COMPARATIVE POPLITEAL AND MESENTERIC COMPUTED

TOMOGRAPHY LYMPHANGIOGRAPHY OF THE CANINE

THORACIC DUCT

IAN R. MILLWARD, ROBERT M. KIRBERGER, PETER N. THOMPSON

Thoracic duct computed tomography (CT) lymphangiograms were performed on seven clinically normal dogs.
The appearance of the thoracic duct system was compared following administration of contrast medium through
a mesenteric lymphatic vessel vs. ultrasound guided percutaneous injection into a popliteal lymph node using
helical and sequential CT acquisition modes. The number of visible thoracic duct branches and the largest
thoracic duct branch cross-sectional area and mean Hounsfield units (HU) were determined from thoracic
vertebra 9 to lumbar vertebra 1. Procedural time and patient discomfort were also assessed. Popliteal ad-
ministration produced a successful thoracic duct lymphangiogram in eight of 11 dogs (73%) after two attempts,
while mesenteric administration was successful in eight of 10 dogs (80%) after a single attempt. Popliteal
lymphography required 46% of the time and was associated with less patient discomfort than mesenteric
lymphangiography. The number of thoracic duct branches seen was not significantly different for either
administration technique (P ¼ 0.256) or CT acquisition mode (P ¼ 0.417). However, the cross-sectional area
and mean HU of the largest thoracic duct branch were greater with mesenteric administration (Po0.001), and
helical image acquisition (Po0.001). The thoracic duct branch number, size, and location were highly variable
between dogs. Percutaneous popliteal lymphography appears to be an acceptable alternative to mesenteric
lymphangiography for the detection of thoracic duct branches in the dog when using either helical or sequential
CT acquisition modes. r 2011 Veterinary Radiology & Ultrasound, Vol. 52, No. 3, 2011, pp 295–301.

Key words: CT, lymphangiography, lymphography, popliteal lymph node, thoracic duct.

Introduction thoracic duct branches could help facilitate curative liga-

C HYLOTHORAX OCCURS SECONDARY to any functional
obstruction of chyle flow through the thoracic duct.1
Treatment for chylothorax should target the underlying
cause. Unfortunately, medical therapy may only result in
palliative management,2,3 and chronic chylothorax is irri-
tating to the pleura and may result in restrictive fibrosing
pleuritis.4,5 Surgery may be considered for patients with
respiratory signs due to chylothorax that do not respond to
medical therapy, and for patients with chylothorax which
have a cause that is amenable to surgical correction.2,3,6–12
Success rates for thoracic duct ligation are o60%,6,9 and
flow through the thoracic duct system continues after
duct ligation if even one small branch is not occluded.10,13,14
The variability that exists in the number, location, and
periodic patency of thoracic duct branches1–3,14–21 may
contribute to the relatively poor rate of success seen fol-
lowing thoracic duct ligation. Improved identification of
From the Department of Companion Animal Clinical Studies (Mill-
ward, Kirberger) and the Department of Production Animal Studies
(Thompson), Faculty of Veterinary Science, University of Pretoria, Pri-
vate Bag X04, Onderstepoort 0110, South Africa.
Address correspondence and reprint requests to Ian R. Millward, at the
above address. E-mail: ian.millward@up.ac.za
Received March 13, 2010; accepted for publication November 12, 2010.
doi: 10.1111/j.1740-8261.2010.01790.x
tion,14,15,21 or enable less invasive procedures such as thor-
ascopic ligation.22,23
Thoracic duct lymphangiography can be used to aid
identification of thoracic duct branches. Contrast medium
can be administered through a mesenteric lym-
phatic,2,7,8,14,15,21,24–27 although injection into a popliteal28
or mesenteric23,29 lymph node can also be used. A com-
parison of these different administration techniques has
not been assessed.
Our purpose was to determine whether ultrasound
guided percutaneous administration of contrast medium
into a popliteal lymph node was an acceptable alternative
to the more invasive administration of contrast medium
into a mesenteric lymphatic vessel. We also aimed to com-
pare helical vs. sequential computed tomography (CT)
scanning in this setting.
Materials and Methods
Twelve healthy mature research beagles were studied.
The dogs were divided randomly into two groups, accord-
ing to whether a mesenteric lymphatic or popliteal lymph
node was injected first. Ten to 21 days after the first pro-
cedure, lymphangiography was repeated using the other
administration route.

295
296 MILLWARD, KIRBERGER, AND THOMPSON
On the day of lymphangiography, all dogs were given
two oral doses of corn oil at 2 ml/kg per dose, with the last
dose being at least 2 h before CT imaging. All animals had
an intravenous catheter placed in a cephalic vein and were
premedicated with carprofen 30 min before induction.
Immediately before induction they were given morphinew
and diazepamz intravenously so as to minimize the
risk of morphine-induced vomiting of the corn oil. Anes-
thesia was induced with propofoly and maintained with
isoflurane.z
For dogs undergoing mesenteric lymphatic catheteriza-
tion, the ventral abdomen was routinely prepared for a
midline celiotomy. The ileum, cecum, and ascending colon
were exteriorized and a large chyle filled mesenteric lym-
phatic was identified. A 22 G intravenous catheter was in-
serted and directed toward the cisterna chyli. The catheter
was held in place with a 5-0 nylon suture before a contrast
medium-filled extension set was attached to the catheter
and sutured to the adjacent intestinal loop. Exteriorized
organs were returned to the peritoneal cavity, and the
celiotomy closed temporarily. The end of the extension set
was left projecting from the incision and was attached to
the skin using a finger trap suture pattern.
After transfer to the CT suite the dogs were placed on
the CT table, head first into the gantry, in sternal recum-
bency with the pelvic limbs extended. Iohexol 300 mg I/
mlk was administered at 1 ml/kg over 1 min, and image
acquisition initiated immediately after administration of
the contrast medium. After completion of the CT study,
the dogs were returned to the surgery suite to have the
lymphatic catheter and extension set removed.
After being anesthetized, the popliteal lymphography
dogs had the popliteal area of the pelvic limbs prepared for
surgery before transfer to the CT suite. The dogs were in
sternal recumbency on the CT table, head first into the
gantry, with the pelvic limbs extended and a small foam
pad under the medial aspect of the stifle to aid ultrasound
guided needle placement. The popliteal lymph node was
identified sonograpically using a 9 MHz linear array trans-
ducer. A 25 G hypodermic needle was guided into the
popliteal lymph node and 1 ml/kg of iohexol was injected
at 1.67 ml/min using a syringe pump, while monitoring
sonographically for contrast medium spillage. The volume
of any spillage was estimated and added to the total vol-
ume of contrast medium injected. CT image acquisition
was initiated immediately after completion of the contrast
medium injection, and mediolateral and craniocaudal
Rimadyl, Pfizer, Sandton, South Africa.
wMorphine Sulfate, Fresenius Kabi, Midrand, South Africa.
zTranject, Merck Generics RSA Pty. Ltd., Modderfontein, South
Africa.
yFresenius Kabi, Midrand, South Africa.
zIsofor, Safe Line Pharmaceuticals, Johannesburg, South Africa.
kOmnipaque, Nycomed Inc., New York, NY.
2011
radiographs of the injected stifle were acquired to further
assess the volume of contrast medium spillage.
The animals were monitored for 5 days after the pro-
cedure and the pain level assessed using a previously de-
scribed30 cumulative pain score system. The minimum
cumulative pain score was 0 and the maximum was 23.
When the pain score for any single category was at the top
of the scale (2 or 3), or the cumulative pain score exceeded
10, then analgesia of morphine and carprofen was used.
For ethical reasons mesenteric lymphangiography was
limited to a single attempt and popliteal lymphography to
two separate attempts, spaced between 10 and 21 days
apart.
The CT procedure for both groups was identical using a
dual slice Siemens Emotion Duo CT scanner,# with the
exception being that image acquisition was initiated ap-
proximately 50 s later in the popliteal administration dogs
due to the time required to move the ultrasound equipment
away from the CT gantry.
A lateral topogram was acquired in a craniocaudal
direction and the helical CT field was defined from mid-
vertebral body of thoracic vertebra 8 (T8) to mid-vertebral
body of lumbar vertebra 2 (L2). The sequential CT fields
were set visually on the lateral topogram for the mid-
vertebral body site of T9–L1 and oriented to ensure that
the slice was perpendicular to the vertebral canal.
Immediately after contrast medium administration the
helical CT acquisition was initiated in a caudocranial
direction, while using a forced inspiratory breath hold
to minimize thoracic movement. A sharp kernel with a
spinal window, using 3 mm slices with 50% overlap and a
pitch of 1.55 was used, and images were reconstructed to
1.5 mm slices. Use of the Siemens CARE doses # for the
helical CT, resulted in exposure settings of 35 mA s and
130 kV.
Sequential CT was initiated within 30 s of completing the
helical CT scan, using the same algorithm and windows,
and the slice thickness was set at 1 mm. Tilting of the gan-
try occurred to ensure that each mid-vertebral body dual
slice was oriented perpendicular to the vertebral canal as
assessed on the lateral topogram. A forced inspiratory
breath hold was not used as motion was not a problem
with the short duration of each slice. Exposures averaged
77 mA s and 130 kV. The duration of each procedure and
CT examination time was recorded, as well as any prob-
lems experienced.
CT images were viewed in the bone window, and were
assessed independently by the primary investigator (I.R.M.)
and a board certified radiologist (R.M.K.) who were un-
aware of the route of contrast medium administration.
For helical CT the mid-vertebral site was defined as the
slice that was midway between the most cranial and most
#Siemens AG, Erlangen, Germany.
Vol. 52, No. 3 THORACIC DUCT CT LYMPHANGIOGRAPHY 297

caudal edges of the pedicle for each vertebra, while for Table 1. Dog CT Status Summary
sequential CT the single slice, which contained the max-
Dog Study
imum number of thoracic duct branches was used for in- Number Weight (kg) Mesenteric CT Popliteal CT Order
terpretation. At each mid-vertebral slice from T9–L1, both 1 11.0 ‘ | P-M
investigators assessed the number of thoracic duct 2 13.0 — ‘‘ P-
branches visible. The largest thoracic duct branch was as- 3 13.6 ‘ — M-
4 13.3 | | M-P
sessed for total cross-sectional area, and the mean Houns- 5 12.8 | | M-P
field units (HU) was measured by centering the largest 6 10.5 | ‘| P-M
circular region of interest within the largest thoracic duct 7 10.0 | | P-M
8 12.0 — ‘ P-
branch. The thoracic duct branch locations were recorded 9 10.2 | | P-M
relative to the aorta. 10 11.3 | ‘| P-M
For the number of visible thoracic duct branches the 11 15.2 | ‘‘ M-P
12 14.8 | | M-P
effects of contrast medium administration method, CT ac-
quisition mode and vertebral site were estimated using a CT data included in the study. |, thoracic duct CT study successfully
zero-truncated Poisson regression model. Clustering of ob- completed; ‘, thoracic duct CT study failed; P, popliteal CT lympho-
servations within an animal was accounted for by model- graphy; M, mesenteric lymphangiography; CT, computed tomography.
ling the animal as a fixed effect. An interaction term
between administration method and CT acquisition mode
was tested and retained in the model if significant at
Po0.05. hyperechoic to surrounding soft tissue. All popliteal lymph
For the continuous outcomes (largest thoracic duct nodes had sonographic evidence of contrast medium leak-
branch cross-sectional area and mean HU) the data were age, which appeared as hypoechoic pockets in the peri-
analyzed using multiple linear regression, including the nodal tissues. Perinodal contrast medium leakage was
observer as an additional independent variable. Because of minimal in successful studies due to the ability to correct
the largest thoracic duct branch cross-sectional area not the needle placement as soon as extranodal leakage was
being normally distributed it was log transformed to detected. None of the dogs where injection of contrast
achieve normality before fitting the statistical models. The medium was performed but the lymph node was not de-
mid-vertebral slice of T9 of each individual was used as the tected sonographically had a successful thoracic duct
reference level for all analyses. Statistical analysis was per- lymphangiogram.
formed using Stata 10.1 statistical software and the level Mesenteric thoracic duct lymphangiography was suc-
of significance was set at a ¼ 0.05 for all analyses. cessful in eight of 10 dogs (80%) after a single attempt,
with the failures being due to kinking or displacement of
the mesenteric catheter during dog transportation or po-
Results sitioning (Table 1).
Twelve dogs were initially included in the trial, of which The average time for percutaneous popliteal lympho-
eleven animals underwent popliteal, and 10 mesenteric ad- graphy was 61 min compared with 113 min for mesenteric
ministration of contrast medium (Table 1). Successful tho- lymphangiography. Dogs undergoing popliteal lympho-
racic duct CT lymphangiograms were obtained using both graphy occupied the CT suite for an average of 31 min,
contrast administration techniques in seven dogs. Two compared with 20 min for mesenteric lymphangiography
dogs were removed from the study due to lost mesenteric (Table 2).
catheter functionality, and three for not having a detectable Postprocedural analgesia was required in two of 10 dogs
lymph node on ultrasound during the procedure. undergoing mesenteric lymphangiography and none of the
Five of the popliteal trial dogs (45%) failed to have a 11 dogs undergoing popliteal lymphography.
lymph node detected on the initial ultrasound examination. The number of thoracic duct branches detected ranged
After a delay of between 10 and 21 days, four of the dogs from one to five at the various sites (Table 3), however, the
with previously nondetectable lymph nodes were reassessed number did not vary significantly within an individual be-
and two still had nondetectable popliteal lymph nodes. tween mesenteric and popliteal techniques (P ¼ 0.256), or
Therefore, popliteal lymphography was successful in six of helical and sequential CT acquisition modes (P ¼ 0.417).
11 dogs (55%) on the first attempt and eight of 10 dogs Popliteal lymphangiography allowed detection of 93%
(80%) after two attempts (Table 1). of the number of thoracic duct branches seen with mes-
During administration of contrast medium into the pop- enteric lymphangiography (Table 3), however, this differ-
liteal lymph node, the lymph node expanded and become ence was not statistically significant (P ¼ 0.256). Using
helical CT allowed detection of 95% of the number of
StataCorp, College Station, TX. thoracic duct branches seen with sequential CT (Table 3),
298 MILLWARD, KIRBERGER, AND THOMPSON 2011

Table 2. Computed Tomographic Start Times and Intervals

Helical Sequential
Dog Admin
Number Method Start End Start End
4 P 23.07 0.24 0.18 1.02
M 10.48 0.25 0.27 1.43
5 P 22.45 0.25 0.16 1.23
M 28.48 0.24 0.19 1.32
6 P 22.05 0.20 0.18 1.07
M 20.51 0.24 0.21 1.32
7 P 30.03 0.21 0.33 2.03
M 19.52 0.24 0.25 1.32
9 P 37.05 0.21 0.29 1.22
M 19.59 0.26 0.24 2.09
10 P 34.21 0.24 0.16 1.24
M 09.14 0.24 0.24 1.52
Fig. 1. Mean attenuation of the largest thoracic duct branch as a function
12 P 36.26 0.24 0.16 1.19
of anatomic site for helical and sequential scanning modes. Error bars in-
M 10.08 0.24 0.21 1.12
dicate one standard deviation. Mean attenuation decreased significantly at
the T12–L1 sites for sequential scanning. TD, thoracic duct.
Time from the start of the topogram. P, popliteal; M, mesenteric.

however, this difference was not statistically significant
(P ¼ 0.417).
Compared with T9 there was a significant reduction in
the number of thoracic duct branches seen as the thoracic
duct system was followed caudally from T11 (P ¼ 0.002) to
L1 (Po0.001).
Mesenteric administration of contrast medium and the
helical CT acquisition mode both resulted in the largest
thoracic duct branch having a significantly greater cross-
sectional area (Po0.001) and a higher mean HU
(Po0.001).
At L1 the largest thoracic duct branch had a significantly
greater cross-sectional area (Po0.001) than at other ver-
tebral sites with a range of 0.1–4.2 mm2 and a mean of
1.67 mm2. When compared with the T9 site the mean HU
value for the largest thoracic duct branch remained rela-
tively constant throughout the T9–T11 vertebral sites, but
declined significantly from T12 (P ¼ 0.029) to L1
(P ¼ 0.001) (Fig. 1). Variation in the mean HU at the
different vertebral sites for the two CT modalities was
particularly notable. With helical CT the HU was relatively
high throughout all assessed vertebral sites, while with
sequential CT the mean HU was high at the three most
cranial vertebral sites but then decreased from T12 to L1
(Fig. 1).
Within the T10–T12 vertebral segment, 87% of thoracic
duct branches were in the right hemithorax between the 12
o’clock and 3 o’clock positions relative to the aorta, while
13% were in the left hemithorax between the 10 o’clock
and 12 o’clock positions (Fig. 2). However, the number,
size, and location of thoracic duct branches in some dogs
was highly variable, even over a distance as short as a half
vertebral body length (Fig. 3).
Discussion
Placing dogs in sternal recumbency, which is more nor-
mal physiologically than dorsal recumbency, minimizes the
impact of positioning on the cardiovascular system.31,32
We believe that sternal recumbency would similarly min-

Table 3. Number of Thoracic Duct Branches Visible Using CT Lymphangiography in Dogs

Number of Visible Thoracic Duct Branches

Mesenteric Popliteal
Helical Sequential Helical Sequential
Mid-Vertebral Body Site Mean SD Ran Mean SD Ran Mean SD Ran Mean SD Ran
T9 2.57 0.98 1–4 3.29 1.11 1–4 2.57 0.54 2–3 2.71 0.76 2–4
T10 2.14 0.90 1–3 2.43 0.98 1–4 2.00 1.00 1–3 2.00 0.58 1–3
T11 2.00 1.00 1–4 1.86 1.46 1–5 1.71 0.76 1–3 1.43 0.54 1–2
T12 1.43 0.54 1–2 1.57 0.79 1–3 1.43 0.54 1–2 1.14 0.38 1–2
T13 1.14 0.38 1–2 1.14 0.38 1–2 1.14 0.38 1–2 1.14 0.38 1–2
L1 1.00 0.00 1 1.57 0.79 1–3 1.29 0.49 1–2 1.57 1.13 1–4
Total number 72 80 71 70

T9–T13, thoracic vertebrae 9–13; L1, lumbar vertebra 1; SD, standard deviation; Ran, range; CT, computed tomography.
Vol. 52, No. 3 THORACIC DUCT CT LYMPHANGIOGRAPHY
Fig. 2. Thoracic duct branch location relative to the aorta for the T10–
T12 vertebral segment using a clock face analogy. TD, thoracic duct.
imize any positional effect on chyle flow, which could result
in a lack of contrast medium flow through collapsed
thoracic duct branches. Sternal recumbency facilitates ac-
cess to the popliteal area for administration of contrast
medium but it also results in dogs undergoing mesenteric
lymphangiography lying on the celiotomy incision and the
protruding administration line.
Radiographic contrast medium has traditionally been
diluted for lymphangiography to reduce viscosity,1,2,7,8,21,22
and is usually given at 0.5–1 ml/kg1,2,7,14,15,22,25 when per-
forming mesenteric lymphangiography. Contrast medium
is generally given as a rapid bolus,1,2,7,8,14,22 and imaging of
the thoracic duct is normally initiated immediately after
completing contrast medium administration.2,7,14,15,21 For
Fig. 3. Helical computed tomography images of T11 from the mid-vertebral body site (A) moving caudally in 1.5 mm increments (B–H). Note the rapid
branching of the thoracic duct system and the variable size of the branches over a 12 mm distance.
299
popliteal lymphography, an infusion rate of 2 ml/min was
not associated with significant leakage or lymph node reac-
tion when using 300 mg I/ml iohexol at a rate of 1 ml/kg.28
Because of limitations of available syringe pumps, we used
an infusion rate of 1.67 ml/min of undiluted 300 mg I/ml
iohexol at 1 ml/kg and the findings were similar, except that
some contrast medium leakage was noted in all dogs.
The optimum time for radiography after administration
of contrast medium into a mesenteric lymphatic vessel has
not been determined. For popliteal lymphography, the
thoracic duct was seen more clearly on radiographs per-
formed within 2 min of contrast medium administration.28
For administration of contrast medium into a mesenteric
lymph node, the optimum time after contrast medium ad-
ministration for radiographic visualization of the thoracic
duct was also o2 min.23 The concentration of contrast
medium within the thoracic duct decreased significantly
after approximately 2 min for both administration tech-
niques (Fig. 1 and Table 2), however, the effect of time
on the degree of thoracic duct branch dilation or the num-
ber of thoracic duct branches identified could not be
assessed. Therefore, it would appear that diagnostic imag-
ing should be performed within 2 min of contrast medium
administration.
There was no statistical difference in the number of tho-
racic duct branches detected with mesenteric vs. popliteal
lymphangiography. However, it should be noted that pop-
liteal administration detected only 93% (141) of the number
of thoracic duct branches that mesenteric administration
did (152) (Table 3). The potential risks associated with
missing a thoracic duct branch when performing
surgical ligation have been eluded to previously.10,13,14
300 MILLWARD, KIRBERGER, AND THOMPSON
While not statistically significant in this study, if popliteal
lymphangiography does actually detect fewer thoracic duct
branches then this may have a severe consequence on
clinical results. Based on the observed variation in our data,
the number of dogs in our study (seven) would have been
sufficient to detect a 17% difference in the average number
of thoracic duct branches detected. At least 27 animals
would have been required to detect a 9% difference.
Popliteal lymphangiography resulted in thoracic duct
branches that were smaller and contained less contrast me-
dium. This may be important when radiography is used
rather than CT to assess the thoracic duct. The finding that
popliteal lymphography may have identified fewer thoracic
duct branches, which were of smaller diameter and con-
tained a lower concentration of contrast medium may have
been due to the administration of contrast medium at a
more peripheral site, the slower rate of infusion, adminis-
tration into a lymph node rather than directly into a lym-
phatic vessel, or the 50 s delay that occurred while that
ultrasound equipment was moved away from the CT gantry.
There was no statistical difference in the number of tho-
racic duct branches seen with helical vs. sequential CT ac-
quisition. However, the cross-sectional area of, and
concentration of contrast medium within the thoracic duct
branches was greater with helical acquisition. These differ-
ences are likely due to helical acquisition being initiated
immediately after contrast medium administration com-
pared with a delay of approximately 45 s when using se-
quential CT (Table 2). The effect of this delay on the lower
thoracic duct branch cross-sectional area, and mean HU, is
supported by the mean HU being high throughout the du-
ration of the helical CT scan but only during the first half of
the sequential CT scan (Fig. 1). Despite these advantages
for helical acquisition it actually detected only 95% (143) of
the thoracic duct branches found with sequential CT (150)
(Table 3). Mesenteric administration combined with se-
quential acquisition resulted in detection of the greatest
number of thoracic duct branches (80) (Table 3) and this
could support that the positive pressure breath hold used
for the helical acquisition may have led to collapse of some
thoracic duct branches. Therefore, helical or sequential ac-
quisition modes may be used but the use of a forced in-
spiratory breath hold should be avoided and other means
considered to minimize respiratory movements.
The greater cross-sectional area of the largest thoracic
duct branch at L1 seen with both administration methods
may have been a reflection the inclusion of the cranial end
of the cisterna chili. The start of the thoracic duct is defined
as where the cisterna chyli achieves its smallest diameter
and this is usually within the L1 segment.16
The intercostal approach that is used classically for liga-
tion of the thoracic duct, allows limited exposure. There-
fore, the T10–T12 vertebral segment of the thoracic duct
system becomes important for surgical planning, because
2011
complete ligation of all thoracic duct branches is criti-
cal.10,13,14 Within this segment, 87% of thoracic duct
branches were to the right of the aorta and 13% to the left.
This differs from the classic description of the thoracic duct
being on the right dorsal border of the aorta from T6 to
L1,2,3,10,14,16 but is consistent with the high degree of vari-
ability in thoracic duct anatomy and position that has been
reported.2,3,14,15,16,19 In these dogs, the left sided branches
may not have been visible from a right-sided surgical ap-
proach and may have been missed. Because of all of the
thoracic duct branches being located between the 10
o’clock and 3 o’clock positions, en bloc ligation of all tis-
sues dorsal to the aorta10–12 may have incorporated all
thoracic duct branches as long as all tissue on the lateral
aspect of the aorta was included.
Mesenteric lymphangiography required extra surgical
time, and extreme care was needed while manipulating the
dog to avoid compromising the implanted catheter. Oth-
erwise mesenteric lymphangiography was simple and quick
to perform.
Many of the dogs did not have a palpable popliteal
lymph node, which was unexpected. We also found it
difficult to identify a popliteal lymph node with ultrasound
in some dogs. Therefore, dogs being considered for pop-
liteal lymphography should have ultrasonographic assess-
ment of the popliteal lymph node before the procedure. If a
popliteal lymph node cannot be identified then another
route of administration should be used. Also, ultrasound
guidance facilitated needle placement into the popliteal
lymph node, which will be important if the lymph node is
not palpable. Sonography also allowed early detection and
correction of any contrast medium spillage. Contrast me-
dium in the lymph node increased its echogenicity but
extravasated contrast medium was hypoechoic; we do not
know the reason for this difference.
The time required for popliteal lymphography was ap-
proximately 50% of that required for mesenteric lympho-
graphy but dogs undergoing popliteal lymphography
occupied the CT suite for almost twice as long. This may
have a cost implication if fees are based on an hourly rate.
No dog undergoing popliteal lymphography experienced
notable discomfort after the procedure, while two (20%) of
the dogs undergoing mesenteric lymphangiography exhib-
ited postoperative pain that required analgesia.
In conclusion popliteal lymphography appears to be an
acceptable alternative to mesenteric lymphangiography
when using either helical or sequential CT for visualiza-
tion of the thoracic duct, assuming that a detectable pop-
liteal lymph node is present. However, this study may lack
suitable power to have detected small differences in the
number of thoracic duct branches detected, which may be
of critical importance when considering surgical ligation of
the thoracic duct. Popliteal lymphography requires less
time, causes less discomfort, and may incur lower cost.
Vol. 52, No. 3 THORACIC DUCT CT LYMPHANGIOGRAPHY
REFERENCES
1. Fossum TW, Birchard SJ. Lymphangiographic evaluation of exper-
imentally induced chylothorax after ligation of the cranial vena cava in dogs.
Am J Vet Res 1986;47:967–971.
2. Birchard SJ, Fossum TW. Chylothorax in the dog and cat. Vet Clin
North Am: Small Anim Pract 1987;17:271–283.
3. Berg J. Chylothorax in the dog and cat. Compend Contin Educ Pract
Vet 1982;4:986–991.
4. Fossum TW, Evering WN, Miller MW, Forrester SD, Palmer DR,
Hodges CC. Severe bilateral fibrosing pleuritis associated with chronic
chylothorax in five cats and two dogs. J Am Vet Med Assoc 1992;201:
317–324.
5. Willauer CC, Breznock EM. Pleurovenous shunting technique for
treatment of chylothorax in three dogs. J Am Vet Med Assoc 1987;191:1106–
1109.
6. Birchard SJ, Smeak DD, McLoughlin MA. Treatment of idiopathic
chylothorax in dogs and cats. J Am Vet Med Assoc 1998;212:652–657.
7. Fossum TW, Mertens MM, Miller MW, et al. Thoracic duct ligation
and pericardectomy for treatment of idiopathic chylothorax. J Vet Intern
Med 2004;18:307–310.
8. Hayashi K, Sicard G, Gellasch K, Frank JD, Hardie RJ, McAnulty
JF. Cisterna chyli ablation with thoracic duct ligation for chylothorax: re-
sults in eight dogs. Vet Surg 2005;34:519–523.
9. Birchard SJ, Smeak DD, Fossum TW. Results of thoracic
duct ligation in dogs with chylothorax. J Am Vet Med Assoc 1988;193:
68–71.
10. Monnet E. Pleura and pleural space. In: Slatter D (ed): Textbook
of small animal surgery, 3rd ed. Philadelphia: WB Saunders, 2002;
387–405.
11. MacDonald NJ, Noble PJM, Burrow RD. Efficacy of en bloc
ligation of the thoracic duct: descriptive study in 14 dogs. Vet Surg
2008;37:696–701.
12. Viehoff FW, Stokhoff AA. En bloc ligation of the thoracic duct in
twelve dogs with idiopathic chylothorax. Tijdschr Diergeneeskd
2003;128:278–283 (in Dutch).
13. Carobbi B, White RAS, Romanelli G. Treatment of idiopathic
chylothorax in 14 dogs by ligation of the thoracic duct and partial pericar-
dectomy. Vet Rec 2008;163:743–745.
14. Birchard SJ, Cantwell HD, Bright RM. Lymphangiography and li-
gation of the canine thoracic duct: a study in normal dogs and three dogs
with chylothorax. J Am Anim Hosp Assoc 1982;18:769–777.
15. Kagan KG, Breznock EM. Variations in the canine thoracic duct
system and the effects of surgical occlusion demonstrated by rapid aqueous
lymphography, using an intestinal lymphatic trunk. Am J Vet Res 1979;40:
948–958.
16. Bezuidenhout AJ. The lymphatic system. In: Evans HE (ed): Miller’s
anatomy of the dog, 3rd ed. Philadelphia: WB Saunders, 1993;717–757.
17. Marais J, Fossum TW. Ultrastructural morphology of the canine
thoracic duct and the cisterna chyli. Acta Anat (Basel) 1988;133:309–312.
301
18. Bannykh S, Mironov A, Jr, Bannykh G, Mironov A. The morphol-
ogy of valves and valve-like structures in the canine and feline thoracic duct.
Anat Embryol 1995;192:265–274.
19. De Freitas V, Piffer CR, Zorzetto NL, Seullner G, Martins SR. On
the topography of the ductus thoracicus in the dog. Anat Anz 1981;149:451–
454.
20. Lob G, Seifert J, Hauck R, Lissner J, Brendel W. Lymph flow fol-
lowing ligation of the ductus thoracicus in dogs. Anat Anz 1975;137:134–137
(in German).
21. Esterline ML, Radlinsky MG, Biller DS, Mason DE, Roush JK,
Cash WC. Comparison of radiographic and computed tomography lymph-
angiography for identification of the canine thoracic duct. Vet Radiol
Ultrasound 2005;46:391–5.
22. Radlinsky MG, Mason DE, Biller DS, Olsen D. Thorascopic
visualisation and ligation of the thoracic duct in dogs. Vet Surg 2002;31:
138–146.
23. Brisson BA, Holmberg DL, House M. Comparison of mesenteric
lymphadenography performed via surgical and laparoscopic approaches in
dogs. Am J Vet Res 2006; 67:168–173.
24. Martin RA, Leighton D, Richards S, Barber DL, Cordes DO, Sufit
E. Transdiaphragmatic approach to the thoracic duct ligation in the cat. Vet
Surg 1988;17:22–26.
25. Sicard GK, Waller KR, McAnulty JF. The effect of cisterna chyli
ablation combined with thoracic duct ligation on abdominal lymphatic
drainage. Vet Surg 2005;34:64–70.
26. Fossum TW, Birchard SJ. Lymphangiographic evaluation of exper-
imentally induced chylothorax after ligation of the cranial vena cava in dogs.
Am J Vet Res 1986;47:967–971.
27. Pardo AD, Bright RM, Walker MA, Patton CS. Transcatheter tho-
racic duct embolisation in the dog. An experimental study. Vet Surg
1989;18:279–285.
28. Naganobu K, Ohigashi Y, Akiyoshi T, Hagio M, Miyamoto T,
Yamaguchi R. Lymphography of the thoracic duct by percutaneous injec-
tion of iohexol into the popliteal lymph node of dogs: experimental study
and clinical application. Vet Surg 2006;35:377–381.
29. Johnson EG, Wisner ER, Kyles A, Koehler C, Marks SL. Com-
puted tomographic lymphography of the thoracic duct by mesenteric lymph
node injection. Vet Surg 2009;38:361–367.
30. Deneuche AJ, Dufayet C, Goby L, Fayolle P, Desbois C. Analgesic
comparison of meloxicam or ketofen for orthopedic surgery in dogs. Vet
Surg 2004;33:650–660.
31. Bornscheuer A, Mahr KH, Botel C, Goldmann R, Gnielinski M,
Kirchner E. Cardiopulmonary effects of lying position on anesthetized and
mechanically ventilated dogs. J Exp Anim Sci 1996;38:20–27.
32. Scrivani PV, Yeager AE, Dykes NL, Scarlett JM. Influence of pa-
tient positioning on sensitivity of mesenteric portography for detecting
anomalous portosystemic blood vessels in dogs: 34 cases (1997–00). J Am
Vet Med Assoc 2001;219:1251–1253.