Original Article pISSN 1738-2637 / eISSN 2288-2928 J Korean Soc Radiol 2016;75(6):480-486 https://doi.org/10.3348/jksr.2016.75.6.480
Discrepancy between Fluoroscopic Arthrography and Magnetic Resonance Arthrography in Patients with Arthroscopically Confirmed Supraspinatus Tendon Tears: The Additional Benefit of Cine Fluoroscopic Arthrography Images 관절경으로 확인된 극상건 파열 환자들에게 있어서 투시 관절 조영술과 자기공명 관절 조영술의 불일치: 영화 투시 관절 조영술의 추가적 이점 Seok Hahn, MD1,2, Young Han Lee, MD1, Jin-Suck Suh, MD1* Department of Radiology, Research Institute of Radiological Science, Medical Convergence Research Institute, and Severance Biomedical Science Institute, Yonsei University College of Medicine, Seoul, Korea 2 Department of Radiology, Inje University College of Medicine, Haeundae Paik Hospital, Busan, Korea 1
Purpose: To determine the additional diagnostic benefits of fluoroscopic arthrography (FA) in patients with full-thickness supraspinatus tendon (SST) tears by comparing FA images with magnetic resonance arthrography (MRA) images. Materials and Methods: This study included FA and MRA images of 53 patients who were confirmed to have full-thickness SST tears by arthroscopy. In the FA analysis, the presence of contrast leakage into the subacromial-subdeltoid bursa was recorded. In the MRA analysis, contrast leakage, retraction of a torn tendon, width and length of the tear, and supraspinatus atrophy were evaluated. Patients were divided into the concordant group or the discordant group based on the presence of contrast leakage to compare the characteristics of SST tears. We used Fisher’s exact test and two-sample t-test for the comparison. Results: Of the 53 patients, 34 were included in the concordant group and 19 were included in the discordant group. In the concordant group, the grades of retraction were higher than those in the discordant group; the width and length of the tears were larger. Muscle atrophy was more severe in the concordant group. Conclusion: A full-thickness SST tear did not always exhibit contrast leakage on FA, particularly small SST tears or tears with low-grade retraction. FA can provide diagnostic information regarding the severity of full-thickness SST tears by itself.
INTRODUCTION Magnetic resonance arthrography (MRA) of the shoulder is widely used despite the pain and discomfort associated with its
Index terms Shoulder Arthrography Magnetic Resonance Imaging Fluoroscopy Received December 15, 2015 Revised May 3, 2016 Accepted June 6, 2016 *Corresponding author: Jin-Suck Suh, MD Department of Radiology, Research Institute of Radiological Science, Medical Convergence Research Institute, and Severance Biomedical Science Institute, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 03722, Korea. Tel. 82-2-2228-7420 Fax. 82-2-393-3035 E-mail:
[email protected] This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Intra-articular injection into the glenohumeral joint is an important procedure performed prior to shoulder MRA (7). Currently, the use of digital radiology allows the entire procedure of fluoroscopic arthrography (FA) to be recorded as cine images.
use, because it provides high sensitivity and specificity for rota-
Arthrographic findings showing contrast leakage into the
tor cuff tears (1-4). It can also provide clear discrimination be-
subacromial-subdeltoid bursal space can be used to investigate
tween rotator cuff tears and small anatomical variations (5, 6).
the pathologic condition of supraspinatus tendon (SST) tears,
480
Copyrights © 2016 The Korean Society of Radiology
Seok Hahn, et al
such as a full-thickness tear. However, in some cases, unexpect-
tion, Otawara, Japan). The injection was performed with the pa-
ed discrepancies occur between arthrography and MRA, show-
tient supine in a straight anteroposterior position. The injection
ing no contrast leakage on arthrography but a full-thickness SST
consisted of approximately 20 cc of the following mixture: 2 cc
tear on MRA.
of ioxitalamic acid (Telebrix 30 Meglumine, Guerbet, Aulnay-
To the best of our knowledge, no study has examined the dif-
sous-Bois, France), 0.08 cc gadopentetate dimeglumine (Magn-
ferences between arthrography and MRA for shoulder patholo-
evist, Bayer Schering Pharma AG, Berlin, Germany), and 18 cc
gy. The purpose of this study is to determine if there are addi-
normal saline. Standard sterile management was applied in all
tional diagnostic benefits of FA in patients with arthroscopically
procedures. Approximately 16–18 cc of contrast mixture was in-
confirmed full-thickness SST tears by comparing FA images
jected into the glenohumeral joint and cine images were ob-
with MRA images.
tained during the injection period by FA. After arthrography, patients were escorted to the MRI room,
MATERIALS AND METHODS
and a shoulder MRA was performed within 20 minutes. One of
Patient Selection
used (Achieva or Achieva TX, Philips Healthcare, Best, the Neth-
the three 3T MRI systems with a dedicated shoulder coil were
Our retrospective study was approved by the hospital’s Insti-
erlands, n = 25; Discovery MR 750, GE Healthcare, Milwaukee,
tutional Review Board. From September 2012 to March 2013,
WI, USA, n = 8; and Trio, Siemens Healthcare, Erlangen, Ger-
we identified 150 patients who underwent arthroscopy for shoul-
many, n = 20). Conventional 2D images, including fat saturation
der pain and had an SST tear reported in their arthroscopic op-
T1-weighted axial images (repetition time/echo time, 690/20 ms
eration notes. Ninety-seven patients were excluded because of
in Philips, 570/7 ms in Siemens, and 580/10 ms in GE; slice
the following criteria: 1) no available preoperative MRA and FA
thickness/interslice gap, 3/0.3 mm; field of view, 140 × 140 mm),
data, including conventional magnetic resonance image (MRI)
oblique coronal and oblique sagittal images (repetition time/echo
without arthrography (n = 22), lack of FA (n = 14), and magnet-
time, 640/10 ms in Philips, 530/7 ms in Siemens, and 580/10 ms
ic resonance (MR) imaging performed in other hospitals (n =
in GE; slice thickness/interslice gap, 3/0.3 mm; field of view, 140
57), and 2) partial-thickness SST tear (n = 4). After exclusion,
× 140 mm), T2-weighted oblique coronal images (repetition
53 patients (age range, 35–76 years; mean age, 64.3 years) with
time/echo time, 3300/70 ms in Philips, 3500/80 ms in Siemens,
an arthroscopically confirmed full-thickness SST tear were en-
and 3300/70 ms in GE; slice thickness/interslice gap, 3/0.3 mm;
rolled in this study. Of them, seven patients had an articular-sid-
field of view, 140 × 140 mm), and T1-weighted oblique sagittal
ed partial thickness tear of the infraspinatus tendon (IST) and
images (repetition time/echo time, 690/8 ms in Philips, 620/8 ms
two patients had a full thickness tear of the IST, In their opera-
in Siemens, and 530/8 ms in GE; slice thickness/interslice gap,
tion records, there was no fibrous adhesion at the tear site or in
3/1 mm; field of view, 140 × 140 mm), were obtained.
the adjacent area. The mean interval between MR examination and arthroscopic surgery was 32.7 days (range, 27–36 days). Of
Imaging Interpretation
the patients enrolled, 21 were men (39.6%) and 32 were women
Two musculoskeletal radiologists (one radiologist with more
(60.4%), and 39 patients (73.6%) affected on the right side and
than 8 years of experience in musculoskeletal imaging and one
14 patients (26.4%) affected on the left side were evaluated.
radiologist with musculoskeletal radiology fellowship) independently reviewed both FA images and MRA images in a random
Fluoroscopic Arthrography and Magnetic Resonance
order within a two-week time span. They kept a two-week in-
Arthrography
terval between evaluation of each set of FA and MRA images.
All patients provided written informed consent for the pro-
They were blinded to radiologic reports, arthroscopic findings,
cedure. A 23-gauge needle was inserted into the glenohumeral
and any clinical information. In both FA and MRA analyses,
joint via an anterior approach under pulsed fluoroscopic guid-
presence of contrast leakage into the subacromial-subdeltoid
ance (Zexira DREX-ZX80, Toshiba Medical Systems Corpora-
bursa was checked. We measured the width of tears on fat satu-
jksronline.org
J Korean Soc Radiol 2016;75(6):480-486
481
Discrepancy between Fluoroscopic Arthrography and Magnetic Resonance Arthrography in Patients with Supraspinatus Tendon Tears
ration T1-weighted oblique coronal images, and the length of
General Electric Healthcare, Chicago, IL, USA).
tears on fat saturation T1-weighted oblique sagittal images. We categorized SST retraction into 3 grades using a published meth-
Statistical Analyses
od for determining the Patte score on T2-weighted oblique cor-
Fisher’s exact test was used to compare categorical variables,
onal images: little, humeral head level, and glenoid level (8). We
and two-sample t-test was used to compare continuous variables
then assessed supraspinatus muscle atrophy using the tangent
between the groups. SPSS software version 20.0 (IBM Corp.,
sign and occupation ratio. We considered that the tangent sign
Armonk, NY, USA) was used for statistical analyses. Findings
was absent when the superior margin of the supraspinatus mus-
were considered statistically significant when the p-value was less
cle was superior to the line tangential to the coracoid and scapu-
than 0.05.
lar spine (9, 10). Occupation ratio was measured according to the method described by Thomazeau et al. (11) and Khoury et al. (12), which was the ratio between the cross section of the supra-
RESULTS
spinatus muscle belly and that of its fossa on the T1-weighted
Of the 53 patients, 34 (64.2%) were included in the concor-
oblique sagittal image crossing through the medial border of the
dant group (Fig. 1), and the remaining 19 (35.8%) were includ-
coracoid process of the scapula. Lines were drawn as close as
ed in the discordant group (Fig. 2, Table 1). None of the patients
possible to the supraspinatus outer margin, inner margins of
showed leakage on FA only. The mean values of the width and
the coracoid process and scapular spine, and superior limits of
length of SST were significantly larger in Group 1 than in Group
the supraspinatus fossa.
2 (p = 0.01 and 0.02, respectively). With respect to the SST retrac-
We accepted consensus for the presence and grading and
tion grade, there was a significant difference between the two
considered the average of the two readers’ values for measure-
groups (p = 0.03). Both the tangent sign and occupation ratio
ments. Finally, we divided the patients into the following two
showed significant differences between the two groups (p < 0.01,
groups: Group 1, concordant group (leakage into the subacro-
both).
mial-subdeltoid bursa on both MRA and FA) and Group 2, dis-
All patients with tendon retraction at the glenoid level (8/53,
cordant group (leakage on MRA, but no leakage on FA). All
15.1%) belonged to Group 1. There were a total of 16 patients with
measurements were performed using a picture archiving and
confirmed massive SST tears. Of them, 15 patients who had mas-
communication system (PACS, Centricity Radiology RA 1000;
sive SST tears (15/53, 28.3%) belonged to Group 1. One patient
A B C Fig. 1. A 75-year-old male patient in Group 1 (concordant group). A. Contrast leakage into the subacromial-subdeltoid bursa is shown on a FA spot image (arrows). B, C. Oblique coronal and sagittal T1-weighted fat saturation MRA show a full-thickness SST tear (width: 23.6 mm, length: 37.6 mm) with retraction of the tendon (asterisk) at the level of the glenohumeral joint. FA = fluoroscopic arthrography, MRA = magnetic resonance arthrography, SST = supraspinatus tendon
482
J Korean Soc Radiol 2016;75(6):480-486
jksronline.org
Seok Hahn, et al
with a massive SST tear who belonged to Group 2 also had a
fluoroscopic guidance. In general, this intra-articular injection
full-thickness subscapularis tendon tear (Fig. 3).
could be regarded as a brief procedural step prior to MRA. However, more information can be obtained from the digitally stored cine FA images. In practice, we usually find concordant results
DISCUSSION
between FA and MRA, but discordant results were observed in
To evaluate SST tears using shoulder MRA, direct arthrogra-
some patients. We reviewed the literature and identified the dif-
phy is a necessary and important procedure for glenohumeral in-
ferences between FA and MRA, but we did not find any reports
jection. It can be uncomfortable for patients because of pain, time
specifically related to this subject. In this study, we tried to deter-
consumption, and radiation if the procedure is performed under
mine if there are any differences between FA and MRA findings
A B C Fig. 2. A 48-year-old female patient in Group 2 (discordant group). A. There is no contrast leakage into the subacromial-subdeltoid bursa on a FA spot image. B, C. Oblique coronal and sagittal T1-weighted fat saturation MRA show a full-thickness SST tear (width: 13.4 mm, length: 10.1 mm) with little retraction of the tendon (asterisk). FA = fluoroscopic arthrography, MRA = magnetic resonance arthrography, SST = supraspinatus tendon Table 1. Comparison between Group 1 and Group 2 Sex Male Female Age* Side Right Left Retraction grade Little Humeral head level Glenoid level Width (mm) of tear* Length (mm) of tear* Tangent sign Yes No Occupational ratio (%)* *Mean ± standard deviation.
jksronline.org
Group 1 (n = 34)
Group 2 (n = 19)
12 22 65.5 ± 7.9
9 10 63.5 ± 7.9
24 10
15 4
p -Value 0.558
0.375 0.748
0.03 1 25 8 24.5 ± 9.8 25.1 ± 9.4
4 15 0 16.4 ± 6.5 16.8 ± 7.3
21 13 43.8 ± 15.1
2 17 66.6 ± 15.8
J Korean Soc Radiol 2016;75(6):480-486
0.01 0.02 < 0.05
< 0.05
483
Discrepancy between Fluoroscopic Arthrography and Magnetic Resonance Arthrography in Patients with Supraspinatus Tendon Tears
A B Fig. 3. A 62-year-old female patient in Group 2 (discordant group). A. There is no contrast leakage into the subacromial-subdeltoid bursa on a FA spot image. B. An axial T1-weighted fat saturation MRA shows a subscapularis tear (arrow). A moderate amount of contrast leaked primarily into the extraarticular space via the tear site. FA = fluoroscopic arthrography, MRA = magnetic resonance arthrography
and to observe if there are additional benefits of FA images.
can cause a change in the pressure gradient between the joint and
We compared the FA and MRA findings between the con-
the bursa. The pressure change allows a greater flow of the con-
cordant and discordant groups with respect to confirmed full-
trast into the bursal space; therefore, leakage appears on MRA
thickness SST tears. As mentioned in the results, the concordant
images. Conversely, if the defect is large, the valve-like action is
group showed not only wider and longer SST tears but also high-
not sufficient to interrupt the flow. This mechanism can explain
er grades of SST retraction than the discordant group. Because
the difference between the concordant and discordant groups.
the dimensions of SST tears and retraction of SST reflect the ov-
The results of our study can be helpful in situations when only
erall defect size of SST tears, our results indicate that if SST tears
shoulder FA s available. Corticosteroid or nonsteroidal anti-in-
are large, the possibility of leakage on FA images is high. The
flammatory drug injection into the glenohumeral joint under flu-
concordant group also had more patients with a positive tangent
oroscopic guidance is one of these clinically useful instances. Flu-
sign and showed a lower occupation ratio than the discordant
oroscopic guidance is generally used for glenohumeral injection
group. A positive tangent sign and a low occupation ratio dem-
and it has several advantages such as a wider view of bony struc-
onstrate atrophy of the supraspinatus, which can be correlated
tures or confirmation of successful injection when a mixture with
with the chronicity of the tear. We think that patients in the con-
contrast media is used.
cordant group had more advanced supraspinatus atrophy and it would take longer for detection after the event of a SST tear.
There are many conditions that can cause shoulder pain such as osteoarthritis, adhesive capsulitis, rotator cuff disease, and la-
We can explain the difference between the concordant and
bral pathology (13). Intra-articular corticosteroid injection has
discordant groups by flap tears or a one-way check valve mecha-
shown a short- or medium-term therapeutic effect because cor-
nism. If the SST tear is small, the edge of the torn tendon or fi-
ticosteroids are powerful anti-inflammatory drugs (14). If con-
brosis that develops after the tear can act as a check valve that
trast leakage is found in the subacromial-subdeltoid bursa on
blocks the flow of injected contrast from the glenohumeral joint
FA images during injection, we can infer that the patient has a
cavity into the subacromial-subdeltoid bursa during FA. On the
full-thickness rotator cuff tear and it is relatively large. Moreover,
other hand, during preparation for MRA, the physical movement
we can inform the patient regarding his or her current shoulder
required (e.g., walking, raising the affected arm, or lying down)
status and recommend the appropriate treatment rather than
484
J Korean Soc Radiol 2016;75(6):480-486
jksronline.org
Seok Hahn, et al
corticosteroid injection. However, there are several limitations to this study. First, it
3. Elentuck D, Palmer WE. Direct magnetic resonance arthrography. Eur Radiol 2004;14:1956-1967
was a retrospective study, but all patients had full-thickness tears
4. Jacobson JA, Lin J, Jamadar DA, Hayes CW. Aids to suc-
confirmed by arthroscopic surgery. Second, we focused on only
cessful shoulder arthrography performed with a fluoro-
full-thickness SST tears because it is difficult to detect partial-
scopically guided anterior approach. Radiographics 2003;
thickness tears with FA images. Third, the injection into the gle-
23:373-378; discussion 379.
nohumeral joint consisted of only approximately 16–18 cc of con-
5. Hodler J, Kursunoglu-Brahme S, Snyder SJ, Cervilla V, Kar-
trast mixture, and the unique case (a case of a massive SST tear
zel RP, Schweitzer ME, et al. Rotator cuff disease: assess-
in the discordant group) in which leakage was not visible on FA
ment with MR arthrography versus standard MR imaging
images occurred due to this reason (Fig. 3). The injected contrast
in 36 patients with arthroscopic confirmation. Radiology
mixture leaked into another site of severe injury in this patient;
1992;182:431-436
hence, leakage into the subacromial-subdeltoid bursa did not oc-
6. Flannigan B, Kursunoglu-Brahme S, Snyder S, Karzel R, Del
cur. If more amount of the contrast mixture was injected into the
Pizzo W, Resnick D. MR arthrography of the shoulder:
joint, the leakage would have been visible on FA images. Fourth,
comparison with conventional MR imaging. AJR Am J
two patients had a full thickness tear of the IST in our study.
Roentgenol 1990;155:829-832
The situation that contrast leakage occurs into the subacromial-
7. Jbara M, Chen Q, Marten P, Morcos M, Beltran J. Shoulder
subdeltoid bursa via the IST tear site cannot be excluded. Final-
MR arthrography: how, why, when. Radiol Clin North Am
ly, a small number of variables were used to compare the con-
2005;43:683-692, viii
cordant and discordant groups. In conclusion, not all full-thickness SST tears showed con-
8. Patte D. Classification of rotator cuff lesions. Clin Orthop Relat Res 1990;(254):81-86
trast leakage on FA images, especially small SST tears or tears
9. Mellado JM, Calmet J, Olona M, Esteve C, Camins A, Pérez
with low-grade retraction. Because the concordance between FA
Del Palomar L, et al. Surgically repaired massive rotator
and MRA is more frequent in patients with larger SST tears and
cuff tears: MRI of tendon integrity, muscle fatty degener-
greater worsening of supraspinatus atrophy, FA can provide di-
ation, and muscle atrophy correlated with intraoperative
agnostic information to determine the severity of full-thickness
and clinical findings. AJR Am J Roentgenol 2005;184:
SST tears by itself and it has the potential to be used as another
1456-1463
imaging modality to evaluate these tears.
10. Warner JJ, Higgins L, Parsons IM 4th, Dowdy P. Diagnosis and treatment of anterosuperior rotator cuff tears. J Shoul-
Acknowledgments
der Elbow Surg 2001;10:37-46
This work was supported by a National Research Foundation
11. Thomazeau H, Rolland Y, Lucas C, Duval JM, Langlais F. At-
(NRF) grant funded by the Korea government, Ministry of Sci-
rophy of the supraspinatus belly. Assessment by MRI in 55
ence, ICT & Future Planning (MSIP, 2015R1A2A1A05001887).
patients with rotator cuff pathology. Acta Orthop Scand 1996;67:264-268
REFERENCES
12. Khoury V, Cardinal E, Brassard P. Atrophy and fatty infiltration of the supraspinatus muscle: sonography versus
1. Spick C, Szolar DH, Reittner P, Preidler KW, Tillich M. MR arthrography of the shoulder: do we need local anesthesia? Eur J Radiol 2014;83:980-983 2. de Jesus JO, Parker L, Frangos AJ, Nazarian LN. Accuracy of MRI, MR arthrography, and ultrasound in the diagnosis of rotator cuff tears: a meta-analysis. AJR Am J Roentgenol 2009;192:1701-1707 jksronline.org
J Korean Soc Radiol 2016;75(6):480-486
MRI. AJR Am J Roentgenol 2008;190:1105-1111 13. Hegedus EJ, Zavala J, Kissenberth M, Cook C, Cassas K, Hawkins R, et al. Positive outcomes with intra-articular glenohumeral injections are independent of accuracy. J Shoulder Elbow Surg 2010;19:795-801 14. Malfair D. Therapeutic and diagnostic joint injections. Radiol Clin North Am 2008;46:439-453, v
485
Discrepancy between Fluoroscopic Arthrography and Magnetic Resonance Arthrography in Patients with Supraspinatus Tendon Tears
관절경으로 확인된 극상건 파열 환자들에게 있어서 투시 관절 조영술과 자기공명 관절 조영술의 불일치: 영화 투시 관절 조영술의 추가적 이점 한 석1,2 · 이영한1 · 서진석1* 목적: 관절경으로 확인된 극상건 완전 파열 환자들에 대해서 투시 관절 조영술 영상과 자기공명 관절 조영술 영상을 비교 함으로써 투시 관절 조영술의 추가적인 진단적 이점을 알고자 하는 것이다. 대상과 방법: 이 연구에는 관절경을 통해 확인된 53명의 극상건 완전 파열 환자들이 포함되었다. 투시 관절 조영술 분석 에서는 견봉하-삼각근하 점액낭으로의 누출을 기록하였다. 자기공명 관절 조영술의 분석에서는 견봉하-삼각근하 점액낭 으로의 조영제 누출, 파열건의 수축, 파열의 너비과 길이, 극상근의 위축 등을 평가하였다. 극상건 파열의 특징을 비교하기 위해 환자들을 조영제 누출 유무에 따라 일치군과 불일치군으로 나누었다. 비교를 위해 피셔의 정확검정과 두 표본 t -검 정을 사용하였다. 결과: 총 53명의 환자 중 34명이 일치군이었고, 19명이 불일치군이었다. 일치군이 파열건 수축의 정도가 불일치군보다 높 았으며, 파열의 너비와 길이도 일치군이 불일치군보다 컸다. 근육 위축은 일치군에서 더 심했다. 결론: 극상건 완전 파열은 투시 관절 조영술에서 조영제 누출이 항상 보이지 않았는데, 특히 극상건 파열이 작거나 파열건 수축이 심하지 않는 경우 보이지 않았다. 투시 관절 조영술은 이 검사만으로 극상건 완전 파열의 심한 정도에 대해 진단적 정보를 제공할 수 있다. 1
연세대학교 의과대학 영상의학교실, 2인제대학교 의과대학 해운대백병원 영상의학과
486
J Korean Soc Radiol 2016;75(6):480-486
jksronline.org