Skin Research and Technology 2007; 13: 385–389
Printed in Singapore  All rights reserved
doi: 10.1111/j.1600-0846.2007.00241.x
In vivo visualization of hyaluronic acid injection by
high spatial resolution T2 parametric magnetic
resonance images
D. Gensanne1,2, G. Josse2, A. M. Schmitt2, J. M. Lagarde2 and D. Vincensini1
1
Bioinorganic Medicine Chemistry Laboratory, Therapeutic and Diagnostic Imaging, Université Paul Sabatier Toulouse, Cedex, France, and
2
Skin Research Center, Pierre Fabre Research Institute, Toulouse, Cedex , France
Background/purpose: In recent years, increasing use of
injectable resorbable fillings has been reported for facial
wrinkle treatment. However, the physiological processes
involved such as the localization and subsequent diffusion
of the injected product in skin tissues are poorly documen-
ted.This may be noninvasively achieved using quantitative
magnetic resonance imaging (MRI), which is duly presented
in this pilot study.
Methods: Hyaluronic acid (HA) was injected intradermally
in the forearm of a young male volunteer. High-resolution
MRI scans using a surface antenna were performed just
after injection, and after 2, 4 and 9 months. Morphological
images were compared with transverse relaxation time (T 2 )
images computed from a pixel-by-pixel analysis.
Results: On high-resolution morphological MR images the
HA injection is barely visible, but with quantitative MRI the
(HA) is a polysaccharide
H YALURONIC ACID
that is a ubiquitous component of all mam-
malian connective tissue. As human skin ages,
there are known alterations in HA metabolism
(1). HA matrices are extremely viscoelastic while
preserving a high level of hydratation.
In several medical applications, synthetic HA
products have been shown to possess many
interesting properties particularly in connective
tissue augmentation of facial skin. Injected into
the dermis, HA is a resorbable biocompatible
material that is removed by enzyme degradation
(hyaluronidases) and reabsorbed over a period
of 6–12 months, depending on the amount and
of the type of the HA bulking agent implanted.
As a consequence, injections must be repeated
at intervals of a few months (2). The ideal in-
jectable material for wrinkle treatment should not
only offer aesthetic results and a long-lasting
effect, but it should also be safe and biodegrad-
& 2007 The Authors
Journal compilation & 2007 Blackwell Munksgaard
Skin Research and Technology
zone of injection is clearly seen. This is due to HA having a
distinctly different transverse relaxation time, T 2
600 ms,
compared with dermal and hypodermal tissues, 35 and
80 ms, respectively.
Conclusion: These preliminary results demonstrate the
ability of the T 2 images for in vivo visualization of the filler
agent and also for characterization of tissue modifications.
In addition, the diffusion and progressive degradation of the
filler agent can be monitored by T 2 measurements over time.
Key words: filler agent – hyaluronic acid – quantitative MR
images
& Blackwell Munksgaard, 2007
Accepted for publication 6 January 2007
able with minimal complications and low diffu-
sion (3, 4).
Efficacy of HA injection is generally assessed
visually by the physician. However, some histo-
logical studies have been published in order to
better understand the tissue modification (5, 6).
To our knowledge demonstration of techniques
for noninvasive imaging of the injected product
has not been reported.
Recent developments in MR make skin ima-
ging possible due to a resolution of up to 150 mm
on a 4 cm field of view (7). Figure 1 shows the
dermis as a dark band, whereas the hypodermis
appears as a bright band, with details of hypo-
dermal fasciae clearly visible. As MRI contrast
depends not only on the tissue properties but also
on the acquisition procedure, standard MR
images give only morphological information. In
contrast, parametric MRI via signal processing
gives more information on the state of the tissue.
385
Gensanne et al.
septae dermis
hypodermis
muscle
Fig. 1. In vivo magnetic resonance image of the skin from the thigh
acquired with a high-sensitivity-dedicated surface coil. The imaging
was acquired on a 1.5 T MR scanner by using a spin-echo sequence
(FSE): TR/TE 5 1400/20 ms, matrix image 5 1282, field of
view 5 502 mm2 and slice thickness 5 2 mm.
This involves the determination of the relaxation
time, T2, from the magnetic signal decay. The T2
value is related to the chemical composition and
the physical environment (8). Such an approach
has been used in the detection of malignant
growth in breast cancer (9).
In this paper, we describe the use of high-
resolution T2 parametric MRI for visualizing
HA injection. Preliminary results relative to the
diffusion and the resorption of the injected pro-
duct over time are presented.
Material and Methods
HA injection procedure
For this study HA commercial filler agent was
used (HA, Juvederm 3, Allergan Inc., Irvine, CA,
USA). The injections were given intradermally
using a 27–30 gauge needle with the bevel edge
of the needle facing upward. An amount of
0.05 mL was injected, corresponding to the usual
quantity of the filler agent used in aesthetic plastic
surgery. To prevent leakage, the injection was
stopped just before needle extraction.
A preliminary ex vivo injection of HA in pig’s ear
skin was given to assess the ability of the quanti-
tative MRI technique for detecting HA. Following
this, an in vivo analysis was made to evaluate
tissue reaction following injection of the filler
agent into the forearm skin of a 30-year-old man.
Quantitative MR imaging
This study was performed on a Signa MR scan
(General Electric, Milwaukee, WI, USA) operat-
ing at 1.5 T. Images were acquired using a multi-
spin-echo sequence (FSE : TR/TE 5 1400/20 ms
386
with 16 echos) with a 60 mm diameter dedicated
surface coil. This allows the detection of small
detail, such as the dermis and subcutaneous fat
structure (10, 11).
For 1282 matrix images, a field of view of
50 mm2 and a slice thickness of 2 mm give a
2
spatial resolution of 0.3 mm3. The total acquisition
time was about 3 min for six contiguous slices.
Transverse relaxation times (T2) were calcu-
lated by fitting the experimental MR signal in-
tensity (Fig. 2) to the following theoretical
expression, according to a nonlinear least-squares
w2 minimization algorithm (12) based on the
Levenberg–Marquardt method (13):
S ¼ S0 expðt=T2 Þ
Relaxometry T2 maps were computed from the
pixel-by-pixel analysis of the morphological
images filtered by the selective blurring filter (14).
In the ex vivo study, images were acquired before
and just after the injection. For the in vivo study the
distribution of HA was evaluated on early post-
injection images performed 12 h after HA injec-
tion. The resorption of the filler agent was quanti-
fied on late images acquired 2, 4 and 9 months
after HA injection. Anatomical markers were used
to allow accurate repositioning of the MR images
between successive acquisition procedures.
1.0
0.8
Signal (a.u.)
0.6
Hyaluronic acid (HA)
Hypodermis
0.4
Dermis
0.2
0.0
0 50 100 150 200
t (ms.)
Fig. 2. Relaxation behaviours of dermis (4), hypodermis (  ) and
pure hyaluronic acid (Juvederm 3) sample (&) measured on a 1.5 T
MR scanner with a multi-spin-echo sequence. The transverse relaxa-
tion times (T2) were determined by fitting a monoexponential signal
model (——) to the relaxation decay: T2, dermis 5 35 ms, T2, hypoder-
mis 5 80 ms, T2, HA 5 605 ms. HA, hyaluronic acid.

(a)
Before injection
MR morphological
dermis
images
hypodermis
cartilage
10mm
(b)
T2-parametric
images
Fig. 3. (a) Ex vivo pig’s ear skin magnetic resonance (MR) morphological images acquired before and after injection of hyaluronic acid (HA) filler
agent (0.05 cm3). Comparison of MR images show a slight swelling of the injection area without significant changes in structure. (b) Corresponding
T2 quantitative MR images. Results show that HA filler agent is readily seen on T2 map.
Results
Figure 3a shows the ex vivo morphological and
parametric T2 images from the pig’s ear skin
before injection of the HA filler agent. After injec-
tion, the filler agent is not readily seen on the
morphological image that only shows a swelling
of the subcutaneous tissue at the site of injection
(Fig. 3b). Conversely, the distribution of the filler
agent in the skin is easily seen on the quantitative
MR images because of the large difference in T2
values between filler agent and skin.
When the same imaging procedure was carried
out on the human forearm (Fig. 4), the injected
product was clearly visible on the T2 image, being
located at the dermis/hypodermis interface. Ow-
ing to the limited spatial resolution of MRI with
regard to dermal thickness, the precise assess-
ment of injected material in this tissue is difficult.
T2 measurement values for dermis and hypoder-
mis are in agreement with data from the literature
using monoexponential relaxation models (15).
With regard to the evaluations made just after
injection, and 2, 4 and 9 months afterwards, the
same two blood vessels were located on these
images, ensuring the accurate relocation of the
injected site. The HA product disappeared over
time with traces still present after 9 months (Fig. 4).
To quantify HA resorption with time, three
parameters were calculated from the T2-images:
lateral extension, volume and mean T2 (Fig. 5). It
can be seen from Fig. 5a that the HA first tends to
diffuse at the dermis/hypodermis interface and
presents a maximum extension at about 4
months. The volume, estimated from the multi-
In vivo visualization of HA injection by MR images
After injection
injection area
T2 (ms)
700
600
500
400
300
200
100
slice scans, regularly decreases over time (Fig. 5b).
Similarly the mean T2 decreases towards the
basal skin value (Fig. 5c). These results show
that the HA product diffuses laterally at first,
and is then progressively resorbed or degraded.
Discussion
It has been shown that high-resolution parametric
MRI, as opposed to morphological acquisition, is
an efficient method to visualize HA injection.
Parametric analysis is necessary not only for high-
lighting the injected zone but also for monitoring
change with time. Antenna with greater sensitivity
would enable a better dermis/hypodermis inter-
face visualization, as suggested in previous work
(7). Moreover, high-resolution quantitative imaging
techniques should be useful for measuring the
conjunctive infiltration and thus evaluating the
integration of the material. The technique would
also be useful for detecting foreign body granu-
loma, fibrosis or for quantifying microvasculariza-
tion within the filler substance. This method, which
works for any filler agent that has different relaxa-
tion times from subcutaneous tissue, could also be
used for the follow-up of deep subcutaneous injec-
tion of filler agent for lipoatrophy associated with
the human immunodeficiency virus (16).
It would also be interesting to complete this
experimental model with other techniques, such
as high-frequency ultrasound and/or histology,
in order to evaluate inflammatory reactions and
the permanence of various filler substances un-
der investigation.
387
Gensanne et al.

(a) Before injection dermis 10 (a)

hypodermis 8

Lateral diffusion (mm)

vessels muscle 6

4
(b) 12 hours later

0
0 2 4 6 8 10
T2 (ms)
Time after injection (month)
600
70
500 (b)
(c) 2 months later
60
400
50
300

Volume (mm3)
40
200
30
100
20
(d) 4 months later 0
10

0
0 2 4 6 8 10
Time after injection (month)

700
(c)

600
(e) 9 months later
500

400
T2 (ms)

300
10mm
200

100
Time

Fig. 4. In vivo T2 maps of the subcutaneous tissues acquired before (a) 0

and after the hyaluronic acid (HA) filler agent injection (0.05 cm3). 0 2 4 6 8 10
Figures (b–e) show the change in HA distribution after 12 h and 2, 4 Time after injection (month)
and 9 months post-injection.
Fig. 5. Graphs of the variation with time of (a) the lateral diffusion
and (b) the volume occupied by the hyaluronic acid in subcutaneous
tissue. (c) Graph of the decrease of the average T2 value measured
Conclusion within the injected area.

The aim of this work was to evaluate the efficacy show that the T2 maps represent an efficient
of T2 parametric MR images for noninvasively tool not only for the in vivo visualization of
localizing HA injection. These preliminary results the product but also for characterizing tissue

388

modifications. The method could be useful for
predicting the duration of HA in the dermis and
hypodermis, with regard to possible metabolic
reactions against it. As such, the method could be
an effective way for comparing the longevity of
various filler agents.
The results obtained are promising but have to
be completed by further work, which is in pro-
gress and shall be presented in the future.
References
1. Meyer LJ, Stern R. Age-dependent changes of hyalur-
onan in human skin. J Invest Dermatol 1994; 102: 385–
389.
2. Lemperle G, Morhenn V, Charrier U. Human histology
and persistence of various injectable filler substances for
soft tissue augmentation. Aesthetic Plast Surg 2003; 27:
354–366.
3. Narins RS, Bowman PH. Injectable skin fillers. Clin Plast
Surg 2005; 32: 151–162.
4. Bergeret-Galley C. Comparison of resorbable soft tissue
fillers. Aesthetic Surg J 2004; 24: 33–46.
5. Christensen L, Breiting V, Janssen M, Vuust J, Hogdall E.
Adverse reactions to injectable soft tissue permanent
fillers. Aesthetic Plast Surg 2005; 29: 34–48.
6. Friedland J. Human histology and the persistence of
various injectable filler substances for soft tissue aug-
mentation. Aesthetic Plast Surg 2003; 27: 367.
7. Gensanne D, Josse G, Lagarde JM, Vincensini D. High
spatial resolution quantitative MR images: an experi-
mental study of. Phys Med Biol 2006; 51: 2843–2855.
Epub 24 May 2006.
8. Estilaei M, MacKay A, Whittall K, Mayo J. In vitro
measurements of water content and T2 relaxation times
in lung using a clinical MRI scanner. J Magn Reson Imag
1999; 9: 699–703.
9. Eliat PA, Dedieu V, Bertino C et al. Magnetic resonance
imaging contrast-enhanced relaxometry of breast tu-
In vivo visualization of HA injection by MR images
mors: an MRI multicenter investigation concerning 100
patients. Magn Reson Imag 2004; 22: 475–481.
10. Mirrashed F, Sharp JC, Krause V, Morgan J, Tomanek B.
Pilot study of dermal and subcutaneous fat structures by
MRI in individuals who differ in gender, BMI, and
cellulite grading. Skin Res Technol 2004; 10: 161–168.
11. Querleux B, Cornillon C, Jolivet O, Bittoun J. Anatomy
and physiology of subcutaneous adipose tissue by in
vivo magnetic resonance imaging and spectroscopy:
relationships with sex and presence of cellulite. Skin
Res Technol 2002; 8: 118–124.
12. Press WH, Flannery BP, Teukolsky SA, Vetterling WT.
Numerical recipes in C. Cambridge University Press,
1989; 15: 683–689.
13. Marquardt DW. An algorithm for the estimation of non-
linear parameters. Soc Ind Appl Math J 1963; 11:
431–441.
14. Gensanne D, Josse G, Lagarde JM, Vincensini D. A post-
processing method for multiexponential spin-spin re-
laxation analysis of MRI signals. Phys Med Biol 2005; 50:
3755–3772.
15. Kamman RL, Bakker CJ, van Dijk P, Stomp GP, Heiner
AP, Berendsen HJ. Multi-exponential relaxation analysis
with MR imaging and NMR spectroscopy using fat-
water systems. Magn Reson Imag 1987; 5: 381–392.
16. Borelli C, Kunte C, Weisenseel P, Thoma-Greber E,
Korting HC, Konz B. Deep subcutaneous application
of poly-L-lactic acid as a filler for facial. Skin Pharmacol
Physiol 2005; 18: 273–278. Epub August 19 2005.
Address:
Dr. Gensanne
Bioinorganic Medicine Chemistry Laboratory
Therapeutic and Diagnostic Imaging
Université Paul Sabatier
118, route de Narbonne
31062 Toulouse cedex
Bât. 3SC
France
Tel: 133 562 488 515
Fax: 133 262 488 523
e-mail: gensanne@hotmail.com
389