| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 All authors: Department of Radiology, CHU Strasbourg, Ave. Molière, Strasbourg, France 67000.
Received April 2, 2006;
accepted after revision October 31, 2006.
Address correspondence to T. Moser
(thomas.moser{at}chru-strasbourg.fr).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. The feasibility of combining MDCT arthrography and MR arthrography and performing them with an optimized contrast solution was evaluated in vitro and in vivo. Forty-five consecutively enrolled subjects with suspected wrist ligament tears underwent MRI and a combined MDCT and MR arthrographic procedure. Two observers reviewed the images for evidence of tears and cartilage abnormalities of the scapholunate and lunotriquetral ligaments and triangular fibrocartilaginous complex. Interobserver agreement was determined with kappa statistics, and the diagnostic accuracy of each technique was calculated.
RESULTS. A 1:1 solution of 2.5 mmol/L
tetraazacyclododecanetetraacetic acid (DOTA)-gadolinium and 300 mg I/mL
iopamidol provided adequate contrast enhancement for both in vitro and in vivo
MDCT arthrographic and MR arthrographic images. Interobserver agreement was
substantial for MRI (
= 0.61) and MR arthrography ( = 0.71) and
almost perfect for MDCT arthrography ( = 0.93). The sensitivity and
specificity of MRI, MDCT arthrography, and MR arthrography for tears of the
scapholunate ligament were 59% and 70%, 95% and 96%, and 68% and 87% for the
first observer and 77% and 83%, 95% and 100%, and 77% and 87% for the second
observer. For tears of the lunotriquetral ligament, these values were 30% and
94%, 100% and 94%, and 60% and 97% for the first observer and 50% and 97%, 90%
and 100%, and 50% and 94% for the second observer. The three techniques
appeared equivalent for complete tears of the scapholunate and lunotriquetral
ligaments, but partial tears were significantly better visualized with MDCT
arthrography. The sensitivity and specificity of MRI, MDCT arthrography, and
MR arthrography for triangular fibrocartilaginous complex tears were 27% and
100%, 100% and 100%, and 82% and 100% for the first observer and 45% and 100%,
100% and 100%, and 82% and 100% for the second observer. For cartilage
abnormalities, these values were 30% and 100%, 100% and 100%, and 30% and 100%
for the first observer and 10% and 100%, 100% and 100%, and 40% and 100% for
the second observer.
CONCLUSION. MDCT arthrography appears more accurate than MRI and MR arthrography, particularly for discerning partial tears of the scapholunate and lunotriquetral ligaments that do not necessitate surgical therapy.
Keywords: arthrography • CT • joint • MRI • wrist
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Triple-compartment wrist arthrography was long considered a reference technique for the evaluation of wrist ligaments [4], but criticism regarding low specificity and the development of MRI progressively led to the abandonment of the technique [5]. Even with thin-section 3D sequences, however, MRI does not perform as well as arthroscopy does, and MRI appears quite limited in the diagnosis of partial tears and cartilage abnormalities [6]. MR arthrography has been found useful, however. In several studies [7, 8], the performance of MR arthrography was almost equivalent to that of arthroscopy in the diagnosis of torn ligaments and of cartilage abnormalities. CT arthrography is performed mostly in Europe and has been investigated in only a few studies [9].
|
|
|
|
This study had a twofold purpose. We first developed an imaging protocol combining MDCT arthrography and MR arthrography with an optimal contrast solution and evaluated the feasibility of the procedure. We then performed MRI, MDCT arthrography, and MR arthrography on a series of 45 patients with posttraumatic wrist pain and determined the diagnostic accuracy of the techniques in the detection of tears and cartilage abnormalities of the wrist ligaments.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In Vitro Study
We prepared 12 phantoms of 2-mL volume with varying proportions of 300 or
370 mg I/mL iopamidol (Iopamiron, Schering-Plough) and 2.5 mmol/L
tetraazacyclododecanetetraacetic acid (DOTA)-gadolinium (Artirem, Guerbet) or
500 mmol/L DOTA-gadolinium (Dotarem, Guerbet). Except for the 500 mmol/L
DOTA-gadolinium, all preparations have been approved in France for articular
injection.
Phantoms 1-5 were composed of 2.5 mmol/L DOTA-gadolinium and 300 mg I/mL iopamidol in the following respective proportions: phantom 1, 100%/0%; phantom 2, 75%/25%; phantom 3, 50%/50%; phantom 4, 25%/75%; phantom 5, 0%/100%. Phantoms 7-11 were composed of 2.5 mmol/L DOTA-gadolinium and 370 mg I/mL iopamidol in the following proportions: phantom 7, 100%/0%; phantom 8, 75%/25%; phantom 9, 50%/50%; phantom 10, 25%/75%; phantom 11, 0%/100%. Phantoms 6 and 12 were composed of 500 mmol/L DOTA-gadolinium and 300 and 370 mg I/mL iopamidol in a 0.5%/99.5% proportion. These two phantoms were prepared to obtain supposed optimal concentrations of DOTA-gadolinium (2.5 mmol/L) and iopamidol (300 or 370 mg I/mL) (Fig. 1A).
The MR arthrographic and MDCT arthrographic parameters described for the patient studies were used to image all of the phantoms at the same time. Image contrast was evaluated qualitatively by means of subjective visual assessment.
Patients
This study was approved by our institutional review board, and informed
consent was obtained from the patients and from the parents of patients who
were minors. Over a 9-month period, we prospectively examined 45 consecutively
enrolled patients. Patients were included by one of six hand surgeons when
wrist ligament tears were suspected on the basis of clinical examination.
Clinical diagnoses, including tears of the scapholunate and lunotriquetral
ligaments and of the TFCC, were recorded for correlation with imaging
findings. The study population contained 30 male patients and 15 female
patients 15-55 years old (mean age, 33.5 years). The right wrist was imaged in
21 (47%) of the cases and the left wrist in 24 (53%) of the cases. After
imaging, follow-up was conducted by hand surgeons for 4-12 months (mean, 7
months), and clinical conclusions and operative findings were recorded.
MRI
MRI was performed with a 1.5-T system (Magnetom Avanto, Siemens Medical
Solutions) with a dedicated wrist coil. Patients were placed in the supine
position with the wrist at the side in a neutral position. A coronal
T1-weighted spin-echo sequence was performed with TR/TE, 586/23; section
thickness, 2 mm without a gap; field of view, 88 x 100 mm; matrix size,
448 x 512; number of acquisitions, 1; acquisition time, 4 minutes 22
seconds. A coronal 3D double-echo steady-state gradient-echo sequence was
performed with selective water excitation; 32.24/7.8; flip angle, 30°;
section thickness, 0.6 mm; field of view, 103 x 150 mm; matrix size, 176
x 256; number of acquisitions, 1; acquisition time, 7 minutes 9 seconds.
A transverse fat-saturated proton density-weighted fast spin-echo sequence was
performed with 5,830/16; section thickness, 2 mm without a gap; field of view,
144 x 144 mm; matrix size, 256 x 320; number of acquisitions, 1;
acquisition time, 3 minutes 36 seconds.
Arthrography
After completion of the MRI study, patients were directed to the
fluoroscopy room, and the dorsal aspect of the wrist was prepared using an
aseptic technique. A 1:1 solution of 300 mg I/mL iopamidol and 2.5 mmol/L
DOTA-gadolinium was mixed. The midcarpal compartment was punctured, and the
solution was injected. If the distal radioulnar joint did not fill
spontaneously, an injection was administered only to this joint when a TFCC
tear was suspected (18 patients). In the absence of radiocarpal communication
after wrist mobilization, an injection was administered to the radiocarpal
compartment. An average total volume of 5 mL of contrast solution was
injected. Posteroanterior radiographs with neutral, radial, and ulnar
deviations and lateral and semisupinated views were obtained during and after
injection.
MDCT Arthrography
Patients were moved to the CT suite and positioned prone with the wrist
lying overhead in pronation. High-resolution acquisition was performed with a
16-MDCT scanner (Somaris Sensation, Siemens Medical Solutions) at 140 kV, 100
mAs, and 0.6 mm of collimation. The time elapsed from the beginning of
arthrography ranged from 5 to 29 minutes (mean, 15 minutes). Reconstructions
were performed in sections 0.6 mm thick with 50% overlap and a high-resolution
kernel, 12-cm field of view, and 512 x 512 matrix size. Transverse,
coronal, and sagittal reformations were obtained. The coronal plane is classic
for wrist imaging, but our experience has shown that transverse images are
essential for analyzing the dorsal and palmar portions of the scapholunate and
lunotriquetral ligaments. Sagittal images are considered of less value but
provided useful information about carpal alignment and the extent of TFCC
tears.
MR Arthrography
MR arthrography was performed with the system and in the position described
for MRI. The time elapsed from the beginning of arthrography ranged from 14 to
78 minutes (mean, 42 minutes). A coronal fat-saturated T1-weighted spin-echo
sequence was performed with 717/23; section thickness, 2 mm without a gap;
field of view, 88 x 100 mm; matrix size, 448 x 512; number of
acquisitions, 1; acquisition time, 5 minutes 21 seconds. A transverse 3D fast
low-angle shot gradient-echo sequence was performed with 32/4.7; flip angle,
60°; section thickness, 1 mm; field of view, 80 x 110 mm; matrix
size, 192 x 256; number of acquisitions, 1; acquisition time, 6 minutes
10 seconds.
Image Evaluation
The contrast enhancement of MDCT and MR arthrographic images was
qualitatively rated by one observer. The following three-point scale was used:
3, good (bright intraarticular contrast enhancement with excellent distinction
of cartilage and ligaments); 2, fair (mild intraarticular contrast enhancement
with adequate depiction of cartilage and ligaments); 1, poor (obscured
differentiation of cartilage and ligaments). From these results, procedures
were divided into three groups as follows: A, good contrast for both
techniques; B, fair contrast for one or both techniques; C, poor contrast for
one or both techniques.
Two observers with 2 and 25 years of experience in musculoskeletal radiology independently reviewed the MR, MDCT arthrographic, and MR arthrographic images for cartilage abnormalities and tears of the scapholunate and lunotriquetral ligaments and the TFCC. To avoid bias, radiologists were blinded to the subjects' sex, age, and initial clinical diagnosis. In addition, sets of MR, MDCT arthrographic, and MR arthrographic images were randomly and separately examined at 21-day intervals. Ligaments were classified as normal (smooth, homogeneous) or torn (intrinsic deposition of liquid or contrast medium). Complete tears of the scapholunate and lunotriquetral ligaments were defined as tears involving the three portions (dorsal, proximal, and palmar) of the ligament, whereas partial ruptures were characterized as tears involving one or two portions.
A small number of abnormalities, such as ligaments not seen and nonspecific modifications in signal intensity and shape, were impossible to classify as a normal finding or a tear. Such abnormalities can be found, for example, when cicatricial fibrous tissue replaces normal ligament. To fulfill the needs of statistical analysis, these abnormalities were classified as partial tears. TFCC tears were classified according to the method of Palmer [12]. For the sake of clarity in our analysis, we essentially differentiated the most common central and peripheral tears. Central tears (Palmer types IIC, IID, and IIE) involve the thinnest part of the TFCC and are mostly considered degenerative, whereas ulnar tears (Palmer type IB) involving ulnar attachment are always traumatic. Other tears are less frequent and were not detected in this study.
Cartilage abnormalities were localized and recorded as present or absent. We did not attempt to compare the extent and depth of the abnormalities visualized with the different techniques. Other pertinent imaging findings, including bone marrow edema, bone fractures, and tendon abnormalities, also were recorded.
Statistical Analysis
The outcomes considered for each observer and each technique were normal,
partially torn, and completely torn scapholunate ligament; normal, partially
torn, and completely torn lunotriquetral ligament; normal, torn on central
side, and torn on ulnar side TFCC; and normal or abnormal cartilage. Kappa
statistics were used to determine interobserver agreement for each technique.
Agreement was defined as almost perfect for a kappa value ranging from 0.81 to
1.00, substantial for a kappa value ranging from 0.61 to 0.80, moderate for a
kappa value ranging from 0.41 to 0.60, and fair for a kappa value ranging from
0.21 to 0.40. To account for sampling uncertainty, all values were calculated
with a 95% CI [13].
Diagnostic accuracy, including sensitivity, specificity, positive predictive value, and negative predictive value, was determined for each technique and each observer for the diagnosis of tears of the scapholunate and lunotriquetral ligaments, TFCC tears, and cartilage abnormalities. We also separately calculated the sensitivity of each technique in the diagnosis of partial and complete tears of the scapholunate and lunotriquetral ligaments and for central and ulnar tears of the TFCC. All percentages were calculated with a 95% CI.
Because our study was aimed at evaluating the performances of different imaging techniques for a wide range of wrist ligament injuries, including those not amenable to surgery, our reference was defined from the imaging, clinical follow-up, and operative findings when available. Results of the analysis performed by two principal observers were compared with findings on conventional arthrographic images. Concordant findings were accepted as the reference. Eight cases with discordant or equivocal findings were reviewed by a third observer who had 10 years of experience in musculoskeletal radiology. In addition, clinical follow-up and operative findings were included to minimize the risk of misinterpretation. Prolonged follow-up was used to document relief of symptoms, which is the rule for partial tears, whereas operative findings were used to confirm complete tears.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In Vivo Study
All subjects completed all three imaging examinations. None had adverse
effects or complications. Image contrast enhancement was adequate in most
cases: 23 (51%) of the procedures were classified group A; 16 (36%), group B;
and six (13%), group C. In only one case was there poor contrast enhancement
on both MR and MDCT arthrographic images. Group C procedures were consistently
associated with underlying synovitis or excessive delay after contrast
injection (generally more than 30 minutes for MDCT arthrography and more than
60 minutes for MR arthrography). Interobserver agreement was substantial for
MRI ( = 0.61; 95% CI, 0.54-0.68) and MR arthrography ( = 0.71;
95% CI, 0.64-0.78) and almost perfect for MDCT arthrography ( = 0.93;
95% CI, 0.89-0.97).
A total of 43 ligament tears and 10 cartilage abnormalities were identified, including 17 partial and five complete scapholunate tears, nine partial tears and one complete tear of the lunotriquetral ligament, and eight central and three ulnar TFCC tears. Surgery was performed in all six cases of complete ligament tear and in two cases of partial tear. Operative findings were consistent with our imaging reference in these cases. Other patients had clinical follow-up results that showed resolution of symptoms under conservative management. Diagnostic accuracy for depiction of tears of the scapholunate and lunotriquetral ligaments, TFCC tears, and cartilage abnormalities is presented in Tables 1, 2, 3 and 4. With MRI and MR arthrography, diagnostic accuracy was unequivocally lower for partial tears than for complete tears. Conversely, MDCT arthrography performed equally well in the diagnosis of partial and complete tears. The initial clinical diagnosis was confirmed in 49% of cases.
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In this study, we developed a procedure combining MDCT arthrography and MR arthrography after one injection of contrast solution. Optimization of this contrast solution by use of products approved for articular use in Europe was achieved through in vitro experiments. A 1:1 solution of 2.5 mmol/L DOTA-gadolinium and 300 mg I/mL iopamidol yielded acceptable contrast enhancement on both MDCT arthrographic and T1-weighted MR arthrographic images. We found that the theoretically optimal solution prepared with 500 mmol/L DOTA-gadolinium resulted in poor contrast enhancement on MR arthrography because of the T1 lengthening caused by iodinated contrast medium, as found by Montgomery et al. [17]. This effect increases with B0 magnetic field intensity [18]. Therefore, the use of iodinated contrast media at higher concentrations, such as 370 mg I/mL iopamidol, is not a good option, particularly with a 3-T MRI system, because such an agent can impair contrast enhancement. On the other hand, we found that MDCT arthrography does not require a high iodine concentration, which can cause beam-hardening artifacts [19]. Contrast enhancement was good with both arthrographic techniques in most patients. Individual analysis of rare cases of insufficient contrast enhancement revealed underlying synovitis or excessively delayed image acquisition, which is the rule for any arthrographic examination and should encourage consistent efforts to minimize delay.
The combined procedure not only is of interest for comparing MDCT arthrography and MR arthrography but also can be useful in other circumstances. MR arthrography of the wrist invariably requires injection of a small volume of iodinated contrast medium for confirmation of needle position in the joint and ideally to obtain an arthrogram. Because of the small capacity of the wrist joint, this volume of iodinated contrast medium often reaches the volume of gadolinium, resulting in dilution identical to that which occurred in our study. Under these conditions, MDCT arthrography requires a few additional minutes of scanning but can provide additional pertinent information not available with MR arthrography. Last, the combined procedure may be particularly beneficial when, for technical or patient-related reasons (e.g., claustrophobia), MR arthrography is stopped or the findings are insufficient for diagnosis.
|
|
|
|
|
|
|
|
|
|
|
Central perforation of the TFCC was far better depicted with MDCT arthrography (Fig. 3A, 3B, 3C, 3D, 3E) than with the other techniques. MDCT arthrography was more specific than MR arthrography in most cases. Cartilage abnormalities were more conspicuously depicted with MDCT arthrography, whereas MR arthrography was insufficiently sensitive (Fig. 4A, 4B, 4C, 4D). Cartilage abnormalities were infrequently associated with bone marrow edema, limiting interest in MRI and MR arthrography for this specific indication. Bone marrow edema also was absent in two cases of bone avulsion detected with MDCT arthrography alone. In a single case of occult scaphoid fracture, MDCT arthrography was less sensitive than MRI, disclosing extensive bone marrow edema. Finally, a major advantage of MDCT arthrography is that analysis of the images appears very straightforward, with almost perfect interobserver agreement, making this the technique preferred by surgeons at our institution.
|
|
|
|
The imaging findings confirmed the initial clinical diagnosis in 49% of the cases. This result differs from those of previous studies [24-27] of arthrography, emphasizing the poor correlation between location of pain and articular communication. Most communications observed in those studies were probably related to degenerative lesions. Inadequate recruitment of subjects also may account for such findings. In contradistinction, our study included a young and homogeneous population with predominantly traumatic lesions.
A limitation of our study was the lack of systematic surgical correlation. Such correlation is ideally obtained with arthroscopy, in which the articular compartments are assessed thoroughly. Use of this technique is not universally accepted, however, and differs greatly among hand surgeons [28]. Furthermore, it is not ethically acceptable to perform systematic arthroscopy, particularly when imaging findings are normal. Our prospective study had the advantage of involving a series of consecutively enrolled patients who had lesions of varying severity; thus the population reflected reality. A bias of retrospective studies based on arthroscopic reports is insufficient representation of wrists with normal or partially torn ligaments, leading to overestimation of the diagnostic accuracy of imaging techniques. Most of the patients in our series had partial ligament tears and therefore underwent conservative therapy. Multiple imaging techniques, analysis by different observers, and long follow-up were expected to reduce the risk of unrecognized significant injury.
In summary, we evaluated MRI, MDCT arthrography, and MR arthrography in the diagnosis of tears and cartilage abnormalities of the wrist ligaments in a series of patients. This evaluation was made possible by the use of an original procedure combining MDCT arthrography and MR arthrography after a single injection of an optimized contrast solution. We found this combined procedure feasible; others also may find is useful because it has the cumulative advantages of both techniques. For example, our procedure can be used in selected cases in which evaluation of both articular cartilage and bone marrow is needed. The imaging techniques performed equally well in the diagnosis of complete tears of the scapholunate and lunotriquetral ligaments. MDCT arthrography, however, proved superior in the diagnosis of partial tears of the scapholunate and lunotriquetral ligaments, TFCC tears, and cartilage abnormalities. Because of its accuracy and reliability, we propose that MDCT arthrography be the primary technique in the diagnosis of tears and cartilage abnormalities of wrist ligaments. MR arthrography currently does not perform as well as MDCT arthrography for these indications but may be advantageous for the study of extraarticular abnormalities. As a consequence, thorough assessment of both articular and nonarticular structures in the wrist may necessitate combining imaging techniques. For example, we routinely combine MDCT arthrography and sonography in a synergistic and cost-effective procedure (unreported personal experience).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. D. Crema, M. D. Marra, A. Guermazi, F. W. Roemer, K. Bohndorf, and N. Jomaah MDCT Arthrography Features of Ulnocarpal Impaction Syndrome Am. J. Roentgenol., November 1, 2009; 193(5): 1376 - 1381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Moser, E. Cardinal, and J.-C. Dosch Partial Versus Partial-Thickness Tears of the Scapholunate and Lunatotriquetral Ligaments Am. J. Roentgenol., November 1, 2009; 193(5): W469 - W469. [Full Text] [PDF]
|
|
|
|
|
|
T. Magee Comparison of 3-T MRI and Arthroscopy of Intrinsic Wrist Ligament and TFCC Tears Am. J. Roentgenol., January 1, 2009; 192(1): 80 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Moser, J.-C. Dosch, A. Moussaoui, X. Buy, A. Gangi, and J.-L. Dietemann Multidetector CT Arthrography of the Wrist Joint: How to Do It RadioGraphics, May 1, 2008; 28(3): 787 - 800. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Taljanovic, J. E. Sheppard, M. D. Jones, D. N. Switlick, T. B. Hunter, and L. F. Rogers Sonography and Sonoarthrography of the Scapholunate and Lunotriquetral Ligaments and Triangular Fibrocartilage Disk: Initial Experience and Correlation With Arthrography and Magnetic Resonance Arthrography J. Ultrasound Med., February 1, 2008; 27(2): 179 - 191. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
