced in 1992 with a dual-slice scanner (CT-Twin; Elscint, Haifa, Israel). Most of the examinations presented in this pictorial essay were produced using a quad-slice scanner (Mx8000; Marconi Medical Systems, Cleveland, OH); a few were produced using a dual-slice machine (Twin; Marconi Medical Systems).

Isotropic Imaging

Narrower slice width yields reformations with higher spatial resolution (Fig. 1A,1B,1C,1D). For narrow slice-width acquisitions, the radiation dose caused by the X-ray penumbra is high in single-slice CT. Because of the geometry of the detectors, multislice CT allows slices thinner than 1 mm with less severe penumbra effects (Fig. 2) than are possible with single-slice CT. Performing CT with thinner slices permits acquisition of MPR images in any plane with high spatial resolution. Radiation exposure can be greater with dual 0.5-mm slice imaging than with quad 1-mm slice imaging. Nonetheless, high-resolution techniques are usually reserved for imaging of the extremities, and extremities do not contain radiosensitive tissue.
Fig. 1A. —Images of 53-year-old woman with advanced degenerative changes in tibiotalar joint were reconstructed from one multislice CT acquisition (dual slice, 0.5-mm slice width, 120kVp, 200 mAs). After scanning, raw data were combined to increase effective slice width. Separate image sets were reconstructed with progressively thicker slices (1.0, 1.5, and 3.0 mm). Each image set was used to produce sagittal reformatted image in same location. Observe degradation of detail in reformats as slice thickness is increased. CT scan obtained with 0.6-mm slice width and 0.3-mm reconstruction interval shows very fine bony details with near-isotropic properties.
Fig. 1B. —Images of 53-year-old woman with advanced degenerative changes in tibiotalar joint were reconstructed from one multislice CT acquisition (dual slice, 0.5-mm slice width, 120kVp, 200 mAs). After scanning, raw data were combined to increase effective slice width. Separate image sets were reconstructed with progressively thicker slices (1.0, 1.5, and 3.0 mm). Each image set was used to produce sagittal reformatted image in same location. Observe degradation of detail in reformats as slice thickness is increased. In CT scan obtained with 1.0-mm slice width and 0.6-mm reconstruction interval, fine details are not as well defined as in A.
Fig. 1C. —Images of 53-year-old woman with advanced degenerative changes in tibiotalar joint were reconstructed from one multislice CT acquisition (dual slice, 0.5-mm slice width, 120kVp, 200 mAs). After scanning, raw data were combined to increase effective slice width. Separate image sets were reconstructed with progressively thicker slices (1.0, 1.5, and 3.0 mm). Each image set was used to produce sagittal reformatted image in same location. Observe degradation of detail in reformats as slice thickness is increased. CT scan obtained with 1.5-mm slice width and 1.0-mm reconstruction interval shows markedly reduced detail when compared with detail shown in A and B.
Fig. 1D. —Images of 53-year-old woman with advanced degenerative changes in tibiotalar joint were reconstructed from one multislice CT acquisition (dual slice, 0.5-mm slice width, 120kVp, 200 mAs). After scanning, raw data were combined to increase effective slice width. Separate image sets were reconstructed with progressively thicker slices (1.0, 1.5, and 3.0 mm). Each image set was used to produce sagittal reformatted image in same location. Observe degradation of detail in reformats as slice thickness is increased. In CT scan obtained with 3.0-mm slice width and 1.5-mm reconstruction interval, stairstep artifacts and blurring are evident.
Fig. 2. —Drawing illustrates geometry of adaptive detector array (M×8000; Marconi Medical Systems, Cleveland, OH) that provides 0.5-mm slices. X-ray beam collimation (not shown) produces 1-mm-thick beam at gantry center. Additional collimators (gray rectangles) cover peripheral detector rows. Numbers indicate detector width in millimeters. Two central 1.0-mm detector rows are each 50% covered, resulting in two channels of 0.5-mm slice-width data. The smaller the prepatient collimator width is, the greater the penumbra radiation relative to the primary beam. Simultaneous acquisition of two or more slices reduces effects of penumbra radiation because penumbra is smaller percentage of width of primary beam. (Primary beam Xrays, solid vertical arrows; penumbra radiation, dashed vertical arrows; gantry rotation direction, dashed curved arrows; scan direction, solid horizontal arrows).
If the acquired voxels have sides of equal dimensions, they are said to be isotropic (Fig. 3A,3B). Using a slice thickness of 0.5 mm and a small field of view (e.g., 10 cm), the slice width nearly coincides with the in-plane pixel size, and isotropic imaging becomes possible [1, 2]. The thin slice width and subsequent isotropic viewing allow the study to be completed in a single acquisition, potentially decreasing the patient's radiation dose. In addition, the scanning and set-up times are reduced for joints, such as the wrist and ankle, that are usually studied in multiple planes. The value of isotropic scanning and isotropic imaging is significant when dealing with trauma. The injured body part can be placed in the gantry in a comfortable position without compromising the study (Figs. 4A,4B and 5A,5B,5C).
Fig. 3A. —Illustrations depict volume data sets. Isotropic (A) and anisotropic (B) volume data sets. If acquired voxels have sides with equal dimensions, voxels are isotropic (i.e., X = Y = Z). If not, voxels are anisotropic (i.e., X = Y ≠ Z). For small joints, isotropic conditions can be achieved using slice thickness of 0.5 mm and small field of view (e.g., 10 cm). Isotropic data allows reformatted images in any plane with spatial resolution identical to original scanning plane.
Fig. 3B. —Illustrations depict volume data sets. Isotropic (A) and anisotropic (B) volume data sets. If acquired voxels have sides with equal dimensions, voxels are isotropic (i.e., X = Y = Z). If not, voxels are anisotropic (i.e., X = Y ≠ Z). For small joints, isotropic conditions can be achieved using slice thickness of 0.5 mm and small field of view (e.g., 10 cm). Isotropic data allows reformatted images in any plane with spatial resolution identical to original scanning plane.
Fig. 4A. —44-year-old woman who fractured left wrist in motor vehicle collision. Coronal (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.3-mm reconstruction interval, 120 kVp, 125 mAs) obtained with forearm in cast show scaphoid (S) impacted into articular surface of comminuted distal radius fracture. Extended coverage into distal radius allows precise quantification of fracture deformity.
Fig. 4B. —44-year-old woman who fractured left wrist in motor vehicle collision. Coronal (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.3-mm reconstruction interval, 120 kVp, 125 mAs) obtained with forearm in cast show scaphoid (S) impacted into articular surface of comminuted distal radius fracture. Extended coverage into distal radius allows precise quantification of fracture deformity.
Fig. 5A. —31-year-old woman with compression fracture of lateral tibial plateau. Coronal (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 174 mAs) show degree of impaction. Fracture compression zone (arrows) in subchondral bone is well depicted.
Fig. 5B. —31-year-old woman with compression fracture of lateral tibial plateau. Coronal (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 174 mAs) show degree of impaction. Fracture compression zone (arrows) in subchondral bone is well depicted.
Fig. 5C. —31-year-old woman with compression fracture of lateral tibial plateau. Three-dimensional surface rendering image presents overview of fracture.
Volume scanning using thin collimation and a small field of view is valuable in many situations other than acute trauma (Figs. 6A,6B,7A,7B,7C,8A,8B,9A,9B). The goal of the volumetric method is to provide high-quality anatomic coronal, sagittal, and axial images of the joint in every study. Although the joint may be scanned in a nonanatomic position, subsequent MPR images can be produced in anatomic planes without sacrificing image quality. We have great confidence in this technique and do not review the original source images.
Fig. 6A. —75-year-old woman with osteoarthritis of knee and large lytic lesion of proximal tibia. Conventional radiograph reveals large lytic process in proximal tibia. CT examination was prompted by possibility of giant geode.
Fig. 6B. —75-year-old woman with osteoarthritis of knee and large lytic lesion of proximal tibia. Coronal reformatted image from multislice CT (dual slice, 1.0-mm slice width, 0.6-mm reconstruction interval, 120 kVp, 200 mAs) shows gas bubbles (open arrow) in lesion as well as communication to joint space (long solid arrow). Gas suspended in viscous fluid of geode likely arose as result of joint vacuum phenomenon. Note small geodes in distal femur (short solid arrow) and other typical degenerative changes. Surgery for bone grafting of cavity confirmed diagnosis.
Fig. 7A. —31-year-old man with known osteochondritis dissecans of medial talar dome. Sagittal (A) and coronal (B) reformatted images from multislice CT (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 195 mAs). Images show deformation of talus as well as small fragments of osteochondral lesion (solid arrows). Incidental exostosis (open arrow, A) is noted.
Fig. 7B. —31-year-old man with known osteochondritis dissecans of medial talar dome. Sagittal (A) and coronal (B) reformatted images from multislice CT (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 195 mAs). Images show deformation of talus as well as small fragments of osteochondral lesion (solid arrows). Incidental exostosis (open arrow, A) is noted.
Fig. 7C. —31-year-old man with known osteochondritis dissecans of medial talar dome. Three-dimensional surface-display image offers good perspective of talar dome and depicts exact location of fragments.
Fig. 8A. —19-year-old man with overgrowth and diffuse periosteal reaction of the second metatarsal bone. Axial (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 150 mAs) show thickened cortex (arrowheads, A) of second metatarsal bone as well as tiny intracortical expansile lesion (arrow, A) consistent with osteoid osteoma, which was later proven at surgical excision. Sagittal image shows exact location of lesion (arrow, B) in relation to both distal and proximal joints.
Fig. 8B. —19-year-old man with overgrowth and diffuse periosteal reaction of the second metatarsal bone. Axial (A) and sagittal (B) reformatted multislice CT images (dual slice, 0.5-mm slice width, 0.2-mm reconstruction interval, 120 kVp, 150 mAs) show thickened cortex (arrowheads, A) of second metatarsal bone as well as tiny intracortical expansile lesion (arrow, A) consistent with osteoid osteoma, which was later proven at surgical excision. Sagittal image shows exact location of lesion (arrow, B) in relation to both distal and proximal joints.
Fig. 9A. —10-year-old boy with decreased range of motion in right elbow after trauma. Coronal (A) and sagittal (B) reformatted images from double-contrast multislice CT arthrogram (dual slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 199 mAs) clearly define relationships of joint capsule, joint cartilage, and bone. Posttraumatic deformity of proximal radius and posterior subluxation of radial head are revealed. Images also show 5-mm defect (arrows) in articular cartilage of radial head extending to bony surface.
Fig. 9B. —10-year-old boy with decreased range of motion in right elbow after trauma. Coronal (A) and sagittal (B) reformatted images from double-contrast multislice CT arthrogram (dual slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 199 mAs) clearly define relationships of joint capsule, joint cartilage, and bone. Posttraumatic deformity of proximal radius and posterior subluxation of radial head are revealed. Images also show 5-mm defect (arrows) in articular cartilage of radial head extending to bony surface.

Principle of Obliquity

When isotropic scanning is not possible because of scanner limitations or because of thicker slice-width selection (>0.5 mm), image quality in subsequent MPR images can be optimized using the “principle of obliquity.” To maximize joint surface visualization on reformatted images, the affected joint should be placed according to the following rules: First, the extremity should be positioned so that the scanning plane is oblique to the joint surfaces of interest. A 45° obliquity is optimal. The joint then should be placed in the center of the scanner gantry to maximize in-plane resolution.
This principle of obliquity (Fig. 10A,10B) causes a larger number of slices to transverse the joint surfaces, thereby better defining the bony anatomy in subsequent multiplanar reformations (Fig. 11A,11B,11C).
Fig. 10A. —Drawings show “principle of obliquity.” Data set from joint placed oblique to scan plane (A) yields better reformatted images of joint surfaces than data set obtained with joint placed perpendicular to scanning plane (B). Quality of reformatted images improves as number of slices traversing joint surfaces is increased.
Fig. 10B. —Drawings show “principle of obliquity.” Data set from joint placed oblique to scan plane (A) yields better reformatted images of joint surfaces than data set obtained with joint placed perpendicular to scanning plane (B). Quality of reformatted images improves as number of slices traversing joint surfaces is increased.
Fig. 11A. —Radiographic ankle phantom containing dry bones embedded in plastic was scanned in three different planes to illustrate principle of obliquity. Identical multislice CT parameters were used (0.5-mm slice width, 120 kVp, 150 mAs, and 100-mm field of view); scanning acquisition plane is indicated on each image (white line). To produce anisotropic data set, second reconstruction was performed from three acquisitions resulting in data sets with images of 1-mm thickness reconstructed at 0.5-mm intervals. Each of three image sets was used to produce a sagittal reformat in approximately same location. The talonavicular (short arrows), tibiotalar (long arrows), and talocalcaneal (curved arrows) joints are displayed for comparison. Axial scanning plane (white line) best depicts surfaces of talonavicular joint (short straight arrow). Joint space between talus and navicular bone is well visualized because scan plane is perpendicular to joint space. Detail of tibiotalar (long straight arrow) and talocalcaneal (curved arrow) joints is of lower quality.
Fig. 11B. —Radiographic ankle phantom containing dry bones embedded in plastic was scanned in three different planes to illustrate principle of obliquity. Identical multislice CT parameters were used (0.5-mm slice width, 120 kVp, 150 mAs, and 100-mm field of view); scanning acquisition plane is indicated on each image (white line). To produce anisotropic data set, second reconstruction was performed from three acquisitions resulting in data sets with images of 1-mm thickness reconstructed at 0.5-mm intervals. Each of three image sets was used to produce a sagittal reformat in approximately same location. The talonavicular (short arrows), tibiotalar (long arrows), and talocalcaneal (curved arrows) joints are displayed for comparison. Coronal scanning plane (white line) best depicts surfaces of tibiotalar (long arrow) and talocalcaneal (curved arrow) joints. Talonavicular (short arrow) joint space is essentially obliterated because scan plane is nearly parallel with joint.
Fig. 11C. —Radiographic ankle phantom containing dry bones embedded in plastic was scanned in three different planes to illustrate principle of obliquity. Identical multislice CT parameters were used (0.5-mm slice width, 120 kVp, 150 mAs, and 100-mm field of view); scanning acquisition plane is indicated on each image (white line). To produce anisotropic data set, second reconstruction was performed from three acquisitions resulting in data sets with images of 1-mm thickness reconstructed at 0.5-mm intervals. Each of three image sets was used to produce a sagittal reformat in approximately same location. The talonavicular (short arrows), tibiotalar (long arrows), and talocalcaneal (curved arrows) joints are displayed for comparison. Oblique scanning plane (white line) depicts all three joints well. For complex joints, oblique scanning plane yields best results. Tibiotalar (long straight arrow), talonavicular (short straight arrow), and talocalcancal (curved arrow) joints are well visualized.

Extended Anatomic Coverage

Anatomic coverage is the display of structures surrounding the region in which the diagnosis is established. Adequate landmarks must be provided if treatment decisions are to be made from CT scans (Figs. 4A,4B and 12A,12B).
Fig. 12A. —11-year-old girl with multiple hereditary exostoses and difficulty with pronation and supination of wrist. Coronal reformatted (A) and three-dimensional shaded-surface—display (B) multislice CT images (dual slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 125 mAs). Multislice Ct technique allows long anatomic coverage with thin slices and detailed visualization.
Fig. 12B. —11-year-old girl with multiple hereditary exostoses and difficulty with pronation and supination of wrist. Coronal reformatted (A) and three-dimensional shaded-surface—display (B) multislice CT images (dual slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 125 mAs). Multislice CT technique allows long anatomic coverage with thin slices and detailed visualization.
For single-slice CT, anatomic coverage can be calculated by multiplying slice width by pitch by exposure time. For example, given a slice width of 1 mm, a pitch of 1, and a 100-sec helical acquisition, the coverage would be 100 mm. Most musculoskeletal applications require greater coverage. Lengthening the coverage by increasing the slice width or the pitch in single-slice CT will degrade the images and hamper the production of high-quality multiplanar reformations.
For multislice CT, calculation of anatomic coverage is more complex and requires knowledge of the number of active detector channels as well as the nominal width of the channels.
Anatomic coverage equals [number of data channels × width of each channel × pitch] × exposure time / gantry rotation time. Given four data channels of 1-mm width each, a pitch of 1, a 100-sec scan, and a 0.5-sec gantry rotation time, the anatomic coverage for multislice CT would be 800 mm [3,4,5].

Large Patients and Metal Hardware

Because of “photon starvation” at the detectors and beam hardening, obese patients and patients with metal hardware cause artifacts in CT. Using higher kilovoltage settings (140 vs. 120 kVp) and higher milliampere-seconds can reduce these artifacts. For single-slice CT, the tube current limits the milliampere-seconds. With multislice CT, the pitch can be reduced so that it is less than 1. This technique results in “overlapping spirals,” causing the effective milliampere-seconds to be increased in the reconstructed images. Use of low pitch settings results in effective tube currents as high as 1500 milliampere-seconds. The trade-offs for using lower pitch are reduced coverage and increased radiation. However, with multislice CT it is possible to achieve a balance between pitch and coverage, enabling radiologists to perform diagnostic examinations (Figs. 13A,13B,14,15A,15B).
Fig. 13A. —27-year-old woman with history of bilateral temporomandibular joint arthroplasties and implant placement. Increasing difficulties in jaw movements prompted CT examination. Curved reformatted (A) and three-dimensional volume-rendering (B) multislice CT images (quad slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 400 mAs). Examination revealed large amount of focal hypertrophic bone around coronoid processes of mandible. Thin-slice scanning technique combined with high milliampere-seconds provides high-quality multiplanar reformatted images showing extensive bony hypertrophy (arrows, A) and metallic implants. Mechanical symptoms are explained by impingement resulting from bony enlargement (arrow, B). Three-dimensional volume rendering gives overview of metallic hardware and can assist in preoperative planning.
Fig. 13B. —27-year-old woman with history of bilateral temporomandibular joint arthroplasties and implant placement. Increasing difficulties in jaw movements prompted CT examination. Curved reformatted (A) and three-dimensional volume-rendering (B) multislice CT images (quad slice, 1-mm slice width, 0.5-mm reconstruction interval, 120 kVp, 400 mAs). Examination revealed large amount of focal hypertrophic bone around coronoid processes of mandible. Thin-slice scanning technique combined with high milliampere-seconds provides high-quality multiplanar reformatted images showing extensive bony hypertrophy (arrows, A) and metallic implants. Mechanical symptoms are explained by impingement resulting from bony enlargement (arrow, B). Three-dimensional volume rendering gives overview of metallic hardware and can assist in preoperative planning.
Fig. 14. —61-year-old woman with bilateral hip prostheses who was referred before surgery for evaluation of remaining “bone stock” surrounding cetabular hardware. Multislice CT was performed with thin slices (2.5-mm slice width, 1.4-mm reconstruction interval), high kilovoltage (140 kVp), and high tube current (750 mAs). Coronal reformatted image shows sparse amount of bone surrounding acetabular components (solid arrows). Wear of plastic joint liner of right hip (open arrow) is shown. Optimal scanning technique minimized metal artifacts.
Fig. 15A. —52-year-old man with severely comminuted intraarticular distal tibial fracture in temporary external fixation. Sagittal reformatted (A) and lateral volume-rendered (B) multislice CT images (1-mm slice width, 0.5-mm reconstruction interval, 140 kVp, 300 mAs) show vertical fracture component in A without significant metal artifacts. Lateral volume-rendered image reveals fracture and external metal stabilizer (arrows, B). CT visualization of bone with this much metal is virtually impossible without highscan techniques possible with multislice CT scanner. Contours of pin through calcaneus and talus also can be seen (arrowheads, B).
Fig. 15B. —52-year-old man with severely comminuted intraarticular distal tibial fracture in temporary external fixation. Sagittal reformatted (A) and lateral volume-rendered (B) multislice CT images (1-mm slice width, 0.5-mm reconstruction interval, 140 kVp, 300 mAs) show vertical fracture component in A without significant metal artifacts. Lateral volume-rendered image reveals fracture and external metal stabilizer (arrows, B). CT visualization of bone with this much metal is virtually impossible without highscan techniques possible with multislice CT scanner. Contours of pin through calcaneus and talus also can be seen (arrowheads, B).

Summary

The advantages of multislice CT are significant, permitting long anatomic coverage combined with thin slice widths at low pitch settings. The thin-slice technique makes isotropic viewing possible. Multi-slice CT also facilitates scanning of obese patients as well as patients with metal hardware. Common imaging parameters for various joints are listed (Table 1). Workstation postprocessing becomes an integral part of the examination.
TABLE 1 Imaging Parameters in Musculoskeletal Multislice CT Protocols
ProtocolSlice Width (mm)IntervalaPitchkVpmAsPosition
Shoulder2.0-2.51.0-1.30.8120-140250+Contralateral arm over head
Elbow1.00.50.7120200Affected arm over head, elbow bent
Wrist and hand0.50.20.8120190Affected arm over head, palm up
Hip2.51.30.6120-140300+ 
Knee0.5-1.00.2-0.50.7-0.8120200+ 
Ankle and foot0.50.20.712019045° Oblique, with coronal head holder
Small joints, limited hardware0.50.20.7140200 
Large joints, large hardware
2.5
1.3
≤0.7
140
Maxb
a
Reconstruction interval.
b
Maximum.

Acknowledgments

We thank Scott A. Persohn for preparing all radiographic illustrations used in this article.

Footnotes

Presented at the annual meeting of the American Roentgen Ray Society, Washington, DC, May 2000.
Address correspondence to K. A. Buckwalter.

References

1.
Kalender WA. Thin-section three-dimensional spiral CT: is isotropic imaging possible? Radiology 1995; 197:578-580
Crossref
PubMed
Google Scholar
2.
Caldemeyer KS, Sandrasegaran K, Nazenin Shinaver C, Mathews VP, Smith RR, Kopecky KK. Temporal bone: comparison of isotropic helical CT and conventional direct axial and coronal CT. AJR 1999; 172:1675-1682
Crossref
PubMed
Google Scholar
3.
Rigauts H. A one year experience with the multislice helical CT. J Belge Radiol 1999; 82:273-276
Google Scholar
4.
Hu H, He HD, Foley WD, Fox SH. Four multidetector-row helical CT: image quality and volume coverage speed. Radiology 2000; 215:55-62
Crossref
PubMed
Google Scholar
5.
McCollough CH, Zink FE. Performance evaluation of a multi-slice CT system. Med Phys 1999; 26:2223-2230
Crossref
PubMed
Google Scholar
View full text|Download PDF