Dynamic MR Imaging of the Pelvic

Floor in Asymptomatic Subjects

Vicky Goh 1 OBJECTIVE. Dynamic MR imaging may be used as an alternative to dynamic cystoproc-
Steve Halligan 1 tography for the evaluation of pelvic floor prolapse and configuration. MR criteria for normal-
Glenda Kaplan 1 ity are derived from proctographic studies because no large MR study of asymptomatic
Jeremiah C. Healy 2 individuals has been performed. Our study aimed to define the normal range of dynamic pel-
vic MR appearances in a large group of asymptomatic individuals.
Clive I. Bartram 1
SUBJECTS AND METHODS. Fifty healthy adult volunteers (25 men and 25 women;
age range, 20–66 years; mean age, 34 years) were prospectively recruited and examined using
dynamic MR imaging. All subjects were interviewed and established as healthy using a vali-
dated questionnaire. Axial, coronal, and sagittal MR imaging was performed at rest and dur-
ing maximum pelvic strain using a static 1.0-T unit and a fast-field-echo sequence, providing
10 slices in 31 sec. Standardized measurements of pelvic configuration were taken, and rest
and strain imaging were compared to determine the range of normal appearances.
RESULTS. Three women developed a cystocele during maximum pelvic strain, two of
whom also showed grade 1 uterocervical prolapse, which was also seen in another woman.
Three men showed posterior pelvic floor descent in excess of 3 cm during straining. No recto-
cele, enterocele, rectal prolapse, or perineal hernia was seen in any subject.
CONCLUSION. The normal range of pelvic organ descent in asymptomatic subjects
seen on dynamic MR imaging included cystocele, uterocervical prolapse, and excessive ano-
rectal junction descent. Although we encountered pelvic prolapse in seven volunteers, it was
infrequent and low grade, suggesting that criteria for abnormality derived from proctography
are generally applicable to MR imaging.

Received May 24, 1999; accepted after revision
August 11, 1999.
Supported by Lister Bestcare, Selby, United Kingdom.
1
Intestinal Imaging Centre, Level 4V, St. Mark’s Hospital,
Northwick Park, Watford Rd., Harrow, London, HA1 3UJ,
United Kingdom. Address correspondence to S. Halligan.
2
Department of Radiology, Chelsea & Westminster
Hospital, Fulham Rd., London, SW10 9NH, United Kingdom.
AJR 2000;174:661–666
0361–803X/00/1743–661
© American Roentgen Ray Society
P
elvic floor weakness and conse-
quent organ prolapse may result in a
variety of symptoms, including
pain, urinary or fecal incontinence, and consti-
pation. The pelvic floor has traditionally been
divided into three compartments—anterior,
middle, and posterior—each attracting its own
specialist interest from urologists, gynecolo-
gists, and proctologists, respectively [1]. Clinical
examination either underestimates or inaccu-
rately diagnoses the site of prolapse in a signifi-
cant proportion of patients, and preoperative
imaging has assumed a prominent role because
of this [2]. Furthermore, it is now apparent that
pelvic floor weakness is usually generalized, so
that the various pelvic floor compartments are
best imaged simultaneously [3]. Conventionally,
this has been achieved using evacuation proc-
tography modified by additional opacification of
the bladder and small bowel, an examination
termed dynamic cystoproctography [4–6]. How-
ever, cystoproctography is a relatively invasive
procedure, involves ionizing radiation, and im-
ages only the lumen of the organs opacified. In
an attempt to overcome these limitations, dy-
namic pelvic MR imaging has been introduced
and is superseding fluoroscopic methods in
some centers [7–9]. An MR diagnosis of abnor-
mality has been based on concordance with es-
tablished proctographic findings. However,
proctographic studies of healthy subjects have
unexpectedly revealed findings often assumed
to be abnormal [10, 11]. To our knowledge, no
corresponding study has been performed for dy-
namic pelvic MR imaging. The aim of this pro-
spective study was to define the normal range of
dynamic pelvic MR appearances in a large
group of healthy volunteers.

AJR:174, March 2000 661


Subjects and Methods
Subjects
Our local ethics committee approved the study.
Fifty adult subjects, 25 men and 25 women, who
were 20–66 years old (mean, 34 years), were re-
cruited prospectively after giving informed written
consent. Seven women were parous (range, one to
three; median parity, two). Twenty-six subjects
were volunteers recruited from hospital personnel;
the remaining 24 were recruited from patients in
the MR unit for musculoskeletal examinations. All
subjects were interviewed before examination and
completed a questionnaire that detailed bowel fre-
quency; the need for laxatives; use of digital ma-
neuvers to assist evacuation; incontinence to urine,
gas, or feces; pelvic pain; organ prolapse; and any
need to consult a physician regarding these symp-
toms. Any abnormal response excluded the subject
from the study. Any patient with a history of pelvic
surgery was also excluded.
MR Imaging
MR imaging was performed with a 1.0-T static
unit (Gyroscan NT 1.0; Philips, Hammersmith,
United Kingdom). No preparation of the subject
was required. The subject lay supine on a water-
proof pad placed on the MR table. Maximum pel-
vic straining was practiced with the patient before
examination; patients were encouraged to bear
down as if emptying their bowels. Images were
obtained in sagittal, axial, and coronal orientations
using the body coil, first with the patient at rest
and then again during maximum pelvic strain. A
T1-weighted fast-field-echo sequence was used
with the following parameters: flip angle, 55°; TR/
TE, 79/240 msec; field of view, 34 cm; slice thick-
ness, 8 mm; interslice gap, 2 mm; matrix size, 256
× 256; and four excitations. This sequence gave 10
slices in 31 sec. The examinations were down-
loaded onto a dedicated workstation (Easyvision;
Philips) for radiologist review.
Image Analysis
All examinations were analyzed by two radiolo-
gists in consensus, who recorded standard measure-
ments of pelvic floor anatomy and also noted the
presence or absence of any structural abnormality.
The pubococcygeal line, defined as the line that
joined the inferior border of the pubic symphysis to
the last coccygeal joint [7, 11], was drawn onto the
midline sagittal resting image. The position of the
bladder base, cervix, and anorectal junction, defined
as the junction of the rectal ampulla and anal canal,
were measured at 90° to the pubococcygeal line.
The anorectal angle, defined as the angle between
the longitudinal axis of the anal canal and the poste-
rior rectal wall, and the levator plate angle, defined
as the angle between the levator plate and the pubo-
coccygeal line, were measured [7]. These values
were then compared with measurements taken from
the corresponding sagittal straining image, and the
change on straining was calculated. A cystocele was
diagnosed if the bladder base descended below the
662
Goh et al.
inferior border of the symphysis pubis [1, 5]. Utero-
cervical prolapse was defined as cervical descent
below the pubococcygeal line (grade 1), to the in-
troitus (grade 2), or to the exterior (grade 3). Ano-
rectal junction descent was defined as excessive if
more than 2.5 cm below the pubococcygeal line at
rest or if more than 3 cm on maximum strain [10–
12]. The axial images were used to calculate the
pelvic floor hiatal area and perimeter at rest and
during maximum strain [7, 8], which were mea-
sured at the level of the most inferior point of the
symphysis pubis. All images at rest and during
maximum strain were also evaluated for the pres-
ence or absence of any other structural abnormality
as follows: enterocele was defined as small bowel
within the rectovaginal septum that reached or
crossed the junction of the upper one third and dis-
tal two thirds of the vagina [13]. Rectocele was de-
fined as an anterior rectal wall bulge exceeding 2
cm [10, 11]. Rectal prolapse and any perineal hernia
through the levator plate were also noted if present.
Statistical Analysis
Descriptive statistics were performed on all
data at rest and during maximum strain. The Stu-
dent’s t test was used to compare parametrically
distributed continuous data, and statistical signifi-
cance was assigned to a p value of less than 0.05.
Calculations were performed using Arcus Quick-
stat Biomedical 1.0 (Research Solutions, Cam-
bridge, United Kingdom).
Results
The examination was well tolerated by all
volunteers. All MR images both at rest and
during maximum strain were considered tech-
nically adequate.
MR Imaging Measurements of Anterior, Middle, and Posterior Pelvic Floor
TABLE 1
Descent in Asymptomatic Men and Women
Mean ± SD
Part Measured
Men (n = 25)
Bladder base (mm)
At rest 30 (7)
On strain 22 (10)
Descent on straining 7 (7)
Cervix (mm)
At rest NA
On strain NA
Descent on straining NA
Anorectal junction (mm)
At rest 0 (10)
On strain –12 (12)
Descent on straining 12 (12)
Note.—All static values refer to distance above pubococcygeal line unless negative. NA = not applicable.
a Student’s t
test.
Mean values with standard deviations (SDs)
for the bladder base, cervix, and anorectal junc-
tion at rest, during maximum strain, and the de-
scent on straining are shown in Table 1. The
bladder base lay above the pubococcygeal line
in all subjects at rest but descended below this
line during maximum strain in three women,
two of whom were parous (one with two and
one with three vaginal deliveries), resulting in a
diagnosis of cystocele (Table 1, Figs. 1 and 2).
Although the mean bladder base was signifi-
cantly higher in men than in women at rest and
during maximum strain, no significant differ-
ence was seen between men and women with
respect to the distance descended. No statistical
difference was noted between any bladder base
value when parous and nulliparous women
were compared.
The cervix descended below the pubococ-
cygeal line in two (one nulliparous and one
with three vaginal deliveries) of the three
women with cystocele, and in a third (who did
not have a cystocele), resulting in a diagnosis
of grade 1 uterocervical prolapse (Figs. 3 and
4). Grade 2 or 3 uterocervical prolapse was
not seen in any woman. No significant differ-
ence was seen when nulliparous and parous
women were compared for any uterocervical
measurement.
The mean anorectal junction position lay on
the pubococcygeal line in both men and
women (Table 1) and was not more than 2.0 cm
below this level in any subject at rest. However,
anorectal junction descent was more than 3.0
cm in three men during maximum strain (Fig.
5). No overall statistical difference was noted
pa
Women (n = 25)
22 (5) <0.0001
12 (11) 0.001
10 (10) 0.33
31 (13)
17 (16)
14 (11)
0 (8) 0.923
–11 (12) 0.71
11 (10) 0.62
AJR:174, March 2000
 MR Imaging of the Pelvic Floor

Fig. 1.—Position of bladder base rel-

ative to pubococcygeal line.
A and B, Ladder plots show position
of bladder base relative to pubococ-
cygeal line at rest and during maxi-
mum strain for 25 asymptomatic men
(A) and 25 asymptomatic women (B).
Note that bladder base descends
below this line in three women, indi-
cating cystocele.
● = nulliparous. ° = multiparous,

A B

Fig. 2.—40-year-old asymptomatic

multiparous woman.
A, Sagittal T1-weighted fast-field-
echo MR image of pelvis at rest
shows pubococcygeal line (black
line) and no apparent abnormality.
B, Sagittal T1-weighted fast-field-
echo MR image of pelvis at strain
shows cystocele diagnosed because
bladder base descent is below pubo-
coccygeal line (white arrow ). Cervix
has also descended below pubococ-
cygeal line (black arrow ).

A B

between men and women (Table 1) or between
parous and nulliparous women.
Mean values at rest and during maximum
strain for the anorectal angle, the levator plate
angle, the pelvic floor hiatal area, and the hi-
atal perimeter in men and women are shown
in Table 2. The anorectal angle during maxi-
mum pelvic strain was significantly more
acute in men than in women, but no signifi-
cant difference was seen either between sexes
or between women of different parity for any
other measurement. No rectocele, enterocele,
rectal prolapse, or perineal hernia was seen in
any subject when all images were examined
for further structural abnormality. An unsus-
pected ovarian dermoid cyst was seen in one
woman, and this diagnosis was subsequently
confirmed surgically.
Discussion
The effects of pelvic floor weakness may not
be localized to one organ or compartment, and
failure to identify all sites of prolapse may lead
to incomplete surgical repair and subsequent
recurrence [1]. Clinical examination tends to
underestimate the degree of prolapse or may
miss it altogether; in one study of 300 women,
enteroceles were revealed on dynamic cysto-
proctography in 111, of which 93 (84%) were
missed clinically [4]. Because of this, many in-
vestigators take the opportunity to modify the
standard proctographic examination by either
administering an oral barium suspension ap-
proximately 1–2 hr before the procedure [2] or,
alternatively, using a vaginal marker so that en-
teroceles can be diagnosed by rectovaginal sep-

AJR:174, March 2000 663


Fig. 3.—Ladder plot shows position of uterocervical junction relative to pubococcygeal line at rest and during
maximum strain for 25 asymptomatic women. Note that uterocervical junction descends below this line in three
°
women, indicating prolapse. = multiparous, ● = nulliparous.
aration. The best choice is probably a barium
paste because a tampon may inhibit prolapse
by splinting the vagina [14]. Dynamic cysto-
proctography is essentially evacuation proctog-
raphy preceded by a cystogram [5]; it extends
the examination into the anterior pelvic floor so
that any cystocele can be diagnosed.
Although evacuation proctography is rapid
and easy to perform, the modifications neces-
sary to image other organs may be time-con-
suming, invasive, and complex, and the
musculature of the pelvic floor itself is not visu-
alized. Furthermore, examination involves irra-
diation, and many patients will be women in
their childbearing years, especially if the exami-
664
Goh et al.
nation is being performed to investigate consti-
pation. In an attempt to overcome these
limitations, MR imaging has been applied to
pelvic floor dynamics with promising results.
Initial reports were necessarily compromised by
slow acquisition times [15, 16], but advancing
MR technology has enabled multislice imaging
during a single straining effort. A study of con-
stipated and anally incontinent women using
this technique revealed unsuspected visceral
prolapse at multiple sites [7]. However, these
studies have recruited limited numbers of con-
trol subjects. Previous proctographic studies of
asymptomatic control subjects have unexpect-
edly shown that many of these individuals show
A B
findings previously considered abnormal [10,
11]. MR criteria for abnormality have been based
on proctographic examinations but it is likely that
values will differ between the two techniques,
primarily because the seated position achieved
during proctography is available only to the few
investigators with access to open-architecture
magnets [17, 18]. Supporting this, a study of 10
women who underwent both proctography and
dynamic MR imaging found good correlation
but poor agreement between corresponding mea-
surements [19]. Because of this, the range of nor-
mal values for dynamic MR imaging needs to be
defined. We have attempted to do this in the larg-
est dynamic MR study to date, to our knowledge.
All subjects examined in this study satisfied
vigorous and validated criteria for health. De-
spite this, we encountered some findings
con–sidered abnormal. We showed significant
differences between men and women with re-
spect to the position of the bladder base, as
would be expected, but a cystocele developed in
three women during maximum pelvic strain,
two of whom also developed grade 1 uterocer-
vical prolapse; the latter was also seen in one
additional woman. However, no large cystocele
was seen, nor was any higher grade of uterocer-
vical prolapse. Similarly, three men showed sig-
nificant pelvic floor descent (≥3 cm) when
judged by proctographic criteria, but this also
was low grade. These findings are broadly in
agreement with those encountered in procto-
graphic studies of asymptomatic individuals in
which most subjects were “normal,” but there is
some crossover with findings considered abnor-
mal [10, 11]. Supporting this, no rectocele, en-
terocele, rectal prolapse, or perineal hernia was
seen in any subject when all images were re-
Fig. 4.—50-year-old asymptomatic
nulliparous woman.
A, Sagittal T1-weighted fast-field-
echo MR image of pelvis at rest
shows pubococcygeal line (black
line) and cervix (arrowhead ). No ap-
parent abnormality is seen.
B, Sagittal T1-weighted fast-field-
echo MR image of pelvis at strain
shows grade 1 uterocervical pro-
lapse (white arrow ) indicated by cer-
vical descent below pubococcygeal
line (black line). Note also cystocele
(black arrow ).
AJR:174, March 2000
 MR Imaging of the Pelvic Floor

Fig. 5.—Position of anorectal junc-

tion relative to pubococcygeal line.
A and B, Ladder plots show position of
anorectal junction relative to pubo-
coccygeal line at rest and during max-
imum strain for 25 asymptomatic men
(A) and 25 asymptomatic women (B).
Note that anorectal junction descent
is excessive (≥3 cm of descent) in
three men.
liparous. ° = multiparous, ● = nul-

A B

MR Imaging Measurements of Anorectal and Levator Plate Angles and

few parous women recruited. The vigorous
TABLE 2 questionnaire excluded many older and parous
Pelvic Floor Hiatus Area and Perimeter in Asymptomatic Men and Women
women from the study, predominantly because
Mean ± SD
Part Measured pa of minor anal incontinence, a symptom strongly
Men (n = 25) Women (n = 25) associated with previous vaginal delivery [21].
Anorectal angle (degrees)
Dynamic pelvic MR imaging is evolving,
and an optimal technique remains to be de-
At rest 101 (13) 106 (12) 0.154
fined. A major disadvantage of the technique
On strain 97 (17) 107 (16) 0.042
may relate to the inability to ensure an ade-
Change on straining –3 (19) 1 (16) 0.357 quate straining effort. In an attempt to mini-
Levator plate angle (degrees) mize this effect, straining was practiced with
At rest 11 (7) 13 (8) 0.531 the subject before the examination, and the
On strain 20 (15) 17 (15) 0.575 volunteer was placed on an absorbent pad to
Change on straining 9 (16) 5 (13) 0.37 minimize the fear of leakage. Despite this,
Pelvic floor hiatus area (mm2) some subjects raised their pelvic floor during
At rest 1920 (241) 2006 (358) 0.336 straining, evidenced by ano-rectal junction
On strain 2477 (1081) 2783 (1399) 0.405 ascent, which suggests pelvic floor contrac-
Change on straining 557 (1177) 777 (1347) 0.554
tion rather than relaxation. To eliminate this
effect, some investigators fill the rectum and
Hiatus perimeter (mm)
bladder and encourage evacuation in the
At rest 175 (13) 183 (20) 0.139
magnet despite the supine position [9]. This
On strain 193 (34) 201 (41) 0.439 approach will provide additional information
Change on straining 17 (37) 18 (37) 0.911 relating to the rate and completeness of rectal
a Student’s t
test. evacuation, important parameters when con-
stipation is being investigated [22]. A seated
viewed. The range of normal values for the an- would diagnosis of a prolapse site. The pelvic position is optimal but only achievable in an
orectal and levator plate angle has also been floor hiatus is a relatively new measurement that open magnet [17, 18].
defined. Although the anorectal angle during is already clinically relevant and is increased in If prolapse is shown, it is tempting to
maximum pelvic strain was significantly more women with pelvic organ prolapse [20]. We attribute symptoms to the prolapse, but the
acute in men, this result is likely to have oc- were unable to find any significant difference prolapse may merely be secondary to an under-
curred because of multiple statistical testing. between the sexes for this measurement, nor lying functional disorder. For example, consti-
These measurements were included for com- were we able to show any significant difference pated patients who strain excessively may
pleteness but are rarely of major clinical signifi- between parous and nulliparous women for any develop cystocele, enterocele, rectocele, rectal
cance and do not influence management, as parameter, possibly because of the relatively prolapse, and pelvic floor descent [23], but sur-

AJR:174, March 2000 665


gery to correct these will not treat the underly-
ing disorder [24]. The results of any dynamic
pelvic examination, be it cystoproctography or
MR imaging, need to be considered in the light
of the patient’s history and other tests so that
therapy is appropriate.
In summary, we examined 50 healthy volun-
teers using dynamic pelvic MR imaging to
define the range of normal appearances en-
countered during this relatively new test. Al-
though we encountered some pelvic floor
prolapse, it was infrequent and of low grade,
suggesting the criteria for abnormality derived
from proctographic examinations are generally
applicable to MR imaging.
References
1. Kelvin FM, Maglinte DDT, Benson JT. Evacua-
tion proctography (defecography): an aid to the
investigation of pelvic floor disorders. Obstet Gy-
necol 1994;83:307–314
2. Kelvin FM, Maglinte DDT, Hornback JA, Benson
JT. Pelvic prolapse: assessment with evacuation
proctography (defecography). Radiology 1992;184:
547–551
3. Maglinte DDT, Kelvin FM, Hale DS, Benson JT.
Dynamic cystoproctography: a unifying diagnos-
tic approach to pelvic floor and anorectal dys-
function. AJR 1997;169:759–767
4. Hock D, Lombard R, Jehaes C, et al. Colpocystodef-
ecography. Dis Colon Rectum 1993;36:1015–1021
5. Kelvin FM, Maglinte DDT, Benson JT, Brubaker
666
Goh et al.
LP, Smith C. Dynamic cystoproctography: a tech-
nique for assessing disorders of the pelvic floor in
women. AJR 1994;163:368–370
6. Halligan S, Spence-Jones C, Kamm MA, Bartram
CI. Dynamic cystoproctography and physiologi-
cal testing in women with urinary stress inconti-
nence and urogenital prolapse. Clin Radiol 1996;
51:785–790
7. Healy JC, Halligan S, Reznek RH, Watson S, Phillips
RKS, Armstrong P. Patterns of prolapse in women
with symptoms of pelvic floor weakness: assessment
with MR imaging. Radiology 1997; 203:77–81
8. Healy JC, Halligan S, Reznek RH, et al. Dynamic
magnetic resonance imaging of the pelvic floor in
patients with obstructed defecation. Br J Surg
1997;84:1555–1558
9. Lienemann A, Anthuber C, Baron A, Kohz P, Re-
iser M. Dynamic MR colpocystorectography as-
sessing pelvic floor descent. Eur Radiol 1997;
7:1309–1317
10. Bartram CI, Turnbull GK, Lennard-Jones JE.
Evacuation proctography: an investigation of rec-
tal expulsion in 20 subjects without defecatory
disturbance. Gastrointest Radiol 1988;13:72–80
11. Shorvon PJ, McHugh S, Diamant NE, Somers S,
Stevenson GW. Defecography in normal volunteers:
results and implications. Gut 1989;30: 1737–1749
12. Parks AG, Porter NH, Hardcastle JD. The syn-
drome of the descending perineum. Proc R Soc
Med 1966;59:477–482
13. Halligan S, Bartram C, Hall C, Wingate J. Entero-
cele diagnosed by simultaneous evacuation proc-
tography and peritoneography: does “defecation
block” exist? AJR 1996;167:461–466
14. Archer BD, Somers S, Stevenson GW. Contrast
medium gel for marking vaginal position during
defecography. Radiology 1992;182:278–279
15. Kruyt RH, Delemarre JBVM, Doornbos J, Vogel
HJ. Normal anorectum: dynamic MR imaging
anatomy. Radiology 1991;179:159–163
16. Yang A, Mostwin JL, Rosenshein NB, Zerhouni
EA. Pelvic floor descent in women: dynamic
evaluation with fast MR imaging and cinematic
display. Radiology 1991;179:25–33
17. Schoenberger AW, Debatin JF, Guldenschuh I, Hany
TF, Steiner P, Krestin GP. Dynamic MR defecogra-
phy with a superconducting, open-configuration
MR system. Radiology 1998;206:641–646
18. Fielding JR, Griffiths DJ, Versi E, Mulkern RV,
Lee M-LT, Jolesz FA. MR imaging of pelvic floor
continence mechanisms in the supine and sitting
positions. AJR 1998;171:1607–1610
19. Healy JC, Halligan S, Reznek RH, et al. Dynamic
MR imaging compared with evacuation proctogra-
phy when evaluating anorectal configuration and
pelvic floor movement. AJR 1997;169:775–779
20. DeLancey JO, Hurd WW. Size of the urogenital
hiatus in the levator ani muscles in normal
women and women with pelvic organ prolapse.
Obstet Gynecol 1998;91:364–368
21. Sultan AH, Kamm MA, Hudson CN, Thomas J,
Bartram CI. Anal sphincter disruption during vag-
inal delivery. N Engl J Med 1993;329:1905–1911
22. Halligan S, Bartram CI, Park HY, Kamm MA.
The proctographic features of anismus. Radiology
1995;197:679–682
23. Spence-Jones C, Kamm MA, Henry MM, Hud-
son CN. Bowel dysfunction: a pathological factor
in uterovaginal prolapse and urinary stress incon-
tinence. Br J Obstet Gynaecol 1994;101:147–152
24. Halligan S. Commentary: imaging anorectal
function. Br J Radiol 1996;69:985–988
AJR:174, March 2000