Hrev_master
Veins and Lymphatics 2018; volume 7:7069
[page 2] [Veins and Lymphatics 2018; 7:7069]
Ultrasonographic, quantitative
comparison of lower extremity
lymphedema versus normal
control. Technical note with
case reports
Vanessa Lôbo de Carvalho,1
Sergio Xavier Salles-Cunha,2
Filipe Amorim Braga,3 Rita de Cássia
dos Santos Moreira,4
Joyce Annenberg Araújo dos Santos,4
Thayná Lucilla Santos de Araújo,4
Guilherme Benjamin Brandão Pitta5
1State University of Health Sciences of
Alagoas (UNCISAL); and Biotechnology
and Health, Northeast Network in
Biotechnology (RENORBIO), Maceió,
AL, Brazil; 2Visiting Professor,
Ultrasonographer, Maceió, AL, Brazil;
and Research Consultant, Itanhaém, SP,
Brazil and Jacksonville Beach, FL,
USA; 3Vascular Surgeon,
Ultrasonographer, Arapiraca, AL,
Brazil; 4Physiotherapy, University
Center Tiradentes, Maceió, AL, Brazil;
5Northeast Network in Biotechnology
(RENORBIO); Federal University of
Rio Grande do Sul (UFRGS); and State
University of Health Sciences of Alagoas
(UNCISAL), Maceió, AL, Brazil
Abstract
Characterization of tissue by ultra-
sonography (CATUS) is a modern-day
research endeavor intended to improve
visual perception and image quantification.
Visual perception increases with color.
Quantification focuses on pixel echo bright-
nesses. A previously presented case report
demonstrated reappearance of lymphatic
channels a few days after manual drainage.
Ultrasonographic images (US) of lym-
phatic leg and foot were quantitated and
compared to a normal extremity based on
proportions of pixels in specific brightness
intervals. Anatomy evaluated included con-
trol-subcutaneous and lymphatic compart-
ments. US with 256 brightness levels were
obtained at the proximal, mid and distal leg
and foot. Control and lymphatic Gray Scale
Medians (GSM) and histograms were com-
pared using t-test and Chi-square statistics.
Average GSM was 97±9 (SD) (82-114,
n=12 images) for control, greater than
51±15 (24-69, n=12) for lymphedematous
leg/foot (P<0.001). Control had >99% of
pixels with brightness in the muscle-fiber
range (41-196), in contrast to 62% for the
lymphatic extremity (P<0.001).
Lymphedema averaged 7%, 3%, 15% and
14% of pixels in blood, blood/fat, fat and
fat/muscle-like regions (0-4, 5-7, 8-26, 27-
40 brightness intervals). Such regions were
visually interpreted as lymphatic channels
or lakes.
Visual perception by colorization is
subjective, but most people perceives
details better, for example, during the day
than at night. Furthermore, US images have
16 times more shades of gray, 256, than that
perceived by the human visual system, 16
on average. Colorization improved percep-
tion of lymphatic channels and lakes by
transforming blood echoes into red and
lymphatic liquid with echoes similar to fat
into yellow.
Pixel proportions in low brightness
intervals were higher in the lymphatic than
in the normal extremity. Lymphedema
severity was quantified. The CATUS tech-
nique may be used to monitor treatment
effects or disease evolution.
Introduction
The utility of ultrasonography (US) to
describe lymphedema, in particular of the
lower extremity, has been demonstrated in a
variety of ways.1-8 We have shown the short
duration of manual drainage by using a
technique being known as characterization
of tissue by ultrasonography (CATUS).9
CATUS is an expansion of Lal’s pixel dis-
tribution analysis of the carotid plaque.10 B-
mode echogenicity intervals were defined
for blood, fat, muscle, fiber and calcium,
according to levels 0-4, 8-26, 41-76, 112-
196 and 211-255 for an US image having
256 brightness amplitudes. The Gray Scale
Median (GSM) is a single number used to
represent an anatomic region. A multi-cen-
ter study demonstrated, for example, that
carotid plaques with GSM < 25 have a high
risk of stroke during stenting.11 CATUS has
been applied to analyze: i) venous throm-
bi;12-15 ii) normal, abnormal and transplant-
ed kidneys;16-18 iii) aneurysms;19 iv) periph-
eral arterial disease;20 and v) lymphedema,9
in addition to the carotid plaques.21,22 This
manuscript is a fundamental description of
the technique applied to a comparison
between US images of: i) the leg and foot of
a traditional lymphedematous lower
extremity versus; ii) a control type, normal
limb (Figure 1).
Materials and Methods
US images were obtained at the medial
aspect of the leg and dorsum of the foot.
The proximal-upper, mid and distal-lower
aspects of the leg were imaged. A Mindray
5 instrument and a 8-12 MHz linear trans-
ducer were employed. A peripheral vein set
up used to study superficial venous reflux
was employed for all images. The original
leg venous set-up of the instrument was
adapted by one of the authors. Usually, a
instrumental venous set-up is designed to
image femoropopliteal veins. The co-author
re-selected items such as frequency range
(8-12 MHz), shallow depth, gain according
to his visual perception (G60 for his instru-
ment), frame rate at 16, and dynamic range
at 70 to optimize his imaging of peripheral,
superficial veins. Such peripheral veins
included saphenous veins and related tribu-
taries and perforating veins. Another co-
author selected this set-up for this investiga-
tion to image subcutaneous, superficial
compartments in the leg. Practically, the
Correspondence: Vanessa Lôbo de Carvalho,
State University of Health Sciences of
Alagoas (UNCISAL); and Biotechnology and
Health, Northeast Network in Biotechnology
(RENORBIO), Rua Professor Manoel Coelho
Neto 201, Jatiúca, Maceió, AL, CEP 57036-
710, Brazil.
E-mail: ftvanessa1981@gmail.com
Key words: Lymphedema; ultrasonography;
brightness quantification.
Acknowledgments: we thank Arthur Ramos
Hospital personnel, Maceio, AL, for allowing
patient recruitment and ultrasound testing in
hospital premises.
Contributions: VLC, design, patient recruit-
ment, data collection, data analysis, manu-
script writer-reviewer; SXSC, design, data
collection, ultrasonographer, data analysis,
manuscript writer; FAB, design, medical con-
sultant, ultrasonographer, vascular laboratory
director, manuscript reviewer; RCSM, JAAS,
TLSA, patient recruitment, patient-care, ultra-
sound assistance, data collection, manuscript
comments; GBBP, design, laboratory director,
research director, manuscript reviewer.
Conflict of interest: the authors declare no
potential conflict of interest.
This work is licensed under a Creative
Commons Attribution 4.0 License (by-nc 4.0).
©Copyright V. Lôbo de Carvalho et al., 2018
Licensee PAGEPress, Italy
Veins and Lymphatics 2018; 7:7069
doi:10.4081/vl.2018.7069
No
n c
om
me
rci
al
us
e o
nly
Article
[Veins and Lymphatics 2018; 7:7069] [page 3]
set-up used was acceptable to two ultra-
sonographers among the authors. Depths
were distinct1 between control, 1.5 cm, and
lymphatic leg, 4.5 cm, due to the larger vol-
ume of the lymphedematous tissue.
Recalibration for uniformity was performed
for each CATUS analysis using black as
zero and similar fascial echoes or focus
arrow as 200. The software program
required redefinition of the 0 and 200
brightness levels. Lal’s original version
used blood and arterial adventitia for a 0-
190 scale.10 This application was not arterial
and did not have standard arterial blood or
adventitia as references. A distinct bright-
ness had to be selected for the new 200
brighness. The author performing the
CATUS analysis trusted the original images
as representatives of the subjects’ condi-
tions and opted for references that did not
alter significantly the original images.
CATUS was performed in anatomic
compartments according to the analyst deci-
sion-making. Normal control images repre-
sented the subcutaneous tissue. The depth
interval 0.1 to 0.6 cm encompassed all but
one image analyzed. Patient´s image repre-
sented regions with lymphatic channels
and/or lakes. Average depth of a rectangular
region analyzed was 0.5 cm to 2.2 cm. The
width encompassed most of the transducer
imaging width available. Figure 2 demon-
strates the versatility of a personalized pro-
gram showing: i) Lal’s criteria; ii) an equal
interval-based echo subdivision; and iii) an
expanded Lal-based tissue subdivision with
14 intervals: non-echogenic, hypo-
echogenic I, II, III and IV, echogenic I to IV,
hyper-echogenic I to IV and saturated. The
later was used for control-lymphedema
comparison. Artificial colors were selected
based on the analyst’s perspective.
Percentages of pixels were calculated
for each interval. Continuous brightness
histograms, GSM, bar histograms and color
imaging were part of the report. Data from
12 images were available for analysis both
of the control subject (right and left) and the
lymphedematous extremity. GSM were
compared by Student’s t-test and histogram
pixel percentage data were compared using
Chi-square statistics available with Excel.
Results
In general, average GSM was 97±9
(SD-standard deviation) (minimum=82-
maximum=114) for control, greater than
51±15 (24-69) for the lymphedematous
leg/foot (P<0.001). Control had >99% of
pixels with brightness in the muscle-fiber
range (41-196), in contrast to 62% for the
lymphatic extremity (P<0.001).
Lymphedema averaged 7%, 3%, 15% and
14% of the pixels in blood, blood/fat, fat
and fat/muscle-like regions (0-4, 5-7, 8-26
and 27-40 brightness intervals). The lowest
GSM values were obtained in the foot of
either subject.
As a technical note, the new brightness
0, based on a black region of the image, was
the same for all 24 images. The new 200
brightness averaged 4% higher for the con-
trol images, 203, versus 194 for the lym-
phedematous extremity.
Discussion
This report described how the CATUS
technique can be useful to distinguish
echogenic differences between a lymphede-
matous extremity and a control limb. This
initial analysis suggested that additional
research could be performed to quantitate
lymphedema and to follow lymphedema
treatment or progression of disease. It also
opened the opportunity to study other types
of edemas in the leg or even in the upper
extremity.
In particular, differences were demon-
strated based on the presence or normal
absence of lymphatic channels or lakes as
shown in Figure 3. Our poster suggested
that manual drainage sequence could be
optimized by ultrasound observations.9 A
reviewer, however, also commented that
fibrosis and hyper-echogenicity should be
better to quantitate lymphedema worsening.
Indeed, Figure 4 probably indicates that the
upper leg of this patient is in worse condi-
tion than the lower leg or foot. Our yet to be
reported experience has raised some
hypotheses that deserve CATUS investiga-
tion such as: i) channels and lakes are more
common in the foot than in the leg; ii) lym-
phedematous legs also have hyper-
echogenic regions; and iii) presence versus
absence of channels may determine better
or worse response to treatment. Fibrosis in
cutaneous, subcutaneous and superficial or
deep compartments without regional lym-
phatic channels should represent worst-case
condition.
Volumetric changes of the subcutaneous
and subfacial, muscular compartments have
been demonstrated long time ago.1 US
images of dilated lymphatic channels have
been published for conditions such as lym-
phatic flow obstruction.2 Worms moving in
US images of the lymphatic tract, the filaria
dance sign, has been an impressive finding.3
Visual impressions by experts allowed
semi-quantitative analysis of skin and
superficial compartments.4-6 Elastography
Figure 1. Photographies of control and lymphedema extremities and corresponding
ultrasound images. The continuous histograms show differences in echogenicity among
two large anatomic regions. Actual comparative analysis was restricted to subcutaneous
control compartment and lymphatic compartment with channels and lakes.
No
n c
om
me
rci
al
us
e o
nly
Article
[page 4] [Veins and Lymphatics 2018; 7:7069]
has been applied to lymphedema.7-8 CATUS
performed under compression conditions
could provide data comparable to elasto-
graphic information.
Echogenicity and artificial colorization
of edematous conditions may be a tissue
research trend in the following decade.23
Gray scale analysis, in general, is a growing
field, particularly in research.24 As an exam-
ple, a peripheral venous thrombus with a
hyperechogenic rock at the tip may be a
more dangerous embolus than a hypoe-
chogenic thrombus.12 Also, hyperechogenic
chronic venous obstruction may be perme-
ated with trapped blood or hypoechogenic
new thrombi causing localized symptoms
for years. These opinions corroborate inves-
tigations that recommend ultrasound as an
Figure 2. Examples of characterization of tissue based on ultrasound (CATUS). From reader’s left to right: i) Lal’s pixel distribution
based on blood, fat, muscle, fiber and calcium; classification default occurred in 32% of pixels with brightness in between such inter-
vals; ii) pixel distribution in intervals with 20 brightness values; and iii) expanded Lal’s classification showing default intervals and sub-
division of muscle and fiber intervals. Artificial colorization was arbitrary and may vary according to the observer’s visual preferences.
Brightness intervals may be defined for each individual application.
No
n c
om
me
rci
al
us
e o
nly
Article
[Veins and Lymphatics 2018; 7:7069] [page 5]
Figure 3. Normal control (reader’s left) and lymphedematous leg comparison. From top to bottom: i) original ultrasound image; ii)
artificial colorization; iii) bar graph histogram representing proportions of pixels according to brightness intervals; iv) table with per-
centages of pixels according to color intervals. Brightness associations: red-blood, yellow-fat, green-muscle, brown-muscle/fiber, light
purple-fiber. Lymphedema presented enlarged lymph channels supposedly filled with blood plasma and/or fatty liquids.
No
n c
om
me
rci
al
us
e o
nly
Article
[page 6] [Veins and Lymphatics 2018; 7:7069]
indicator of treatment postures.25
CATUS was programmed to provide the
user with alternatives and knowledge on
how the analytical process works. It permits
specific analysis of a measurement in ques-
tion. Data collection, however, must avoid
variability. Similar set up and similar US
knobs must be uniform for comparisons.
Re-scaling does not compensate for all vari-
ability issues. A simple application would
be to follow the same patient with the same
technique.
CATUS is a subset of CATIM, charac-
terization of tissue by imaging.
Photographic CATIM of diabetic and
venous ulcers have been published.26,27
CATUS improves perception by artificial
colorization and permits quantitative analy-
sis based on percentages of pixels in specif-
ic brightness intervals. The physiological
logic behind CATUS-CATIM is that the
human brain perceives color better than
gray, like the expression night and day.
Furthermore, the human eye only distin-
guishes, on average, 16 gray levels while an
ultrasound image usually have 16 times
more tons of gray, 256.28
Conclusions
The CATUS technique allowed quanti-
tative comparison showing lower
echogenicity in the lymphatic extremity as
compared to control. Lower echogenic
regions were perceived as channels and/or
lakes. Research is still evolving in several
areas including edema. Artificial coloriza-
tion may improve perception pending indi-
vidual tendencies. Pixel distribution per-
centages upgrade imaging analysis by quan-
tification. Consistent US knobology and
patient follow-up under similar conditions
are recommended.
References
1. Doldi SB, Lattuada E, Zappa MA, et al.
Ultrasonography of extremity lym-
phedema. Lymphology 1992;25:129-
33.
2. Drinan KJ, Wolfson PM, Steinitz D, et
al. Duplex imaging in lymphedema. J
Vasc Technol 1993;17:23-6.
3. Amaral F, Dreyer G, Figueredo JS, et al.
Adult worms detected by ultrasonogra-
phy in human bancroftian filariasis. Am
J Trop Med Hyg 1994;50:753-7.
4. Suehiro K, Morikage N, Murakami M,
et al. Significance of ultrasound exami-
nation of skin and subcutaneous tissue
in secondary lower extremity lym-
Figure 4. Lymphedematous leg. From reader’s left to right: original ultrasound image, artificial colorization and continuous histogram
based on pixel brightness. Brightness associations: red-blood, yellow-fat, green-muscle, brown-muscle/fiber, light purple-fiber. Upper
leg image had a higher gray scale image (GSM) value. Image at the bottom demonstrates worst case condition, or most hypo-echogenic
region, observed in the foot.
No
n c
om
me
rci
al
us
e o
nly
Article
[Veins and Lymphatics 2018; 7:7069] [page 7]
phedema. Ann Vasc Dis 2013;6:180-8.
5. Suehiro K, Morikage N, Murakami M,
et al. Subcutaneous tissue ultrasonogra-
phy in legs with dependent edema and
secondary lymphedema. Ann Vasc Dis
2014;7:21-7.
6. Suehiro K, Morikage N, Murakami M,
et al. A study of increase in leg volume
during complex physical therapy for leg
lymphedema using subcutaneous tissue
ultrasonography. J Vasc Surg Venous
Lymphat Disord 2015;3:295-302.
7. Suehiro K, Morikage N, Murakami M,
et al. Skin and subcutaneous tissue
strain in legs with lymphedema and
lipodermatosclerosis. Ultrasound Med
Biol 2015;41:1577-83.
8. Suehiro K, Kakutani H, Nakamura K, et
al. Immediate changes to skin and sub-
cutaneous tissue strains following man-
ual lymph drainage in legs with lym-
phedema. Ann Vasc Dis 2016;9:30-4.
9. Salles-Cunha SX, Silveira SAF,
Menezes FH. Case report: ultrasound
virtual histology to grade treatment of
lower extremity lymphedema. Poster
presented at the 36th Society for
Vascular Ultrasound Annual Meeting,
Washington, DC, June 7-9, 2012.
10. Lal BK, Hobson RW II, Pappas PJ, et
al. Pixel distribution analysis of B-
mode ultrasound scan images predicts
histologic features of atherosclerotic
carotid plaques. J Vasc Surg
2002;35:1210-17.
11. Biasi GM, Froio A, Diethrich EB, et al.
Carotid plaque echolucency increases
the risk of stroke in carotid stenting: the
Imaging in Carotid Angioplasty and
Risk Of Stroke (ICAROS) study.
Circulation 2004;110:756-62.
12. Salles-Cunha SX. Duplex scanning for
acute venous thrombosis. In: Peter
Gloviczki, ed. Handbook of venous dis-
orders, Guidelines of the American
Venous Forum, 3rd ed. London: Edward
Arnold, Publisher; 2009. pp 129-41.
13. Cassou-Birckholz MF, Engelhorn CA,
Salles-Cunha SX, et al. Assessment of
deep venous thrombosis by grayscale
median analysis of ultrasound images.
Ultrasound Q 2011;27:55-61.
14. Barros FS, Sandri JL, Prezotti BB, et al.
Pulmonary embolism in a rare associa-
tion to a floating thrombus detected by
ultrasound in the basilic vein at the dis-
tal arm. Rev Bras Ecocardiogr Imagem
Cardiovasc 2011;24:89-92.
15. Salles Cunha SX, Varjão de Oliveira
Guimaraes AF. Complementary role of
thermography in the diagnosis of sub-
clavian vein thrombosis: case report.
Poster presented at 40th Congress of the
Brazilian Society of Angiology and
Vascular Surgery, Florianópolis, SC,
September 30 to October 5, 2013.
16. Valiente Engelhorn AL, Engelhorn CA,
Salles-Cunha SX. Initial evaluation of
virtual histology ultrasonographic tech-
niques applied to a case of renal trans-
plant. J Vasc Ultrasound 2015;39:142-4.
17. Valiente Engelhorn AL, Engelhorn CA,
Salles-Cunha SX, et al. Ultrasound tis-
sue characterization of the normal kid-
ney. Ultrasound Q 2012;28:275-80.
18. Valiente Engelhorn AL, Engelhorn CA,
Salles-Cunha SX. Ultrasonographic tis-
sue characterization of kidneys in
patients with unilateral renal artery
stenosis. J Vasc Ultrasound 2016;40:70-
5.
19. Salles-Cunha SX. Technical note:
Ultrasonographic evaluation of aortic
aneurysms treated with endoprosthesis.
J Vasc Bras 2012;11:150-3.
20. Marks NA, Ascher E, Hingorani AP, et
al. Gray-scale median of the atheroscle-
rotic plaque can predict success of
lumen re-entry during subintimal
femoral-popliteal angioplasty. J Vasc
Surg 2008;47:109-16.
21. Menezes FH, Silveira TC, Silveira SAF,
et al. Preliminary comparisons between
in vivo ultrasonographic virtual histol-
ogy and histopathological findings of
endarterectomized carotid plaque. J
Vasc Bras 2013;12:193-201.
22. Barros FS, Pontes SM, Prezotti BB, et
al. Floating thrombus in the internal
carotid artery: Surgical planning
defined by vascular ultrasound. Arq
Bras Cardiol: Imagem Cardiovasc
2013;26:335-40.
23. Ueda-luchi T, Ohno N, Miyati T, et al.
Assessment of the interstitial fluid in
the subcutaneous tissue of healthy
adults using ultrasonography. Sage
Open Medicine 2015;3:2050312115613351.
24. Harris-Love MO, Seamon BA, Teixeira
C, Ismail C. Ultrasound estimates of
muscle quality in older adults: reliabili-
ty and comparison of Photoshop and
ImageJ for the grayscale analysis of
muscle echogenicity. Peer J
2016;4:e1721.
25. Prandoni P, Prins MH, Lensing AW, et
al. AESOPUS Investigators. Residual
thrombosis on ultrasonography to guide
the duration of anticoagulation in
patients with deep venous thrombosis: a
randomized trial. Ann Intern Med
2009;150:577-85.
26. Pereira VHH, Costa Filho EM, Santos
FTA, et al. Photographic image tissue
characterization of the ulcerated diabet-
ic foot during treatment: technical note.
J Vasc Bras 2013;12:303-7.
27. Santos FTA, Rocha CA, Salles-Cunha
SX, et al. Tissue characterization by
photographic imaging during treatment
of chronic venous ulcer: technical note.
J Vasc Bras 2015;14:177-81.
28. Beach KW, Paun M, Primozich JF.
Principles and instruments of diagnostic
ultrasound and Doppler ultrasound. In:
Aburahma AF, Bergan JJ, eds.
Noninvasive vascular diagnosis: a prac-
tical guide to therapy, 2nd ed. London:
Springer-Verlag; 2007. pp 27.
No
n c
om
me
rci
al
us
e o
nly