MDCT Versus Digital Radiography in the Evaluation of Bone Healing in Orthopedic Patients

Transcription

1 MDCT and Digital Radiography of one Healing Musculoskeletal Imaging Original Research C M E D E N T U R I C L I M G I N G JR 2006; 186: X/06/ merican Roentgen Ray Society Y O Christian R. Krestan 1 Helge Noske 2 Violeta Vasilevska 1 Michael Weber 1 Gerd Schueller 1 Herwig Imhof 1 Christian Czerny 1 Krestan CR, Noske H, Vasilevska V, et al. F Keywords: bone healing, digital radiography, MDCT, musculoskeletal imaging, orthopedic surgery DOI: /JR Received March 18, 2005; accepted after revision ugust 31, We thank Professor Mathias Prokop, University of Utrecht, The Netherlands, for assistance with manuscript revision. 1 Department of Radiology, Medical University of Vienna, Waehringer Guertel 18-20, Vienna 1090, ustria. ddress correspondence to C. R. Krestan (christian.krestan@meduniwien.ac.at). 2 Department of Orthopedic Surgery, Medical University of Vienna, Vienna, ustria. MDCT Versus Digital Radiography in the Evaluation of one Healing in Orthopedic Patients OJECTIVE. ssessment of bone healing in orthopedic patients is usually monitored by radiographs in two views. The purpose of our study was to compare multiplanar reconstructions from MDCT data sets with digital radiographs for assessing the extent of bone healing. MTERILS ND METHODS. Forty-three orthopedic patients (19 women, 24 men) who underwent MDCT and radiography after arthrodesis, fractures, or spinal fusions were included in our study. MDCT was performed on an MX 8000IDT scanner and served as the gold standard. The technical parameters were adapted to the anatomic region. bone algorithm for reconstruction was used (3,500/600 H). Multiplanar reconstructions were calculated in two orthogonal planes. ll patients underwent digital radiography on a Multix FD system in two views according to standard procedures. Multiplanar reconstructions and radiographs were analyzed by two musculoskeletal radiologists in a consensus interpretation to determine bone healing using a semiquantitative approach. RESULTS. In 27 patients (63%), MDCT and digital radiography were concordant with regard to the extent of bone healing, whereas in 16 patients (37%) the results were not concordant. In eight patients (19%) digital radiographs underestimated the extent of bone healing, whereas in another eight patients (19%) they overestimated the degree of fusion. CONCLUSION. MDCT using high-quality 2D reformatting is recommended as the primary imaging technique for the evaluation of bone healing. his study evaluates the value of T MDCT compared with digital conventional radiography (using a flatpanel detector) in the diagnosis of bone healing or nonunion in orthopedic patients [1]. To our knowledge, no clinical study has been published using MDCT to monitor bone production. In a preclinical spine fusion model that compared radiography [2], MDCT, and histologic verification, MDCT had a substantially better positive predictive value for the assessment of bone fusion and thus served as the gold standard in this study. one healing is an important biologic process in orthopedic patients. It may take place after fractures, arthrodesis, spondylodiskitis, osteotomies, and distraction of shortened long bones. iologic factors play an important role in the process of bone healing [3]. Osteoprogenitor cells must be transported to the site of fracture or osteotomy. Local factors stimulate bone production, which begins with periosteal and endosteal callus formation and eventually leads to calcification and complete fusion of the bone parts. If a fracture or osteotomy does not heal after 6 8 months, it is considered a nonunion [4]. Several factors predispose a patient to nonunion of bones, including mechanical instability, loss of blood supply, and infections. Proper diagnosis of nonunions is essential for appropriate patient management because inadequate bone healing is a contraindication for patient mobilization. Nonunions can be treated by internal or external fixation or by bone grafting and with electrical or ultrasound stimulation or extracorporeal shock wave treatment. The correct radiologic diagnosis is an important factor in dealing with postoperative orthopedic patients. Generally, radiographs in two views have been used to monitor bone healing in clinical patients. one production has been estimated to occur within 15 weeks after osteotomy; complete bone healing may take 3 6 months or even longer [5]. The reliability of conventional radiographs for the determination of fracture healing has been questioned in previous studies [6]. CT has been used for the monitoring of bone production and fracture healing, and its 1754 JR:186, June 2006

2 MDCT and Digital Radiography of one Healing advantages over conventional radiography in early fracture healing have been reported [7]. The quality of coronal or sagittal multiplanar reformations, which are essential for the assessment of fracture lines or bone bridges perpendicular to the scanning plane, depends on the scanner type and the scanning parameters. To avoid stairstep artifacts in CT, isotropic or near-isotropic resolution is necessary and has become attractive with the introduction of MDCT scanners [8, 9]. Experimental studies have shown that MDCT reduces stairstep artifacts with multiplanar reconstruction when compared with single-detector CT [10]. Materials and Methods Patients This was a retrospective study that identified 46 patients in the radiology information system database who had undergone MDCT and conventional radiography for the evaluation of bone healing. ll patients were referred from the department of orthopedic surgery. Patient histories included fractures, arthrodesis, spondylodesis, spondylodiskitis, osteotomies, and distraction of shortened long bones. natomic regions were as follows: 16 spine (lumbar, thoracic, cervical), 11 long bones (femur, humerus, tibia), nine feet, four hands, one patella, one clavicula, and one sacroiliac joint. The time between MDCT and conventional radiography was 0 13 weeks. fter three patients with time intervals of 7, 9, and 13 weeks were excluded, 43 patients (24 men and 19 women; mean age, 51.9 ± 17.9 years) with time intervals of weeks (mean, 1.3 weeks) remained in the study group. Scanning Techniques MDCT was performed on a 16-MDCT scanner (MX 8000IDT, Philips Medical Systems) with clinically used protocols adapted for the specific anatomic region. Table 1 shows details of the scanning protocols. Data were acquired in the helical mode (16 slices per tube rotation) with thin collimation. From these data, we reconstructed thin axial slices with 50% overlap to yield near-isotropic voxels (almost identical to the length of the voxel in the x, y, and z axes) for further processing. This allows 2D and 3D reconstructions with a resolution similar to the source images that form the basis of good-quality multiplanar reconstructions (MPRs) [11]. MPRs were reconstructed from contiguous axial slices ranging from 1.5 to 3 mm thick, depending on the anatomic region. Digital radiography was performed on a Multix FD system (Siemens Medical Solutions) in two views (e.g., lumbar spine and hip, 70 kv and 40 ms; ankle joint, 55 kv and 6 ms) according to TLE 1: Scanning Parameters for MDCT at Different natomic Regions Parameter Lumbar and Thoracic Spine Cervical Spine Humerus, Hip, Femur, and Tibia nkle, Subtalar, and Hand Joints Collimation (mm) Slice / increment (mm) 2 / 1 2 / 1 2 / / 0.4 kv ms Pitch Rotation time (sec) MPR (mm) 3 / 3 2 / 2 3 / 3 2 /2 (1.5 / 1.5) Window setting (width / length) (H) 3,500 / 600 3,500 / 600 3,500 / 600 3,500 / 600 Matrix Note MPR = multiplanar reconstruction. standard clinical procedures. This system features an X-ray tube with focal spot sizes of 0.6 and 1.0 mm and a flat-panel detector (Trixell Pixium 4600, Siemens Medical Solutions) with a cesium iodide amorphous silicon layer (matrix size, 3 3 K pixel elements; pixel size, 143 µm; and an active area of cm). Evaluation of one Healing Digital radiographs in both views and MPRs, usually reconstructed in two planes perpendicular to the scanning plane, were evaluated on an GF (gfa-gevaert) PCS workstation (IMPX) using a dual-head digital light box system combining two 20.8-inch (52.8 cm) portrait LCD gray-scale flat panel displays (arco) with a resolution of 3,072 2,048 pixels for film interpretation. First, digital radiographs were analyzed without knowledge of the MPRs. Subsequently, the MPRs from MDCT were reviewed independently of the digital radiographs. The MDCT diagnosis was made exclusively from sagittal and coronal MPRs, which were, in all cases, orthogonal to the fracture or arthrodesis plane. Fusion of osseous structures was scored with a semiquantitative approach for both techniques (MDCT, digital radiography) as complete (c), partial (p), and no bone bridging (0). Definitions of fusion were as follows: complete, bone bridges with no gap; partial, some bone bridges with gaps between; and no bridging, no osseous bridges. Two musculoskeletal radiologists assessed all MDCT examinations and digital radiographs in a consensus interpretation. In addition, the diagnostic confidence of digital radiography for evaluating bone bridging using a semiquantitative 3-grade scale (reliable, fair, unreliable) was scored by the same reviewers. Then the percentage of patients in whom MDCT altered the initial digital radiography diagnosis was calculated for these three groups. Statistical nalysis Descriptive statistical analysis was calculated from the data, and the absolute numbers and percentages of concordant and discordant results are given for the three stages of bone healing (complete, partial, none). To assess the agreement between digital radiography and MDCT, Cohen s kappa was calculated and a Fisher s exact test was performed. If a significant difference existed between the groups for the three diagnostic confidence levels (reliable, fair, unreliable) with regard to the percentage of digital radiography diagnosis altered by MDCT, this was calculated with a Fisher s exact test. ll computations were performed using SPSS software, version Results On MDCT of the 43 patients, 14 (32.6%) showed no evidence of bone bridging, 23 (53.5%) showed evidence of partial fusion, and six (14%) showed complete fusion (Table 2). Overall agreement with digital radiography was found in 27 (63%) patients (Figs. 1 and 2). In 16 (37%) patients with disagreement between both techniques, overestimation of the bone healing process on digital radiography TLE 2: Results for Digital Radiography and MDCT for one Fusion in ll 43 Patients Fusion on Fusion on MDCT Digital Radiography None Partial Complete Total None Partial Complete Total Note Values in boldface indicate concordant results. JR:186, June

3 Fig year-old woman with history of subcapitular humeral fracture and nonunion after fixation. and, Coronal multiplanar reconstruction () and axial radiograph () of left shoulder show nonunion (arrows). Fig year-old woman with history of posterior lumbar fusion of L5 S1. and, Coronal multiplanar reconstruction () and anteroposterior radiograph () show partial fusion (arrows) JR:186, June 2006

4 MDCT and Digital Radiography of one Healing Fig year-old man with history of osteomyelitis and nonunion of proximal femur., Coronal multiplanar reconstruction shows no fusion (arrow)., nteroposterior radiograph shows partial fusion (arrow). occurred in eight (19%) patients (Fig. 3) and underestimation in eight (19%) (Figs. 4 and 5). Detailed analysis showed that in CT group (no fusion, 14 patients), agreement occurred in eight patients (57%) and overestimation on digital radiography in six (43%). In CT group (partial fusion, 23 patients), agreement occurred in 16 patients (70%), overestimation in two (9%), and underestimation in five (22%). In CT group C (complete fusion, six patients), agreement occurred in three patients (50%) and underestimation in three (50%). lthough Fisher s exact test indicated a significant correlation between CT and digital radiography (p = 0.007), the resulting kappa value was low (κ = 0.348). In 16 patients, MDCT altered the diagnosis initially made on digital radiography. ecause of the small number of patients, no statistically significant difference was seen among the three groups. The percentage of altered diagnoses ranged from 12.5% to 50%. In patients with fair or unreliable diagnostic confidence ratings, MDCT changed the initial diagnosis based on digital radiography of patients with a good diagnostic confidence level at a higher percentage ( %). In only nine (26.5%) of 43 patients was the diagnostic confidence of digital radiography rated as reliable, with a relatively low percentage (12.5%) of diagnostic change by MDCT (Table 3). In 22 patients (51.2%), metal implants were used by the orthopedic surgeons; however, in all of these patients, assessment of bone fusion was possible with only minor to moderate artifacts that were attributable to the protocol used (low pitch, standard milliampere-seconds) [12] because MDCT allows higher X-ray tube currents, which create better penetration of metallic orthopedic fixation devices [11]. Discussion This study shows that significant correlation exists between digital radiographs and MPRs from a 16-MDCT scanner in the evaluation of bone healing. However, the low kappa value brings into question the clinical value of digital radiography. Overall agreement between both techniques was 63%, with a considerable amount of overestimation (19%) and underestimation (19%) with regard to the extent of bone fusion on digital radiography. MDCT was considered the gold standard of both techniques in this study. Preclinical studies showed that radiography and 4-MDCT yielded high sensitivity and negative predictive value; however, the positive predictive value was poor, especially with fine-detail radiographs. MDCT with reformatted images appeared to be superior to fine-detail radiographs in accurately identifying nonunions [2]. The combined use of radiography and MDCT analysis allows interpretation of the vascularized fibula autograft in patients with bone tumors. bsence of bridges to the allograft indicates unsuccessful vascularization with subsequent risk of graft fracture [13]. CT was used as early as 1986 for the assessment of fracture healing, and CT proved to be superior to conventional radiog- JR:186, June

5 Fig year-old woman with history of L4 L5 fusion and anterolisthesis of L3 L4., Sagittal multiplanar reconstruction shows complete fusion (arrow)., Lateral radiograph shows partial fusion (arrow). raphy for the detection of gaps in callus [14, 15]. In principle, near-isotropic imaging, which is a prerequisite for good-quality MPRs, was possible even with nonhelical CT scanners. With the introduction of helical CT, the higher-quality MPRs became more popular but were still limited by the low scanning range. The huge performance gain achieved with modern MDCT scanners, with scanning times of only several seconds even when using thin collimation made the reconstruction of high-quality MPRs attractive, and they are now widely used in clinical practice. Other imaging techniques that have been used for the assessment of fracture healing include sonography and conventional tomography, which was used long before the introduction of clinical CT scanners. Studies using sonography for the monitoring of callus formation found an accuracy of up to 80% for the detection of pseudarthrosis after posterolateral spinal fusion [16]. Sonography showed advantages compared with conventional radiography in the visualization of the early phases of callus organization and in its progression to bridging new bone formation after fractures of long bones [17]. In pediatric radiology, sonography yielded equivalent diagnostic performance in clavicular fractures but was superior to conventional radiography for detecting early signs of bone healing or pseudarthrosis [18]. However, the penetration of the ultrasound beam is limited by the transmission frequency of the probe; thus, the practical use of diagnostic sonography is limited to superficial bones and cortical bone close to the sonographic probe. Conventional tomography has been used widely in musculoskeletal imaging, but it has been largely replaced by CT techniques [19]. Conventional tomography has been used for many years for the evaluation of the postoperative spine after posterior spinal arthrodesis [20]. Thin-section tomography had good correlation with surgery in the diagnosis of pseudarthrosis after fusions for scoliosis and was superior to anteroposterior, lateral, and oblique radiography [21]. However, conventional tomography also suffers from certain disadvantages. The common linear movement is mechanically easy to produce but will give rise to rather thick tomographic sections and a short blurring path (the length of the tomographic section). If thinner sections are required, more complex movements are needed. ecause conventional tomography does not completely blur out all distracting structures, the inherent lack of sharpness of the conventional tomographic image could make the assessment of bone bridges problematic. Thinner sections of conventional tomography, in particular, suffer from greater background blur. In dentistry radiology, the technique is called orthopantomography and is still widely used, although for practical reasons other conventional tomographic techniques have been mostly replaced by CT, and the commercial availability of conventional tomography scanners has decreased substantially. CT eliminates the blurring problem of conventional tomography and increases the per JR:186, June 2006

6 MDCT and Digital Radiography of one Healing Fig year-old man with sacroiliac joint instability and history of fusion., Coronal multiplanar reconstruction shows partial fusion (arrow)., nteroposterior radiograph shows no fusion (arrow). TLE 3: Percentage of Times MDCT Changed Initial Digital Radiography Diagnosis ased on the Diagnostic Reliability of Digital Radiography in ll 43 Patients Digital Radiography MDCT Changed Digital Radiography Yes No Total Reliable 2 (12.5) 7 (25.9) 9 Fair 8 (50.0) 10 (37.0) 18 Unreliable 6 (37.5) 10 (37.0) 16 Total Note Fisher s exact test, p = Numbers in parentheses are percentages. ceptibility of fracture healing. MDCT has the advantage that the X-ray beam passes through the whole volume of the object in a short time, and, when using isotropic or near-isotropic resolution, volumetric imaging with reconstruction of arbitrary MPRs is useful. nother important advantage of high-performing MDCT scanners is the clinically relevant reduction of motion artifacts because of the low scanning time compared with conventional CT or tomography. potential limitation of CT is the deterioration of image quality due to metal implants. s mentioned in the Results section, more than half the patients (51.1%) in our study had metal devices, which, however, led to only minor artifacts in most patients scanned. Metal artifacts in CT depend on several factors. Titanium hardware, which has been most widely used in orthopedic surgery, causes the least amount of artifacts. The CT technique also has an important impact on the severity of artifacts, with high milliampere-second and high peak kilovoltage settings leading to the reduction of artifacts. With MDCT and low pitches, a high tube current is achieved, which is the basis for good-quality MPRs [22]. Image reconstruction is also important in metal artifact reduction. ecause filtered back-projection is artifact-prone, altered image reconstruction (e.g., projection interpolation algorithm) can reduce metal artifacts [23, 24]. Dedicated metal artifact reduction software has been developed by most manufacturers, and future research could potentially further improve image quality in patients with metallic implants. With 16-MDCT scanners, the trend is to first reconstruct an overlapping secondary raw data set and then to obtain MPRs of axial, coronal, or arbitrarily angulated sections with a predefined section width [9]. one bridges are high-contrast objects and are reliably detected on 1.5- to 3-mm-thick MPRs, depending on the anatomic region, with thicker MPRs preferable for the lumbar spine and somewhat thinner MPRs preferable for the hand region [25, 8]. Reports of secondary MPRs and their impact on fracture diagnosis in acute and nonacute trauma patients have been published for many anatomic regions [12, 26 28]. In spinal trauma patients, MDCT reduces imaging time and patient manipulation and, finally, improves pa- JR:186, June

7 tient outcomes and could be cost-effective compared with conventional radiography [29]. Nonunion is the cessation of both the periosteal and the endosteal healing responses without bone bridging in patients after arthrodesis, spondylodiskitis, or fractures, with major clinical impact for the patient [30]. Nonunions of long bone fractures usually require one or more operative procedures, depending on whether the nonunion is infected [4]. Thus, proper radiologic diagnosis of delayed union or nonunion is essential for effective patient management. Clinical studies with thin-section helical CT showed bridging trabeculation after interbody fusion in 95% of cases, whereas on radiography, this was diagnosed in only 4% of cases [31]. The task in our study was to evaluate the diagnostic performance of MDCT versus digital radiography with regard to bone healing in different anatomic regions using adapted scanning protocols in clinical patients. The limitations of our study include that slightly different protocols were used and different anatomic regions were studied. Furthermore, histopathologic workup or surgical exploration was not available in this study group. However, the results of our study show that the subjective diagnostic confidence of digital conventional radiography was unreliable in a high percentage of patients (Table 3). In a significant number of patients, MDCT did alter the initial diagnosis made by digital radiography (Tables 2 and 3). Patient management by orthopedic surgeons (immobilization or mobilization, including removal or application of plasters and metallic prostheses) relied, in all patients, on the final MDCT diagnosis. Determination of the optimal time between fracture or arthrodesis and MDCT was not the purpose of this study and likely will be decided by referring orthopaedic surgeons. We recommend time intervals similar to those used for the radiographic evaluation, depending on the anatomic location and the underlying disease. lmost two thirds of our study population underwent MDCT between 1 and approximately 6 months after the initial event (fracture, orthopedic surgery). MDCT, using high-quality 2D reformatting, can be recommended as the most reliable imaging technique and should be considered the gold standard of the currently available imaging techniques. Our study shows that digital radiography is an often unreliable diagnostic technique and is not sufficient as the exclusive radiologic method for the management of patients with questionable bone healing. MDCT provides fast and valuable information in assessing bone healing in patients with a clinical suspicion of delayed union or nonunion, especially when the primary radiograph is equivocal. References 1. Volk M, Strotzer M, Holzknecht N, et al. Digital radiography of the skeleton using a large-area detector based on amorphous silicon technology: image quality and potential for dose reduction in comparison with screen-film radiography. Clin Radiol 2000; 55: Yee J, ae HW, Friess D, Robbin M, Johnstone, Yoo JU. ccuracy and interobserver agreement for determinations of rabbit posterolateral spinal fusion. Spine 2004; 29: Einhorn T, Majeska RJ, Rush E, Levine PM, Horowitz MC. The expression of cytokine activity by fracture callus. J one Miner Res 1995; 10: Rodriguez-Merchan EC, Forriol F. Nonunion: general principles and experimental data. Clin Orthop Relat Res 2004; 419: Young JW, Kovelman H, Resnik CS, Paley D. Radiologic assessment of bones after Ilizarov procedures. Radiology 1990; 177: lokhuis TJ, de ruine JH, ramer J, et al. The reliability of plain radiography in experimental fracture healing. Skeletal Radiol 2001; 30: Grigoryan M, Lynch J, Fierlinger L, et al. Quantitative and qualitative assessment of closed fracture healing using computed tomography and conventional radiography. cad Radiol 2003; 10: uckwalter K, Farber JM. pplication of multidetector CT in skeletal trauma. Semin Musculoskelet Radiol 2004; 8: Prokop M. General principles of MDCT. Eur J Radiol 2003; 45 [suppl 1]:S4 S Fleischmann D, Rubin GD, Paik DS, et al. Stair-step artifacts with single versus multiple detector-row helical CT. Radiology 2000; 216: Rydberg J, Liang Y, Teague SD. Fundamentals of multichannel CT. Semin Musculoskelet Radiol 2004; 8: Farber JM. Imaging of the wrist with multichannel CT. Semin Musculoskelet Radiol 2004; 8: Manfrini M, Vanel D, De Paolis M, et al. Imaging of vascularized fibula autograft placed inside a massive allograft in reconstruction of lower limb bone tumors. JR 2004; 182: raunstein EM, Goldstein S, Ku J, Smith P, Matthews LS. Computed tomography and plain radiography in experimental fracture healing. Skeletal Radiol 1986; 15: Schnarkowski P, Redei J, Peterfy CG, et al. Tibial shaft fractures: assessment of fracture healing with computed tomography. J Comput ssist Tomogr 1995; 19: Jacobson J, Starok M, Pathria MN, Garfin SR. Pseudarthrosis: sonography evaluation after posterolateral spinal fusion: work in progress. Radiology 1997; 204: Maffulli N, Thornton. Ultrasonographic appearance of external callus in long-bone fractures. Injury 1995; 26: lab E, Geissler W, Rokitansky. Sonographic management of infantile clavicular fractures. Pediatr Surg Int 1999; 15: Ho C, Sartoris DJ, Resnick D. Conventional tomography in musculoskeletal trauma. Radiol Clin North m 1989; 27: Clader TJ, Dawson EG, assett LW. The role of tomography in the evaluation of the postoperative spinal fusion. Spine 1984; 9: Dawson EG, Clader TJ, assett LW. comparison of different methods used to diagnose pseudarthrosis following posterior spinal fusion for scoliosis. J one Joint Surg m 1985; 67: White LM, uckwalter K. Technical considerations: CT and MR imaging in the postoperative orthopedic patient. Semin Musculoskelet Radiol 2002; 6: Mahnken H, Raupach R, Wildberger JE, et al. new algorithm for metal artifact reduction in computed tomography: in vitro and in vivo evaluation after total hip replacement. Invest Radiol 2003; 38: Watzke O, Kalender W. pragmatic approach to metal artifact reduction in CT: merging of metal artifact reduced images. Eur Radiol 2004; 14: uckwalter K, Rydberg J, Kopecky KK, Crow K, Yang EL. Musculoskeletal imaging with multislice CT. JR 2001; 176: Herzog C, hle H, Mack MG, et al. Traumatic injuries of the pelvis and thoracic and lumbar spine: does thin-slice multidetector-row CT increase diagnostic accuracy? Eur Radiol 2004; 14: Haapamaki VV, Kiuru MJ, Koskinen SK. nkle and foot injuries: analysis of MDCT findings. JR 2004; 183: egemann PG, Kemper J, Gatzka C, Stork, Nolte-Ernsting C, dam G. Value of multiplanar reformations (MPR) in multidetector CT (MDCT) of acute vertebral fractures: do we still have to read the transverse images? J Comput ssist Tomogr 2004; 28: Van Goethem JW, Maes M, Ozsarlak O, van den Hauwe L, Parizel PM. Imaging in spinal trauma. Eur Radiol 2005; 15: Marsh D. Concepts of fracture union, delayed union, and nonunion. Clin Orthop Relat Res 1998; 355[suppl]:S22 S Shah RR, Mohammed S, Saifuddin, Taylor. Comparison of plain radiographs with CT scan to evaluate interbody fusion following the use of titanium interbody cages and transpedicular instrumentation. Eur Spine J 2003; 12: JR:186, June 2006