Hedvig Hricak, Emil A. Tanagho,
MD MD
Female
#{149} Eduardo #{149} Jack
Secaf, MD W. McAninch,
Urethra:
#{149} David
MR
The potential of magnetic resonance (MR) imaging in the evaluation of the female urethra was studied in 64 patients. Spin-echo Ti- and T2weighted images were obtained in all 64 patients, and contrast-enhanced Ti-weighted images were also obtained in 27 patients. Urethral pathologic conditions, established with urethroscopy or histologic examination, or both, included urethral diverticula, inflammatory granuloma, and primary and metastatic neoplasms. On T2-weighted images, all normal urethras demonstrated a characteristic targetlike appearance with differentiation among the outer ring of low signal intensity, the middle zone of higher signal intensity, and the center of low signal intensity. After injection of gadopentetate dimeglumine, the targetlike appearance of the normal urethra was seen on the Ti-weighted images. Urethral diverticula were detected with MR imaging in all nine patients with that diagnosis, and in each, MR imaging demonstrated urethral expansion, distortion of the zonal anatomy, and presence of fluid in the middle zone. Primary or metastatic urethral neoplasms were also detected with MR imaging in every patient with the diagnosis, but differentiation between benign and malignant disease was not possible. Local staging of primary or metastatic malignant disease was correct in eight of the ii patients. In three patients, the inflammatory changes could not be differentiated from tumor invasion, resulting in overestimation of tumor extent.
T
ies,
Genitourinary
851.1214
Urethra,
#{149} Urethra,
diverticula,
studies,
851.1214
851.322,
851.33
system, abnormalities,
851.1491 Urethra,
#{149}
Urethra,
#{149}
neoplasms,
MR stud-
clinical and radiologic evaluation of the female urethra is difficult. Symptoms of urethral diverticula and tumors often mimic those of urinary tract infection, and tumors may not be diagnosed until late in their course (i-4). Urethral cancinoma, therefore, is often at an advanced stage when detected, and therapy decisions are further hampered by limited information about tumor extent provided by either physical examina-
pathologic patients. clinically
tion
ing
or
radiographic
tomographic
1991;
178:527-535
computed
(CT)
MATERIALS
AND
conditions urology verticula mors
44.2)
in
patients
were
included
years,
and
eight
aged
29-59
or physical
thral thras
mass. were
tients with
who had MR imaging
I
The patients selected from
were
From
the
tute
Departments
of Radiology (EAT.,
Urology
CA
partment
94143-0628;
Hospital.
of 12,
sion
received C
RSNA,
St. the
Box (J.J.B.);
San
Francisco
Francisco
(J.W.M.).
revision
requested
1990;
September
reprint 1991
requests
27;
Uni-
School 0628,
San
Mablincknodt
St Louis
Urology,
San
April
Francisco,
Parnassus
of Radiology,
MR
San
(H.H.,
J.W.M.),
of Fran-
Instiand
the
De-
General Received
accepted
to H.H.
June
19; reviOctober
urethral
tumors
(average,
39.7)
in
years.
and
were
subsequently
urobogic evaluation. At clinical and uation, ticula.
nine Seven
surgical
patients of them
pected urethral toms of recurrent and a suggestion
referred pathologic
had had
diverticula urinary of urethral
cal examination. Urethroscopy rography (double-balloon demonstrated
five
the
of the
tients,
seven
both
studies
and
those
urethstudy)
catheter diverticula
but
failed
pa-
was consymptomatwith an
in-
diverticulum no
was used therapy
clinically
associated
in
in two
to demonstrate
of urethral
imaging, urethroscopy the diagnosis, but
were
diversus-
with symptract infection mass at clini-
urethral
patients,
had
eval-
urethral clinically
any anomalies. The diagnosis firmed at surgery in all seven ic patients. In the two patients finding
for
to
suspected
urethra. among with
The these
pa-
infec-
flow
of urine
clinically
(n palpable
2). A urethral mass in 10 of 1 1 patients.
was
Before MR imaging, all 1 1 patients underwent transurethral biopsy, the results of which confirmed malignant tumors in
Urethral
and
of California, 505
review.
tu-
63.3)
tion of the urinary tract (n = 6), urethral bleeding (n = 3), and obstruction to the
of ure-
with normal urea pool of pa-
for
di-
Two patients had an incidental finding of a urethral pathologic condition (suggestive of urethral diverticula) at MR imag-
tients
undergone a pelvic study and whose medical
available
D.W.B.)
versity
findings
28 had
urethral
(average,
secondary
in
from the urethral aged 34-51
primary
34-76
primary carcinoma of the most common symptoms
in
the study. The normal urethra was assessed in 36 patients aged 25-77 (average, 43.9) years with no recorded urinary symptoms
was evaluated of these patients urethral pathologic
years,
1 i aged
was instituted. Eleven patients
Sixty-four
ES.,
#{149}
and were referred clinic with suspected in seven patients
(average,
at MR confirm
METHODS
Patients
charts
MD
condition Twenty-six suspected
cidental
2. Address
Radiology
or
findings (5). The normal appearance of the female urethra by magnetic resonance (MR) imaging has been described (6,7). This study was designed to (a) evaluate the appearance of the normal female urethra on MR images with and without enhancement by gadopentetate dimeglumine and (b) evaluate the potential of MR imaging in the diagnosis of various urethral pathologic conditions and staging of urethral tumors.
851.214. 851.321,
J. Brown,
#{149} Jeffrey
HE
disco,
terms:
MD
Imaging’
Medicine,
Index
W. Buckley,
MD
nine patients and only chronic inflammation and inflammatory granulation tissue in two patients. Nine of the 1 1 patients (including one of two with inflammatory granuloma) underwent surgery. The othen patient tory disease
with biopsy-proved inflammaunderwent a 12-month clini-
cal and MR imaging follow-up remaining patient (with primary malignancy) underwent radiation only. The secondary urethral tumors, ated in eight patients, included
study. The urethral thera-
py
two
evabumeta-
527
b.
a. Figure
1. differentiation 20)
Transa.xial among
image
taken
among
after
an outer
case,
the
Figure
injection
ring
low-signal-intensity
T2-weighted
female rectum
of gadopentetate
bow-signal-intensity
outer
2.
images of the normal urethra, vagina, and
(2,000/60)
dimeglurnine
(large
ring
C.
urethra through is not possible. arrow),
is complete.
(C),
the middle third. On the T2-weighted the
urethra
On the nonenhanced (2,000/60) image
demonstrates zone,
a higher-intensity middle V vagina, R rectum.
and
a tangetlike a central
Ti-weighted (500/20) (b) and on the Ti-weighted
appearance low-intensity
image (a), (500/
with differentiation dot (small arrow).
In this
transax-
ial images of the normal female urethra at the level of the proximal (a) and distal (b) third of the urethra demonstrate the targetlike appearance. The low-signal-intensity muscle (arrow) is thicker in the proximal urethra than elsewhere. B = urinary bladden, V = vagina.
static tumors-one conjunctiva and
noma-and
six tumors
sion-four and The
from
two two
with
and
derwent
tumor
cinoma diagnosis
cell
The
derwent
radiation
In summary,
cal and
MR
underwent
Imaging
the
underwent
imaging
follow-up, therapy.
For
Ti-weighted
the matrix
528
imaging magnet Milwaukee),
images,
sequences were time msec/echo
500-800/20 time msec).
size
of the
varied size
was
Radiology
#{149}
pelvic
between 192
the (nepetiFor T2-
pulse sequences Depending the
34 and
40 cm. The and
5-mm-
gap)
in both
of
transaxial
and sagittal (T2imaged for every
complete low-signal-intensity each of these three levels; of the urethra in the sagittal
patient. In one patient, additional coronal (T2-weighted) images were obtained. In 27 patients, transaxial Ti-weighted images were obtained immediately after the
the cal
intravenous
(d)
injection
Image
Analysis
Normal
Urethra
On MR thra shows appreciated (6).
on
girdle, X 256,
field
(20%
T2-weighted) planes were
of 0.1
opentetate dimeglumine body weight (Magnevist; Wayne, NJ).
before imaging, 1 mg of glucagon administered intramuscularly to each
weighted images, the were 2,000-2,500/50-80. view
seven
Techniques
patient. pulse tion
clini-
and
sections
(Ti- and weighted)
primary or 1 1 underwent
radiation
b.
thick
i9 patients
a 12-month
a.
the un-
therapy.
All 64 patients underwent with a i.5-T superconductive (Signa; GE Medical Systems,
and was
di-
car-
or vagina all had with biopsy and
among
one
unwith
an adjacent
with clinically suspected metastatic urethral tumors, surgery,
vagina. lesions
patients
from
of the cervix confirmed
cervix
carcinoma)
six
invasion
exten-
of the of the metastatic
renal
surgery.
of the carci-
cell
direct
carcinoma
from carcinoma patients with
(melanoma rect
from melanoma from renal
one
urethra
The
mmol
of gad-
per kilogram of Berlex Imaging,
following
of a targetlike mal, middle, thra; (Ii) the
meatus pubis
located inferior at the level of the
the presence signal-intensity
Urethral
urethral following
corded: location size, including
features evaluated:
of the (a)
the
normal
presence
appearance of the proxiand distal thirds of the unepresence of a complete or in-
the
MR
signal
T2-weighted Ti-weighted and the to local
to the labia
symphysis minora;
its
was
of the lesion; maximum
intensity
on
of the and
from
the
pathology
detectwere me-
the lesion diameter; the
Ti-
non-contrast-enhanced contrast-enhanced extent invasion
and or
Conditions abnormality characteristics
larged lymph nodes. tern for use with MR oped
base (urethrovesiof the external
of any morphologic abnormality.
Pathologic
When ed, the
images, the normal female urea targetlike appearance best on T2-weighted axial images were
level of the bladder junction) to the level
ring at (c) the length plane from
and and
images;
disease presence
with respect of en-
A tumor imaging
staging syswas devel-
staging
February
system
1991
4.
5.
Figures
3-5. (3) T2-weighted (2,000/60) transaxiab image of the normal female urethra through the middle of the urethra. The outer lowsignal-intensity ring (arrow) is incomplete posteriorly. (4) T2-weighted (2,000/60) sagittab image of a urethra of normal length (4.0 cm). The urethral length was measured from the bladder neck (long arrow) to the level of the labia minora located inferior to the pubic ramus and clitoris (c) (short arrow). The bladder neck is in normal position. Note the difficulties in defining the level of the external urethral meatus. B = urinary bladder, Li corpus uterus. (5) T2-weighted (2,000/60) sagittal image of a short urethra (2.2 cm) in a 65-year-old woman. Bladder neck (arrow), bow in position, is located at the bevel of inferior ramus of pubis. L leiomyoma, T cervical tumor.
patients, the lower third of the unethra was not included on transaxial images. In all patients, the urethra demonstrated a homogeneous medium-intensity signal (similar to that of striated muscle) on nonenhanced Ti-
weighted
images
(Fig
weighted
images
the
of the
urethra
1). On T2zonal
could
anatomy
be appreciated,
but its appearance varied depending on the level: proximal, middle, on distal third of the urethra. In the proximal third of the urethra, a targetlike appearance was seen in 58%
of patients, cally
In the
proposed
The
by
Grabstald
biostatistical
et al (4)
analysis
evaluation of accuracy suits [TP] + true-negative number of cases), [TP tive
(true-positive results sensitivity
+ false-negative
(TN/[TN
+
itive results]), value (TN/[TN
re-
specificity results]),
value
i).
the [TN]/ (TP/
results]),
false-positive
predictive
(Table
included
(TP/[TP
posi-
+ false-pos-
and negative predictive + false-negative results]).
Contrast-enhanced
Images
Gadolinium-enhanced MR images, obtamed in 27 patients, were evaluated for the appearance of the normal urethra (n 17), urethral diverticuba (n 2), inflammatory granuloma (n i), and primary (n = 3) and metastatic (n 4) cancer. Except for the group of patients with normal urethras, subgroup analysis
Volume
the number of patients in was too small for statistical of the
178
advantages
Number
#{149}
or
2
limitations
each
of the use of gadopentetate In these groups, therefore, tion
of the findings Signal intensity
dimeglumine. only a descnip-
is presented. was assessed
ments regions before
were obtained on operator-defined of interest in Ti-weighted images and after injection of gadopentedimeglumine.
means.
both
and
tate
quantitative
with
visual
Quantitative
expressed as a percentage hancement determined % contrast enhancement trast SI precontnast)/SI where
SI
=
signal
Measure-
data
of contrast by the equation (SI postconprecontrast,
en-
intensity.
RESULTS Normal
Urethras
Nonenhanced Images Diagnostic studies were in 33 of 36 patients. In six
performed of these
33
middle
theme
was
correlation
third
of the
no statistiwith
age.
urethra,
a
tangetlike appearance with a cleanly defined low-intensity central dot was seen in 80% of all women studied and in iOO% of women aged less than 40 years. In the distal third of the urethra, a targetlike appearance was seen in only 30% of patients, again with no statistical correlation with age. The outer low-signal-intensity muscle ring in the proximal third of the urethra was the thickest ring and was complete in all of the patients (Fig 2). The muscle ring in the mid-
die
were
and
significant
third
of the
urethra
was
complete
81% of patients (Fig i). It was indistinct posteriorly in the remaining i9% of patients (Fig 3). The muscle ring in the distal third of the urethra was the thinnest (Fig 2), the complete ring being seen in only 27% of the patients. T2-weighted images obtained in the sagittal plane showed the urethra with medium-low signal intensity, and the zonal anatomy was not appneciated. Urethral length ranged from 2 to 4.5 cm (median, 3.5 cm) (Fig in
Radiology
529
#{149}
Figure 6. Tl-weighted (500/20) images of the normal urethra immediately after injection of gadopentetate dimeglumine in two different patients, both with no recorded urinary symptoms. (a) The urethra has a targetlike appearance. with clear dernonstration of the central low-signal-intensity dot. V = vagina. (b) The central region of the urethra ed by urethral
has high a low-intensity crest.
4). The
signal
measured
shorter theme
in older were too
more
than
intensity ring. Arrow
length
surroundindicates
tended
to be
patients (Fig 5), but few patients aged
60 years
for statistical a.
analysis. Con trast-enhanced After injection meglumine, the
ance
of the
urethra
Ti-weighted
outer sual The only
Images of gadopentetate tangetlike appear-
could
images
ring
showed
ic).
showed
on
The
no appreciable
enhancement of signal quantitative enhancement 7.0% ± 3. The middle
however,
di-
be seen
(Fig
vi-
intensity. was zone,
marked
enhance-
ment of signal intensity (i46% ± 2i.7), compared with that seen on nonenhanced images. The central low-intensity dot showed no enhancement in i4 of 17 patients and enhancement similar to or greater than that of the middle region of the urethra in the remaining three patients (Fig 6). In two of these three patients, a thin, low-intensity ring was seen surrounding the high signal intensity of the central portion. In all of these three patients, the lowsignal-intensity dot was not appreci-
ated
on T2-weighted
Urethral
Conditions
Diverticula Urethral diverticula nosed in nine patients,
with
clinical
sev-
symptoms
ring
of homogeneous signal (Fig 7), and in the other Ti-weighted images an area of lower signal within the enlarged urethra of
lower
preserved 8b).
images, signal
in each
An
area
within
the
inthree inten(Fig
the
intensity
patient
imdi-
Ti-weightan enlarged
urethra tensity patients, showed
T2-weighted
outer was
(Figs middle
7b, por-
tion of the urethra was enlarged and appeared as a region of higher signal intensity (signal intensity similar to
530
Radiology
#{149}
of urethral diverticula. R = rectum, V vagina, B transaxiab image shows enlarged urethra of homogeneous (2,000/60) transaxial image shows preservation of the
weighted (500/20) ty. (b) T2-weighted intensity
c.
MR images
7.
ring
of
the
urethra.
An
area
of
high
signal
intensity
=
surrounds,
180#{176}, the central low-signal-intensity dot (arrow). (c) T2-weighted shows the anterior and posterior diverticula separated by a central now). The diverticulum is located in the middle third of the urethra, restricted by the vaginal attachments.
greater
than
the urinary The central
were diagof whom
In six of nine patients, ed images demonstrated
8). On
Figure
or
and two were asymptomatic. MR aging demonstrated the urethral verticulum in every case.
sity
b.
a.
bladder. (a) Tisignal intensiouter bow-signal-
for
approximately
(2,000/60) sagittab image bow-intensity lumen (anand its cranial extent is
images.
Pathologic
en presented
b.
intensity
that
of
bladder) urethral was
urine
within
ing
(Figs 7b, 8b). dot of low signal
present
in
four
of
nine
patients (Fig 7b). Urethral diverticula were located in the proximal third of the urethra in two patients (Fig 8), the middle third of the urethra in six patients (Fig 7), and the distal third
of the
urethra
in one
patient.
In four
diverticula
the cm
in On
ately
remain-
largest
diameter.
images
obtained
injection the
within
mained
of
low
the
depiction
ing
the
diagnosis.
or of
to
4.0
1.3 immedi-
urethral
tissue
the
whereas
the
ing
8) size
of gadopentetate
enhancement,
urine
types
the
the
after
(Fig The
from
the
showed
anterolateral
lateral
location. ranged
dimeglumine,
with
7) and
in
diverticula
patients, only a single urethral diverticulum was present. In the remaining five patients, two or more diverticula were seen. In three of these five patients, the diverticula were bilateral; in one patient, there were two diverticula both on the same side (Fig 8d); and in the remaining patient, three separate diverticula were detected. In reference to the urethral lumen, three diverticula were located (Fig
were
posterolateral
diverticula signal
ne-
mak-
intensity,
easy
and
facilitat-
Neoplasms Tumor
detection-Thirteen
suspected
metastatic
were
(n
studied.
was
detected
was
not
among
2)
In on
possible the (Figs
patients
primary
various
(ii
urethral
each case, MR images, to
11) on neoplasms
the lesion but it
differentiate malignant
tumor
9, 10) or between
malig-
February
1991
b. Figure
8.
intensity
MR images area (arrow)
c.
of urethral diverticula. in the urethral region.
B
bladder,
R
(b) T2-weighted
=
d.
rectum,
V
(2,000/60)
= vagina. transaxial
(arrow). (C) Following an injection of gadopentetate dimeglumine, a Ti-weighted but the fluid within the diverticulum (arrow) is not enhanced. (d) T2-weighted intensity strand within the diverticulurn (long arrow), indicating the presence sion of the diverticulum located in the middle and proximal third.
(a) Ti-weighted (500/20) image shows preservation
tnansaxiab image shows of the outer bow-intensity
a lowring
(500/20) transa.xiab image shows urethral enhancement, (2,000/60) image obtained in the sagittal plane shows of two diverticula. Small arrow indicates the cephalic
invasion malignancy,
from
a lowexten-
an adjacent primary tumor extension into
the urethra was correctly identified with MR imaging (Fig i2). In all of these patients, the T2-weighted images demonstrated interruption of the
low-signal-intensity
cular
ring.
In studies
hancement meglumine tients, the
a. Figure
9. MR images of squamous cell cancinoma of the urethra (MR imaging and surgical stage C3). T = tumor, R = rectum. In Ti-weighted (600/20) sagittab (a), proton-density (2,200/20) transaxial (b), and T2-weighted (2,200/70) tnansaxial (C) images, the large tumor is seen in the region of the urethra. Long arrow indicates Foley catheter. The tumor is extending into the retropubic adipose tissue (small arrows in b), and, as seen on the T2-weighted image (c), there is complete tumor invasion of the outer muscle ring in its anterior and right lateral portion. The left lateral pant of the muscular ring is not invaded, as it remains of bow signal intensity (open arrow). The bow signal intensity of the vaginal wall indicates it is not invaded. All findings were proved at surgery.
tumor
(Table aging
and
benign
granulation
each
tissue
(Fig ii). On nonenhanced Tiweighted images, all lesions were of low signal intensity and could not be differentiated from the periunethnal muscular
layer
or from
the
adjacent
low signal intensity of the vaginal wall. On T2-weighted ages, however, the tumors an increased signal intensity rupted the the normal
targetlike urethra.
of malignant nign
tumor
anterior imdisplayed and dis-
appearance The appearance
and
that
of be-
inflammatory
granulation
were similar on both weighted images.
Ti-
The mectly
tumor was MR imaging
Volume
location identified
178
of the with
Number
#{149}
2
and
of
tissue
T2conin
case
of nine
primary
tumors.
In
two patients, the tumor was located in the distal urethra, and in the memaining seven, the tumor was locat-
ed in the proximal The inflammatory was located in the The
maximum
was
between
on entire granulation proximal
diameter
2 and
urethra. tissue urethra.
of the
tumor
4 cm in six pa-
tients and greaten than 4 cm in three patients. Following injection of gadopentetate dimeglumine in three patients, the urethral tumors demonstrated signal enhancement and the normal tangetlike appearance of the urethra was disrupted but no additional information was obtained.
In each
of six patients
with
direct
informa-
local extent demonstrated
correctly
in eight
73%)
who
of on
of 1 i patients underwent
2). Overstaging occurred in two
(ac-
surgery
with MR imof three pa-
tients with early disease (stage B) because the complete low-signal-intensity ring of the peniunethral muscular layer could not be detected and infiltration of the vaginal wall could not
be excluded nant
no additional
staging.-The was
MR images curacy,
en-
provided.
Tumor
c.
b.
but
was
mus-
with gadopentetate diperformed in four patumor and urethra were
enhanced tion
outer
involving
(Fig
i3).
Stage
manifested by anterior through the peniumethral layer and obliteration
to the
symphysis
C disease,
spread muscular of fat posterior
pubis,
was
correctly
identified in five patients. In one patient, infiltration of the vaginal wall and extension into the vagina were correctly classified as stage C2 dis-
ease. correctly
Tumors
in four
classified
patients as stage
were C3 dis-
ease, based on demonstration mect extension of the proximally cated urethral tumor into the
of dilourinary
bladder. Manifestations of bladder involvement included wall thickening and increased signal intensity on T2-weighted images indicating contiguous spread of tumor (Fig 10).
Radiology
.
531
a. 10.
MR images
ed (2,000/60) ing the entire vaginal wall
of melanoma
extension in two cases
while
changes
only
congestion Metastatic
two
metastatic
(b) coronal images and urethra and extending itself is not invaded. V
False-positive interpreted and
c.
b.
Figure
vagina,
(500/20) urinary bladder R rectum.
(MR imaging (C)
and
T2-weighted (B). The tumor
and
d.
surgical
(2,000/60) demonstrates
stage
Di).
Ti-weighted
(600/20)
(a) and
T2-weight-
large mass (M) invad(arrows in C), but the
(d) transaxial images show right panavaginal extension
of tumor was (Figs i i, i3),
in inflammation
were disease
patients
Ti-weighted into the
to the urethra
with
seen at surgery. was identified
primary
and
in
meta-
static urethral carcinoma. Inguinal adenopathy was correctly identified in both patients with stage D disease, but in the patient with stage D2 dis-
ease,
there
lignant but not
were
pelvic seen
multiple
nodes on MR
other
found images.
ma-
at surgery
a. Figure
DISCUSSION Evaluation
of the
presents
challenging
diologic
problems.
female
urethra
clinical
and
Clinical
ma-
symp-
toms of both benign and malignant disease mimic each other, and the shortness of the urethra limits the accumacy of urethroscopy or radiographic studies. MR imaging consistently demonstrates urethral anatomy
that
so
differentiation
urethra, vagina, sue is easy (6,7).
Normal
and
among
b. 11.
MR images
of biopsy-proved
c.
granulation
tissue
(MR imaging
incorrect
diagno-
sis of urethral tumor at stage C3). T2-weighted (2,000/60) sagittal (a), Ti-weighted (500/20) transaxial (b), and T2-weighted (2,000/60) transaxial (c) images show localized enlangement of the upper third of the urethra with suggestion of tumor mass (T). On the sagittal image (a), the tumor appears to extend along posterior wall of bladder (B) (arrows), rendering false impression of MR imaging stage C3 disease. On transaxial T2-weighted image (C), the tumor demonstrates high signal intensity, and the normal targetbike appearance of the urethra is not seen. Although the direct tumor invasion of the vagina is not seen, it is believed that lack of low signal intensity in the vaginal wall indicates its invasion.
the
periunethnal
tis-
Urethra
On T2-weighted hanced Ti-weighted
on gadolinium-enimages, the
non-
mal urethra demonstrates a charactenistic tangetlike appearance. The features seen at MR imaging can be
correlated features
with of the
female
urethra
the
known
female
has
histologic
urethra.
three
The
histological-
ly distinct layers (8-13): the outer muscular layer, the middle submucosa (also known as the spongiose enec-
tile sity
tissue layer), and the mucosa. The outer ring of low signal intenseen at MR imaging most proba-
532
Radiology
#{149}
a. Figure
b. 12.
Squamous
cell carcinoma
C.
of the cervix
with
direct
tumor
extension
to the vagina
and urethra seen in transaxial Ti-weighted (550/20) (a), 12-weighted (2,200/70) (b), and gadolinium-enhanced Ti-weighted (550/20) (c) images. The urethra is displaced to the left. The vaginal wall is infiltrated by tumor. Tumor extended to the peniurethral tissue on the right and to a lesser extent on the left. Tumor also has invaded the right lateral wall of the urethra. The uninvolved anterior and left part of the urethral wall is demonstrated on the 12-weighted (b) and gadolinium-enhanced Ti-weighted (c) images (arrow).
February
1991
Table 2 Comparis
on of Surgical
and M R Imaging
in 11 Patients
Staging
MR Findings of Tumor Extension versus Surgery as Standard Patient Age (y)
Histologic
Diagnosis
56 49 59 51 34
Squamous Granulation Clean cell Transitional Squamous
cell tissue
58 76
Squamous Transitional
cell
69 69 64 45
Transitional Squamous Melanoma Squamous
cell cell
Note.-FN
Surgery Surgery Surgery Surgery Surgery
cell cell
and
irradiation
Surgery Surgery
cell
Surgery Surgery Surgery Surgery
cell
false-negative.
=
Management
FP
and and and
false-positive,
irradiation irradiation irradiation
TN
true-negative,
Surgical Stage
MR Imaging Stage
Fat
Urethra
Vagina
Bladder
Nodes
B 0 B B C2
B C3 Cl, C3 Ci,
TN TN TN FP TP
TP FP TP TP TP
TN FP FP F? TP
TN FP TN FP TN
TN TN TN TN TN
C3 C3
C3 C3
TP TN
TP TP
TN TN
TP TP
TN TN
C3 C3 Di D2
C3 C3 Di D2
TP FN TP TP
TP TP TP TP
TN TN TN TN
TP TP TP TP
TN TN TP TP
TP
=
C2 C2
true-positive.
ably corresponds to the submucosa, which contains an elaborate, rich vascular plexus intermixed with bundles of smooth muscle fibers and loosely woven connective tissue (8i3). This gives it a cavernous characten, and it is sometimes referred to as the spongiose enectile tissue of a fe-
male
Figure
13.
Transitional
cell carcinoma
of
the urethra (MR imaging stage C3 and sungical stage B) seen in sagittal 12-weighted (2,000/60) image. T tumor. The urethra is enlarged and of high signal intensity. While the upper two-thirds of the vagina remain of low signal intensity, there appears to be interruption
of
the
low-signal-intensity
stripe inferiorly (arrow), interpreted as invasion. bladder (B) demonstrates sity
with
which was falsely The base of the high
interruption
of
signal
bladder
sections
inten-
wall.
tion
This
was falsely interpreted as tumor invasion (stage C3). Only inflammatory changes found at surgery.
of smooth
muscle
the detrusor bladder) and
ed muscle.
The
striated
ring
Volume
178
Number
#{149}
2
is less
than
the anatomy discrepancy ma’s
mesolu-
problem
in
one
length at MR
value
of imag-
reported
in
literature: 4 cm (9). This may be due to the ureth-
slight
anterior
curvature
and
the
difficulty of identifying the distal end of the urethra in the sagittal plane of imaging.
with
prominent in the middle of the urethra and may be complete or incomplete posteriorly (i3). It also may be less prominent in postmenopausal women (i3). The middle layer of the urethra, which demonstrates a high signal intensity on T2-weighted images, probmost
spatial this
future. At 3.5 cm, the average the normal urethra seen
urinary of stniat-
muscle
improved
overcome
the were
muscle the muslayer of fibers
(contiguous
muscle of the an outer layer
and may
ing
bly corresponds to the outer layer (Fig 2). Histologically, cle layer consists of an inner longitudinal and thin circular
(8).
The central portion of the urethra appears on MR images as a low-signal-intensity dot that is related to mucosa consisting of folded stratified squamous, pseudostnatified columnan, or transitional epithelium, depending on which area of the urethra is being studied. The epithelium has compact cells and little extracellular space. The central dot of low signal intensity was most consistently seen in the middle of the urethra, but in 20% of patients with normal urethras, it was not present. The zonal anatomy of the urethra was not seen in the sagittal plane. This may be due to partial volume effect, and thinner
is
Urethral
Diverticula
MR imaging unique
value
appears in
the
to have diagnosis
a of
ure-
thral diventicula. Urethral diverticula are difficult to diagnose clinically and may go undetected even after urethroscopy, as was the case for two of the symptomatic patients in our series. When a urethral diverticulum
is infected, the symptoms may mimic those from tumor or infection of the bladder or urethra. Although a radiographic study with contrast material performed with a double-balloon catheter (i4,i5) is valuable for the diagnosis of urethral diventicula, it may be technically difficult. In two of our symptomatic patients, it failed to demonstrate the abnormality, whereas
MR imaging
demonstrated
the
le-
case. On nonenhanced T2-weighted MR images, the diverticula are seen as urethral enlargement with a middle zone of high signal intensity and an intact outer ring of low signal intensity. In our series, the outer low-signal-intensity ring, indicating preservation of the musculam layer of the urethra, was present in each case. However, it has been meported that the wall of urethral diverticula may be composed of vaginal mucosa (i6), and therefore this sion
MR cases
in
every
finding are
may
be altered
as more
studied.
The multiplanar capability of MR imaging allows demonstration of the relationship between the extension of the urethral diverticula and the vagina (14). As the distal two-thirds of the urethra are fused with the vagina (by the interdigitation of their surrounding muscular layers), the diverticula in this region, as shown in Figure 7, remain confined. Those located in the proximal urethra (the proximal third of the urethra and the vagina are not fused [i4]) extend cephalad (Fig 8). The diagnosis of urethral diverticula with MR imaging can be reinforced with injection of gadopentetate dimeglumine, after which urine within the diverticula remains
of
low
signal
intensity,
while the remaining urethral tissue is enhanced. Our images were always obtained immediately after injection, before the gadopentetate dimeglu-
Radiology
#{149} 533
mine could be excreted and change the signal intensity of the urine. Because of its accuracy, MR imaging can be used in select cases of suspected urethral diverticula in which clinical and radiographic evaluation fails to demonstrate the abnormality, yet symptoms of postmictunition dribbling and urinary tract infections persist.
Urethral
Carcinoma
Incidence Primary urethra 0.02% of women), women (i ,4, i 7). constitutes transitional
and Biologic Behavior carcinoma of the female is a mane lesion (less than all malignant tumors in which usually occurs in aged more than 40 years Squamous cell carcinoma 40%-75% of the cases, and cell carcinoma and ade-
nocarcinoma
are
the
next
most
com-
mon types (i-4). Melanoma is among the more common primary malignancies that metastasize to the unethra (i8). Clinical diagnosis of urethral canci-
noma
is difficult.
present those
Patients
often
with symptoms mimicking of chronic infection of the tract, and a mass lesion may
nary be appreciated
even
unnot
by an expeni-
enced clinician. Therefore, the diagnosis can be missed for a long time, and disease is often at an advanced stage by the time of diagnosis. In the largest reported series, 44 of 74 patients were first seen with advanced stage C and D disease (1). The main prognostic variables of urethral carcinoma are tumor size and location and stage of the disease at the time of diagnosis (i7). Because tumor size (maximum tumor diameten) has been reported to be an im-
pontant
factor
classification tens of less
greaten
for patient into
than
than
4 cm has
diamecm, and
been
suggested
(3). Determining important (i7). cal
correlations
tween
staging
tumor Although have
and
location no been
is also statisti-
made
location
be-
of ure-
thnal lesions, lesions confined to the distal urethra are generally of an earlien stage (i7). Once tumor location is determined, urethral carcinoma may be classified into two groups:
(a) proximal
(posterior)
tumor
in-
volving the portion of the urethra close to the bladder neck and (b) distal (anterior) urethral tumor confined to the distal third of the urethra (4). The tumors are often classified as “entire” when any portion of the urethra other than the distal one-
534
Radiology
#{149}
common at first examination reported to occur in iO%-15% tients (i,i7).
and are of pa-
appearance of benign granuloma and that of malignant tumor are similar
MR
on
Imaging Studies Radiologic studies presently available provide rather poor evaluation of urethral carcinoma. Primary lesions may be detected with umethrogmaphy, but this method fails to reveal local spread of disease and does not add to the information gained at cystoumethroscopy. CT has been used to
evaluate
local
spread
of tumor,
but
difficulty in differentiating the urethra from vagina and therefore detecting the involvement of the antenon vaginal wall has been a limitation of this modality (5). Further difficulties lie in the CT evaluation of the base of the urinary bladder because tumor and bladder wall have similar CT attenuation. Furthermore, the bladder base is difficult to evaluate in the axial plane of imaging.
data
are available
on ultra-
sound evaluation of urethral ma, probably because of the ty in transabdominal scanning structure located just posterior symphysis
pubis.
carcinodifficulof a to the
Endovaginal
scan-
ning may circumvent this problem. Evaluation of lymphadenopathy has been attempted with both CT and lymphangiognaphy, but again, limit-
ed data However,
are available because
lymph
node
the
in the literature. presence of
metastasis
is a grave
prognostic factor, lymphangiography has been advocated as a routine cedune before deciding on the therapy to be used (3).
MR imaging
may
play
images.
though tected in
our
any
an impor-
tant mole in the evaluation of urethral carcinoma, especially in staging. The value of MR imaging in the detection of urethral tumor is limited, as the
patients
(stage 0) A) disease. nosis
of
with
or only Once
and
stage.
Estimation
diameter three clinically
gories
(>2
nation
of
in
situ
(stage diagcarcinoma is obMR imaging can be tumor size, loca-
urethral
maximum within
either
submucosal the histologic
tamed, however, used to demonstrate tion,
al-
carcinoma was deimages in each patient the study did not in-
study,
dude
of
the
of the tumor proposed
cate-
cm, 2-4 cm,
MD MD
Female
#{149} Eduardo #{149} Jack
Secaf, MD W. McAninch,
Urethra:
#{149} David
MR
The potential of magnetic resonance (MR) imaging in the evaluation of the female urethra was studied in 64 patients. Spin-echo Ti- and T2weighted images were obtained in all 64 patients, and contrast-enhanced Ti-weighted images were also obtained in 27 patients. Urethral pathologic conditions, established with urethroscopy or histologic examination, or both, included urethral diverticula, inflammatory granuloma, and primary and metastatic neoplasms. On T2-weighted images, all normal urethras demonstrated a characteristic targetlike appearance with differentiation among the outer ring of low signal intensity, the middle zone of higher signal intensity, and the center of low signal intensity. After injection of gadopentetate dimeglumine, the targetlike appearance of the normal urethra was seen on the Ti-weighted images. Urethral diverticula were detected with MR imaging in all nine patients with that diagnosis, and in each, MR imaging demonstrated urethral expansion, distortion of the zonal anatomy, and presence of fluid in the middle zone. Primary or metastatic urethral neoplasms were also detected with MR imaging in every patient with the diagnosis, but differentiation between benign and malignant disease was not possible. Local staging of primary or metastatic malignant disease was correct in eight of the ii patients. In three patients, the inflammatory changes could not be differentiated from tumor invasion, resulting in overestimation of tumor extent.
T
ies,
Genitourinary
851.1214
Urethra,
#{149} Urethra,
diverticula,
studies,
851.1214
851.322,
851.33
system, abnormalities,
851.1491 Urethra,
#{149}
Urethra,
#{149}
neoplasms,
MR stud-
clinical and radiologic evaluation of the female urethra is difficult. Symptoms of urethral diverticula and tumors often mimic those of urinary tract infection, and tumors may not be diagnosed until late in their course (i-4). Urethral cancinoma, therefore, is often at an advanced stage when detected, and therapy decisions are further hampered by limited information about tumor extent provided by either physical examina-
pathologic patients. clinically
tion
ing
or
radiographic
tomographic
1991;
178:527-535
computed
(CT)
MATERIALS
AND
conditions urology verticula mors
44.2)
in
patients
were
included
years,
and
eight
aged
29-59
or physical
thral thras
mass. were
tients with
who had MR imaging
I
The patients selected from
were
From
the
tute
Departments
of Radiology (EAT.,
Urology
CA
partment
94143-0628;
Hospital.
of 12,
sion
received C
RSNA,
St. the
Box (J.J.B.);
San
Francisco
Francisco
(J.W.M.).
revision
requested
1990;
September
reprint 1991
requests
27;
Uni-
School 0628,
San
Mablincknodt
St Louis
Urology,
San
April
Francisco,
Parnassus
of Radiology,
MR
San
(H.H.,
J.W.M.),
of Fran-
Instiand
the
De-
General Received
accepted
to H.H.
June
19; reviOctober
urethral
tumors
(average,
39.7)
in
years.
and
were
subsequently
urobogic evaluation. At clinical and uation, ticula.
nine Seven
surgical
patients of them
pected urethral toms of recurrent and a suggestion
referred pathologic
had had
diverticula urinary of urethral
cal examination. Urethroscopy rography (double-balloon demonstrated
five
the
of the
tients,
seven
both
studies
and
those
urethstudy)
catheter diverticula
but
failed
pa-
was consymptomatwith an
in-
diverticulum no
was used therapy
clinically
associated
in
in two
to demonstrate
of urethral
imaging, urethroscopy the diagnosis, but
were
diversus-
with symptract infection mass at clini-
urethral
patients,
had
eval-
urethral clinically
any anomalies. The diagnosis firmed at surgery in all seven ic patients. In the two patients finding
for
to
suspected
urethra. among with
The these
pa-
infec-
flow
of urine
clinically
(n palpable
2). A urethral mass in 10 of 1 1 patients.
was
Before MR imaging, all 1 1 patients underwent transurethral biopsy, the results of which confirmed malignant tumors in
Urethral
and
of California, 505
review.
tu-
63.3)
tion of the urinary tract (n = 6), urethral bleeding (n = 3), and obstruction to the
of ure-
with normal urea pool of pa-
for
di-
Two patients had an incidental finding of a urethral pathologic condition (suggestive of urethral diverticula) at MR imag-
tients
undergone a pelvic study and whose medical
available
D.W.B.)
versity
findings
28 had
urethral
(average,
secondary
in
from the urethral aged 34-51
primary
34-76
primary carcinoma of the most common symptoms
in
the study. The normal urethra was assessed in 36 patients aged 25-77 (average, 43.9) years with no recorded urinary symptoms
was evaluated of these patients urethral pathologic
years,
1 i aged
was instituted. Eleven patients
Sixty-four
ES.,
#{149}
and were referred clinic with suspected in seven patients
(average,
at MR confirm
METHODS
Patients
charts
MD
condition Twenty-six suspected
cidental
2. Address
Radiology
or
findings (5). The normal appearance of the female urethra by magnetic resonance (MR) imaging has been described (6,7). This study was designed to (a) evaluate the appearance of the normal female urethra on MR images with and without enhancement by gadopentetate dimeglumine and (b) evaluate the potential of MR imaging in the diagnosis of various urethral pathologic conditions and staging of urethral tumors.
851.214. 851.321,
J. Brown,
#{149} Jeffrey
HE
disco,
terms:
MD
Imaging’
Medicine,
Index
W. Buckley,
MD
nine patients and only chronic inflammation and inflammatory granulation tissue in two patients. Nine of the 1 1 patients (including one of two with inflammatory granuloma) underwent surgery. The othen patient tory disease
with biopsy-proved inflammaunderwent a 12-month clini-
cal and MR imaging follow-up remaining patient (with primary malignancy) underwent radiation only. The secondary urethral tumors, ated in eight patients, included
study. The urethral thera-
py
two
evabumeta-
527
b.
a. Figure
1. differentiation 20)
Transa.xial among
image
taken
among
after
an outer
case,
the
Figure
injection
ring
low-signal-intensity
T2-weighted
female rectum
of gadopentetate
bow-signal-intensity
outer
2.
images of the normal urethra, vagina, and
(2,000/60)
dimeglurnine
(large
ring
C.
urethra through is not possible. arrow),
is complete.
(C),
the middle third. On the T2-weighted the
urethra
On the nonenhanced (2,000/60) image
demonstrates zone,
a higher-intensity middle V vagina, R rectum.
and
a tangetlike a central
Ti-weighted (500/20) (b) and on the Ti-weighted
appearance low-intensity
image (a), (500/
with differentiation dot (small arrow).
In this
transax-
ial images of the normal female urethra at the level of the proximal (a) and distal (b) third of the urethra demonstrate the targetlike appearance. The low-signal-intensity muscle (arrow) is thicker in the proximal urethra than elsewhere. B = urinary bladden, V = vagina.
static tumors-one conjunctiva and
noma-and
six tumors
sion-four and The
from
two two
with
and
derwent
tumor
cinoma diagnosis
cell
The
derwent
radiation
In summary,
cal and
MR
underwent
Imaging
the
underwent
imaging
follow-up, therapy.
For
Ti-weighted
the matrix
528
imaging magnet Milwaukee),
images,
sequences were time msec/echo
500-800/20 time msec).
size
of the
varied size
was
Radiology
#{149}
pelvic
between 192
the (nepetiFor T2-
pulse sequences Depending the
34 and
40 cm. The and
5-mm-
gap)
in both
of
transaxial
and sagittal (T2imaged for every
complete low-signal-intensity each of these three levels; of the urethra in the sagittal
patient. In one patient, additional coronal (T2-weighted) images were obtained. In 27 patients, transaxial Ti-weighted images were obtained immediately after the
the cal
intravenous
(d)
injection
Image
Analysis
Normal
Urethra
On MR thra shows appreciated (6).
on
girdle, X 256,
field
(20%
T2-weighted) planes were
of 0.1
opentetate dimeglumine body weight (Magnevist; Wayne, NJ).
before imaging, 1 mg of glucagon administered intramuscularly to each
weighted images, the were 2,000-2,500/50-80. view
seven
Techniques
patient. pulse tion
clini-
and
sections
(Ti- and weighted)
primary or 1 1 underwent
radiation
b.
thick
i9 patients
a 12-month
a.
the un-
therapy.
All 64 patients underwent with a i.5-T superconductive (Signa; GE Medical Systems,
and was
di-
car-
or vagina all had with biopsy and
among
one
unwith
an adjacent
with clinically suspected metastatic urethral tumors, surgery,
vagina. lesions
patients
from
of the cervix confirmed
cervix
carcinoma)
six
invasion
exten-
of the of the metastatic
renal
surgery.
of the carci-
cell
direct
carcinoma
from carcinoma patients with
(melanoma rect
from melanoma from renal
one
urethra
The
mmol
of gad-
per kilogram of Berlex Imaging,
following
of a targetlike mal, middle, thra; (Ii) the
meatus pubis
located inferior at the level of the
the presence signal-intensity
Urethral
urethral following
corded: location size, including
features evaluated:
of the (a)
the
normal
presence
appearance of the proxiand distal thirds of the unepresence of a complete or in-
the
MR
signal
T2-weighted Ti-weighted and the to local
to the labia
symphysis minora;
its
was
of the lesion; maximum
intensity
on
of the and
from
the
pathology
detectwere me-
the lesion diameter; the
Ti-
non-contrast-enhanced contrast-enhanced extent invasion
and or
Conditions abnormality characteristics
larged lymph nodes. tern for use with MR oped
base (urethrovesiof the external
of any morphologic abnormality.
Pathologic
When ed, the
images, the normal female urea targetlike appearance best on T2-weighted axial images were
level of the bladder junction) to the level
ring at (c) the length plane from
and and
images;
disease presence
with respect of en-
A tumor imaging
staging syswas devel-
staging
February
system
1991
4.
5.
Figures
3-5. (3) T2-weighted (2,000/60) transaxiab image of the normal female urethra through the middle of the urethra. The outer lowsignal-intensity ring (arrow) is incomplete posteriorly. (4) T2-weighted (2,000/60) sagittab image of a urethra of normal length (4.0 cm). The urethral length was measured from the bladder neck (long arrow) to the level of the labia minora located inferior to the pubic ramus and clitoris (c) (short arrow). The bladder neck is in normal position. Note the difficulties in defining the level of the external urethral meatus. B = urinary bladder, Li corpus uterus. (5) T2-weighted (2,000/60) sagittal image of a short urethra (2.2 cm) in a 65-year-old woman. Bladder neck (arrow), bow in position, is located at the bevel of inferior ramus of pubis. L leiomyoma, T cervical tumor.
patients, the lower third of the unethra was not included on transaxial images. In all patients, the urethra demonstrated a homogeneous medium-intensity signal (similar to that of striated muscle) on nonenhanced Ti-
weighted
images
(Fig
weighted
images
the
of the
urethra
1). On T2zonal
could
anatomy
be appreciated,
but its appearance varied depending on the level: proximal, middle, on distal third of the urethra. In the proximal third of the urethra, a targetlike appearance was seen in 58%
of patients, cally
In the
proposed
The
by
Grabstald
biostatistical
et al (4)
analysis
evaluation of accuracy suits [TP] + true-negative number of cases), [TP tive
(true-positive results sensitivity
+ false-negative
(TN/[TN
+
itive results]), value (TN/[TN
re-
specificity results]),
value
i).
the [TN]/ (TP/
results]),
false-positive
predictive
(Table
included
(TP/[TP
posi-
+ false-pos-
and negative predictive + false-negative results]).
Contrast-enhanced
Images
Gadolinium-enhanced MR images, obtamed in 27 patients, were evaluated for the appearance of the normal urethra (n 17), urethral diverticuba (n 2), inflammatory granuloma (n i), and primary (n = 3) and metastatic (n 4) cancer. Except for the group of patients with normal urethras, subgroup analysis
Volume
the number of patients in was too small for statistical of the
178
advantages
Number
#{149}
or
2
limitations
each
of the use of gadopentetate In these groups, therefore, tion
of the findings Signal intensity
dimeglumine. only a descnip-
is presented. was assessed
ments regions before
were obtained on operator-defined of interest in Ti-weighted images and after injection of gadopentedimeglumine.
means.
both
and
tate
quantitative
with
visual
Quantitative
expressed as a percentage hancement determined % contrast enhancement trast SI precontnast)/SI where
SI
=
signal
Measure-
data
of contrast by the equation (SI postconprecontrast,
en-
intensity.
RESULTS Normal
Urethras
Nonenhanced Images Diagnostic studies were in 33 of 36 patients. In six
performed of these
33
middle
theme
was
correlation
third
of the
no statistiwith
age.
urethra,
a
tangetlike appearance with a cleanly defined low-intensity central dot was seen in 80% of all women studied and in iOO% of women aged less than 40 years. In the distal third of the urethra, a targetlike appearance was seen in only 30% of patients, again with no statistical correlation with age. The outer low-signal-intensity muscle ring in the proximal third of the urethra was the thickest ring and was complete in all of the patients (Fig 2). The muscle ring in the mid-
die
were
and
significant
third
of the
urethra
was
complete
81% of patients (Fig i). It was indistinct posteriorly in the remaining i9% of patients (Fig 3). The muscle ring in the distal third of the urethra was the thinnest (Fig 2), the complete ring being seen in only 27% of the patients. T2-weighted images obtained in the sagittal plane showed the urethra with medium-low signal intensity, and the zonal anatomy was not appneciated. Urethral length ranged from 2 to 4.5 cm (median, 3.5 cm) (Fig in
Radiology
529
#{149}
Figure 6. Tl-weighted (500/20) images of the normal urethra immediately after injection of gadopentetate dimeglumine in two different patients, both with no recorded urinary symptoms. (a) The urethra has a targetlike appearance. with clear dernonstration of the central low-signal-intensity dot. V = vagina. (b) The central region of the urethra ed by urethral
has high a low-intensity crest.
4). The
signal
measured
shorter theme
in older were too
more
than
intensity ring. Arrow
length
surroundindicates
tended
to be
patients (Fig 5), but few patients aged
60 years
for statistical a.
analysis. Con trast-enhanced After injection meglumine, the
ance
of the
urethra
Ti-weighted
outer sual The only
Images of gadopentetate tangetlike appear-
could
images
ring
showed
ic).
showed
on
The
no appreciable
enhancement of signal quantitative enhancement 7.0% ± 3. The middle
however,
di-
be seen
(Fig
vi-
intensity. was zone,
marked
enhance-
ment of signal intensity (i46% ± 2i.7), compared with that seen on nonenhanced images. The central low-intensity dot showed no enhancement in i4 of 17 patients and enhancement similar to or greater than that of the middle region of the urethra in the remaining three patients (Fig 6). In two of these three patients, a thin, low-intensity ring was seen surrounding the high signal intensity of the central portion. In all of these three patients, the lowsignal-intensity dot was not appreci-
ated
on T2-weighted
Urethral
Conditions
Diverticula Urethral diverticula nosed in nine patients,
with
clinical
sev-
symptoms
ring
of homogeneous signal (Fig 7), and in the other Ti-weighted images an area of lower signal within the enlarged urethra of
lower
preserved 8b).
images, signal
in each
An
area
within
the
inthree inten(Fig
the
intensity
patient
imdi-
Ti-weightan enlarged
urethra tensity patients, showed
T2-weighted
outer was
(Figs middle
7b, por-
tion of the urethra was enlarged and appeared as a region of higher signal intensity (signal intensity similar to
530
Radiology
#{149}
of urethral diverticula. R = rectum, V vagina, B transaxiab image shows enlarged urethra of homogeneous (2,000/60) transaxial image shows preservation of the
weighted (500/20) ty. (b) T2-weighted intensity
c.
MR images
7.
ring
of
the
urethra.
An
area
of
high
signal
intensity
=
surrounds,
180#{176}, the central low-signal-intensity dot (arrow). (c) T2-weighted shows the anterior and posterior diverticula separated by a central now). The diverticulum is located in the middle third of the urethra, restricted by the vaginal attachments.
greater
than
the urinary The central
were diagof whom
In six of nine patients, ed images demonstrated
8). On
Figure
or
and two were asymptomatic. MR aging demonstrated the urethral verticulum in every case.
sity
b.
a.
bladder. (a) Tisignal intensiouter bow-signal-
for
approximately
(2,000/60) sagittab image bow-intensity lumen (anand its cranial extent is
images.
Pathologic
en presented
b.
intensity
that
of
bladder) urethral was
urine
within
ing
(Figs 7b, 8b). dot of low signal
present
in
four
of
nine
patients (Fig 7b). Urethral diverticula were located in the proximal third of the urethra in two patients (Fig 8), the middle third of the urethra in six patients (Fig 7), and the distal third
of the
urethra
in one
patient.
In four
diverticula
the cm
in On
ately
remain-
largest
diameter.
images
obtained
injection the
within
mained
of
low
the
depiction
ing
the
diagnosis.
or of
to
4.0
1.3 immedi-
urethral
tissue
the
whereas
the
ing
8) size
of gadopentetate
enhancement,
urine
types
the
the
after
(Fig The
from
the
showed
anterolateral
lateral
location. ranged
dimeglumine,
with
7) and
in
diverticula
patients, only a single urethral diverticulum was present. In the remaining five patients, two or more diverticula were seen. In three of these five patients, the diverticula were bilateral; in one patient, there were two diverticula both on the same side (Fig 8d); and in the remaining patient, three separate diverticula were detected. In reference to the urethral lumen, three diverticula were located (Fig
were
posterolateral
diverticula signal
ne-
mak-
intensity,
easy
and
facilitat-
Neoplasms Tumor
detection-Thirteen
suspected
metastatic
were
(n
studied.
was
detected
was
not
among
2)
In on
possible the (Figs
patients
primary
various
(ii
urethral
each case, MR images, to
11) on neoplasms
the lesion but it
differentiate malignant
tumor
9, 10) or between
malig-
February
1991
b. Figure
8.
intensity
MR images area (arrow)
c.
of urethral diverticula. in the urethral region.
B
bladder,
R
(b) T2-weighted
=
d.
rectum,
V
(2,000/60)
= vagina. transaxial
(arrow). (C) Following an injection of gadopentetate dimeglumine, a Ti-weighted but the fluid within the diverticulum (arrow) is not enhanced. (d) T2-weighted intensity strand within the diverticulurn (long arrow), indicating the presence sion of the diverticulum located in the middle and proximal third.
(a) Ti-weighted (500/20) image shows preservation
tnansaxiab image shows of the outer bow-intensity
a lowring
(500/20) transa.xiab image shows urethral enhancement, (2,000/60) image obtained in the sagittal plane shows of two diverticula. Small arrow indicates the cephalic
invasion malignancy,
from
a lowexten-
an adjacent primary tumor extension into
the urethra was correctly identified with MR imaging (Fig i2). In all of these patients, the T2-weighted images demonstrated interruption of the
low-signal-intensity
cular
ring.
In studies
hancement meglumine tients, the
a. Figure
9. MR images of squamous cell cancinoma of the urethra (MR imaging and surgical stage C3). T = tumor, R = rectum. In Ti-weighted (600/20) sagittab (a), proton-density (2,200/20) transaxial (b), and T2-weighted (2,200/70) tnansaxial (C) images, the large tumor is seen in the region of the urethra. Long arrow indicates Foley catheter. The tumor is extending into the retropubic adipose tissue (small arrows in b), and, as seen on the T2-weighted image (c), there is complete tumor invasion of the outer muscle ring in its anterior and right lateral portion. The left lateral pant of the muscular ring is not invaded, as it remains of bow signal intensity (open arrow). The bow signal intensity of the vaginal wall indicates it is not invaded. All findings were proved at surgery.
tumor
(Table aging
and
benign
granulation
each
tissue
(Fig ii). On nonenhanced Tiweighted images, all lesions were of low signal intensity and could not be differentiated from the periunethnal muscular
layer
or from
the
adjacent
low signal intensity of the vaginal wall. On T2-weighted ages, however, the tumors an increased signal intensity rupted the the normal
targetlike urethra.
of malignant nign
tumor
anterior imdisplayed and dis-
appearance The appearance
and
that
of be-
inflammatory
granulation
were similar on both weighted images.
Ti-
The mectly
tumor was MR imaging
Volume
location identified
178
of the with
Number
#{149}
2
and
of
tissue
T2conin
case
of nine
primary
tumors.
In
two patients, the tumor was located in the distal urethra, and in the memaining seven, the tumor was locat-
ed in the proximal The inflammatory was located in the The
maximum
was
between
on entire granulation proximal
diameter
2 and
urethra. tissue urethra.
of the
tumor
4 cm in six pa-
tients and greaten than 4 cm in three patients. Following injection of gadopentetate dimeglumine in three patients, the urethral tumors demonstrated signal enhancement and the normal tangetlike appearance of the urethra was disrupted but no additional information was obtained.
In each
of six patients
with
direct
informa-
local extent demonstrated
correctly
in eight
73%)
who
of on
of 1 i patients underwent
2). Overstaging occurred in two
(ac-
surgery
with MR imof three pa-
tients with early disease (stage B) because the complete low-signal-intensity ring of the peniunethral muscular layer could not be detected and infiltration of the vaginal wall could not
be excluded nant
no additional
staging.-The was
MR images curacy,
en-
provided.
Tumor
c.
b.
but
was
mus-
with gadopentetate diperformed in four patumor and urethra were
enhanced tion
outer
involving
(Fig
i3).
Stage
manifested by anterior through the peniumethral layer and obliteration
to the
symphysis
C disease,
spread muscular of fat posterior
pubis,
was
correctly
identified in five patients. In one patient, infiltration of the vaginal wall and extension into the vagina were correctly classified as stage C2 dis-
ease. correctly
Tumors
in four
classified
patients as stage
were C3 dis-
ease, based on demonstration mect extension of the proximally cated urethral tumor into the
of dilourinary
bladder. Manifestations of bladder involvement included wall thickening and increased signal intensity on T2-weighted images indicating contiguous spread of tumor (Fig 10).
Radiology
.
531
a. 10.
MR images
ed (2,000/60) ing the entire vaginal wall
of melanoma
extension in two cases
while
changes
only
congestion Metastatic
two
metastatic
(b) coronal images and urethra and extending itself is not invaded. V
False-positive interpreted and
c.
b.
Figure
vagina,
(500/20) urinary bladder R rectum.
(MR imaging (C)
and
T2-weighted (B). The tumor
and
d.
surgical
(2,000/60) demonstrates
stage
Di).
Ti-weighted
(600/20)
(a) and
T2-weight-
large mass (M) invad(arrows in C), but the
(d) transaxial images show right panavaginal extension
of tumor was (Figs i i, i3),
in inflammation
were disease
patients
Ti-weighted into the
to the urethra
with
seen at surgery. was identified
primary
and
in
meta-
static urethral carcinoma. Inguinal adenopathy was correctly identified in both patients with stage D disease, but in the patient with stage D2 dis-
ease,
there
lignant but not
were
pelvic seen
multiple
nodes on MR
other
found images.
ma-
at surgery
a. Figure
DISCUSSION Evaluation
of the
presents
challenging
diologic
problems.
female
urethra
clinical
and
Clinical
ma-
symp-
toms of both benign and malignant disease mimic each other, and the shortness of the urethra limits the accumacy of urethroscopy or radiographic studies. MR imaging consistently demonstrates urethral anatomy
that
so
differentiation
urethra, vagina, sue is easy (6,7).
Normal
and
among
b. 11.
MR images
of biopsy-proved
c.
granulation
tissue
(MR imaging
incorrect
diagno-
sis of urethral tumor at stage C3). T2-weighted (2,000/60) sagittal (a), Ti-weighted (500/20) transaxial (b), and T2-weighted (2,000/60) transaxial (c) images show localized enlangement of the upper third of the urethra with suggestion of tumor mass (T). On the sagittal image (a), the tumor appears to extend along posterior wall of bladder (B) (arrows), rendering false impression of MR imaging stage C3 disease. On transaxial T2-weighted image (C), the tumor demonstrates high signal intensity, and the normal targetbike appearance of the urethra is not seen. Although the direct tumor invasion of the vagina is not seen, it is believed that lack of low signal intensity in the vaginal wall indicates its invasion.
the
periunethnal
tis-
Urethra
On T2-weighted hanced Ti-weighted
on gadolinium-enimages, the
non-
mal urethra demonstrates a charactenistic tangetlike appearance. The features seen at MR imaging can be
correlated features
with of the
female
urethra
the
known
female
has
histologic
urethra.
three
The
histological-
ly distinct layers (8-13): the outer muscular layer, the middle submucosa (also known as the spongiose enec-
tile sity
tissue layer), and the mucosa. The outer ring of low signal intenseen at MR imaging most proba-
532
Radiology
#{149}
a. Figure
b. 12.
Squamous
cell carcinoma
C.
of the cervix
with
direct
tumor
extension
to the vagina
and urethra seen in transaxial Ti-weighted (550/20) (a), 12-weighted (2,200/70) (b), and gadolinium-enhanced Ti-weighted (550/20) (c) images. The urethra is displaced to the left. The vaginal wall is infiltrated by tumor. Tumor extended to the peniurethral tissue on the right and to a lesser extent on the left. Tumor also has invaded the right lateral wall of the urethra. The uninvolved anterior and left part of the urethral wall is demonstrated on the 12-weighted (b) and gadolinium-enhanced Ti-weighted (c) images (arrow).
February
1991
Table 2 Comparis
on of Surgical
and M R Imaging
in 11 Patients
Staging
MR Findings of Tumor Extension versus Surgery as Standard Patient Age (y)
Histologic
Diagnosis
56 49 59 51 34
Squamous Granulation Clean cell Transitional Squamous
cell tissue
58 76
Squamous Transitional
cell
69 69 64 45
Transitional Squamous Melanoma Squamous
cell cell
Note.-FN
Surgery Surgery Surgery Surgery Surgery
cell cell
and
irradiation
Surgery Surgery
cell
Surgery Surgery Surgery Surgery
cell
false-negative.
=
Management
FP
and and and
false-positive,
irradiation irradiation irradiation
TN
true-negative,
Surgical Stage
MR Imaging Stage
Fat
Urethra
Vagina
Bladder
Nodes
B 0 B B C2
B C3 Cl, C3 Ci,
TN TN TN FP TP
TP FP TP TP TP
TN FP FP F? TP
TN FP TN FP TN
TN TN TN TN TN
C3 C3
C3 C3
TP TN
TP TP
TN TN
TP TP
TN TN
C3 C3 Di D2
C3 C3 Di D2
TP FN TP TP
TP TP TP TP
TN TN TN TN
TP TP TP TP
TN TN TP TP
TP
=
C2 C2
true-positive.
ably corresponds to the submucosa, which contains an elaborate, rich vascular plexus intermixed with bundles of smooth muscle fibers and loosely woven connective tissue (8i3). This gives it a cavernous characten, and it is sometimes referred to as the spongiose enectile tissue of a fe-
male
Figure
13.
Transitional
cell carcinoma
of
the urethra (MR imaging stage C3 and sungical stage B) seen in sagittal 12-weighted (2,000/60) image. T tumor. The urethra is enlarged and of high signal intensity. While the upper two-thirds of the vagina remain of low signal intensity, there appears to be interruption
of
the
low-signal-intensity
stripe inferiorly (arrow), interpreted as invasion. bladder (B) demonstrates sity
with
which was falsely The base of the high
interruption
of
signal
bladder
sections
inten-
wall.
tion
This
was falsely interpreted as tumor invasion (stage C3). Only inflammatory changes found at surgery.
of smooth
muscle
the detrusor bladder) and
ed muscle.
The
striated
ring
Volume
178
Number
#{149}
2
is less
than
the anatomy discrepancy ma’s
mesolu-
problem
in
one
length at MR
value
of imag-
reported
in
literature: 4 cm (9). This may be due to the ureth-
slight
anterior
curvature
and
the
difficulty of identifying the distal end of the urethra in the sagittal plane of imaging.
with
prominent in the middle of the urethra and may be complete or incomplete posteriorly (i3). It also may be less prominent in postmenopausal women (i3). The middle layer of the urethra, which demonstrates a high signal intensity on T2-weighted images, probmost
spatial this
future. At 3.5 cm, the average the normal urethra seen
urinary of stniat-
muscle
improved
overcome
the were
muscle the muslayer of fibers
(contiguous
muscle of the an outer layer
and may
ing
bly corresponds to the outer layer (Fig 2). Histologically, cle layer consists of an inner longitudinal and thin circular
(8).
The central portion of the urethra appears on MR images as a low-signal-intensity dot that is related to mucosa consisting of folded stratified squamous, pseudostnatified columnan, or transitional epithelium, depending on which area of the urethra is being studied. The epithelium has compact cells and little extracellular space. The central dot of low signal intensity was most consistently seen in the middle of the urethra, but in 20% of patients with normal urethras, it was not present. The zonal anatomy of the urethra was not seen in the sagittal plane. This may be due to partial volume effect, and thinner
is
Urethral
Diverticula
MR imaging unique
value
appears in
the
to have diagnosis
a of
ure-
thral diventicula. Urethral diverticula are difficult to diagnose clinically and may go undetected even after urethroscopy, as was the case for two of the symptomatic patients in our series. When a urethral diverticulum
is infected, the symptoms may mimic those from tumor or infection of the bladder or urethra. Although a radiographic study with contrast material performed with a double-balloon catheter (i4,i5) is valuable for the diagnosis of urethral diventicula, it may be technically difficult. In two of our symptomatic patients, it failed to demonstrate the abnormality, whereas
MR imaging
demonstrated
the
le-
case. On nonenhanced T2-weighted MR images, the diverticula are seen as urethral enlargement with a middle zone of high signal intensity and an intact outer ring of low signal intensity. In our series, the outer low-signal-intensity ring, indicating preservation of the musculam layer of the urethra, was present in each case. However, it has been meported that the wall of urethral diverticula may be composed of vaginal mucosa (i6), and therefore this sion
MR cases
in
every
finding are
may
be altered
as more
studied.
The multiplanar capability of MR imaging allows demonstration of the relationship between the extension of the urethral diverticula and the vagina (14). As the distal two-thirds of the urethra are fused with the vagina (by the interdigitation of their surrounding muscular layers), the diverticula in this region, as shown in Figure 7, remain confined. Those located in the proximal urethra (the proximal third of the urethra and the vagina are not fused [i4]) extend cephalad (Fig 8). The diagnosis of urethral diverticula with MR imaging can be reinforced with injection of gadopentetate dimeglumine, after which urine within the diverticula remains
of
low
signal
intensity,
while the remaining urethral tissue is enhanced. Our images were always obtained immediately after injection, before the gadopentetate dimeglu-
Radiology
#{149} 533
mine could be excreted and change the signal intensity of the urine. Because of its accuracy, MR imaging can be used in select cases of suspected urethral diverticula in which clinical and radiographic evaluation fails to demonstrate the abnormality, yet symptoms of postmictunition dribbling and urinary tract infections persist.
Urethral
Carcinoma
Incidence Primary urethra 0.02% of women), women (i ,4, i 7). constitutes transitional
and Biologic Behavior carcinoma of the female is a mane lesion (less than all malignant tumors in which usually occurs in aged more than 40 years Squamous cell carcinoma 40%-75% of the cases, and cell carcinoma and ade-
nocarcinoma
are
the
next
most
com-
mon types (i-4). Melanoma is among the more common primary malignancies that metastasize to the unethra (i8). Clinical diagnosis of urethral canci-
noma
is difficult.
present those
Patients
often
with symptoms mimicking of chronic infection of the tract, and a mass lesion may
nary be appreciated
even
unnot
by an expeni-
enced clinician. Therefore, the diagnosis can be missed for a long time, and disease is often at an advanced stage by the time of diagnosis. In the largest reported series, 44 of 74 patients were first seen with advanced stage C and D disease (1). The main prognostic variables of urethral carcinoma are tumor size and location and stage of the disease at the time of diagnosis (i7). Because tumor size (maximum tumor diameten) has been reported to be an im-
pontant
factor
classification tens of less
greaten
for patient into
than
than
4 cm has
diamecm, and
been
suggested
(3). Determining important (i7). cal
correlations
tween
staging
tumor Although have
and
location no been
is also statisti-
made
location
be-
of ure-
thnal lesions, lesions confined to the distal urethra are generally of an earlien stage (i7). Once tumor location is determined, urethral carcinoma may be classified into two groups:
(a) proximal
(posterior)
tumor
in-
volving the portion of the urethra close to the bladder neck and (b) distal (anterior) urethral tumor confined to the distal third of the urethra (4). The tumors are often classified as “entire” when any portion of the urethra other than the distal one-
534
Radiology
#{149}
common at first examination reported to occur in iO%-15% tients (i,i7).
and are of pa-
appearance of benign granuloma and that of malignant tumor are similar
MR
on
Imaging Studies Radiologic studies presently available provide rather poor evaluation of urethral carcinoma. Primary lesions may be detected with umethrogmaphy, but this method fails to reveal local spread of disease and does not add to the information gained at cystoumethroscopy. CT has been used to
evaluate
local
spread
of tumor,
but
difficulty in differentiating the urethra from vagina and therefore detecting the involvement of the antenon vaginal wall has been a limitation of this modality (5). Further difficulties lie in the CT evaluation of the base of the urinary bladder because tumor and bladder wall have similar CT attenuation. Furthermore, the bladder base is difficult to evaluate in the axial plane of imaging.
data
are available
on ultra-
sound evaluation of urethral ma, probably because of the ty in transabdominal scanning structure located just posterior symphysis
pubis.
carcinodifficulof a to the
Endovaginal
scan-
ning may circumvent this problem. Evaluation of lymphadenopathy has been attempted with both CT and lymphangiognaphy, but again, limit-
ed data However,
are available because
lymph
node
the
in the literature. presence of
metastasis
is a grave
prognostic factor, lymphangiography has been advocated as a routine cedune before deciding on the therapy to be used (3).
MR imaging
may
play
images.
though tected in
our
any
an impor-
tant mole in the evaluation of urethral carcinoma, especially in staging. The value of MR imaging in the detection of urethral tumor is limited, as the
patients
(stage 0) A) disease. nosis
of
with
or only Once
and
stage.
Estimation
diameter three clinically
gories
(>2
nation
of
in
situ
(stage diagcarcinoma is obMR imaging can be tumor size, loca-
urethral
maximum within
either
submucosal the histologic
tamed, however, used to demonstrate tion,
al-
carcinoma was deimages in each patient the study did not in-
study,
dude
of
the
of the tumor proposed
cate-
cm, 2-4 cm,
