http://fai.sagepub.com/Foot & Ankle International

http://fai.sagepub.com/content/early/2014/07/11/1071100714540890.1The online version of this article can be found at:

 DOI: 10.1177/1071100714540890

published online 11 July 2014Foot Ankle IntFabrice Colin, Tamara Horn Lang, Lukas Zwicky, Beat Hintermann and Markus Knupp

Subtalar Joint Configuration on Weightbearing CT Scan  

Published by:

http://www.sagepublications.com

On behalf of: 

  American Orthopaedic Foot & Ankle Society

can be found at:Foot & Ankle InternationalAdditional services and information for    

  http://fai.sagepub.com/cgi/alertsEmail Alerts:

 

http://fai.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Jul 11, 2014OnlineFirst Version of Record  

- Jul 18, 2014OnlineFirst Version of Record >>

by guest on August 29, 2014fai.sagepub.comDownloaded from by guest on August 29, 2014fai.sagepub.comDownloaded from

Foot & Ankle International® 1 –6© The Author(s) 2014Reprints and permissions: sagepub.com/journalsPermissions.navDOI: 10.1177/1071100714540890fai.sagepub.com

Basic Science Research Article

Introduction

While the ankle joint has been examined and described in numerous morphological and biomechanical studies, only little is known about the subtalar (ST) joint. However, it is well accepted that the ST joint consists of 3 articulating fac-ets: anterior, intermediate anterior, and posterior.23 The pos-terior talar articular surface is strongly concave in the sagittal plane and usually flat or minimally concave in the coronal plane.23 However, in cadaveric and radiological studies, large variations of the ST joint morphology have been described.6,12,14,16

Inman12 described the ST joint as a torque transmitter, in which rotation of the leg is converted into pronation and supination of the foot and vice-versa. The orientation and shape of the ST joint therefore has an important impact on the biomechanical properties of the hindfoot. However, to the best of our knowledge, the radiological anatomy and shape of the ST joint has not been examined and analyzed in detail in an asymptomatic cohort, without any hindfoot pathology.

Plain weightbearing radiographs play an important role in the diagnosis and assessment of hindfoot pathologies.22

However, newer imaging modalities such as computed tomography (CT) allow more detailed imaging of the inter-tarsal relationship and are even considered superior in eval-uating various bone pathologies.3,15 Several groups have developed devices that simulate weightbearing during CT scan examination in a seated position.7,11,17,26,27 These stud-ies generally showed that weightbearing had a significant effect on the anatomical changes of the foot configuration in both asymptomatic and pathological foot conditions. However, none of these studies analyzed the morphology of the ST joint. Furthermore, these seated weightbearing CT scan examinations were done using an imaging modality not designed to collect weightbearing images, but neverthe-less showed the importance of using weightbearing devices. Nowadays, weightbearing CT scan modalities that allow

540890 FAIXXX10.1177/1071100714540890Foot & Ankle InternationalColin et alresearch-article2014

1Clinic of Orthopaedic & Traumatology Surgery, Nantes, France2Clinic of Orthopaedic Surgery, Liestal, Switzerland

Corresponding Author:Markus Knupp, MD, Clinic of Orthopaedic Surgery, Kantonsspital Baselland, Rheinstrasse 26, CH-4410 Liestal, Switzerland. Email: markus.knupp@ksbl.ch

Subtalar Joint Configuration on Weightbearing CT Scan

Fabrice Colin, MD1, Tamara Horn Lang, PhD2, Lukas Zwicky, MSc2, Beat Hintermann, MD2, and Markus Knupp, MD2

AbstractBackground: Standard values that describe the morphology of the subtalar (ST) joint have previously been obtained from cadaveric studies or by using conventional unloaded radiographs. It is known that these parameters differ significantly from those measured in vivo and in loaded images, limiting the diagnostic value of the previously published morphological parameters in the literature. However, the morphology of the ST joint clearly affects its function. The objective of this study was to determine the morphology of the posterior facet of the ST joint using loaded computed tomography (CT) images and to describe the different configurations found in asymptomatic patients.Methods: A weightbearing CT scan was performed on 59 patients without any history of hindfoot and ankle pathology. The shape of the posterior facet and the subtalar vertical angle (SVA) were measured in 3 different coronal planes of the ST joint.Results: The posterior facet was concave in 88% and flat in 12%. The posterior facet was oriented in valgus in 90% and varus in 10% when measured in the middle coronal plane. However, the SVA changed depending on which coronal plane it was measured in.Conclusion: We believe it is important to get a better insight into the morphological parameters of the ST joint.Clinical Relevance: Knowledge of subtalar joint morphology could help clarify why certain failures have occurred in reconstructive hindfoot surgery and thus might help plan the surgical procedure to reduce these failures in the future.

Keywords: weightbearing CT scan, subtalar joint, morphology

by guest on August 29, 2014fai.sagepub.comDownloaded from

2 Foot & Ankle International

true weightbearing images are available, and images taken represent the physiological hindfoot morphology found in an upright weightbearing position.

The shape of the subtalar joint is of importance since the morphology of a joint has an effect on the joint function. Knowing the morphology of the ST joint in asymptomatic patients may serve as a tool to understand the evolution of hindfoot deformity and/or degenerative wear of the ankle and subtalar joint. The hypothesis of this study was that the posterior facet of the ST joint adopted different shapes and orientation on the coronal plane on weightbearing CT in an asymptomatic cohort.

Methods

Population

This study was conducted in accordance with the Declaration of Helsinki and Guidelines for Good Clinical Practice. The protocol was approved by the Ethics Committee. A total of 59 healthy volunteers informed about the radiation expo-sure (male, 34; female, 25; mean age, 45 years; range, 18-74) without hindfoot and ankle pathologies after clinical evaluation were included in this study. Patients with previ-ous history of hindfoot and ankle trauma and/or surgery were excluded from this study.

Radiographic Technique

CT images were collected starting from 10 cm proximal to the tibiotalar joint to the sole of the foot with the subjects standing upright and fully weightbearing using a cone beam CT (Planmed Verity, Planmed Oy, Helsinki, Finland; 0.2 mm slice thickness, 1 mm slice interval). The CT scan had an anode voltage of up to 96 kV, 7.5 mA, 40 mAs, FOV 13 × 16 cm, and according to a previous study, the maximal estimated radiation dose, using the default adult exposure parameters, was 12.6 μSv.19 Images were saved as Digital Imaging and Communications in Medicine (DICOM) files, and 3D Multi-Planar Reconstruction (3D-MPR) was per-formed using an image processing software (OsiriX MD, Pixmeo, Geneva, Switzerland).

Measurement Methods

To define the 3 different coronal planes used, the length of the ST joint was measured on a sagittal plane that included the projection of a line connecting the center of the heel and the head of the second metatarsal. A cut through the middle of the ST joint on the sagittal plane was defined as the middle coronal plane used to measure the subtalar vertical angle (SVA). The third anterior coronal plane was defined as 5 mm anterior to the middle coronal plane and the third posterior coronal plane as 5 mm posterior to the middle coronal plane (Figure 1).

The shape of the posterior talar articular surface was determined as being concave or flat on the coronal plane

(Figures 2, 3). The articular surface was considered flat if the radius of a circle whose curvature was adjusted/adapted to the articular surface was more than 60 mm.

The orientation of the ST joint was measured using the SVA. Van Bergeyk et al defined the SVA as being the incli-nation between a line drawn connecting the medial and lat-eral border of the posterior facet of the ST joint and a vertical line on the coronal plane that is perpendicular to the floor.26 Values less than 90 degrees were defined as varus and more than 90 degrees as a valgus configuration (Figure 2).26 To assess the screw-shaped surface of the ST joint, the SVA was measured in 3 different coronal planes (anterior, middle, and posterior planes).

All of the measurements taken were done by a trained orthopaedic surgeon (FC).

Statistical Methods

Normal distribution of the data was checked using a Kolmogorov-Smirnov test. A Student t test was used to compare 2 groups. To compare more than 2 groups, a 1-way ANOVA followed by a Bonferroni’s post hoc comparisons test was performed. The degree of correlation was described using the Pearson correlation coefficient. The level of sig-nificance was defined as P < .05. All statistical data analysis was performed using SPSS 21.0 (IBM SSPS, Armonk, New York, USA).

Results

The shape of the articular surface of the ST joint was con-sidered on each coronal cut. The geometry of the articular

Figure 1. A sagittal cut of a weightbearing CT scan was used to define the anterior, middle, and posterior coronal planes used to measure the configuration of the subtalar (ST) joint. This sagittal plane was defined as running through the center of the heel and the head of the second metatarsal. Using a 3D Multi-Planar Reconstruction, the middle coronal plane was defined as being the middle of the ST joint length. The anterior and posterior coronal cuts were defined as 5 mm anterior and 5 mm posterior to the middle cut.

by guest on August 29, 2014fai.sagepub.comDownloaded from

Colin et al 3

surface did not change depending on the coronal plane ana-lyzed and was found to be concave in 52 patients (88%) and flat in 7 patients (12%). The SVA of the posterior facet var-ied according to the coronal plane it measured: anterior, middle, or posterior. Forty-two patients (71%) presented a varus-valgus-valgus configuration (SVA anterior in varus, SVA middle, and SVA posterior in valgus) (Figure 4). Eleven patients (19%) presented a valgus-valgus-valgus configuration. Six patients (10%) presented a varus-varus-valgus configuration. The SVA measured in the middle coronal plane of the posterior facet was in valgus in 53 patients (90%) and in varus in 6 patients (10%). The range of the SVA varied from 67.7 degrees measured in the ante-rior coronal plane to 122.7 degrees measured in the poste-rior coronal plane. The SVA measured in the anterior and posterior planes highly correlated with the SVA measured in the middle plane (r = 0.85 and r = 0.82, respectively; P < .001) (Figure 5).

The SVA measured in the middle coronal plane was sig-nificantly higher in patients with a flat posterior facet (flat: 103.9 degrees, SD = 4.4 degrees; concave: 97.4 degrees, SD = 6.2; P = .01) (Figure 6).

Discussion

In this study, weightbearing CT scans were used in patients with asymptomatic clinical morphology of the hindfoot to assess the ST joint configuration in the coronal plane. We found that the ST joint is either shaped concave or flat with an SVA angle ranging from a varus to a valgus orientation when measured in the coronal plane. The majority of the volunteers who participated had a concave shaped ST joint, which was valgus oriented when looking at the posterior facet joint line. The novelty of this study lies in the patient cohort used and the method applied to describe the mor-phology of the ST joint: (1) asymptomatic volunteers were used and (2) weightbearing CT scans were used. To our knowledge this has not been done previously.

Earlier anatomical studies have shown that the posterior facet is concave like a ball and socket.1 In 1941 Manter20 and later others2,12,23 described the ST joint in cadavers and found that the posterior facet was an oblique, screw-shaped surface in the coronal plane.2,12,23 Subsequent radiological studies showed that the posterior facet was concave, flat, or

Figure 2. A coronal image taken of a hindfoot using a weightbearing CT scan device. Illustrated is the subtalar vertical angle (SVA), defined as the inclination of the line connecting the medial and lateral aspects of the talus with a vertical line, which is perpendicular to the ground. In this individual, on middle coronal cut of the subtalar (ST) joint, the shape was concave and the SVA was oriented in valgus.

Figure 3. An example of an individual with a flat shaped facet on the middle coronal cut of the subtalar (ST) joint of a weightbearing CT scan.

by guest on August 29, 2014fai.sagepub.comDownloaded from

4 Foot & Ankle International

slanted on the coronal plane.6,13,14 Farsetti et al6 were the first to use CT scans to assess the shape of the ST joint on the coronal plane. However, this study was conducted in patients with clubfeet. Of the 108 clubfeet examined, only 12 (11%) showed a concave shape of the ST joint.6

In contrast, in the present study, we found that in an asymptomatic cohort the shape of the ST joint surface was

Figure 5. Line diagram of the subtalar vertical angle (SVA) related to its position. The SVA increases in valgus when the measurement was performed posteriorly. Figure 6. Subtalar vertical angle (SVA) on the middle coronal

plane related to the shape of the subtalar (ST) joint. The SVA on the middle coronal plane of the ST joint in the flat shape group was significantly higher in valgus (flat: 103.9 degrees, SD = 4.4 degrees; concave: 97.4 degrees, SD = 6.2; P = .01).

Figure 4. This is an example of a varus-valgus-valgus subtalar (ST) joint configuration found on the 3 coronal planes measured using a weightbearing CT scan. The subtalar vertical angle (SVA) anterior, middle, and posterior was oriented in (a) varus, (b) valgus, (c) valgus.

by guest on August 29, 2014fai.sagepub.comDownloaded from

Colin et al 5

concave in 88% of the ST joints measured. Only 12% of the patients measured showed a flat ST joint shape. Interestingly, these patients who had a flat ST joint additionally had a more pronounced valgus configuration than the patients with a concave ST joint shape. Thus, we believe that patients with a flat ST joint shape have a limited capacity to com-pensate for hindfoot deformities, particularly those with a flatfoot or an ankle valgus deformity. In contrast, in patients with a concave shaped ST joint, the joint itself may be able to compensate coronal plane deformities. This is supported by data from anatomical and functional studies.10,18,25 Recently, Hayashi et al10 described the complex interaction of the ankle and the ST joint in malaligned feet in the coro-nal plane. They suggested that the ST joint is able to com-pensate for varus in the distal tibia with a progressive valgus inclination. These data further suggest that patients with a concave shaped ST joint thus may more likely benefit from conservative treatment options such as ankle/foot orthosis or insoles than patients with a flat shaped ST joint. However, further studies are necessary to elucidate this.

A further parameter used to describe the ST joint was the SVA. Van Bergeyk et al26 were the first to describe the SVA in order to assess the hindfoot alignment in chronic lateral ankle instability. Inter- and intraobserver reliability were measured and documented. They reported that the ST joint in the control group (no hindfoot pathophysiology) was ori-ented in valgus in a majority of the patients examined. Similar to Van Bergeyk et al, our study showed that the ori-entation of the ST joint, in a majority of the cases, was in valgus (measured in the middle coronal plane: in valgus in 90% and in varus in 10%). However, in contrast to our study, they did not take into consideration that the SVA is dependent on the coronal plane in which it is measured (anterior, middle, or posterior). For example, when mea-sured in the anterior plane, the SVA in a majority of the patients (48 patients, 81%) was in varus. Our data therefore show that depending on the coronal plane in which the SVA was measured, the amount of varus/valgus inclination var-ied markedly.

It is known that the osseous ankle joint configuration could be a risk factor to develop chronic ankle instability and ankle osteoarthritis.2,8-10,24 However, the role and the influence of the ST joint on hindfoot disease and instability is still unclear. Our data suggest that the function and the compensatory role of the ST joint could highly depend on the orientation and the shape of the articular surface. This is important to know, since hindfoot osteotomies not only affect the tibiotalar joint but also the ST joint.4 We strongly believe that the ankle and the ST joint are intimately linked and thus some of the failures in reconstructive hindfoot sur-gery may be due to the ST joint configuration.

To our knowledge, this is the first weightbearing CT study describing the ST joint configuration on an asymp-tomatic cohort. A high number of patients were included in attempt to be representative of the general population. A

reliable and reproducible angle independent of the long axis of the tibia to determine the inclination of the ST joint was used.26 The CT scans used in our study presented several advantages: the images collected, with patients being in an upright, weightbearing position, resembled the physiologi-cal situation of how the ST joint is normally stressed. The image quality of the portable extremity CT scanner was considered excellent, according to Martinez et al,21 and fur-thermore, the detail and accuracy of CT provides the best method nowadays for evaluating hindfoot pathology and alignment.5

Nevertheless, this study presented some limitations. Although full weightbearing CT scans have the advantage of being high resolution images where the ST joint can be scanned in a loaded, upright position and thus under physi-ological stress, it is a new method for examining the foot, and thus little previous data are available.3 Therefore, no comparison can be made with our results. Furthermore, although we included only patients without hindfoot and ankle pathology, we did not take into consideration objec-tive data such as the radiological inclination of the tibial plafond and/or the hindfoot alignment as an exclusion cri-terion. The reliability of the SVA angle was measured in simulated weightbearing position in a previous study and not in a true upright position as in our study.26 This could be a potential bias. Instead of the radiation exposure esti-mated at 12.6 μSv, which was widely lower than the maxi-mal limit of radiation exposure allowed for diagnostic use, this exposure presented a risk potential for asymptomatic volunteers.

In conclusion, a majority of the asymptomatic patients had a concave shaped ST joint in a valgus orientation in the coronal plane. However, variations were found and may have a significant impact on the biomechanics of both the hind- and the midfoot. For the authors, the data suggest that the morphology of the posterior facet may be more involved in hindfoot pathologies than previously expected.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

1. Bonnel F, Teissier P, Maestro M, et al. Biometry of bone com-ponents in the talonavicular joint: a cadaver study. Orthop Traumatol Surg Res. 2011;97(suppl 6):66-73.

2. Bonnel F, Toullec E, Mabit C, Tourne Y, Sofcot. Chronic ankle instability: biomechanics and pathomechanics of liga-ments injury and associated lesions. Orthop Traumatol Surg Res. 2010;96(4):424-432.

by guest on August 29, 2014fai.sagepub.comDownloaded from

6 Foot & Ankle International

3. Collan L, Kankare JA, Mattila K. The biomechanics of the first metatarsal bone in hallux valgus: a preliminary study utilizing a weight bearing extremity CT. Foot Ankle Surg. 2013;19(3):155-161.

4. Davitt J, Beals T, Bachus K. The effects of medial and lateral displacement calcaneal osteotomies on ankle and subtalar joint pressure distribution. Foot Ankle Int. 2001;22(11):885-889.

5. Ellis SJ, Deyer T, Williams BR, et al. Assessment of lateral hindfoot pain in acquired flatfoot deformity using weightbear-ing multiplanar imaging. Foot Ankle Int. 2010;31(5):361-371.

6. Farsetti P, De Maio F, Russolillo L, Ippolito E. CT study on the effect of different treatment protocols for clubfoot pathol-ogy. Clin Orthop Relat Res. 2009;467(5):1243-1249.

7. Ferri M, Scharfenberger AV, Goplen G, Daniels TR, Pearce D. Weightbearing CT scan of severe flexible pes planus deformities. Foot Ankle Int. 2008;29(2):199-204.

8. Frigg A, Frigg R, Hintermann B, Barg A, Valderrabano V. The biomechanical influence of tibio-talar containment on stability of the ankle joint. Knee Surg Sports Traumatol Arthrosc. 2007;15(11):1355-1362.

9. Frigg A, Magerkurth O, Valderrabano V, Ledermann HP, Hintermann B. The effect of osseous ankle configuration on chronic ankle instability. Br J Sports Med. 2007;41(7):420-424.

10. Hayashi K, Tanaka Y, Kumai T, Sugimoto K, Takakura Y. Correlation of compensatory alignment of the subtalar joint to the progression of primary osteoarthritis of the ankle. Foot Ankle Int. 2008;29(4):400-406.

11. Ikoma K, Noguchi M, Nagasawa K, et al. A new radiographic view of the hindfoot. J Foot Ankle Res. 2013;6(1):48.

12. Inman V, Stiehl J. Inman’s Joints of The ankle, 2nd ed. Baltimore, MD: William and Wilkins; 1991.

13. Ippolito E, Fraracci L, Caterini R, Di Mario M, Farsetti P. A radiographic comparative study of two series of skeletally mature clubfeet treated by two different protocols. Skeletal Radiol. 2003;32(8):446-453.

14. Ippolito E, Fraracci L, Farsetti P, Di Mario M, Caterini R. The influence of treatment on the pathology of club foot. CT study at maturity. J Bone Joint Surg Br. 2004;86(4):574-580.

15. Johnson PT, Fayad LM, Frassica FJ, Fishman EK. Computed tomography of the bones of the foot: neoplastic disease. J Comput Assist Tomogr. 2009;33(3):436-443.

16. Johnston CE II, Hobatho MC, Baker KJ, Baunin C. Three-dimensional analysis of clubfoot deformity by computed tomography. J Pediatr Orthop B. 1995;4(1):39-48.

17. Kido M, Ikoma K, Imai K, et al. Load response of the tarsal bones in patients with flatfoot deformity: in vivo 3D study. Foot Ankle Int. 2011;32(11):1017-1022.

18. Knupp M, Stufkens SA, van Bergen CJ, et al. Effect of supra-malleolar varus and valgus deformities on the tibiotalar joint: a cadaveric study. Foot Ankle Int. 2011;32(6):609-615.

19. Koivisto J, Kiljunen T, Wolff J, Kortesniemi M. Assessment of effective radiation dose of an extremity CBCT, MSCT and conventional X ray for knee area using MOSFET dosemeters. Radiat Prot Dosimetry. 2013;157(4):515-524.

20. Manter J. Movements of the subtalar and transverse tarsal joints. Anat Rec. 1940;80:397.

21. Martinez S, Herzenberg JE, Apple JS. Computed tomogra-phy of the hindfoot. Orthop Clin North Am. 1985;16(3):481-496.

22. Saltzman CL, el-Khoury GY. The hindfoot alignment view. Foot Ankle Int. 1995;16(9):572-576.

23. Sarrafian SK. Biomechanics of the subtalar joint complex. Clin Orthop Relat Res. 1993;290:17-26.

24. Schaefer KL, Sangeorzan BJ, Fassbind MJ, Ledoux WR. The comparative morphology of idiopathic ankle osteoarthritis. J Bone Joint Surg Am. 2012;94(24):e181.

25. Ting AJ, Tarr RR, Sarmiento A, Wagner K, Resnick C. The role of subtalar motion and ankle contact pressure changes from angular deformities of the tibia. Foot Ankle. 1987;7(5):290-299.

26. Van Bergeyk AB, Younger A, Carson B. CT analysis of hind-foot alignment in chronic lateral ankle instability. Foot Ankle Int. 2002;23(1):37-42.

27. Zhang Y, Xu J, Wang X, et al. An in vivo study of hindfoot 3D kinetics in stage II posterior tibial tendon dysfunction (PTTD) flatfoot based on weight-bearing CT scan. Bone Joint Res. 2013;2(12):255-263.

by guest on August 29, 2014fai.sagepub.comDownloaded from

SCIENTIFIC ARTICLE

CTarthrography of the wrist using a novel, mobile, dedicatedextremity cone-beam CT (CBCT)

Seppo K. Koskinen & Ville V. Haapamäki & Jari Salo &

Nina C. Lindfors & Mika Kortesniemi & Lauri Seppälä &

Kimmo T. Mattila

Received: 11 June 2012 /Revised: 17 August 2012 /Accepted: 27 August 2012# ISS 2012

AbstractPurpose To evaluate the feasibility and intra- and interob-server agreement of CBCT arthrography of wrist ligaments,triangular fibrocartilaginous complex (TFCC), and to assessthe sensitivity (SE), specificity (SP), accuracy (ACC), andpositive and negative predictive value (PPV, NPV) of CBCTarthrography in the diagnosis of scapholunate (SLL) andlunotriquetral (LTL) ligament tears, TFCC, and cartilageabnormalities of the scaphoid and lunate with theircorresponding radial surfaces (scaphoid and lunate fossa)using a novel, mobile, dedicated extremity CBCT scanner.Materials and methods Fifty-two consecutively enrolledsubjects (26 M, 26 F, mean age 38 years, range 18–66 years)with suspected wrist ligament tears underwent CBCT-arthrography before normally scheduled MR arthrog-raphy.An extremity CBCT was used for imaging with iso-tropic voxel size of 0.4 × 0.4 × 0.4 mm3. Subsequent routine1.5 T MRI was performed using a dedicated wrist coil.Twoobservers reviewed the anonymized CBCT images twice for

contrast enhancement (CE) and technical details (TD), fortears of the SLL, LTL, and TFCC. Also, cartilage abnormal-ities of the scaphoid and lunate with their correspondingradial surfaces (scaphoid and lunate fossa) were evaluated.Inter- and intraobserver agreement was determined usingweighted kappa statistics. Since no surgery was performed,MRI served as a reference standard, and SE and SP, ACC,PPV, and NPV were calculated.Results Intra- and interobserver kappa values for both read-ers (reader 1/reader 2; first reading/second reading) with95 % confidence limits were: CE 0.54 (0.08–1.00)/ 0.75(0.46–1.00); 0.73 (0.29–1.00)/ 0.45 (0.07–0.83), TD 0.53(0.30–0.88)/ 0.86 (0.60–1.00); 0.56 (0.22–0.91)/ 0.67(0.37–0.98), SLL 0.59 (0.25–0.93)/ 0.66 (0.42–0.91); 0.31(0.06–0.56)/ 0.49 (0.26–0.73), LTL 0.83 (0.66–1.00)/ 0.68(0.46–0.91); 0.90 (0.79–1.00)/ 0.48 (0.22–0.74); TFCC(0.72–1.00)/ (0.79–1.00); 0.65 (0.43–0.87)/ 0.59 (0.35–0.83), radius (scaphoid fossa) 0.45 (0.12–0.77)/ 0.64(0.31–0.96); 0.58 (0.19–0.96)/ 0.38 (0.09–0.66), scaphoid

S. K. Koskinen :V. V. Haapamäki :M. KortesniemiDepartment of Radiology, HUS Helsinki Medical Imaging Center,Helsinki University Central Hospital,Helsinki, Finland

J. SaloDepartment of Orthopedic and Trauma Surgery, HelsinkiUniversity Central Hospital,Helsinki, Finland

J. SaloDepartment of Orthopedics, Traumatology and Hand Surgery,Kuopio University Hospital, UEF,Kuopio, Finland

N. C. LindforsDepartment of Orthopedic and Hand Surgery, Helsinki UniversityCentral Hospital,Helsinki, Finland

L. SeppäläPlanmed Oy,Helsinki, Finland

K. T. MattilaDepartment of Diagnostic Radiology, Turku University CentralHospital,Turku, Finland

S. K. Koskinen (*)Department of Radiology, HUS Helsinki Medical Imaging Center,Helsinki University Hospital, Töölö Trauma Center,Topeliuksenkatu 5, PL 266,Helsinki 00029 Finlande-mail: seppo.koskinen@helsinki.fi

Skeletal RadiolDOI 10.1007/s00256-012-1516-0

0.43 (0.12–0.74)/ 0.76 (0.55–0.96); 0.37 (0.00–0.75)/ 0.32(0.04–0.59), radius (lunate fossa) 0.68 (0.36–1.00)/ 0.42(0.00–0.86); 0.62 (0.29–0.96)/ 0.51 (0.12–0.91), and lunate0.53 (0.16–0.90)/ 0.68 (0.44–0.91); 0.59 (0.29–0.88)/ 0.42(0.00–0.84), respectively.

The overall mean accuracy was 82–92 % and specificitywas 81–94 %. Sensitivity for LTL and TFCC tears was 76–83, but for SLL tears it was 58 %. For cartilage abnormal-ities, the accuracy and negative predictive value were high,90–98 %.Conclusions A dedicated CBCT extremity scanner is anew method for evaluating the wrist ligaments andradiocarpal cartilage. The method has an overall accu-racy of 82–86 % and specificity 81–91 %. For cartilageabnormalities, the accuracy and negative predictive val-ue were high.

Keywords Wrist . Arthrogram . Cone beam CT . Trauma .

Musculoskeletal

Introduction

Magnetic resonance arthrography is an established tech-nique for detecting ligament and cartilage injuries of thewrist [1–3]. However, the technique is rather expensiveand time-consuming, and MRI in general has severalcontraindications. Also, spatial resolution of MRI canbe a limiting factor, especially in areas of convex orconcave joint facets with thin cartilage [4, 5]. Therefore,alternative techniques such as CT arthrography (CTA)have received more attention [6]. CTA is a much fastertechnique and has excellent spatial resolution [7]. More-over, the ability for multiplanar reformats makes it evenmore versatile. In addition to conventional CT scanners,cone-beam CT (CBCT) has been used for dental imagingsince the late 1990s [8, 9]. The clinical indications in-clude dentoalveolar imaging, such as assessing dentoal-veolar trauma and preoperative assessment of impactedteeth, and preoperative planning for maxillofacial surgery[10]. This method has recently been implemented inorthopedic imaging to study finger and wrist fracturesand scaphoid fracture after screw fixation [11, 12]. Morerecently, dedicated extremity CBCT scanners have beenintroduced [13, 14]. This application offers an attractivealternative, with high spatial resolution, easy installation,and low radiation dose [4, 11–15] compared to conven-tional CT scanners. Also, the ability to image the lowerextremity during weight bearing, i.e., while the patientstands, opens new possibilities to study degenerativejoint disease of the knee, ankle, and foot (Tuominen etal., Weight bearing CT-imaging of the lower extremity,unpublished).

In this study, we wanted to expand the dedicatedextremity CBCT scanner’s capabilities even further,i.e., with CT arthrography. Therefore, the purpose ofthis study was to evaluate the feasibility and intra- andinterobserver agreement of CT arthrography of scapho-lunate and lunotriquetral ligaments, triangular fibrocarti-laginous complex, and cartilage abnormalities using anovel, dedicated extremity CBCT. To our knowledge,this is the first report of wrist arthrography using adedicated extremity CBCT scanner.

Materials and methods

Fifty-three consecutively enrolled subjects were offereda CBCT arthrogram immediately prior to their routinely

Fig. 1 CBCT image (88 kVp, 8 mA) of six phantoms of 2-ml volumewith varying proportions of 240 mg I/ml iohexol and 2.5 mmol/l tet-raazacyclododecanetetraacetic acid (DOTA)-gadolinium in the follow-ing respective proportions: 1 phantom 1,100 %/0 %; phantom 2, 75 %/25 %; phantom 3, 50 %/50 %; phantom 4, 25 %/75 %; phantom 5, 0 %/100 %, and phantom 6100 % isotonic NaCl. Using the mixture of 50/50 (3), there was no decrease in attenuation values as there was withthe concentration of 25/75 (4)

Fig. 2 The hand and wrist can be placed freely in any desired position

Skeletal Radiol

scheduled MRI arthrogram. The time period was7 months, from November 2010 to May 2011. Onepatient refused to participate in the study and hencethe final study group was comprised of 52 subjects(26 M, 26 F, mean age 38 years, range 18–66 years).Approval was obtained from the hospital’s ethics com-mittee, and written informed consent was obtained fromeach subject.

The indication for MR arthrography was suspectedscapholunate (SL) ligament tear in 17 subjects, lunotri-quetral (LT) ligament tear (n01), SL and LT ligamenttear (n08), SL and triangular fibrocartilage complex(TFCC) tear (n02), LT and TFCC tear (n02), TFCCtear (n019). In addition, radiocarpal joint cartilage as-sessment was requested in six cases, and evaluation ofdistal radioulnar joint (DRUJ) in two cases. Fivepatients were operated on within the last 12 months(Brunelli’s tenodesis for scapholunate instability, STT-arthrodesis + TFCC fixation, fixation of distal radiusfracture with volar plate (n02), resection of pisiformbone), and one patient had undergone a diagnostic wristarthroscopy 12 months earlier. Moreover, one patienthad avascular necrosis with fragmentation of the lunate.

Since the patients underwent subsequent MRIarthrography of the wrist, we first studied the amountof CT and MR contrast used in the mixture to obtainoptimal contrast. We prepared six phantoms of 2-mlvolume with varying proportions of 240 mg I/mliohexol (Omnipaque, GE Healthcare, Oslo, Norway)

and 2.5 mmol/l tetraazacyclododecanetetraacetic acid(DOTA)-gadolinium (Artirem, Guerbet, France). Phan-toms 1–5 were composed of 240 mg I/ml iohexol and2.5 mmol/l DOTA-gadolinium in the following respec-tive proportions: phantom 1,100 %/0 %; phantom 2,75 %/25 %; phantom 3, 50 %/50 %; phantom 4,25 %/75 %; phantom 5, 0 %/100 %, and phantom 6100 % isotonic NaCl. The mixture of 1:1 was selected,because it provided adequate contrast for both in vitroand in vivo CBCT (Fig. 1) and MR arthrographyimages.

A mean of 2.2 ml (range 1.5–3.0 ml) of 1:1 solutionof 2.5 mmol/l Gd- DOTA and 240 mg I/ml iohexol wasinjected into the radiocarpal joint under palpationguidance.

A novel extremity CBCT scanner (Planmed Verity,Planmed Oy, Helsinki, Finland) was used to image the wrist(Fig. 2). The dimensions of the scanner are (L × W × H):185 × 76 × 160 cm, weight app. 350 kg, maximum power-consumption 1.5 kVAwith no external cooling needed. Thescanner uses a Toshiba X-ray tube with a tungsten target,anode voltage up to 96 kV, anode current 1–12 mA, dualfiltration 0.5 mm-Cu + 2.5 mm-Al, and pulsed X-ray radi-ation. The scanner has a 20 × 25-cm flat-panel amorphoussilicon detector. The field of view (FOV) is approximately13 × 16 cm, and 300 projection images were acquired overan angle of 210° with a scan time of 18 s and reconstructiontime of 30–120 s. The isotropic voxel size was0.4 × 0.4 × 0.4 mm3. Based on previous tests [14], 88 kVp

Table 1 Intra- and interobserverkappa values for readers 1 (R1)and 2 (R2) for contrast enhance-ment (CE), technical details(TD), for tears of scapholunate(SLL) and lunotriquetral (LTL)ligament, and triangular fibro-cartilaginous complex (TFCC)and for cartilage for scaphoidand lunate with correspondingradial surface. R1.1 and R2.1denote the first and secondreader’s first reading, and R1.2and R2.2 the second readings,respectively. The 95 % confi-dence intervals are shown inparentheses

Intraobserver Interobserver

CE R1 0.54 (0.08–1.00) R1.1-R2.1 0.73 (0.29–1.00)

R2 0.75 (0.46–1.00) R1.2-R2.2 0.45 (0.07–0.83)

TD R1 0.53 (0.30–0.88) R1.1-R2.1 0.56 (0.22–0.91)

R2 0.86 (0.60–1.00) R1.2-R2.2 0.67 (0.37–0.98)

SLL R1 0.59 (0.25–0.93) R1.1-R2.1 0.31 (0.06–0.56)

R2 0.66 (0.42–0.91) R1.2-R2.2 0.49 (0.26–0.73)

LTL R1 0.83 (0.66–1.00) R1.1-R2.1 0.90 (0.79–1.00)

R2 0.68 (0.46–0.91) R1.2-R2.2 0.48 (0.22–0.74)

TFCC R1 0.86 (0.72–1.00) R1.1-R2.1 0.65 (0.43–0.87)

R2 0.91 (0.79–1.00) R1.2-R2.2 0.59 (0.35–0.83)

RADIUS (scaphoid fossa) R1 0.45 (0.12–0.77) R1.1-R2.1 0.58 (0.19–0.96)

R2 0.64 (0.31–0.96) R1.2-R2.2 0.38 (0.09–0.66)

SCAPHOID R1 0.43 (0.12–0.74) R1.1-R2.1 0.37 (0.00–0.75)

R2 0.76 (0.55–0.96) R1.2-R2.2 0.32 (0.04–0.59)

RADIUS (lunate fossa) R1 0.68 (0.36–1.00) R1.1-R2.1 0.62 (0.29–0.96)

R2 0.42 (0.00–0.86) R1.2-R2.2 0.51 (0.12–0.91)

LUNATE R1 0.53 (0.16–0.90) R1.1-R2.1 0.59 (0.29–0.88)

R2 0.68 (0.44–0.91) R1.2-R2.2 0.42 (0.00–0.84)

Skeletal Radiol

and 8 mA were used, giving a DAP of 716 mGycm2. Thesmall size enabled the scanner to be installed in the generalX-ray room.

After CBCT, the patients were transferred to the MRIsuite within 15 min post-injection. For MRI, we used a1.5-T scanner (Signa HD, GE Medical Systems, Mil-waukee, WI, USA) using the following imaging param-eters: coronal T1 SE (TR/TE 620/13 ms, slice thickness2.5 mm/2.7 space, matrix 256 × 192), axial T2 FSE fat-saturated MR images (TR/TE eff. 2800/44 ms, slicethickness 3 mm/3.3 mm space, ETL 8, matrix320 × 192), and coronal T2-w fat-saturated FSE (TR/TE eff. 2760/68 ms, ETL 10, slice thickness 2.5 mm/2.7space, matrix 288 × 192). FOV was 10 cm in eachsequence.

The images were transferred to a 27″ iMac computer(Mac OS 10.6.4, Cupertino, CA, USA) and analyzed usingAycan OsirixPro (v. 1.3, English Edition, January 2010,Aycan Digital Systems GmbH, Würzburg, Germany) soft-ware. Before the final analysis, the studies were anonymizedand assigned a random ID number generated using freewareobtained at http://www.random.org/.

A month later, two radiologists with more than5 years of experience in musculoskeletal (MSK) traumaimaging using CT and MRI reviewed the CBCT imagestwice with a minimum 2-week interval between evalua-tions, for contrast enhancement (CE; 10good, 20fair;30poor) and technical details (TD; 10good, 20motion,30suboptimal contrast injection). Also, evidence of tearsof scapholunate (SLL) and lunotriquetral (LTL) liga-ments (10no tear, 20completely torn, 30partial tear)and triangular fibrocartilaginous complex (TFCC) (10no tear, 20 torn) and cartilage abnormalities (10normal,20 thinning, 30exposed subchondral bone) for scaphoidand lunate with corresponding radial surface (scaphoidand lunate fossa) were evaluated.

For MRI analysis, we used standard clinical workstations(Agfa DS3000, IMPAX 5.3, Agfa-Gaevert, Mortsel, Bel-gium) with 2-megapixel monitors (Barco Inc., Kortrijk,Belgium).

Inter- and intraobserver agreement was determinedusing weighted kappa statistics. In this study, it wasdefined that kappa-values 0.01–0.20 mean slight agree-ment, 0.21–0.40 fair agreement, 0.41–0.60 moderateagreement, 0.61–0.80 substantial agreement, and 0.80–0.99 almost perfect agreement [16]. Since no surgerywas performed, the consensus MRI reading served as areference standard, and sensitivity (SE), specificity (SP),accuracy (ACC), and positive and negative predictivevalues (PPV, NPV) were calculated.

Statistical analyses were done using a commercial soft-ware package SAS/STAT v.9.2 (SAS Institute Inc., Cary,NC, USA).

Results

Each CBCT imaging study was technically successful. Intwo patients, the contrast was suboptimal; in one case mostlikely due to previous fracture fixation, in the other case dueto technical difficulties during injection. Also, in twopatients with radial volar titanium plate and screws, themetal artifact caused less diagnostic problems in the assess-ment of both the wrist cartilage and ligament structures inCBCT images compared to the MRI.

Interobserver agreement for SL ligament was fair. Mod-erate agreement was seen on articular cartilage for radius,scaphoid, and lunate, and substantial agreement for LT

Fig. 3 Coronal (a), sagittal (b), and axial (c) images of the wrist. Voxelsize 0.4 × 0.4 × 0.4 mm3. Note the excellent visualization of cartilagesurface in radial and lunate cartilage (arrow, b)

Skeletal Radiol

ligament and TFCC. Intraobserver agreement for TFCC wasalmost perfect (Table 1).

Image quality is demonstrated in Fig. 3We found 11 SLL tears (ten total, one partial), seven LTL

tears (five total, two partial), and 25 TFCC degenerations ortears according to Palmer classification (type 1A; 16, 1B; 2,2A; 2, 2B 3; 2C; 2 and 2D 2) (Figs. 4, 5, and 6). For cartilageabnormalities, there were four cartilage abnormalities (thin-ning or subchondral bone exposure) for radial surface facingscaphoid (scaphoid fossa) and in scaphoid, six in lunate, andfive in corresponding radial surface (lunate fossa).

The mean values for SE, SP, ACC, PPV, and NPV forSLL tears were 56, 91, 83, 67, and 89 %; for LTL tears 83,

81, 82, 44, and 96 %; for TFCC tears 76, 90, 87, 83, and87 %, respectively. The results for each reader are presentedin Table 2.

The mean values for SE, SP, ACC, PPV, and NPV forcartilage abnormalities for radial surface facing were 69, 94,92, 53, and 97 %; for scaphoid 71, 94, 92, 61, and 97 %; forlunate fossa 63, 92, 90, 28, and 98 %; and for lunate 70, 92,90, 52, and 97 %, respectively. The results for each readerare presented in Table 3.

Discussion

CT arthrography is an alternative in patients when MRIis contraindicated or when MRI is not available.

Fig. 5 TFC 1A (arrowhead) and LTL (arrow) tear. a Coronal CT-image (voxel size 0.4 × 0.4 × 0.4 mm3) with b corresponding T1-w fat-saturated SE MR image (TR/TE 380/13 ms, FOV10cm, slice thickness2.5 mm/2.7 space, matrix 256 × 192). SLL is intact (block arrow)

Fig. 4 TFC 1A tear (arrow). a Coronal CT-image (pixel size0.4 × 0.4 × 0.4 mm3) with b corresponding T2-w fat-saturated FSEMR image (TR/TE eff. 2,760/68 ms, ETL 10, FOV 10 cm, 3 mm/3.2space, matrix 288 × 192). Arrowhead in a and b denotes contrast indistal radioulnar joint indicating communication between radiocarpaland distal radioulnar joint compartments

Skeletal Radiol

Moreover, multi-detector CT (MDCT) arthrography hasbeen reported to be more accurate than MRI or MRarthrography, especially in detecting partial tears of SLand LT ligaments [6]. In addition, experimental flat-panel C-arm CT arthrography has recently been reported[17]. In the current study, we used a novel, dedicatedambulatory cone-beam CT. Each study was technicallysuccessful, with overall accuracy of 82–86 %. Thismethod offers exciting possibilities for orthopedic imag-ing in general. High spatial resolution, low radiation

dose, and easy installation provide a potential one-stopshop for various orthopedic problems, such as subtlefracture detection and post-traumatic evaluation of frac-ture consolidation, especially when osteosynthetic mate-rial has been used [12]. Moreover, this technique allowsimaging in a comfortable sitting position or even inpatients lying on a hospital bed.

The spatial resolution used (0.4 mm3) is superior tostandard MR imaging where slice thickness is usually2–3 mm and in-plane resolution 0.5–0.7 mm although

Fig. 6 SL-tear (arrow) in a 63-year-old male. a Coronal CT-image, b blue line shows theplane to create the axial plane inc. Corresponding coronal fat-saturated T1 SE (TR/TE 620/13 ms, FOV 10 cm, slice thick-ness 2.5 mm/2.7 space, FOV10 cm, matrix 256×192 d) and eaxial T2 FSE fat-saturated MRimages (TR/TE eff. 2800/44 ms, slice thickness 3 mm/3.3 mm space, ETL 8, FOV10 cm, matrix 320 × 192).Contrast leakage is seen tomidcarpal joint (asterisk, a). Noperforation in TFC (arrowhead;a, d) is seen. Note also the ex-cellent visualization of cartilage(block arrow; a), whereas thesoft-tissue contrast (c) is sub-optimal compared to MRI (e).Curved arrow in c, e indicatesthe median nerve

Skeletal Radiol

sub-millimeter isotropic voxels can also be obtainedwith 3.0-T scanners with gradient echo sequences, suchas VIBE (volume interpolated breath-hold examination[18]). However, compared to MRI, the imaging timeusing CBCT is very short (18-s scan time) and patientpositioning is easy, making it less prone to motionartifacts. Moreover, wrist CBCT arthrography could bean alternative in postoperative patients with radial volarplate fixation due to relative absence of metallic arti-facts compared to MRI. The spatial resolution of thededicated extremity CBCT is similar to MDCT scanners.However, due to limitations in low-contrast sensitivityof CBCT due to, e.g., scatter, soft-tissue contrast islower compared to MDCT data. General challenges inCBCT contrast detectability, including scatter, beamhardening, truncation, and limited number of projec-tions, can be partially corrected with calculation meth-ods as reported for C-arm applications [19]. In the caseof wrist imaging with FOV of 13 × 16 cm, the trunca-tion effect is not a concern with the dedicated extremityCBCT.

According to dentomaxillofacial studies, the radiationdoses of CBTC examinations are significantly lowercompared to conventional MDCT scans [20–23]. The

CBCT technique has also been previously used to detectfinger fractures with equal accuracy as MDCT but withsignificantly less radiation [10]. Also, three cases ofCBCT imaging of the wrist, including one with intra-articular contrast, has been reported [11]. The scannersused, however, represent a different design and technol-ogy than that used in the current study. Due to thetarget size and geometry of the X-ray beam in thescanner, conventional computed tomography dose index(CTDI) measures are not applicable to estimate theradiation exposure to the patient. Dose area product(DAP) also has limitations because the radiation beamarea is wider than the wrist area. For typical exposurevalues used in the study (88-kV tube voltage, 8-mAtube current), DAP of 716 mGy*cm2 was measured.For the wrist imaging, a conversion coefficient of0.01 mSv/Gycm2 [22] was used, providing an effectivedose estimate of 7 μSv.

Table 2 Sensitivity (SENS), specificity (SPES), accuracy (ACC),positive predictive value (PPV), and negative predictive value (NPV)for scapholunate (SL), lunotriquetral (LT) ligament, and triangularfibrocartilaginous complex (TFCC) tears for readers 1 and 2. R1.1and R2.1 denote first and second reader’s first reading, and R1.2 andR2.2 the second readings, respectively

R1.1 R1.2 R2.1 R2.2

SL

SENS (%) 36 64 64 67

SPES (%) 95 98 88 85

ACC (%) 81 90 82 80

PPV (%) 67 88 58 57

NPV (%) 84 91 90 89

LT

SENS (%) 100 75 86 71

SPES (%) 86 85 81 74

ACC (%) 88 84 82 73

PPV (%) 54 50 43 31

NPV (%) 100 95 97 94

TFCC

SENS (%) 74 71 83 80

SPES (%) 87 91 90 87

ACC (%) 82 84 88 88

PPV (%) 78 80 83 80

NPV (%) 84 85 90 93

Table 3 Sensitivity (SENS), specificity (SPES), accuracy (ACC),positive predictive value (PPV), and negative predictive value (NPV)for cartilage abnormalities for scaphoid and lunate with correspondingradial surface for readers 1 and 2. R1.1 and R2.1 denote first andsecond reader’s first reading, and R1.2 and R2.2 the second readings,respectively

R1.1 R1.2 R2.1 R2.2

RADIUS (scaphoid fossa)

SENS (%) 50 50 75 100

SPES (%) 94 91 94 98

ACC (%) 90 88 92 98

PPV (%) 40 40 50 80

NPV (%) 96 96 98 100

SCAPHOID

SENS (%) 40 43 100 100

SPES (%) 89 91 96 100

ACC (%) 84 88 96 100

PPV (%) 29 43 71 100

NPV (%) 93 95 100 100

RADIUS (lunate fossa)

SENS (%) 100 50 67 33

SPES (%) 89 91 90 96

ACC (%) 90 90 88 92

PPV (%) 29 20 29 33

NPV (%) 100 98 98 96

LUNATE

SENS (%) 60 40 100 80

SPES (%) 95 93 85 96

ACC (%) 92 88 86 94

PPV (%) 60 40 42 67

NPV (%) 95 93 100 98

Skeletal Radiol

The inter- and intraobserver agreement was in mostcases substantial. The fair interobserver agreement in SLligament can be explained by the heterogeneity of ourmaterial with a substantial number with previous sur-gery and still-existing metal implants. In these cases, thedistorted anatomy and postoperative changes interferedwith the analysis even in retrospect. However, the studygroup was comprised of unselected material and isrepresentative of a typical busy orthopedic practice.Also, the differences between readers most likely reflectthe personal differences. In order to increase the speci-ficity, reader 1 has, in general, lower sensitivity, and ifthe number of pathologic conditions is small, e.g., fivein cartilage in lunate fossa (Table 3), one discordantreading may lead to substantial change in sensitivity.In addition, with an increasing prevalence, the valuesof the sensitivity and specificity of both observers haveto increase in order to maintain a certain kappa value[24]. Therefore, a low kappa value, combined with ahigh prevalence of one of the categories, cannot beinterpreted easily, and in such a case, the number ofobservations should be extended to make the prevalenceof the categories more equally distributed [24].

The visualization of cartilage surface was excellent(Fig. 3), and provides a potentially good method of detect-ing cartilage injuries in the radiocarpal as well as other smalljoints, where cartilage thickness is in the order of 1 mm orless. While MR arthrography is good for detecting cartilagedefects [25], arthroscopy is needed for a reference standard.It should be noted, however, that in case of orthopedichardware, the visualization of cartilage with CBCT wasbetter than with MR arthrography (Fig. 7)

In a previous study, CT arthrography was shown to have avery high (>90 %) sensitivity and specificity for SL, LT, andTFCC tears [6]. In our study, the corresponding numbers wereslightly smaller, probably reflecting differences in the studypopulation. They were also able to study partial tears, whereasin our study the amount of partial tears as well as the cartilageabnormalities was too small for a more detailed analysis.

The dedicated extremity CBCT scanner’s low radiationdose, easy installation, and easy use open new possibilitiesfor imaging of the radiocarpal joint. The use of CBCTarthrography and subsequent MR arthrography combinesthe inherent strengths of both methods; excellent visualiza-tion of bony structures and soft tissue contrast, respectively.Concomitant reading of both studies would potentially leadto increased diagnostic confidence and performance.

CT arthrography is recommended together with MRarthrography in order to increase the diagnostic accuracyof foveal tears and to assess the associated bone fragments[26].

In conclusion, the dedicated CBCT extremity scanneris a new method for evaluating wrist ligaments. Mod-erate agreement was seen on articular cartilage forradius, scaphoid and lunate, and substantial agreementfor LT ligament and TFCC. Interobserver agreement forSL ligament was fair. Intraobserver agreement forTFCC was almost perfect. The method has an accuracyof 82–86 % and specificity of 81–91 %. Sensitivity forLT and TFCC tears was 77–83 %, but for SLL tears itwas 58 %. For cartilage abnormalities, the accuracy andnegative predictive value were high, 90–98 %.

Fig. 7 In a patient with volar radial plate, ligament structures (SL-ligament; asterisk), wrist cartilage surface (especially in lunate, arrow-head) and a scaphoid cartilage defect (arrow) are better delineated onthe coronal CT-image due to fewer metal artifacts (a) compared tocorresponding T1-w fat-saturated SE MR image (b)

Skeletal Radiol

Acknowledgments Sharon J. Kuong, MD, is acknowledged for herlinguistic help.

Conflict of interest The authors acknowledge that within the past3 years they have received benefits which might result in a conflict ofinterest or the appearance of a conflict. Nature of benefit: ResearchConsultans (authors 1,3–5,7), employment (author 6).

References

1. Zlatkin MB, Chao PC, Osterman AL, Schnall MD, Dalinka MK,Kressel HY. Chronic wrist pain: evaluation with high-resolutionimaging. Radiology. 1989;173:723–9.

2. Smith DK. MR imaging of normal and injured wrist ligaments.Magn Reson Imaging Clin N Am. 1995;3:229–48.

3. Golimbu CN, Firooznia H, Melone CP, Rafii M, Weinreb J, LeberC. Tears of the triangular fibrocartilage of the wrist: MR imaging.Radiology. 1989;173:731–3.

4. Hobby JL, Tom BD, Bearcroft PW, Dixon AK. Magnetic reso-nance imaging of the wrist: diagnostic performance statistics. ClinRadiol. 2001;56:50–7.

5. Haims AH, Moore AE, Schweitzer ME, Morrison WB, Deely D,Culp RW, et al. MRI in the diagnosis of cartilage injury in thewrist. Am J Roentgenol. 2004;182:1267–70.

6. Moser T, Dosch JC, Moussaoui A, Dietemann JL. Wrist ligamenttears: evaluation of MRI and combined MDCT and MR arthrog-raphy. Am J Roentgenol. 2007;188:1278–86.

7. Cerezal L, de Dios Berná-Mestre J, Canga A, et al. MR and CTarthrography of thewrist. SeminMusculoskelet Radiol. 2012;16:27–41.

8. Mozzo P, Procacci C, Tacconi A, et al. A new volumetric CTmachine for dental imaging based on the cone-beam technique:preliminary results. Eur Radiol. 1998;8:1558–64.

9. Arai Y, Tammisalo E, Iwai K, Hashimoto K, Shinoda K. Devel-opment of compact computed tomographic apparatus for dentaluse. Dentomaxillofac Radiol. 1999;28:245–8.

10. De Vos W, Casselman J, Swennen GRJ. Cone-beam computerizedtomography (CBCT) imaging of the oral and maxillofacial region:a systematic review of the literature. Int J Oral Maxillofac Surg.2009;38:609–25.

11. Faccioli N, Foti G, Barillari M, Atzei A, Mucelli RP. Finger fracturesimaging: accuracy of cone-beam computed tomography and multi-slice computed tomography. Skeletal Radiol. 2010;39:1087–95.

12. De Cock J, Mermuys K, Goubau J, van Petegem S, Houthoofdt B,Casselman JW. Cone-beam computed tomography: a new low-

dose, high-resolution imaging technique of the wrist, presentationof three cases with technique. Skeletal Radiol. 2012;41:93–6.

13. Zbijewski W, De Jean P, Prakash P, et al. A dedicated cone-beamCT system for musculoskeletal extremities imaging: design, opti-mization, and initial performance characterization. Med Phys.2011;38:4700–13.

14. Mattila KT, J. Kankare J, Kortesniemi M, et al. Cone beam CT forextremity imaging. EPOS Abstract, ECR 2011, Vienna March 3–7,2011. doi:10.1594/ecr2011/C-0297.

15. Biswas D, Bible JE, Bohan M, Simpson AK, Whang PG, GrauerJN. Radiation exposure from musculoskeletal computerized tomo-graphic scans. J Bone Joint Surg Am. 2009;91:1882–9.

16. Landis JR, Koch GG. The measurement of observer agreement forcategorial data. Biometrics. 1997;33:159–74.

17. Guggenberger R, Fischer MA, Hodler J, Pfammatter T, AndreisekG. Flat-panel CT-arthrography. Feasibility study and comparisonto multidetector CT arthrography. Invest Radiol. 2012;47:312–8.

18. Zheng Z-Z, Shan H, Li X. Fat-suppressed 3D T1-weighted gradient-echo imaging of the cartilage with a volumetric interpolated breath-hold examination. Am J Roentgenol. 2010;194:W414–9.

19. Heigl B, Kowarschik M. High-Speed> Reconstruction for C-ArmComputed Tomography. In Proceedings of the 9th InternationalMeeting on Fully Three-Dimensional Image Reconstruction inRadiology and Nuclear Medicine, pp. 25–28, Lindau, Germany,July 2007.

20. Tsiklakis K, Donta C, Gavala S, Karayianni K, Kamenopoulou V,Hourdakis CJ. Dose reduction in maxillofacial imaging using low-dose cone-beam CT. Eur J Radiol. 2005;56:413–7.

21. Roberts JA, Drage NA, Davies J, Thomas DW. Effective dosefrom cone-beam CT examinations in dentistry. Br J Radiol.2009;82:35–40.

22. Koong B. Cone-beam imaging: is this the ultimate imaging mo-dality? Clin Oral Implants Res. 2010;21:1201–8.

23. Hart D, Wall BF. Radiation Exposure of the UK population fromMedical and Dental X-ray Examinations; Report NRPB-W4; Na-tional Radiological Protection Board, UK; 2002.

24. Donker DK, Hasman A, van Geijn HP. Interpretation of low kappavalues. Int J Biomed Comp. 1993;33:55–64.

25. Scheck RJ, Romagnolo A, Hierner R, Pfluger T, Wilhelm K, HahnK. The carpal ligaments in MR arthrography of the wrist: correla-tion with standard MRI and wrist arthroscopy. J Magn ResonImaging. 1999;9:468–74.

26. Omlor G, Jung M, Grieser T, Ludwig K. Depiction of the triangu-lar fibro-cartilage in patients with ulnar-sided wrist pain: compar-ison of direct multi-slice CT arthrography and direct MRarthrography. Eur Radiol. 2009;19:147–51.

Skeletal Radiol