| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Institute for Diagnostic Radiology, University Hospital Zurich, Raemistrasse
100, 8091 Zurich, Switzerland.
2 Department of Radiology, Orthopedic University Hospital Balgrist, Zurich,
Switzerland.
3 Guerbet AG, Zurich, Switzerland.
Received May 29, 2006;
accepted after revision August 31, 2006.
Address correspondence to D. Weishaupt
(dominik.weishaupt{at}usz.ch).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Our local ethics committees and the national drug administration approved this multicentric study. We prospectively studied 11 shoulders, 11 hips, and 10 wrists. After the intraarticular gadolinium injection, patients underwent a baseline MR arthrography protocol (time point [TP] 1) and subsequent MRI at another four time points (TP 2-TP 5) up to 240 minutes. The course of contrast-to-noise ratio (CNR) over time was calculated. Three observers assessed the degree of visualization of different intraarticular structures and the overall image quality at each time point using a 3-point scale and a 5-point scale, respectively.
RESULTS. For all joints, CNR measurements showed peak CNR at TP 1 (21 minutes) and TP 2 (45 minutes) with a subsequent, near-logarithmic decline of CNR values over time. Visualization of different anatomic structures decreased over time. Overall image quality was insufficient for diagnostic purposes at TP 3 (96 minutes) in three (27%) of 11 shoulders and in three (27%) of 11 hips. In two (20%) of 10 wrists, image quality was insufficient at TP 2 (45 minutes).
CONCLUSION. For MR arthrography, the degree of visualization of intraarticular structures depends on the time elapsed between contrast injection and MRI. MR arthrography of the shoulder and hip should be performed within 90 minutes, and MR arthrography of the wrist should be performed within 45 minutes, after intraarticular injection.
Keywords: arthrography • hip • MR • musculoskeletal imaging • shoulder • wrist
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Intraarticular injection of gadolinium-based contrast agents is considered to be a safe procedure [8]. The prevalence of joint infection rates after arthrography is reported to be 0.003% [9]. Although the technique of direct MR arthrography is increasingly used by radiologists, there is little in the literature regarding the temporal behavior of the injected gadolinium compound. Apart from experimental results in animals [10, 11], only a few studies have addressed the kinetics of the intraarticularly injected gadolinium [12-14]. To our knowledge, only one study has evaluated the time behavior of an intraarticularly injected gadolinium compound in humans. In the study by Wagner et al. [14], the temporal behavior of intraarticularly administered gadolinium was evaluated over time in four shoulders, four knees, and four hips of human volunteers after direct MR arthrography.
In this study we sought to assess the temporal behavior of an intraarticular gadoliniumbased contrast agent in the shoulder, hip, and wrist in symptomatic patients. MR arthrograms obtained at different time points after gadolinium injection were assessed not only with regard to quantitative parameters but also with regard to the degree of visibility of various anatomic structures at each time point.
The purpose of this study was to prospectively investigate the contrast dynamics over time and the relationship between the visualization of intraarticular structures and the time elapsed between the intraarticular injection of the contrast agent and MRI in symptomatic patients.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patients
During an 8-month period (February to September 2005), a total of 973
direct MR arthrographic examinations of the shoulder, 136 direct MR
arthrographic examinations of the wrist, and 343 direct MR arthrographic
examinations of the hip were performed at the two institutions involved in
this study.
Patients were included in the study only if there was a clinical indication for MR arthrography and if the patients were willing to participate in the study protocol, which consisted of the routine protocol of direct MR arthrography (baseline examination) and subsequent MR examinations at different time points (see following text). To homogenize the study population, only those patients with suspected internal derangements of the shoulder, hip, or wrist joint were included in whom synovial hypertrophy or synovitis was not expected. Patients with inflammatory rheumatologic disorders or with suspected septic arthritis were excluded.
From these 1,452 patients who were considered eligible for the study, we included consecutively 11 patients who were scheduled for direct MR arthrography of the shoulder, 11 patients who were scheduled for direct MR arthrography of the hip, and 10 patients who were scheduled for direct MR arthrography of the wrist (19 men, 13 women; mean age, 36 years; age range, 18-64 years).
MR Protocol
After intraarticular injection of the contrast agent under fluoroscopic
guidance, all patients underwent MRI according to the routine MR arthrographic
protocol for each joint. The MR protocol started with a fat-suppressed and a
non-fat-suppressed T1-weighted spin-echo sequence in all joints. After these
two sequences (time point [TP] 1, 20 minutes), the other sequences of the
individual MR arthrographic protocol were performed. After acquiring all the
sequences of the routine protocol, another pair consisting of a fat-suppressed
and a non-fat-suppressed T1-weighted spin-echo sequence was obtained (TP 2, 45
minutes). Another three pairs of fat-suppressed and a non-fat-suppressed
T1-weighted spin-echo sequences were repeated 90 minutes (TP 3), 180 minutes
(TP 4), and 240 minutes (TP 5) after the intraarticular injection of the
contrast agent. The sequence parameters are displayed in
Table 1. Sequences were
oriented in the transverse plane for the shoulder and in the coronal plane for
the hip and wrist. Sequence parameters were unchanged throughout each
consecutive pair of T1-weighted spin-echo sequences.
|
In addition to these sequences, our routine protocols for MR arthrography examination at baseline (between TP 1 and TP 2) of the shoulder, hip, and wrist included standard T2-weighted fast spin-echo sequences and STIR or proton density-weighted sequences in the sagittal, coronal, and transverse planes. In the hip, fast imaging sequences using steady-state precession (true fast imaging with steady-state free precession [FISP] or steady-state free precession [SSFP]) were also performed. However, not all these sequences were used for image analysis in our study.
MRI was performed using one of three 1.5-T scanners (Avanto or Symphony, Siemens Medical Solutions; or Signa Excite HD, GE Healthcare). For MR arthrography of the shoulder, a dedicated shoulder array coil was used. Similarly, for MR arthrography of the wrist, dedicated 4-channel high-resolution wrist coils were used. For MR arthrography of the hip, an 8-channel phased-array torso-pelvic coil or a flexible surface coil (total imaging matrix, TIM) was used.
Intraarticular Contrast Injection
Intraarticular injections were performed under fluoroscopic guidance using
aseptic technique. In each case, 1 mL of mepivacaine (Scandicain 2%,
AstraZeneca) was administered before the insertion of a 19-gauge needle (hip,
shoulder) or 23-gauge needle (wrist) into the joint space. The intraarticular
injection protocol, including the volume of the intraarticularly injected
contrast agent, was similar to those reported in the literature for shoulder
[13,
15,
16], hip
[16], and wrist
[1] MR arthrography.
For shoulder and hip joints, a maximum of 1 mL of an iodinated contrast agent (200 mg I/mL of iopamidol, Iopamiro 200, Bracco Diagnostics; or 200 mg I/mL of sodium and ioxaglate meglumine, Hexabrix 200, Guerbet) was administered for control of correct needle placement. Intraarticular injection was performed using an anterosuperior approach to the joint space with a total of 12 mL of gadoterate meglumine (0.0025 mmol/mL of gadolinium with an osmolality of 285 mosm/kg of H2O [Artirem], Guerbet) for the shoulder joint and 10 mL of gadoterate meglumine for the hip joint. In the wrist, a doublecompartment injection technique was used (distal radioulnar and midcarpal compartments) with a minimal amount of iodinated contrast agent. A maximum of 1 mL of gadoterate meglumine was injected into the distal radioulnar joint and 3-4 mL into the midcarpal compartment. The injection was stopped when fluoroscopic images showed sufficient distention of the individual joint compartment.
In all patients, care was taken not to introduce air into the joint, and any joint effusion was aspirated before instilling the contrast material. Epinephrine was not used to avoid any influence on the temporal behavior of the gadolinium-based contrast agent. All patients tolerated the procedure well with no adverse reactions or complications related to the intraarticular injection.
All patients were asked to avoid active motion of the joint between the intraarticular injection of the contrast material and baseline MRI and between baseline MRI and the subsequent MR examinations at different time points.
Image Analysis
Quantitative image analysis—Signal intensity was assessed by
measuring regions of interest (ROIs) in the joint space (intraarticular fluid)
and in the adjacent muscle tissue (shoulder, deltoid muscle; hip, iliopsoas
muscle; wrist, pronator quadratus muscle) at each time point on both
non-fat-suppressed and fat-suppressed images (image sets). Signal intensity of
the intraarticular fluid was measured in the axillary recess of the shoulder
adjacent to the articular surface of the humeral head
[14], in the recess of the hip
[14], and in the recess of the
distal radioulnar joint or adjacent to the radial styloid in the radiocarpal
joint space of the wrist [17].
ROIs were placed at the identical locations on each image set. Moreover, to
avoid volume artifacts (e.g., in late image sets), placement of the ROIs was
repeated three times, and calculation of contrast-to-noise ratio (CNR) was
averaged on the basis of these measurements of signal intensity. The circular
ROI measured 5-10 mm2 for the shoulder, 3-8 mm2 for the
hip, and 1-5 mm2 for the wrist. SD of the background noise was
computed for all sequences. For comparison purposes, CNR of the intraarticular
signal was calculated using the following equation
[13]:
 |
Qualitative image analysis—Three experienced musculoskeletal radiologists, blinded to all clinical and radiologic data, evaluated all MR images in random order, first independently and then in consensus. Randomization was performed with regard to patients and the different time points and with regard to fat-suppressed versus non-fat-suppressed MR images. Three separate interpretation sessions for shoulder, hip, and wrist MR arthrograms were performed. MR images were evaluated on a dedicated PACS workstation using commercially available PACS software (I-SoftView, version 5.2, Cedara Software).
All image sets (i.e., non-fat-suppressed and fatsuppressed images of each joint) were analyzed with regard to three visual aspects: joint distention, sharpness of joint borders, and contrast between intraarticular fluid and surrounding structures. Each of these aspects was scored as poor, fair, or good. The exact definitions of the scores and their adaptation for each joint are presented in Table 2. Presence and cause of artifacts (e.g., air, motion, inhomogeneous fat suppression, and infolding or susceptibility artifacts) were noted.
|
In addition, overall image quality was assessed. Overall image quality
reflected a general impression of each image set with regard to the use for
diagnostic purposes. A 5-point scale (1-5) was used. A score of 5 meant
excellent image quality (perfect contrast between relevant structures,
well-delineated borders, no artifacts), and a score of 1 meant inferior image
quality (insufficient contrast between the joint and surrounding structures,
severe blurring, or severe artifacts)
[7]. A cutoff value of 3 was
defined for the differentiation between sufficient and insufficient quality.
We assumed that at least 90% of all image sets should be of sufficient quality
(score value 
3) to render a specific time point clinically useful.
Statistical Analysis
Before this study, sample size estimation was performed. A sample size of
32 patients was considered to be comparable to other pharmacokinetic studies
[18-20]
and to other studies with regard to the temporal behavior of intraarticularly
administered gadolinium-based contrast agents
[14]. Data are presented in a
descriptive fashion by reporting the original raw data in absolute and
relative numbers and by using graphical and pictorial formats. CNR values at
each time point are summarized by using clustered box plots to show the
structure of the original raw data (e.g., the median, quartiles, and extreme
values) and the decline of the CNR values over time. Data regarding overall
image quality in the shoulder, hip, and wrist are presented as means of the
overall image quality at each time point. For all computations, SPSS software
was used (release 12.0.1, SPSS).
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Quantitative Analysis
Figure 1A,
1B shows the CNR values
obtained in the shoulder, hip, and wrist at each time point on T1-weighted
spin-echo images with and without fat suppression. Box plots showed similar
trends with decreasing CNR over time for each joint. The peak median CNR was
observed at TP 1 in most joints and image sets, followed by a near-logarithmic
decline of CNR values during follow-up. Minimum CNR values showed a trend
toward 0 over time (no contrast could be measured between the joint fluid and
the surrounding muscle tissue). Absolute CNR values were slightly higher with
fat suppression when compared with images sets without fat suppression.
Qualitative Image Analysis
Table 3 gives an overview of
the frequency distribution of the different quality ratings. There were a
decreasing number of good or fair ratings over time for each of the visual
aspects (joint distention, sharpness of anatomic structures, and contrast
between two adjacent anatomic structures).
|
Artifacts were observed in 24 (7.5%) of 320 image sets. Minor motion artifacts were present in 10 image sets, infolding and ringing artifacts occurred in five image sets, and small air bubbles in the joint fluid were detected in nine image sets. None of these 24 image sets was considered to be of insufficient image quality due to the observed artifact.
Figure 2A, 2B, 2C shows the mean overall image quality for direct MR arthrography. Both image sets (with and without fat suppression) showed similar trends, with a decrease of the overall image quality over time. Spin-echo imaging with fat suppression was consistently superior to spinecho imaging without fat suppression (Fig. 3A, 3B).
|
|
|
|
|
3 for both the
fatsuppressed and the non-fat-suppressed sequences. This percentage decreased
over time (at TP 3, 8/10 [80%] fat-suppressed sequences and 4/10 [40%]
non-fat-suppressed sequences; at TP 4, 4/8 [50%] fat-suppressed sequences and
1/8 [13%] non-fat-suppressed sequences; and at TP 5, 3/7 [43%] fat-suppressed
sequences and 0/7 [0%] non-fat-suppressed sequences). In the hip, overall
image quality was rated 3 in 10/11 (90%) MR arthrographies at TP 1 for
both fat-suppressed and non-fat-suppressed sequences (at TP 2, 9/10 [90%,
fat-suppressed] and 9/10 [90%, non-fat-suppressed]; at TP 3, 7/10 [70%] and
6/10 [60%]; at TP 4, 5/10 [50%] and 4/10 [40%]; and at TP 5, 1/9 [11%] and 1/9
[11%]) (Fig. 4A,
4B,
4C,
4D,
4E). Corresponding results for
image quality of the wrist were as follows: TP 1, 9/10 (90%, fat suppressed)
and 8/10 (80%, non-fat-suppressed); TP 2, 8/10 (80%) and 5/10 (50%); TP 3,
8/10 (80%) and 4/10 (40%); TP 4, 7/10 (70%) and 1/10 (10%); and TP 5, 3/10
(30%) and 1/10 (10%).
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
From a practical point of view, knowledge of temporal behavior is important because the amount of fluid in the joint and the signal characteristics of the injected contrast agent determine the quality of an MR arthrogram.
Kopka et al. [13] investigated the relationship between MRI and the time point of the contrast injection. In a prospective study, Wagner et al. [14] performed MR examinations of four shoulder, hip, and knee joints in healthy volunteers at various time points after intraarticular injection of gadopentetate dimeglumine using standard doses. CNR measurements were performed. The maximum CNR decrease (53%) occurred 1 hour after intraarticular gadolinium injections in the shoulder, whereas the maximum CNR decrease occurred 2 hours after injection in the hip joint. In the knee joint, the maximum CNR decrease (86%) was found between 3.5 and 6.25 hours.
Using spectrometric methods, Hajek et al. [11] measured residual quantities of gadolinium in the cartilage and synovium after MR arthrography with gadopentetate dimeglumine in an animal model. It was shown that gadopentetate dimeglumine is resorbed by the synovium within a few hours. Using histologic sections of human joint cartilage, Engel [28] showed that, 14 hours after the intraarticular administration of gadolinium, no measurable traces of gadopentetate dimeglumine could be established in the specimen.
In our study, we showed a near-logarithmic decrease of the CNR within the first 45-90 minutes for all evaluated joints. The decline of the CNR is reflected by a visual decrease in the image quality. The image quality for diagnostic purpose becomes critical 45 minutes after contrast injection for the wrist joint and after 90 minutes for the shoulder and hip joints. With regard to our statement that at least 90% of all image sets at a specific time point should be sufficient for diagnostic purposes, MRI of the wrist joint should be performed within 45 minutes, and MRI of the shoulder and hip joints should be performed within 90 minutes, after intraarticular contrast injection.
To our knowledge, the influence of using a fat-suppression technique on the image quality at different time points has not been systematically evaluated for MR arthrography. At each time point, we obtained a pair of a fat-suppressed and a non-fat-suppressed T1-weighted spin-echo sequences. We noted similar results for the different quality ratings and for the temporal behavior of CNR for fat-suppressed and for non-fat-suppressed MRI. However, fat-suppressed spin-echo sequences are more sustainable in longer delays between the injection and the MRI than are non-fatsuppressed spin-echo sequences.
Our study contributes to the optimization of patient scheduling for direct MR arthrography. Because patients usually must be transferred from the fluoroscopy suite to the MR imager after the intraarticular injection, image quality of MR arthrograms may be insufficient if the time delay between the injection and T1-weighted MRI is too long.
We acknowledge several limitations of our study. One limitation may be that we did not include patients with inflammatory rheumatologic disorders. Because of the expected synovial hyperperfusion in patients with inflammatory rheumatology disorders, the diffusion of the intraarticularly administrated gadolinium chelate in MR arthrography may even be faster, resulting in a shorter time window during which MRI should be performed. On the other hand, direct MR arthrography is not typically performed in inflammatory disease. IV contrast injection may be superior by showing the hypertrophic synovial membrane.
Our study could also be criticized in that we evaluated only the temporal behavior of the signal characteristics on non-fat-suppressed or fat-suppressed spin-echo sequences. However, conventional T1-weighted spin-echo or fast spin-echo sequences are still considered the most useful sequences for MR arthrography, particularly when the T1 effect of the injected contrast agent is taken into consideration [1, 16]. Other studies have suggested the use of steady-state free precision sequences and other gradient-echo sequences for imaging internal joint disorders [29, 30]. Finally, we did not rate the value of the individual image sets obtained at different time points for potential differences in making a final diagnosis. However, because at the different time points only one type of sequence—that is, spin-echo sequences—and only one imaging plane were available, such an imaging analysis was not possible.
In conclusion, for MR arthrography, visualization of intraarticular anatomic structures depends on the time elapsed between the contrast injection and MRI. For MR arthrography of the shoulder and hip, MRI should be performed within 1.5 hours after contrast injection, and MR arthrography of the wrist should be performed within 45 minutes.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Saupe, M. Zanetti, C. W. A. Pfirrmann, T. Wels, C. Schwenke, and J. Hodler Pain and Other Side Effects after MR Arthrography: Prospective Evaluation in 1085 Patients Radiology, March 1, 2009; 250(3): 830 - 838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Andreisek, J. M. Froehlich, J. Hodler, D. Weishaupt, V. Beutler, C. W. A. Pfirrmann, C. Boesch, and D. Nanz Direct MR Arthrography at 1.5 and 3.0 T: Signal Dependence on Gadolinium and Iodine Concentrations--Phantom Study Radiology, June 1, 2008; 247(3): 706 - 716. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
