27.5.jum.online.

qxp:Layout 1 4/16/08 9:38 AM Page 729

Article

Three-Dimensional Ultrasonographic
Application for Analyzing Synovial
Hypertrophy of the Knee in Patients
With Osteoarthritis
Ji Hyeon Ju, MD, Kwi Young Kang, MD, In Je Kim, MD,
Jang Uk Yoon, MD, Ho-Youn Kim, MD, PhD,
Sung-Hwan Park, MD, PhD

Objective. The purpose of this study was to evaluate 3-dimensional (3D) ultrasonography for the visu-
alization of intra-articular synovial hypertrophy in patients with osteoarthritis. Methods. Knee joints of
22 patients with osteoarthritis were examined by 3D ultrasonography, and their synovial fluids were
analyzed. Ultrasonographic image patterns, vascular endothelial growth factor and transforming
growth factor β concentrations in synovial fluid, and serum inflammatory markers were analyzed.
Results. It was possible to visualize the intra-articular structure by 3D ultrasonography. Three-
dimensional ultrasonography revealed various interesting stereoscopic views of the synovial structures,
and the patterns of synovial proliferation ranged from simple proliferations to complex shrubby struc-
tures. Patients with a more complex and proliferative pattern in the synovium tended to have higher
C-reactive protein concentrations, but this difference was not significant (P = .09). The concentrations
of vascular endothelial growth factor and transforming growth factor β in synovial fluid were signifi-
cantly higher in patients with complex hypertrophy (P < .05). Conclusions. Three-dimensional technolo-
gy was useful in delineating the shape of the synovium. It may have a possible impact on future imaging
in rheumatology. Key words: osteoarthritis; synovial hypertrophy; 3-dimensional ultrasonography.

Abbreviations
CRP, C-reactive protein; ESR, erythrocyte sedimentation
rate; NSAID, nonsteroidal anti-inflammatory drug; OA,

D
osteoarthritis; 3D, 3-dimensional; TGF-β, transforming
growth factor β; 2D, 2-dimensional; VAS, visual analog
scale; VEGF, vascular endothelial growth factor
Received October 26, 2007, from the Division of
Rheumatology, Department of Internal Medicine,
Kangnam St Mary’s Hospital, College of Medicine,
Catholic University of Korea, Seoul, Korea. Revision
requested November 19, 2007. Revised manuscript
accepted for publication December 26, 2007.
We thank Lim Ri-Kyung for expert help with
the 3-dimensional gynecologic ultrasonographic
technique and Gyl-Sun Kim, MD, for manuscript
preparation. This work was supported by grant
R11-2002-098-05003-0 from the Science Research
Center/Engineering Research Center program of
the Ministry of Science and Technology/Korea
Science and Engineering Foundation to the
Rheumatism Research Center, Catholic University
of Korea.
Address correspondence to Sung-Hwan Park,
MD, PhD, Division of Rheumatology, Department
of Internal Medicine, Center for Rheumatic
Disease, Kangnam St Mary’s Hospital, College of
Medicine, Catholic University of Korea, 505 Banpo-
Dong, Seocho-Ku, Seoul 137-040, Korea.
E-mail: rapark@catholic.ac.kr
uring the last decade, musculoskeletal ultra-
sonography has become an established imag-
ing technique for diagnosis and follow-up in
patients with rheumatic diseases.1 The nonin-
vasiveness, portability, relatively low cost, absence of ion-
izing radiation, and feasibility in performing repeated
examinations make ultrasonography particularly useful
for therapeutic monitoring. The real-time capability of
ultrasonography allows dynamic assessment of joint and
tendon movements, which can often aid detection of
structural deformities.2
Newer ultrasonographic techniques currently in use
include color and power Doppler imaging, which provide
color-coded tissue maps. The amount of color fractions is
related to the vascularity and may be useful in visualizing
vascular tissues such as the synovium for the diagnosis
and treatment of soft tissue inflammation.3 Three-
dimensional (3D) ultrasonography can also be used to
visualize synovial inflammation, which by the conven-

© 2008 by the American Institute of Ultrasound in Medicine • J Ultrasound Med 2008; 27:729–736 • 0278-4297/08/$3.50
27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 730
Synovial Hypertrophy of the Knee in Osteoarthritis
tional 2-dimensional (2D) method are character-
ized by the combined existence of accumulated
joint effusion and a thickened synovium.4
Conventional 2D ultrasonographic features are
used as the reference standards for assessing syn-
ovial inflammation. Typical findings of synovitis
are a diffuse nodular hypoechoic hypertrophic
synovium with a thickness of greater than 4 mm
on the medial longitudinal plane parallel to the
quadriceps tendon with the knee semiflexed at
45°. The morphologic characteristics of inflam-
matory synovitis are evaluated on a 3-point cate-
gorical scale: absent, nodular, or diffuse.5
According to our experience with ultrasonogra-
phy, the inflamed synovium in human joints
shows various patterns of hypertrophy from a flat
monotonous surface to proliferative shapes such
as treelike or thread ball–like patterns. It seems
that patients with more proliferative features have
larger amounts of synovial fluid and stronger
signals on power Doppler imaging. However,
because conventional ultrasonography tends to
show the object with sectional images, it has been
difficult for beginners or nonprofessionals per-
forming ultrasonography to delineate the accu-
rate cubital pattern of synovial hypertrophy.
Three-dimensional ultrasonography has recent-
ly been used in obstetric practice because it has
the advantages of providing both the obstetrician
and the mother with more precise information
about the fetal surface and real images of the fetal
face.6 Three-dimensional ultrasonography is also
used more frequently as a diagnostic method for
detecting fetal malformations in early pregnancy.
A sufficient amount of amniotic fluid is needed for
the obstetrician to perform 3D ultrasonography of
the fetus. Three-dimensional ultrasonography
reveals the surface of the examined fetus, and the
examiner can perceive a lifelike image of the 3D
reconstructed face and limbs of the fetus. We
thought that this 3D technology might be
applied for delineating the contour of the synovi-
um providing that the amount of joint effusion is
sufficient. The adoption of 3D technology for
visualization of joint-related structures has been
reported for cases of olecranon bursitis and
extensor tenosynovitis.7,8
We found that it was possible to reconstruct the
surface of the synovium with 3D ultrasonogra-
phy, especially in patients with relatively large
730
amounts of joint effusion. Three-dimensional
ultrasonography may provide new information
on synovial morphologic characteristics that
could not be obtained with previous noninvasive
methods.
Materials and Methods
This study was approved by the Institutional
Review Board of Kangnam St Mary’s hospital
(permission number KCMC07OT057). We per-
formed 3D ultrasonography of the knee in 22
patients with osteoarthritis (OA) between
October 2005 and October 2006. Knee OA was
defined as knee pain plus at least 5 of the follow-
ing 9 characteristics: age older than 50 years,
stiffness for less than 30 minutes, crepitus, bone
tenderness, bone enlargement, no palpable
warmth, erythrocyte sedimentation rate (ESR) of
less than 40 mm/h, rheumatoid factor–negative
status, and a synovial fluid sign of OA (clear, vis-
cous, or white blood cell count <2000/mm3). The
main inclusion criteria were age older than 18
years, primary knee OA according to the
American College of Rheumatology definition,
radiographic Kellgren-Lawrence grades of 1
through 4 (0, no feature of OA; 1, minute osteo-
phyte, doubtful importance; 2, definite osteo-
phyte, unimpaired joint space; 3, moderate
diminution of the joint space; and 4, joint space
greatly impaired with sclerosis of the subchon-
dral bone), symptoms of at least 6 months’ dura-
tion, and pain intensity at study entry of greater
than 30 mm on a 100-mm visual analog scale
(VAS). We usually use a VAS to determine how
painful the condition is. The patient is asked to
point on the scale from 0 to 100 mm (0, no pain;
100, extreme pain). The exclusion criteria includ-
ed any known cause of secondary OA, surgery on
the study joint within the previous 12 months,
intra-articular corticosteroid injections within
the previous 4 weeks, or intra-articular radionu-
clide injections within the previous 3 months.
Ultrasonography
After informed consent was obtained, the
patients were examined by a rheumatologist,
who collected all of the clinical and examination
findings. Another rheumatologist who was expe-
rienced in musculoskeletal ultrasonography per-
J Ultrasound Med 2008; 27:729–736
27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 731
formed an ultrasonographic examination of the
studied knee. The suprapatellar pouch was
scanned in extended OA knee joints. The
patients were scanned with an Accuvix IV sys-
tem (Medison Co, Ltd, Seoul, Korea) using a lin-
ear array transducer with a center frequency of
12 MHz (L12-5/38 mm, HDI 5000; Philips
Medical Systems, Bothell, WA), and the 2D
images of the joints were gathered. The patients
were scanned again with a 3D transducer
(S-VNW6-12; GE Healthcare Kretztechnik, Zipf,
Austria), which moved mechanically in one
direction (freehand technique) to obtain a
sequence of 2D images for extracting the 3D
images. The acquired data were stored digitally
on a hard disk as a cine loop in which the 3D
synovium could be viewed as the image rotated,
which enhanced depth perception and gave a
true 3D image. Synovitis was defined as hypoe-
choic synovial hypertrophy with a thickness of
greater than 4 mm and a diffuse or nodular
appearance. Joint effusion was evaluated by
measuring the maximal depth of the suprapatel-
lar recesses in millimeters in the sagittal plane
parallel to the quadriceps tendon.
Data Evaluation
On study entry, clinical and laboratory disease
activity markers were assessed; these comprised
physical examination findings, vital signs, and
measures of disease activity such as the
patient’s global assessment of knee pain, the
physician’s VAS score, and a health assessment
questionnaire. Global inflammatory activity was
assessed by ESR and C-reactive protein (CRP)
concentrations.
Sandwich Enzyme-Linked Immunosorbent
Assay
Vascular endothelial growth factor (VEGF) and
transforming growth factor β (TGF-β) concen-
trations in culture supernatants were measured
by a sandwich enzyme-linked immunosorbent
assay as described previously.9 Briefly, mono-
clonal antibodies to VEGF and TGF-β (R&D
Systems, Inc, Minneapolis, MN) were added to
a 96-well plate (Nunc A/S, Roskilde, Denmark).
A standard curve was drawn by plotting optical
density versus the log of the concentration of
recombinant cytokines.
J Ultrasound Med 2008; 27:729–736
Ju et al
Statistics
The data were analyzed by the Student t test with
SPSS version 11 software (SPSS Inc, Chicago, IL)
to examine the differences in clinical parameters
and ultrasonographic findings between groups
with different patterns of hypertrophy. The data
are presented as means. P < .05 was considered
statistically significant.
Results
Patient Characteristics
The patients’ characteristics and drug histories
were investigated. The mean age of the patients
was 61 years; the mean disease duration was 8.1
years; and the mean overall treatment duration
was 3.9 years. The most common Kellgren-
Lawrence grade was 3 (53% of patients), followed
by 2 (16%), 1 (16%), and 4 (1%). Patients took
various nonsteroidal anti-inflammatory drugs
(NSAIDs). The drugs that the patients took most
frequently for pain control were NSAIDs in 17
(77%) and sole analgesics in 5 (23%). The ESR and
serum CRP concentrations were checked at the
time of ultrasonography, and the mean values
were 25.4 mm/h and 0.77 mg/dL, respectively.
Patterns of Synovial Hypertrophy
We found that it was possible to reconstruct the
surface of the synovium with 3D ultrasonography,
especially in patients with relatively large amounts
of joint effusion (Figure 1). We could change the
virtual viewpoint freely to capture the most per-
ceptive 3D features of the synovium. In our analy-
sis, we could categorize the patterns of synovial
hypertrophy into 2 groups. In group 1, the surface
of the synovium appeared simple or elevated, and
in group 2, the surface showed complex and pro-
liferative features. In some patients, the surface of
the synovium appeared smooth with a tiny gran-
ular pattern (Figure 2A), and some areas of the
synovium had focal or nodular hypertrophy
showing elevated lumps (Figure 2B). These 2 mor-
phologic types constituted group 1. A villous pat-
tern of synovial hypertrophy was frequently seen
during ultrasonography and appeared as a grassy
appearance (Figure 2C). Some patients who had
profuse joint effusion and more severe arthritic
symptoms showed complex and branching syn-
ovial hypertrophy or a shrubby pattern (Figure
731
27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 732
Synovial Hypertrophy of the Knee in Osteoarthritis
2D). The grassy and shrubby types of synovial
hypertrophy constituted group 2 (complex and
proliferative). We compared the clinical charac-
teristics between the 2 groups, which showed no
statistical significance (Table 1).
Differences in Serum Inflammatory Markers
Between Simple Hypertrophy (Group 1) and
Complex Hypertrophy (Group 2)
The ESR and CRP are surrogate markers of
inflammation and are usually within the normal
range in patients with OA.10 There was no statis-
tical significance in the clinical and 2D ultra-
sonographic parameters between groups 1 and
2 except for synovial concentrations of VEGF
and TGF-β (Table 2). Serum ESR and CRP con-
centrations did not differ significantly between
Figure 1. Various 3D reconstructions of an OA synovium. A–H, The amount of synovial fluid varies from a small to large volume, and
various shapes of synovial hypertrophy are shown on 3D ultrasonography. Three-dimensional images are more informative than flat
conventional 2D images.
732
patients with OA and synovial hypertrophy in
groups 1 and 2 (Figure 3, A and B). Patients with
a more complex and proliferative pattern in the
synovium tended to have higher CRP concen-
trations, but this difference was not significant
(P = .09).
The VEGF and TGF-β concentrations were
higher in the synovial fluid of patients in group
2 than in group 1. We used an enzyme-linked
immunosorbent assay to measure the concen-
trations of cytokines accumulated in the syn-
ovial fluid of the patients. The concentrations of
VEGF and TGF-β were significantly higher in
patients with complex hypertrophy (group 2;
Figure 3, C and D). The VEGF concentration in
synovial fluid was correlated with the tentative
grade of synovial hypertrophy (P < .05).
J Ultrasound Med 2008; 27:729–736
27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 733

Ju et al

Figure 2. Tentative grading of synovial hypertrophy of OA Discussion

knee joints. Various features of synovial hypertrophy are clas-
sified by the degree of hypertrophy. A, Plain (I) and simple ele-
vation/hump (II) patterns are thought to indicate less prolifer-
ation (group 1). B, More complex and actively growing pat-
terns of the synovium are classified as grassy (III) and shrubby
(IV), indicating more proliferation (group 2).
A center. The hypertrophied synovium had various
Musculoskeletal ultrasonography is now widely
used to assess joint abnormalities and is useful
for evaluating the synovium of arthritic joints.
We have observed many cases of synovial hyper-
trophy in both OA and rheumatoid knees in our
features, which ranged from a smooth inner sur-
face or a simple elevation to villous polyps or
weblike structures or treelike branching pat-
terns. We grouped the various patterns of syn-
ovial hypertrophy into 2 groups: synovia with
simple hypertrophy or tiny granular or simple
elevations were classified into group 1, and those
with complex hypertrophy with grasslike or
shrubby appearances were classified into group
2. To determine the clinical significance of these
shapes, we compared the clinical parameters
and laboratory findings between these tentative
groups. The complex structures of synovial

Table 1. Patient Characteristics

Group 1 Group 2
Characteristic (n = 11) (n = 11) P
Age, y 58 62 .39
Sex, male/female, n 2/9 2/9
Disease duration, y 9.9 7.4 .47
Treatment duration, y 3.1 4.6 .42
B Kellgren-Lawrence scale, 1/2/3/4, n 2/3/5/1 2/1/7/0
Medication
Simple analgesics only, n 4 3
NSAIDs, n 7 8

Table 2. Differences in Clinical and Laboratory Findings Between

Groups 1 and 2

Group 1 Group 2
Finding (n = 11) (n = 11) P
Kellgren-Lawrence scale, 3 3
most common grade
Ultrasonography
Synovial thickness, mm 5.2 5.5 .53
Synovial fluid, mm 4.3 5.5 .39
Osteophytes, mm 3.2 2.4 .28
Blood
ESR, mm/h 16.7 34.1 .19
CRP, mg/dL 0.19 1.3 .09
Synovial fluid
VEGF, pg/mL 311.3 515.9 .02
TGF-β, pg/mL 80.8 140.2 .01

J Ultrasound Med 2008; 27:729–736 733

27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 734

Synovial Hypertrophy of the Knee in Osteoarthritis

hypertrophy reflected a more severe inflamma- hypertrophy may be important in treating

tory status in both the synovial fluid and serum
markers of inflammation.
Rheumatoid arthritis involves synovitis as the
main pathophysiologic process, and the degree of
synovial hypertrophy directly reflects the severity
of synovitis. In rheumatoid arthritis, more severe
synovitis is generally associated with a worse
outcome.11 If this relationship also exists in OA,
accurate assessment of the status of synovial
inflammation that is manifested in the form of
patients with OA. Power Doppler techniques can
help visualize the amount of blood flow, which
correlates with the status of inflammation.12
Direct visualization with arthroscopy is the best
way of assessing inner synovial hypertrophy.13
However, arthroscopic tools have limitations
because they require an operating room, are
invasive techniques, and cost more and are thus
used infrequently as diagnostic tools. We rea-
soned that application of 3D ultrasonography

Figure 3. Differences between groups 1 and 2 for clinical parameters in serum and synovial fluid. A, The ESR was nonsignificantly
higher in group 2 than in group 1 (P = .19). B, The serum CRP concentration was nonsignificantly higher (P = .09) and the range of
values was wider in group 2 than group 1. C, The concentration of VEGF, a mediator of vessel growth, was significantly higher in
group 2 than group 1 (P = .02). D, The TGF-β concentration was significantly higher in group 2 than group 1 (P = .01).

A B

C D

734 J Ultrasound Med 2008; 27:729–736

27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 735
may overcome some of these problems. Three-
dimensional ultrasonography is now practiced
widely in Korean obstetric clinics because par-
ents want to see the face of their fetus in advance
and to detect fetal abnormalities. With the help
of an obstetric technician, we performed 3D
ultrasonography on knee joints of patients with
OA, which gave us many interesting images of
the inner joint structures. Synovial hypertrophy
can be delineated well by 3D ultrasonography
because joint effusion plays a role similar to that
of amniotic fluid by functioning as a good trans-
mitter for the ultrasound beam.
Adequate blood flow provided by synovial
angiogenesis is required for development of syn-
ovial hypertrophy. Angiogenesis is integral to
development of an inflammatory pannus, and
the formation of new blood vessels augments the
supply of nutrients and oxygen to the inflamma-
tory cell mass, therefore contributing to the per-
petuation of synovitis.14,15 The concentration of
VEGF, a marker of angiogenesis, was significantly
higher in the knee joints of patients with OA and
complex hypertrophy. The ESR and serum CRP
concentrations were insignificantly higher in
patients with complex patterns of hypertrophy.
Provided that morphologic changes reflect
functional differences, 3D ultrasonography,
which can show the inner synovial surface in
detail, should be useful in assessing the prognosis
and in choosing treatment modalities. Patients
with OA who have massive synovial proliferation
with effusion might be treated more vigorously
than patients with no synovial abnormalities.
Thus, patients with a more complex pattern of
synovial hypertrophy should be treated different-
ly than patients with simple hypertrophy.
Our study was limited by the small number of
patients, the short period of follow-up, and the
fact that 3D ultrasonography is not a well-estab-
lished diagnostic tool for rheumatologic dis-
eases. Indirect visualization of the synovium by
3D ultrasonography might be overcome if we
can perform a comparison study of direct visual-
ization with arthroscopy. To address these limita-
tions, large-scale prospective cohort studies are
needed to confirm our preliminary data. Finally,
3D ultrasonography might provide a new way of
understanding synovial morphologic character-
J Ultrasound Med 2008; 27:729–736
Ju et al
istics that could not be obtained with previous
noninvasive methods. We think that 3D ultra-
sonography can provide meaningful informa-
tion as an initial method for assessing synovial
hypertrophy in patients with OA.
References
1. Peterfy CG. New developments in imaging in rheumatoid
arthritis. Curr Opin Rheumatol 2003; 15:288–295.
2. Backhaus M, Burmester GR, Gerber T, et al. Guidelines for
musculoskeletal ultrasound in rheumatology. Ann Rheum
Dis 2001; 60:641–649.
3. Walther M, Harms H, Krenn V, Radke S, Faehndrich TP,
Gohlke F. Correlation of power Doppler sonography with
vascularity of the synovial tissue of the knee joint in patients
with osteoarthritis and rheumatoid arthritis. Arthritis
Rheum 2001; 44:331–338.
4. Fiocco U, Ferro F, Vezzù M, et al. Rheumatoid and psoriatic
knee synovitis: clinical, grey scale, and power Doppler ultra-
sound assessment of the response to etanercept. Ann
Rheum Dis 2005; 64:899–905.
5. D’Agostino MA, Conaghan P, Le Bars M, et al. EULAR
report on the use of ultrasonography in painful knee
osteoarthritis, part 1: prevalence of inflammation in
osteoarthritis. Ann Rheum Dis 2005; 64:1703–1709.
6. Bega G, Lev-Toaff A, Kuhlman K, Kurtz A, Goldberg B,
Wapner R. Three-dimensional ultrasonographic imaging in
obstetrics: present and future applications. J Ultrasound
Med 2001; 20:391–408.
7. Ju JH, Yoon CH, Kim HY, Park SH. Clinical images: visualiza-
tion of the inner synovial surface with three-dimensional
ultrasonography. Arthritis Rheum 2007; 56:214.
8. Ju JH, Kwok SK, Seo SH, Yoon CH, Kim HY, Park SH.
Visualization of extensor digitorum tenosynovitis with
three-dimensional ultrasonography. Rheumatology (Oxford)
2007; 46:886–887.
9. Asturias JA, Arilla MC, Aguirre M, et al. Quantification of
profilins by a monoclonal antibody-based sandwich ELISA.
J Immunol Methods 1999; 229:61–71.
10. Dougados M. Clinical assessment of osteoarthritis in clinical
trials. Curr Opin Rheumatol 1995; 7:87–91.
11. Massardo L, Gabriel SE, Crowson CS, O’Fallon WM,
Matteson EL. A population-based assessment of the use of
orthopedic surgery in patients with rheumatoid arthritis.
J Rheumatol 2002; 29:52–56.
12. Wakefield RJ, Brown AK, O’Connor PJ, Emery P. Power
Doppler sonography: improving disease activity assess-
ment in inflammatory musculoskeletal disease. Arthritis
Rheum 2003; 48:285–288.
13. Kane D, Veale DJ, FitzGerald O, Reece R. Survey of
arthroscopy performed by rheumatologists. Rheumatology
(Oxford) 2002; 41:210–215.
735
27.5.jum.online.qxp:Layout 1 4/16/08 9:38 AM Page 736

Synovial Hypertrophy of the Knee in Osteoarthritis

14. Strunk J, Heinemann E, Neeck G, Schmidt KL, Lange U. A

new approach to studying angiogenesis in rheumatoid
arthritis by means of power Doppler ultrasonography and
measurement of serum vascular endothelial growth factor.
Rheumatology (Oxford) 2004; 43:1480–1483.

15. Taylor PC. Serum vascular markers and vascular imaging in

assessment of rheumatoid arthritis disease activity and
response to therapy. Rheumatology (Oxford) 2005; 44:
721–728.

736 J Ultrasound Med 2008; 27:729–736