Spiral CT allows rapid acquisition of volumetric data of any body part of interest. In contrast to conventional CT, spiral CT acquires images at an angled plane of section during simultaneous patient translation and x-ray exposure. ¹ This continuous acquisition of images is enabled because of the slip-ring technology of the gantry construction which replaced the former electrical cabling, allowing for continuous rotation of the source-detector assembly. ² Transaxial images are then reconstructed based on interpolation from the volumetric data set. This allows for reconstruction of images at arbitrary intervals, providing overlapping images of the area of interest and the potential for high quality multiplanar and 3D reconstructions.

The advantages of spiral CT include the elimination of respiratory misregistration and diminished patient motion artifact due to the continuous scanning ability and rapid acquisition of images during a single breath-hold. Because imaging of bones and joints does not routinely require breath-holding, this relative advantage over conventional CT is less for musculoskeletal applications than for applications in the chest and abdomen. For musculoskeletal imaging the major advantage is in reconstruction of overlapping image sections for 2D and 3D reformation. This attribute of spiral CT maximizes longitudinal resolution without increasing radiation exposure. ³ Multiplanar reconstructions can be useful in the evaluation of traumatic injuries in anatomically complex areas, characterization of complex fractures and dislocations, and assessment of intraarticular bodies. ^4,5 They can be applied toward preoperative planning as well as to postoperative or postreduction analysis and have relevance in both the adult and pediatric populations.

The early disadvantages of spiral CT included limited x-ray tube heat capacity, resulting in relatively low mA settings for continuous scanning, as well as limited computing power. However, with recent advances in both hardware and software these limits have vanished. Current generation scanners produce high tube current for exposure durations of up to 60 seconds, and gantry rotation periods between 0.5 and 0.8 seconds per image are becoming routine. Many manufacturers are also now providing machines with multiple detectors, providing for faster throughput, increased longitudinal coverage, or increased longitudinal spatial resolution (thinner collimation) during a given acquisition. High quality spiral CT images are thus routinely feasible.

Paramount to the success of spiral imaging is the correct application of scanning parameters, such as collimation, pitch, and reconstruction interval. The collimation, or slice thickness, varies depending upon the body part being examined. Compared with body applications, the longitudinal coverage required for musculoskeletal imaging usually is limited. In general, we favor the use of relatively narrow collimator settings (3 to 5 mm in the larger body parts such as the hips and pelvis, and 1 to 2 mm in the wrists and hands) and extend the scan duration or pitch as necessary to provide for longitu-dinal coverage.

Pitch is defined as the speed of table translation divided by the collimation. For musculoskeletal applications pitch rarely needs to be increased above 1.5. As pitch increases, there is the potential to decrease longitudinal resolution due to slice profile broadening, ² and to increase image artifacts.

Reconstruction intervals, or slice indices, vary depending upon the body part being imaged and the goal of the examination. Finer reconstruction intervals result in superior multiplanar reformation and 3D rendering. For 2D reformations, reconstruction should occur at approximately 50% overlap to maximize longitudinal resolution. We typically utilize the high frequency or bone algorithm for reconstruction in this setting. For 3D reformations, reconstruction should occur at 50 to 75% overlap and should utilize the standard algorithm (for example, a 3 mm collimation reconstructed every 1 mm). ⁶ While prolonging the scan reconstruction time and data storage requirements, all generated images do not need to be filmed. We film every other to every third image and review all images on a soft-copy workstation. The quality of 2D and 3D reformations improves with reconstruction of overlapping sections. ^6,7

We now discuss acquisition and reconstruction parameters for specific body regions. The parameters currently in use are shown in table 1. For all applications, attempts should be made to carefully position the patient and to limit the field of view to the anatomy of interest. Intravenous contrast is typically not utilized in the evaluation of trauma or articular disorders but can be useful in the evaluation of muscle or soft tissue for infection or neoplasm. ⁵

Imaging of the upper extremity can be challenging due to the anatomic complexity of the joints involved as well as patient positioning. In the shoulder, spiral CT can be applied to evaluate pathology of the glenohumeral or acromioclavicular joints or the scapula. Because of the considerable x-ray attenuation due to the normal shoulders and upper thorax, it is important to use relatively high tube current settings to optimize image quality. We use 250 to 300 mAs and 120 kVp. Axial images are acquired with a 3 mm collimation and a pitch of 1 to 1.5. Reformations are made in the oblique coronal and oblique sagittal planes with respect to the glenohumeral joint (figure 1). The reconstructions allow ready identification and characterization of glenoid fractures (figure 2), and humeral head fractures and dislocations (figure 3). These reconstructions can be critical in the evaluation of injuries in which a patient's positioning and limited range of motion make interpretation of axial images alone difficult (figure 4), and often they can aid the orthopedic surgeon in spatial conceptualization. CT arthrography of the shoulder has been widely used in the evaluation of Hill-Sachs deformities, labral tears, and capsular injuries; however, in general, CT has been surpassed by MR arthrography. ^5,8

Spiral CT of the elbow can provide pertinent information not readily available by radiography alone, and may alter clinical decisions. ⁹ For evaluation of chronic trauma and impingement syndromes, we image the elbow in both flexed and extended positions (figure 5) with 1 to 3 mm collimation, a pitch of 1 to 1.5, and 1 to 1.5 mm reconstruction intervals. Sagittal and coronal recon-structions help evaluate or visualize fractures and loose bodies identified on radiographs, and can diagnose radio graphically "occult" fractures in patients with elbow effusions. Our orthopedic surgeons find that 2D and 3D reconstructions are helpful in planning surgery and in communicating with patients. In order to evaluate which fragments move with which parent bone, reformations are typically acquired in more than one elbow position (figures 5 and 6).

The wrist and hand are optimally evaluated in axial, coronal, and sagittal planes with a narrow collimation of 1 mm, a pitch of 1, and reconstruction intervals of 1 mm. ¹⁰ With such thin primary sections, creation of overlapping reconstructions usually is not necessary. Two planes generally are imaged directly, with one being perpendicular to the area of interest (figure 7). Images in the third anatomic plane usually are reconstructed from the axial images. Distal radial and ulnar fractures are well characterized on these reconstructed images. For the evaluation of scaphoid fractures, images are to be obtained in an oblique sagittal plane along the long axis of the scaphoid (figure 8), or with direct coronal imaging. ¹⁰ Because many scaphoid fractures run in the plane of section for transaxial scanning to the wrist, the findings can be subtle and reformations in the sagittal or coronal plane can be misleading if negative. In addition to the evaluation of fractures, CT also can aid in the assessment of degenerative arthritis, such as that arising from chronic carpal instability (figure 9).

The pelvis and hips are readily amenable to characterization by spiral CT. Typically, 3 mm collimation images with a pitch of 1 to 1.5 are reconstructed at 1.5 to 2 mm intervals. If the study is abnormal, coronal and sagittal reformations are routinely generated. In trauma, spiral CT with 2D and 3D reformats can reproduce plain radiographic views, thereby limiting radiation exposure while simultaneously providing optimal visualization. Some investigators have used these reformatted images as replacements for the standard Judet and Letournel views. ¹¹ CT is useful in the evaluation of pelvic and acetabular fractures (figure 10), and in assessment of intraarticular bodies (figure 11). CT also is excellent for depiction of osseous sacral and sacroiliac joint disorders (figure 12) and in the evaluation of some developmental pediatric disorders, such as developmental dysplasia of the hip (figure 13). For pediatric patients CT can assess osteocartilaginous causes of frequent dislocations, such as iliopsoas interposition and femoral anteversion, and can be useful for imaging of patients in casts. ¹²

Although MRI is the standard modality for imaging of the knee, CT is useful in the evaluation of fractures for surgical planning. Axial images are acquired with a 2 to 3 mm collimation, a pitch of 1 to 1.5, and are reconstructed at 1 to 1.5 mm intervals. Sagittal and coronal reformations can help characterize tibial plateau fractures and the extent of depression (figure 14). Additionally, CT is valuable in the evaluation of the postoperative or postreduction knee, which may be casted, subsequently limiting plain film evaluation. Other investigators have used spiral CT in the preoperative planning of a patient with multiple exostoses ¹³ and in the detection of insufficiency fractures in patients with knee allografts. ¹⁴

Much like the wrist and hand, the ankle and foot usually can be oriented in more than one plane relative to the scanner gantry, thereby allowing direct axial and coronal sections. In most cases of fracture characterization, 2 to 3 mm collimation is adequate, although a 1 mm collimation may be useful for fracture detection. This is particularly true of subtle fractures of the smaller foot bones, such as the navicular or cuneiforms. The multiplanar reformation capabilities of spiral CT are especially useful in the evaluation of complex fractures in which there are multiple fracture planes. Complex tibial plafond fractures are readily characterized by the extent of the fracture line, degree of displacement, and disruption of the ankle mortise (figure 15). Pretorius et al were able to classify distal tibial fractures into those that required acute reduction and those that needed delayed definitive arthroplasty based on interpretation of spiral CT images. ¹¹ Similarly,
multiplanar calcaneal fractures are delineated with coronal and sagittal reconstructions (figure 16). ^15,16 Corbett et al have devised a classification system of calcaneal fractures to help guide in the decisions behind conservative versus surgical management. ¹⁷ Smaller body parts such as the digits are best imaged in the axial or coronal planes with a 1 mm collimation, a pitch of 1, and l mm reconstructions (figure 17).

Critical to the interpretation of spiral CT images is an understanding of accompanying artifacts. Stair step artifact is a common artifact visualized as a disruption of inclined surfaces in a regular, stair step fashion. ² It is most often encountered when oblique surfaces are nearly parallel to the transverse plane rather than aligned with the direction of table motion (figure 18).

Three-dimensional image rendering drawbacks include those associated with volume formation, tissue classification, and image projection. ¹⁸ In volume formation, artifacts can occur in preprocessing, acquisition, or data editing. Figure 19 demonstrates a 3D rendered image in which noise from the 1 mm acquisition is manifest as linear disruptions in the bone contour. Moreover, this image demonstrates a common pitfall, in which a markedly displaced glenoid fracture fragment was inadvertently edited completely out of the image.

The future of spiral CT is expanding rapidly and includes advances such as quantitative prosthetic modeling and virtual arthroscopy. Three-dimensional reconstructions of spiral CT images are being used in bone modeling of pathology for presurgical and preradiation therapy, as well as in surgical simulations such as virtual osteotomies. ^5,18 These 3D models are used to guide the manufacturing of prostheses, both off-the-shelf and custom made, and the volumetric data acquired with spiral CT also may be applied in the assessment of post prosthetic fit by evaluating tissue displacement. ¹⁹ Perspective volume rendering is another new application of spiral CT which allows visualization of volumetric data sets from an internal viewpoint, thereby simulating endoscopy. These techniques have been applied to the vascular system, tracheobronchial tree, and gastrointestinal tract. ²⁰ In the musculoskeletal system such techniques have the potential to aid in surgical planning and in communications with the referring physicians. ²¹

High quality spiral CT of the musculoskeletal system requires tailoring the examination to the body part being imaged and to the clinical question being addressed. The collimation, pitch, tube current, reconstruction interval, and reconstruction algorithm have important influence on image quality. If chosen appropriately, excellent quality and clinical value can routinely be achieved. AR