Ji-Bin Archie Philip
Liu, MD #{149}Larry S. Miller, MD A. Alexander, MD #{149}Laurence J. Klenn, MD #{149}Carl L. Millward,
Transnasal Morphologic
catheter-based ultratransducers can be inserted into the esophagus transnasally to evaluate esophageal wall structures. Studies were performed in two sheep esophagus specimens in vitro, in 17 healthy human subjects, and in 16 patients with esophageal abnormalities (eight with achalasia, four with scieroderma, three with esophageal carcinoma, and one with esophagitis). In the sheep specimens, endoluminal US delineated seven layers of the esophageal wall; these results correlated closely with histologic findings. Real-time US of the normal esophageal wall was performed during resting and swallowing. Muscles at the lower esophageal sphincter (LES) were shown to be thicker than muscles in the body of the esophagus. Thickening of the muscular layers at the LES in achalasia, dilated blood vessels within the submucosa in esophagitis, and fibrotic changes within the muscular layers in scieroderma were demonstrated. Extramural structures adjacent to the esophagus were also seen. These preliminary results suggest that transnasal esophageal US may become an important diagnostic tool in evaluation of the esophagus. Index terms: Achalasia, 71.745 #{149}Catheters and catheterization, technology #{149} Esophagitis, 71.291 #{149} Esophagus, diseases, 71.291, 71.321, 71.613, 71.745 . Esophagus, US, 71.12981 #{149} Sderoderma, 71.613 #{149}Ultrasound (US), technology 1992;
B. Goldberg,
MD
184:721-727
P
I From the Division of Diagnostic Ultrasound, Department of Radiology (J.B.L., B.B.G., R.I.F., A.A.A., L.N.), Department of Gastroenterology (L.S.M., D.O.C.), and Department of Pathology and Cell Biology (P.J.K., C.L.M.), Thomas Jef-
University
Hospital,
Rick I. Feld, MD 0. Castell, MD
#{149}
Seventh
Floor,
Main
Bldg. 132 S 10th St. Philadelphia, PA 19107-5244. Received August 9, 1991; revision requested October 7; revision received March 30, 1992; accepted April 7. Address reprint requests tOJ.B.L. C RSNA, 1992
Preliminary
of an endoluminab ultrasound (US) transducer into the rectum to evaluate the wall of the bowel (in an attempt to identify malignant tumors) was first reported by Wild and Reid in 1956 (1). Since that time, a number of researchers have been working on the further development of endoluminal sonography. In 1976, Lutz and R#{246}sch(2) reported using a US probe passed through the accessory channel of an endoscope. They obtained A-mode images to evaluate the cystic versus solid characteristics of lesions deforming the gastric wall. In the early 1980s, the concept of combining a flexible endoscope with a US transducer to image the gastrointestinal tract was developed and results were published by Fukuda et al (3) and DiMagno et al (4). With use of relatively high-frequency transducers (7-12.5 MHz), it was possible to image not only the walls of the gastrointestinal tract but, also structures adjacent to the tract LACEMENT
(5). This
approach
has
proved
useful
in preoperative diagnosis and staging of esophageal and gastric cancers (6,7). In 1989, Silverstein et al (8) deveboped a 20-MHz linear-array US probe that could be passed through the biopsy channel of a flexible endoscope to evaluate the gastrointestinal tract wall. More recently, ROsch and Classen (9) used a 7.5-MHz, 3.7-mmdiameter US probe in an attempt to image the esophagus, stomach, and duodenum by means of the working channel of a gastroscope. Over the past several years, flexible US transducer-containing (4.8-9 F), originally
ferson
MD Donald
#{149}
US ofthe Esophagus: and Function Studies’
High-frequency sound (US)
Radiology
#{149} Barry
Needleman, BS
designed
catheters for intra-
vascular US applications, have become available (10,11). During the past 2 years, these miniature catheter-based transducers have been used to evabuate nonvascular lumina, including the genitourinary tract, endometrial canab, bile ducts, and bronchial tree (1216). We report our use of these transducers to image the esophagus.
MATERIALS US
AND
METHODS
System The imaging
available Milpitas, catheters length)
system
used
in this study
is
commercially (IVUS; Diasonics, Calif). Specially developed 6.2-F (2 mm in diameter, 95 cm in containing 20-MHz transducers
are used
(Sonicath; Medi-tech/Boston SciWatertown, Mass) (Fig 1). The singbe-ebement US transducer is mounted on the end of a wire (core), which is conentific,
nected to a motor on the US instrument. The 1 .35-mm-diameter core is inserted into the flexible catheter, and the motor is rotated to produce a 360#{176} transaxiab real-
time image.
Sterile
introduced
eter
water
between
to eliminate
(0.5-1.0
the
core
air, which
mL) is
and
the
could
with transmission of the ultrasound The transducer sends and receives trasound pendicubar
cath-
interfere
beam. the ub-
signal at an angle 10#{176} from perto the long axis of the catheter
(Fig 2). The operating
frequency
of 20
MHz results in an axial resolution mm and a penetration of about
of 0.1 2.0 cm.
Real-time images are recorded on videotape for later evaluation, and individual frozen images are stored on a digital imager disk system (3M Medical Imaging Systems,
St Paul).
Animal
Studies
Whole sheep
esophagus
were
specimens
studied
in vitro.
specimens were placed The transducer-containing inserted
into
the
lumen
from the proximal was
advanced
ious
layers
imaged
two
fresh
in a water bath. catheter was of the
end.
from The
esophagus
As the catheter
at intervals of 3 cm, the varof the esophageal wall were and measured. The echogenicity
of the various
layers
of the esophagus
was
also evaluated. To eliminate near-field transducer artifacts, 15 mL of saline was injected into the lumen of the esophagus
and the imaging procedure After sonographic imaging, were fixed in 10% formalin
was repeated. the specimens and imbedded
CSM = circular smooth muscle, LES = lower esophageal sphincter, LSM = longitudinal smooth muscle, TM = total muscle. Abbreviations:
721
in paraffin. Histologic cross sections were prepared that corresponded to the US image
planes.
Finally,
tified
wall structures
those
depicted
Human
the histologically
were
iden-
correlated
with
on the US images.
Studies
Seventeen healthy volunteers (10 men and seven women, 21-85 years old with a mean age of 40 years) and 16 patients (seven men and nine women, 32-73 years old with a mean age of 43 years) were included in the study. Each of the 16 patients had a known diagnosis: Eight had achalasia, four had scleroderma, three had esophageal carcinoma, and one had reflux esophagitis. All four scberoderma patients had heartburn, and three had dysphagia. The time course of the disease was 1-15 years. The diagnosis of scleroderma was based on typical clinical features and results of serologic tests. All achalasia patients had their disease diagnosed on the basis of typical manometric findings (high residual pressures in the lower esophageal sphincter [LES] and absent peristabsis) and radiographic appearances (barium esoph-
agogram showing dilatation of the esophageal body and narrowing at the LES). Clinical symptoms included dysphagia, regurgitation,
cough. With
weight
the subject
a 16-F nasogastric through the was advanced
loss,
and
in a sitting
nocturnal
position,
tube was inserted
nose. As the nasogastric to the pharyngobaryngeal
tube
level, the subject was asked to drink 5-10 mL of water to aid the passage of the tube into the esophagus. As soon as the nasogastric tube entered the stomach (an event easily recognized by means of aspiration
of gastric ing
the
contents), US
through
the catheter
transducer
the
tube
was
contain-
1. II
Figures
1, 2.
(1) Endoluminal
US transducer
Figure
3.
Schematic
the catheter-based transnasally (NG) into
the
representation
transducer
(Fig 3). The
through the esophagus.
nasogastric
subject
were
slowly
withdrawn from the stomach to the LES. The esophageal wall was imaged during the resting state at the level of the LES and 5-10 cm above the LES. The subject was then asked to drink 5-10 mL of water
while real-time US was performed to evaluate swallowing. In the 17 healthy subjects and the eight achalasia patients, three images were selected from the level of the LES. In addi-
digitally
stored
subjects
three
the area
in a computer
im-
5-10
selected
cm
were
system
the CSM,
722
the LSM, and an intermuscubar tissue layer) was measured at octants by two independent U.B.L., L.S.M.). In the control
Radiology
#{149}
was determined by using a Pearson correbation coefficient. Intraobserver variability was determined by means of repeat measurement on the original images by each
independent
investigator
the first reading. carried out with
tion coefficient.
3 months
Analysis of the use of a Pearson
Muscle
layer
after
data was correla-
thicknesses
at the LES in control subjects were compared with muscle layer thicknesses at the LES in achalasia patients. The Student test was used for analysis of the data.
(Mi-
croSonic, Indianapolis). On each image, the thickness of the circular smooth muscle (CSM), longitudinal smooth muscle (LSM), and total muscle (TM) (including connective eight radial investigators
wire
pa-
the LES. Interobserver variability between the mean wall thickness measurements made by each independent investigator
tion, in the 17 healthy
of a drive
was
The US imaging tube
ages were selected from above the LES. All images
end
tube
manometric
nasogastnc
the
shows
patients, muscle layer thicknesses at the LES were compared with the thicknesses of the corresponding layers 5-10 cm above
examination.
on
(T) passed
tients, and the scleroderma patients, the distance from the nose to the LES region was previously established by means of and
is mounted
inserted
then placed in a supine position. For all volunteers, the achalasia
catheter
(arrow)
(core), which is housed in a 6.2-F catheter. Sterile water (0.5-1.0 mL) is injected with a 27gauge needle through the sealed catheter tip to eliminate any air around the transducer. (2) The transducer (T) rotates clockwise, producing a 360#{176} cross-sectional real-time US image at an angle of 10#{176} from the perpendicular. A, Longitudinal view; B, cross-sectional view corresponding to the scanning plane.
RESULTS Animal
Studies
With use of the 20-MHz catheterbased transducer, it was possible to delineate seven layers of the esophageab wall in the two sheep specimens.
Correlation with the innermost sented
of the histologic US images showed layer (hyperechoic) the
squamous
findings that the repre-
epithelium
September
and
1992
b. Figure
4.
(a) The
20-MHz
wall. (b) Close correlation
US transducer
between
C.
(T) located
within
the cross-sectional
the
saline-distended
sheep
esophagus
delineates
seven
layers
of the
esophageal
histologic
slice and the US image in a can be seen. (c) Schematic drawing identifies esophageal wall structures in cross section. In a-c, I = squamous epithelium and lamina propria, 2 = muscularis mucosae, 3 = submucosa, circular muscle, 5 = intermuscular connective tissue, 6 = longitudinal muscle, 7 = adventitia.
cosa),
a hyperechoic
hypoechoic
and
longitudinal),
intermuscular between hyperechoic
p.
Catheter-based US transducer (T) positioned within the LES region (a) and 8 cm above the LES (b) in a healthy volunteer delineates the various normal layers of the esophageal wall. Note that the thicknesses of the muscular layers in the LES region are different from the thicknesses of those in the body of the esophagus. 1 = mucosa, 2 = submucosa, 3 = circular muscle, 4 = intermuscular connective tissue, 5 = longitudinal muscle, 6 = adventitia. Figure
5.
the
adventitia Volume
(Fig 184
3
Subjects
In all control subjects, the crosssectional US images were obtained at various bevels, including the LES region and the middle and upper portions of the esophagus. There was a moderately echogenic mucosa (including
lamina
4).
#{149} Number
Control
the
squamous
propria,
and
epithelium,
muscularis
mu-
layers
a thin connective two muscle adventitia
4
=
two (circular
echogenic tissue
layers, (Fig
layer
and
a
5). Be-
cause the esophageal lumen in the resting state is collapsed and folded together, the muscubaris mucosae could not be clearly seen. When the esophageal lumen was distended with water, it was possible to delineate seven esophageal wall layers similar to those seen in vitro in the sheep esophagus. Measurements of muscle thickness at the LES in the control group are shown in Table 1. Measurements of muscle thickness 5-10 cm above the LES in the control group are shown in Table 2. The CSM, LSM, and TM were significantly thicker at the LES than in the esophageal body (P < .001 for all three comparisons) (Fig 5). Interobserver
lamina propria, the second layer (hypoechoic) the muscularis mucosae, the third layer (hyperechoic) the submucosa, the fourth layer (hypoechoic) the circular muscle, the thin fifth layer (hyperechoic) the intermuscular connective tissue, the sixth layer (hypoechoic) the longitudinal muscle, and the seventh layer (hyperechoic) the
submucosa,
muscular
the
variability
for
mea-
surements at the LES was r = .98 and P < .01 for the CSM, r = .73 and P < .01 for the LSM, and r = .98 and P < .01 for the TM. Interobserver variability for measurements in the body of the esophagus was r = .94 and P < .01 for the CSM, r = .67 and P < .01 for the LSM, and r = .97 and P < .01 for the TM. Intraobserver variability for measurements at the LES for investigator 1 was r = .99 and P < .01 for the CSM, r = .89 and P < .01 for the LSM, and r = .96 and P < .01 for the TM. Intraobserver variability for measurements in the body of the esophagus for investigator 1 was r = .94 and P < .01 for the CSM, r = .71 and P < .01 for the LSM, and r = .98 and P < .01 for the TM. Intraobserver variability for measurements at the LES for investigator 2 was r = .98 and P < .01 for the CSM, r = .84 and P < Radiology
#{149} 723
Figure
6.
(a) Catheter-based
(T) located within shows a hypoechoic x 0.49 cm (cursors), The predominantly an artery (pulsation real-time
imaging).
US transducer
the distal esophageal body structure measuring 0.91 which is a lymph node. anechoic area (arrows) is could be seen during (b) Cross-sectional
US
image obtained at the gastroesophageal lion shows a hypoechoic region (arrows) jacent to the esophageal wall, representing the diaphragm.
juncad-
.01 for the LSM, and r = .95 and P .01 for the TM. Intraobserver variability
for
measurements
in the
body