genitourinary imaging pelvic floor dysfunction · 2010. 12. 12. · pelvic floor dysfunction:...

Pelvic Floor Dysfunction:Assessment with Combined Analysisof Static and Dynamic MR ImagingFindings1

Rania F. El Sayed, MDSahar El Mashed, MDAhmed Farag, MDMedhat M. Morsy, MDMohamed S. Abdel Azim, MD

Purpose: To prospectively analyze static and dynamic magnetic res-onance (MR) images simultaneously to determine whetherstress urinary incontinence (SUI), pelvic organ prolapse(POP), and anal incontinence are associated with specificpelvic floor abnormalities.

Materials andMethods:

This study had institutional review board approval, andinformed consent was obtained from all participants.There were 59 women: 15 nulliparous study controlwomen (mean age, 25.6 years) and 44 patients (mean age,43.4 years), who were divided into four groups accordingto chief symptom. Static T2-weighted turbo spin-echo im-ages were used in evaluating structural derangements;functional dynamic (cine) balanced fast-field echo imageswere used in detecting functional abnormalities and re-cording five measurements of supporting structures. Find-ings on both types of MR images were analyzed together todetermine the predominant defect. Analysis of varianceand the Bonferroni t test were used to compare groups.

Results: In the four patient groups, POP was associated with leva-tor muscle weakness in 16 (47%) of 34 patients, with levelI and II fascial defects in seven (21%) of 34 patients, andwith both defects in 11 (32%) of 34 patients. SUI wasassociated with defects of the urethral supporting struc-tures in 25 (86%) of 29 patients but was not associatedwith bladder neck descent. Levator muscle weakness maylead to anal incontinence in the absence of anal sphincterdefects. Measurements of supporting structures were sig-nificant (P � .05) in the identification of pelvic floor laxity.

Conclusion: Combined analysis of static and dynamic MR images ofpatients with pelvic floor dysfunction allowed identificationof certain structural abnormalities with specific dysfunc-tions.

� RSNA, 2008

1 From the Departments of Radiology (R.F.E.S., S.E.M.),Surgery (A.F.), Anatomy (M.M.M.), and Urology (M.S.A.A.),Faculty of Medicine, Cairo University, Kaser El Aini Street,Cairo, Egypt 11511. Received June 28, 2007; revisionrequested August 31; revision received December 10;accepted January 21, 2008; final version accepted March3. Address correspondence to R.F.E.S. (e-mail:[email protected]).

� RSNA, 2008

ORIG

INAL

RESE

ARCH

�GE

NITO

URIN

ARY

IMAG

ING

518 Radiology: Volume 248: Number 2—August 2008

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Female pelvic floor dysfunction is ageneral term applied to a wide va-riety of clinical conditions, most

commonly stress urinary incontinence(SUI), pelvic organ prolapse (POP), andanal incontinence (1).

Static magnetic resonance (MR) im-aging is utilized to delineate compo-nents of the pelvic organ support sys-tem, including the anal sphincter com-plex (2–9). Dynamic (cine) MR imagingwith fast sequences enables functionalevaluation to assess pelvic floor relax-ation and pelvic organ descent (10–16).However, the precise anatomic causesof POP and SUI remain somewhat un-clear. POP has been attributed both todamage to the levator ani muscle (10)and to an endopelvic fascial defect (17);however, some believe that it is still un-clear which of these factors is more re-sponsible (18–20). Similarly, SUI hasbeen attributed to urethral hypermo-bility (21,22), to unequal movement ofthe urethral walls (23), and to defectsin the urethral supporting structures(6,5,24,25).

Because of these controversies, treat-ment is often started regardless of thespecific anatomic lesion involved. A re-port from Olsen et al (26) indicated that29% of the procedures performed forincontinence and prolapse are repeat

surgeries, suggesting the need for ad-vances in both diagnosis and manage-ment of these disorders. DeLancey (27)emphasized the need for specific tests topinpoint the specific anatomic defect re-sponsible for pelvic floor dysfunction ineach patient.

Thus, the purpose of our study wasto prospectively analyze static and dy-namic MR images simultaneously to de-termine whether SUI, POP, and analincontinence are associated with spe-cific pelvic floor abnormalities.

Materials and Methods

PatientsThe study group consisted of 59 women:15 volunteer nulliparous women with amean age of 25.6 years (range, 22–35years), who served as a control group,and 44 women with a mean age of 43.4years (range, 18–62 years), a parityrange of 0 to 7, and clinical symptoms ofpelvic floor dysfunction. The volunteerswere employees of our institution whowere recruited during the study periodand who did not have lower genitouri-nary tract symptoms. The women withpelvic floor dysfunction symptoms pre-sented consecutively between January2004 and January 2005 to the cliniciansinvolved in our study and were referredto a radiologist (R.F.E.S.) for assess-ment. The study was approved by ourinstitution’s review board, and in-formed consent was obtained from allparticipants. All volunteers and patientsunderwent both clinical examinationand MR imaging.

The control group was subdividedinto two subgroups, volunteer subgroupA and volunteer subgroup B, accordingto their MR imaging findings. MR im-ages for volunteer subgroup A (n � 9)showed neither anatomic defects nor

pelvic organ descent. In contrast, MRimages for volunteer subgroup B (n � 6)demonstrated both anatomic defectsand pelvic organ descent below thepubococcygeal line (PCL).

Patients were classified into one offour groups, depending on their chiefsymptom: POP without SUI (group A,n � 10), SUI without POP (group B,n � 10), SUI with bladder and/or gen-ital prolapse (group C, n � 16), andanal incontinence associated with POP(group D, n � 8). Three of the eightpatients in group D reported SUI in ad-dition to their main disorder. Amongthe members of the four patient groups,five had undergone prior pelvic or ano-rectal surgery. Four patients had under-gone various procedures for SUI: twopatients from group A and one patientfrom group C had undergone transvag-inal taping, one patient from group Dhad undergone bladder neck suspen-sion, and one patient from group D hadhad undergone hemorrhoidectomy.

Clinical ExaminationAll control group members were inter-viewed and were found to be healthy,with no symptoms of lower genitouri-nary abnormalities, by using a validatedquestionnaire (28).

A urogynecologist (M.S.A.A., with 15years of experience in clinical evaluationand management of pelvic floor dysfunc-tion) examined the patients in groups A,B, and C, and a coloproctologist (A.F.,

Published online before print10.1148/radiol.2482070974

Radiology 2008; 248:518–530

Abbreviations:PCL � pubococcygeal linePOP � pelvic organ prolapseSUI � stress urinary incontinence

Author contributions:Guarantors of integrity of entire study, R.F.E.S., S.E.M.,M.S.A.A.; study concepts/study design or data acquisitionor data analysis/interpretation, all authors; manuscriptdrafting or manuscript revision for important intellectualcontent, all authors; manuscript final version approval, allauthors; literature research, R.F.E.S., M.M.M.; clinicalstudies, R.F.E.S., A.F., M.S.A.A.; statistical analysis,R.F.E.S.; and manuscript editing, all authors

Authors stated no financial relationship to disclose.

Advances in Knowledge

� Combined analysis of static anddynamic MR images of the pelvicfloor reveals that certain anatomicdefects on static images are asso-ciated with specific functional ab-normalities on dynamic images.

� It is possible to differentiatewhether prolapse is due to defectsin the endopelvic fascia, to levatormuscle weakness, or to abnormal-ities in both fascia and muscles.

� Stress urinary incontinence is as-sociated with structural defects inthe urethral supporting structuresrather than with bladder neckdescent.

� In the absence of an anal sphinc-ter defect, anal incontinence isassociated with marked levatormuscle weakness.

Implication for Patient Care

� The association between preciseanatomic defects in the pelvic or-gan support system and specificpelvic floor dysfunction allows adefect-specific approach to pelvicfloor dysfunction for each patient.

GENITOURINARY IMAGING: MR Imaging of Pelvic Floor Dysfunction El Sayed et al

Radiology: Volume 248: Number 2—August 2008 519

with 19 years of experience) examinedpatients in group D and documentedsymptoms and physical findings. Duringpelvic examination, the Valsalva maneu-ver was performed, and the results ofthe clinical urinary stress test, as well asthe degree and type of prolapse, wererecorded and graded as follows: grade 0indicated normal findings; grade I, min-imal descent; grade II, descent halfwayto the introitus; grade III, descent to theintroitus; and grade IV, descent beyondthe introitus (29). Clinical severity ofanal incontinence was assessed by usingthe Pescatori score for anal inconti-

nence (30). Anal sphincter tone, analsphincter defects, levator ani tender-ness, and defects and degree of relax-ation of the puborectalis muscle wereassessed with rectal examination (31).

MR ImagingMR imaging was performed with thepatient supine in a 1.5-T MR imagingunit (Gyroscan PowerTrak 6000; Phil-ips Medical Systems, Best, the Nether-lands) by using a pelvic phased-arraycoil. No oral or intravenous contrastagent was administered. All patientshad undergone a rectal enema with

warm water the night before the MRimaging examination and were asked tovoid 2 hours before the examination.The rectum was opacified with 90–120mL of ultrasonographic gel (Aquasonic;Parker Laboratories, Fairfield, NJ) in allbut eight patients, who either had apainful perianal condition (three pa-tients had perianal fissures, and one hadan acute attack of hemorrhoids) or re-fused the instillation of gel (four pa-tients). The MR images of these patientswere still of diagnostic quality; com-ment on rectocele was included in ourreport in five of these eight patients in

Figure 1

Figure 1: Normal static and dynamic MR imaging findings in 20-year-old woman in control group. (a) Axial T2-weighted turbo spin-echo image (5000/132) at level of proxi-mal urethra (U) shows normal urethral supporting structures. Arrows�puborectalis slings, �� space of Retzius, V� vagina. (b) Axial balanced fast-field echo image (9/4) ofanal sphincter. The consecutive layers from the lumen outward include the innermost high-signal-intensity layer (the combined mucosa and submucosa), the low-signal-intensitylayer (the submucosal smooth muscle [arrowhead]), the internal anal sphincter (white arrow), and the deep external anal sphincter (black arrows). (c–e) Dynamic (c) sagittal,(d) axial, and (e) coronal balanced fast-field echo MR images (5/1.6) obtained at maximum straining. In c, there is no POP; the H-line and the M-line are illustrated. In d, the widthof the levator hiatus corresponds to the dashed line. In e, there is no excessive increase of the iliococcygeus angle (plotted lines).



whom the rectum was clearly delineatedby air.

Static images of the pelvis were firstacquired in three planes by using T2-weighted turbo spin-echo sequences(repetition time msec/echo time msec,5000/132; field of view, 240–260 mm;section thickness, 5 mm; gap, 0.7 mm;number of signals acquired, two; flip an-gle, 90°; matrix, 512 � 512; acquisitiontime, 3.12 minutes for each sequence).In addition, T2-weighted balanced fast-field echo images (9.0/4.0; field of view,220 mm; section thickness, 3 mm; num-ber of signals acquired, eight; flip angle,45°; matrix, 512 � 512; acquisitiontime, 2.12 minutes) of the anal sphinc-ter complex were obtained in the con-trol group volunteers and in patients ingroup D. In this sequence, section ori-entation was parallel and perpendicularto the plane of the anal canal (8,9).

Dynamic MR imaging was performed

in the sagittal, axial, and coronal planesby using the balanced fast-field echo se-quence (5.0/1.6; field of view, 300 mm;section thickness, 6–7 mm; gap, 0.7mm). As a modification of a previoustechnique (14), in each plane we ac-quired five sections during six phases;each phase took 10 seconds. These sixphases were acquired (a) with the pa-tient at rest, (b) during contraction ofthe pelvic floor (the patient was in-structed to squeeze the buttocks as iftrying to prevent the escape of urine),(c) during mild straining, (d) duringmoderate straining, (e) during maxi-mum straining, and (f) during a re-peated maximum straining sequence toensure a maximal Valsalva maneuver(the patient was instructed to beardown as much as she could, as thoughshe were constipated and trying to defe-cate).

All participants were trained re-

garding these instructions before goingto the MR imaging examination. A radi-ologist (R.F.E.S.) attended each MR im-aging examination to ensure the compli-ance of the woman’s response by ob-serving the movement of the anteriorabdominal wall, as well as the move-ment of the pelvic organs, to minimizevariations between examinations.

Data AnalysisThe MR imaging data sets were as-sessed by a radiologist (R.F.E.S., with 7years of experience interpreting pelvicfloor MR images). The radiologist wasaware of each patient’s clinical historyand diagnosis. All measurements wereobtained with software (DicomWorks,version 1.3.5; http://dicom.online.fr/).A consensus reading of the MR imagesfor patients in groups A, B, and C wasperformed by the radiologist (R.F.E.S.)and the urogynecologist (M.S.A.A.),

Figure 2

Figure 2: Schematic presentation of MR imaging report.



and a consensus reading of the MR im-ages for patients in group D was per-formed by the radiologist and the colo-proctologist (A.F.).

Analysis of static MR images.—Analysis of static images in the controland patient groups was based on scru-tiny of the urethral supporting system,the vaginal supporting system, and theanal sphincter complex.

The urethral supporting structuresconsist of ligaments (32), level III fascialsupport (18), and the puborectalis mus-cle (4). On images obtained in the axialplane, urethral ligament abnormalitieswere classified as distorted when inter-nal architectural changes with wavinessof the ligaments were seen and as adefect when there was discontinuity ofthe ligament with visualization of thetorn parts (5,6). A level III fascial defectwas recognized by the “drooping mus-tache sign,” which is formed by fat inthe prevesical space against the bilat-eral sagging of the detached lower thirdof the anterior vaginal wall from thearcus tendineus fascia (7). A puborec-talis muscle defect was recognized bythe loss of the normal symmetric ap-pearance of the muscle slings or disrup-tion of the attachment of the muscle tothe pubic bone (4–6).

Vaginal supporting structures in-clude level I fascia (which suspend theupper portion of the vagina from thepelvic wall), level II fascia (which attachthe midportion of the vagina more di-rectly to the pelvic wall) (19), and theiliococcygeus muscle. In the axial plane,a defect in the fascia was visualized assagging of the fluid-filled posterior uri-nary bladder wall due to the detach-ment of the vaginal supporting fasciafrom the lateral pelvic wall, known asthe “saddlebags sign” (7). The iliococcy-geus muscle was assessed for loss of thenormal symmetric appearance of itsmuscle slings or disruption of its attach-ment to the obturator internus musclein the coronal plane.

For evaluation of the anal sphinctercomplex, the normal thickness of eachmuscle layer and the total thickness ofthe anal sphincter (measured from theinner border of the internal anal sphinc-ter to the outer border of the external

Figure 3

Figure 3: Static and dynamic MR images in two women in control group (one 20-year-old woman [a] andone 25-year-old woman [b–d]). (a, b) Static axial T2-weighted turbo spin-echo images (5000/132) show(a) normal level I vaginal fascial support, compared with (b) central (dashed arrow) and paravaginal (solidarrows) level I fascial defects. UB � urinary bladder. (c) Dynamic cine midsagittal and (d) parasagittal bal-anced fast-field echo MR images (5/1.6) obtained at maximum straining; d (obtained routinely within the fivesections in each dynamic sequence) demonstrates uterine descent (UD) and the ileococcygeus muscle (ar-row) better than c alone does.

Table 1

Findings Regarding Pelvic Organ Support System Defects on Static Axial MR Imagesin Control Group

VolunteerSubgroup

Urethral Supporting Structures

Vaginal Supporting Structures

Anal Sphincter

LigamentsLevel IIIFascia

PuborectalisMuscle

Level I and II Vaginal EndopelvicFascia

IleococcygeusMuscle External InternalParavaginal Central

Paravaginaland Central

A (n � 9) 0 0 0 0 0 0 0 0 0B (n � 6) 0 0 0 2 0 4 0 0 0

Note.—Data are numbers of volunteers with a defect in the given structure.



anal sphincter) were measured in thecontrol group. Anal sphincter lesionswere classified according to the muscleinjured (the internal or external analsphincter or the puborectalis muscle)and according to lesion type (defectand/or scarring). A sphincteric defectwas defined as discontinuity of the mus-cle ring; scarring was defined as a low-signal-intensity deformation of the nor-mal pattern of the muscle layer (33).

Analysis of dynamic MR images.—In the sagittal plane, the PCL, whichextends from the inferior border of thesymphysis pubis anteriorly to the tip ofthe coccyx posteriorly, was used as thereference line. For each participant, thedescent of the bladder neck, bladderbase, uterus, and anorectal junction be-low the PCL (15) was recorded. In thepatient groups, loss of urine through

the urethra at maximum straining wasalso recorded. (However, absence ofurine loss during MR imaging was notconsidered to invalidate the patient’sreport of SUI.)

Other measurements in the sagittalplane during maximum straining in-cluded the H-line, which extends fromthe inferior aspect of the pubic symphy-sis to the anorectal junction; the M-line,which drops as a perpendicular linefrom the PCL to the posterior aspect ofthe H-line (11); and the levator plateangle, which is enclosed between thelevator plate and the PCL (12). In theaxial and coronal planes, respectively,the width of the levator hiatus (14) andthe iliococcygeus angle (2) were mea-sured at rest and during maximumstraining. These five measurements ofsupporting structures were all consid-

ered to reflect the status and the weak-ness of the levator ani (Fig 1).

Combined analysis of static and dy-namic MR images.—Findings obtainedfrom static and dynamic MR images inthe same patient were then analyzedsimultaneously to determine whether aparticular anatomic defect detected onstatic images was associated with a spe-cific dysfunction on dynamic images.The most marked type of pelvic sup-porting system defect was reported asthe predominant defect (Fig 2).

Statistical AnalysisThe mean and standard deviation ofthe measurements for the control sub-groups and for each patient groupwere calculated.

The mean and standard deviation ofthe measurements of supporting struc-

Table 2

Findings Regarding Dysfunction and Measurements of Supporting Structures on Dynamic Cine MR Images in Control Group

VolunteerSubgroup

Pelvic Organ Descent (cm)*

Width of Levator Hiatusin Axial Plane (cm)

Ileococcygeus Angle inCoronal Plane (degrees)

Bladder Neck Bladder Base Uterus Other†H-Line(cm)*

M-Line(cm)*

Levator PlateAngle(degrees)* At Rest

AtMaximumStraining At Rest

AtMaximumStraining

A (n � 9) No descentbelow PCL(9/9)

No descentbelowPCL (9/9)

No descent belowPCL (9/9)

None (9/9) 5.8 � 0.5 1.3 � 0.5 11.7 � 4.8 3.3 � 0.4 4.5 � 0.7 20.9 � 3.5 33.4 � 8.2

B (n � 6) 0.8 � 0.4 (6/6) 0.9 � 0.5 (6/6) 1.1 � 0.2 (4/6) None (6/6) 6.1 � 0.8 1.9 � 0.8 33 � 12.9 3.1 � 0.4 5.9 � 1.3 24 � 1.8 40.1 � 8.6

Note.—No volunteer in the control group had SUI, defined as loss of urine at maximum straining, on dynamic cine MR images. Data are means � standard deviations, with numbers in parenthesesindicating in how many of the total number of volunteers mean pelvic organ descent was calculated.

* As measured in sagittal plane at maximum straining.† Including peritoneocele, enterocele, rectocele, and anorectal junction descent.

Table 3

Findings Regarding Pelvic Organ Support System Defects on Static Axial MR Images in Patient Groups

Patient Group

Urethral Supporting StructuresVaginal Supporting Structures

Anal Sphincter

LigamentLevel IIIFascia

PuborectalisMuscle

Level I and II Vaginal Endopelvic FasciaIliococcygeusMuscle External InternalParavaginal Central

Paravaginal andCentral

A: POP only (n � 10) 0 0 0 6 1 2 1 0 0B: SUI only (n � 10) 0 9 0 3 1 4 0 0 0C: SUI and POP (n � 16) 4 8 1 9 1 6 0 1 (Thinning) 0D: Anal incontinence and POP (n � 8)* 0 2 1 4 0 2 2 6 (Defects)† 3 (Scarring)

Note.—Data are numbers of patients with a defect in the given structure.

* Three of these patients also reported SUI in addition to their main disorder.† Two patients had intact sphincters.



tures in volunteer subgroup A werecompared with those in volunteer sub-group B and with those in patientgroups by using software (InStat, ver-sion 3.05; GraphPad Software, San Di-ego, Calif). We used one-way analysis ofvariance to determine the statistical dif-ference between these measurements.Whenever a measurement was shownto be statistically significant, we thenused the Bonferroni t test to deter-mine which group’s measurementswere statistically significant comparedwith those for volunteer subgroup A.Differences were considered signifi-cant at P � .05 and highly significant atP � .001.

Results

Combined Analysis of Static and DynamicMR ImagesControl group.—On static axial MR im-ages, a central type of fascial defect wasdetected in combination with a paravag-inal defect in four members of volunteersubgroup B (Table 1). This type of de-fect led to sagging of the midportion ofthe posterior wall of the urinary bladder(Fig 3).

Combined analysis of static and dy-namic MR images in volunteer subgroupB (Tables 1 and 2) demonstrated thatPOP was mainly due to the presence ofdefects in level I and level II endopelvicfascia, because levator muscle weak-ness, as indicated by all measurementsof supporting structures in this sub-group except the levator plate angle,was of no significant statistical differ-ence from that in volunteer subgroup A.

The thicknesses of the muscle layersof the anal sphincter complex in thecontrol group were as follows: externalmuscle layer, 2.07 mm � 0.45 (stan-dard deviation); internal muscle layer,4.16 mm � 0.72; and longitudinal mus-cle layer, 1.04 mm � 0.4. The totalthickness of the anal sphincter was 9.07mm � 2.59.

Patient group A.—In the patientgroup with POP but without SUI, POPwas found to be related to a predomi-nant fascial defect in three patients, tolevator muscle weakness in two pa-

Tabl

e4

Find

ings

Rega

rdin

gDy

sfun

ctio

nan

dM

easu

rem

ents

ofSu

ppor

ting

Stru

ctur

eson

Dyna

mic

Cine

MR

Imag

esin

Patie

ntGr

oups

Patie

ntGr

oup

Pelvi

cOr

gan

Desc

ent(

cm)*

Wid

thof

Leva

torH

iatus

inAx

ialPl

ane

(cm

)Ili

ococ

cyge

usAn

gle

inCo

rona

lPlan

e(d

egre

es)

Blad

derN

eck

Blad

derB

ase

Uter

usOt

her†

H-Li

ne(c

m)*

M-L

ine

(cm

)*Le

vato

rPlat

eAn

gle

(deg

rees

)*At

Rest

AtM

axim

umSt

rain

ing

AtRe

stAt

Max

imum

Stra

inin

g

A:PO

Pon

ly(n

�10

)1

�0.

7(7

/10)

1.4

�1

(7/1

0)3.

1�

1(6

/10)

77

�0.

72.

9�1.

240

.6�

16.9

4.2

�0.

67.

2�

1.7

24�

1.8

57.9

5�

14B:

SUIo

nly

(n�

10)

0.6

�0.

3(5

/10)

0.7

�0.

4(5

/10)

Nout

erin

ede

scen

t0

5.95

�0.

72.

4�

1.1

17.9

�1.

13.

2�

0.3

4.4

�0.

825

.1�

1.8

35.9

�9.

2C:

SUIa

ndPO

P(n

�16

)1.

9�

0.5

(16/

16)

2.3

�0.

9(1

6/16

)1.

7�

0.7

(11/

16)

47.

9�

0.6

3.7

�1.

150

.5�

16.8

4.2

�0.

57.

4�

0.8

24.8

�20

61�

7D:

Anal

inco

ntin

ence

and

POP

(n�

8)‡

2�

0.8

(8/8

)2.

9�

1.2

(8/8

)4.

1�

1.7

(8/8

)5

10.9

�1.

95.

98�

2.1

77.6

�18

.94.

5�

0.4

7.95

�0.

827

.5�

5.2

73.2

�10

.4

Note

.—No

patie

ntin

grou

pA,

seve

npa

tient

sin

grou

pB,

12pa

tient

sin

grou

pC,

and

two

patie

nts

ingr

oup

Dha

dSU

I,de

fined

aslo

ssof

urin

eat

max

imum

stra

inin

g,on

dyna

mic

cine

MR

imag

es.U

nles

sot

herw

ise

spec

ified

,dat

aar

em

eans

�st

anda

rdde

viat

ions

,with

num

bers

inpa

rent

hese

sin

dica

ting

inho

wm

any

ofth

eto

taln

umbe

rof

patie

nts

mea

npe

lvic

orga

nde

scen

tw

asca

lcul

ated

.

*As

mea

sure

din

sagi

ttalp

lane

atm

axim

umst

rain

ing.

†In

clud

ing

perit

oneo

cele

,ent

eroc

ele,

rect

ocel

e,an

dan

orec

talj

unct

ion

desc

ent;

data

are

num

bers

ofpa

tient

sw

ithon

eor

mor

eof

thes

eot

her

dysf

unct

ions

.‡

Thre

eof

thes

epa

tient

sal

sore

porte

dSU

Iin

addi

tion

toth

eir

mai

ndi

sord

er.



tients, and to both abnormalities in fivepatients (Tables 3, 4; Fig 4). Urinaryloss did not accompany the bladderneck descent that was detected in sevenpatients; none of these patients re-ported SUI.

Patient group B.—In the patient groupwith SUI but without POP, level III fas-cial anatomic defects in the urethral

supporting structures were detected innine patients (Table 3). Dynamic sagit-tal images revealed a characteristic ver-tical direction of bladder neck move-ment (Fig 5). In one patient, no grossdefects of urethral supporting struc-tures were detected.

Patient group C.—In the patientgroup with SUI and bladder and/or gen-

ital prolapse, on static axial MR images,all patients were seen to have anatomicdefects at the vaginal fascial levels. Forthe urethral supporting system, defectsinvolving the muscular and ligamentousstructures were detected in five of 16patients (Fig 6), whereas eight of 16patients had a level III fascial defect (Ta-ble 3). In three of 16 patients, atypicaldistortion of the ligaments and level IIIfascia was detected.

POP was due to a fascial defect infour patients, to levator muscle weak-ness in six patients, and to both defectsin six patients. SUI was almost alwaysassociated with defects in one of theurethral supporting structures.

Patient group D.—In the patientgroup with anal incontinence and POPwith or without SUI, on static MR im-ages, structural defects of the analsphincter were detected in six patients(Fig 7), three of whom also reportedSUI. These three patients also had ana-tomic defects of the urethral supportingstructures. On images obtained with dy-namic sequences, there was marked le-vator muscle weakness (Table 4). Com-bined analysis of static and dynamic MRimages suggested that marked levatormuscle weakness was the main factorleading to POP in all patients in thisgroup. This weakness was suggested asthe cause of anal incontinence in thetwo patients who had no anal sphincterdefects (Fig 8).

Statistical AnalysisThe mean pelvic organ descent andthe mean measurements of supportingstructures in the control and patientgroups (with standard deviations) arepresented in Tables 1 and 2 and Ta-bles 3 and 4, respectively.

Analysis of variance revealed thatmeasurements of supporting structureswere of highly significant difference(P � .001) between the six groups. De-tailed values of t tests and the exact Pvalues for each measured criterion ofvolunteer subgroup A versus othergroups are presented in Tables 5 and 6.The differences in measurements ofsupporting structures between patientgroups C and D were either significant(P � .05) or highly significant (P �

Figure 4

Figure 4: Static and dynamic MR images in two patients with POP (one 27-year-old woman [a, b] and one54-year-old woman [c, d]). (a, c) Static axial T2-weighted turbo spin-echo images (5000/132). (b, d) Corre-sponding dynamic sagittal balanced fast-field echo images (5/1.6) at maximum straining. In a, a bilateralasymmetric level I paravaginal fascial defect (arrows) is more severe on the right. In b, a sagging levator plate(arrow) and uterine descent (UD) can be seen; combined analysis of static and dynamic imaging findings forthis patient revealed a predominant fascial defect. In c, a bilateral symmetric level I paravaginal fascial defect(arrows) is seen. In d, bladder neck descent (dashed arrow), cystocele (Cy), uterine descent (UD), sigmoido-cele (S), and the excessive sagging of the levator plate (solid arrow) relative to the fascial defect indicate thatlevator muscle weakness is the predominant defect responsible for POP (this patient did not report SUI evenafter cystocele repair).



.001). Measurements for patient groupA were also significantly different, ex-cept for the iliococcygeus angle at rest(P � .056). The measurements for vol-unteer subgroup B and for patient groupB were not significantly different fromthose for volunteer subgroup A, exceptfor the levator plate angle for volunteersubgroup B (P � .0123) and the iliococ-cygeus angle at rest for patient group B(P � .004) (Tables 5 and 6).

The frequency of defects and themost common predominant defects wereas follows: In the four patient groups(n � 44), 34 patients reported POP, andlevator muscle weakness was the pre-dominant defect, being seen in 16(47%) of 34 patients. Level I and II fas-cial defects were seen in seven (21%) of34 patients, and both muscular and fas-cial defects were seen in 11 (32%) of 34patients. Of the 29 women with SUI, 25(86%) had an injury involving one of theurethral supporting structures. Themost frequent injury was a level III fas-cial defect, detected in 19 (76%) ofthese 25 women.

Discussion

In our study, we found a noteworthyrelationship between static and dy-namic MR imaging findings. This rela-tionship provided insight into the asso-ciation between MR imaging findingsand pelvic floor symptoms.

In patients with POP, we were ableto differentiate whether POP was dueto endopelvic fascial defects, tomarked levator muscle weakness, orto both by comparing volunteers insubgroup B and patients in group D.Static and dynamic MR images in vol-unteer subgroup B showed that POPwas caused mainly by defects in level Iand level II endopelvic fascia, as leva-tor muscle weakness was mild in thiscontrol subgroup. In contrast, group Dpatients demonstrated a more ad-vanced degree of muscle weakness rel-ative to the fascial defect; these find-ings indicated that levator muscleweakness, rather than defects in theendopelvic fascia, was the main factorresponsible for POP. The MR imagingfindings in volunteer subgroup B also

indicated that absence of symptoms isnot necessarily associated with ab-sence of structural defects or pelvicorgan descent.

In previous work (10), POP hasbeen attributed to damage to the levatormuscle. DeLancey (18,19), however,described the interaction between pel-vic floor muscles and endopelvic fasciaand maintained that it was not possibleto determine whether damage to muscleor damage to fascia is responsible forprolapse, because these two aspects ofpelvic support are intimately interde-pendent.

Our results demonstrate that SUIis associated with the presence of de-fects in the urethral supporting struc-tures. We found mainly level III en-dopelvic fascial defects on static MR im-ages but no bladder neck hypermobilityon dynamic MR images. These findingswere supported by the fact that SUI wasneither accompanied by bladder neckdescent in seven patients in group A norassociated with the advanced degree ofbladder neck descent encountered ingroup D. In addition, in the latter group,only the three patients who had SUI alsohad defects in the urethral supportingstructures on static MR images. At the

same time, the presence of defects inlevel I and level II endopelvic fascia inpatients with SUI could not be consid-ered a contributing factor to the pa-tients’ symptoms, because the existenceof similar defects in volunteer subgroupB and in group A patients was not ac-companied by SUI.

In agreement with our findings isthe hammock hypothesis suggested byDeLancey (24), who reported that lossof level III endopelvic fascial support atthe vesical neck is one of the factorsresponsible for SUI. Our findings dem-onstrate at imaging what had been hy-pothesized. Furthermore, in more re-cent studies (5,6), static MR imagingrevealed a higher prevalence of lesionsof the urethral supporting system in pa-tients with SUI than in age-matchedstudy control subjects.

In contrast to our findings, Klutkeet al (3), who used static MR imaging,and others (14,16), who used dynamicMR imaging, have suggested urethralhypermobility as one of the causes ofSUI. Yang et al (10), who used onlydynamic MR imaging, did not resolvethe paradoxical lack of urinary inconti-nence in some patients with large cysto-cele because they believed it was un-

Figure 5

Figure 5: Static and dynamic MR images in 48-year-old patient with SUI. (a) Static axial T2-weighted turbospin-echo image (5000/132) at level of proximal urethra shows a level III endopelvic fascial defect, indicatedby the “drooping mustache” sign (arrows), misalignment of the urethra (U), and distorted H-shaped vagina(V). (b) Dynamic sagittal balanced fast-field echo image (5/1.6) shows loss of urine during straining; the blad-der neck movement was noted to be in a vertical (arrow) and not rotational direction. This type of movementwas frequently associated with level III endopelvic fascial defects. SP � symphysis pubis, UB � urinarybladder.



clear whether the cause was kinking ofthe urethra or other factors.

In addition to pinpointing the under-lying structural abnormality of thesedisorders, our approach has severalclinical implications. In SUI, combinedanalysis of static and cine MR imagesrevealed an association between level IIIfascial defects and vertical bladder neckmovement. We believe that identifica-tion of a predominant anatomic defecton static images and its effect on thekinematics of the pelvic organ in dy-

namic sequences could be of value,making it possible in the future to useMR imaging to choose the most appro-priate surgical technique from amongthe available therapeutic options for SUI(34,35). For POP, many types of cysto-cele are currently lumped under thissingle term and can be documented onsagittal dynamic MR images obtained atmaximum straining. However, it is notpossible from these midline images tosee the specific differentiating structuraldefects (27). Therefore, instead of the

clinician focusing on a “dropped blad-der,” combined analysis of both kinds ofMR images can provide the clinicianwith complete mapping of the site andtype of defects. Such information mayhelp the clinician decide on physiother-apy for a patient with global muscleweakness and normal fascia but on sur-gical repair for a patient with a focalbreak in the fascia and/or a muscle tear.Thus, the suggested approach allows in-tegration of static and dynamic MR im-aging findings so that clinicians canmore precisely identify the underlyinganatomic defect responsible for symp-toms in individual patients with pelvicfloor dysfunction (even allowing differ-entiation of the underlying anatomic de-fect when any two patients have thesame symptoms) and can thus decide ona more tailored treatment of the under-lying abnormality.

Our study used phased-array coilsin MR imaging in patients with analincontinence. The different patternsand sites of injuries diagnosed in pa-tients in group D in our study are com-parable to those in previous reports(33,36). In group D patients with analincontinence, static MR imagingshowed no sphincter defect in two pa-tients, whereas dynamic images re-vealed marked muscle weakness in allmembers of this group. Combined anal-ysis of both types of MR images sug-gested that in the absence of an analsphincter defect, anal incontinencecould be associated with marked levatormuscle weakness.

In contrast to other reported find-ings (2,14), we observed the width ofthe levator hiatus and the iliococcygeusangle to be statistically significant in theidentification of pelvic floor laxity. Thus,we recommend that these measure-ments be added to the supporting mea-surements that have been found usefulin the quantification of pelvic organ de-scent (14).

A limitation of our study was theheterogeneity of the patient groups.However, it was important to includepatients with different clinical symp-toms of pelvic dysfunction so that wecould identify differences in the underly-ing anatomic derangement. For the

Figure 6

Figure 6: Static and dynamic MR images in two patients (one 49-year-old woman [a, b] and one 42-year-old woman [c, d]) with both SUI and POP. (a) Static axial T2-weighted turbo spin-echo image (5000/132) ofproximal urethra shows muscular defect, reflected by detachment of the puborectalis muscle (arrows) from thepubic bone on both sides. SP � symphysis pubis, V � vagina. (b) Dynamic axial balanced fast-field echoimage (5/1.6) at maximum straining shows width of levator hiatus to be 7.55 cm (dashed line). (c) Static axialT2-weighted image obtained with same sequence as a shows distortion of periurethral ligaments (arrow-heads) on the right side, with the ligament receding backward compared with the normal left side (arrow). U �urethra. (d) Dynamic coronal image at maximum straining obtained with same sequence as b shows an ilio-coccygeus angle (enclosed between the two dashed lines) of 71.0°.



Figure 7

Figure 7: Static MR images in two patients (one 54-year-old woman [a] and one 33-year-old woman [b, c]) with both anal incontinence and POP. (a) Axial balanced fast-fieldecho image (9/4) of anal sphincter complex. Arrow� scarring of external anal sphincter, arrowhead� scarring of internal anal sphincter. (b) Axial and (c) coronal MR imagesobtained with the same sequence as a show fraying and an anterior defect (�) of the internal anal sphincter with bulging mucosa (white arrow), plus a defect of the external analsphincter (black arrows). The brace in c indicates a defect of the deep part of the external anal sphincter as well as the puborectalis muscle.

Table 5

Comparison of H-Line, M-Line, and Levator Plate Angle Measurements in Volunteer Subgroup A with Those in Volunteer Subgroup Band Patient Groups

Group Compared withVolunteer Subgroup A

H-Line M-Line Levator Plate Anglet Value P Value 95% CI (cm) t Value P Value 95% CI (cm) t Value P Value 95% CI (degrees)

Volunteer subgroup B 0.5488 .592 �1.585, 1.045 0.9289 .369 �2.326, 1.126 2.904 .0123 �40.994, 1.706Patient group

A 2.891 .010 �2.386, �0.094 2.895 .010 �1.630, 0.125 4.520 �.001 �46.095, �11.845B 0.349 .730 �1.296, 0.996 1.918 .072 �2.584, 0.422 0.9720 .344 �23.355, 10.895C 5.399 �.001 �3.139, �1.061 4.720 �.001 �3.774, �1.046 6.677 �.001 �54.340, �23.280D 11.332 �.001 �6.352, �3.928 7.910 �.001 �6.301, �3.119 9.727 �.001 �84.040, �47.820

Note.—Measurements were obtained in the sagittal plane at maximum straining. CI � confidence interval.

Table 6

Comparison of Levator Hiatus and Iliococcygeus Angle Measurements in Volunteer Subgroup A with Those in Volunteer Subgroup Band Patient Groups

Group Compared withVolunteer Subgroup A

Width of Levator Hiatus in Axial Plane Iliococcygeus Angle in Coronal PlaneAt Rest At Maximum Straining At Rest At Maximum Straining

t Value P Value 95% CI (cm) t Value P Value 95% CI (cm) t Value P Value 95% CI (degrees) t Value P Value 95% CI (degrees)

Volunteer subgroup B 0.6292 .540 �0.4870, 0.7870 2.539 .055 �2.873, 0.07342 2.121 .059 �7.118, 0.8183 1.306 .214 �20.254, 6.954Patient group

A 4.427 .0004 �1.475, �0.364 5.679 �.001 �4.014, �1.446 2.433 .056 �6.609, 0.309 5.524 �.001 �36.380, �12.660B 0.5293 .603 �0.44453, �0.665 0.291 .744 �1.144, 1.424 3.282 .004 �7.709, �0.790 0.545 .593 �14.280, 9.440C 4.669 .0001 �1.384, �0.376 6.652 �.001 �4.065, �1.735 3.398 .002 �7.127, �0.852 6.849 �.001 �38.324, �16.815D 5.550 .000056 �1.807, �0.632 6.786 �.001 �4.808, �2.092 4.835 .0002 �10.279, �2.961 8.464 �.001 �52.272, �27.188

Note.—CI � confidence interval.



same reason, the clinical data weremade available to the radiologist whoanalyzed the images: to allow determi-nation of whether combined analysis ofboth types of MR images could detectdifferences in the pathogenesis ofthese clinical problems and to deter-mine whether analyzing both types ofimages together could provide all of

the information necessary for diagno-sis and relevant to treatment. We rec-ognize that the utility of demonstrat-ing the specific abnormality in the pel-vic supporting structures in themanagement and treatment of pelvicfloor dysfunction still must be validated,just as the reproducibility of our pro-posed analytic approach has to be de-

termined. Nevertheless, it is worthmentioning that promising results havebeen reported with respect to the abilityof imaging to change patient care in41.6% (37), 41% (38), and 75% (39) ofpatients with different spectra of pelvicfloor dysfunction.

In conclusion, combined analysisof both static and dynamic MR imagesin patients reporting SUI, POP, andanal incontinence provides comple-mentary information and allows iden-tification of certain structural abnor-malities and their association withspecific pelvic floor dysfunctions. Thisanalytic approach gives insight intothe diagnosis of these complex disor-ders and may also allow for a defect-specific approach to disease manage-ment and surgical technique.

Acknowledgments: Special thanks to ProfessorAhmed Samy, MD, PhD, and the professors inthe Radiology Department at Cairo UniversityHospitals for providing R.F.E.S. the time andfacilities needed to conduct research for thisstudy; to Professor Tahany Elzainy, PhD, for as-sistance in organizing the information in themanuscript; to Amany Elbasmy, MD, PhD, whoensured the accuracy of our statistics; to KarenKinkel, MD, PD, Jurgen Futterer, MD, PhD,Professor Jelle Barentsz, MD, PhD, and Profes-sor Parvati Ramchandani, MD, for their effortsand their review of the manuscript; and to Ka-tharine O’Moore-Klopf, BA, for editorial assis-tance.

References1. Bump RC, Norton PA. Epidemiology and

natural history of pelvic floor dysfunction.Obstet Gynecol Clin North Am 1998;25:723–746.

2. Singh K, Reid WM, Berger LA. Magneticresonance imaging of normal levator anianatomy and function. Obstet Gynecol 2002;99:433–438.

3. Klutke C, Golomb J, Barbaric Z, Raz S. Theanatomy of stress incontinence: magneticresonance imaging of the female bladderneck and urethra. J Urol 1990;143:563–566.

4. DeLancey JO. Fascial and muscular abnor-malities in women with urethral hypermobil-ity and anterior vaginal wall prolapse. Am JObstet Gynecol 2002;187:93–98.

5. Stoker J, Rociu E, Bosch JL, et al. High-resolution endovaginal MR imaging in stressurinary incontinence. Eur Radiol 2003;13:2031–2037.

6. Kim JK, Kim YJ, Choo MS, Cho KS. The

Figure 8

Figure 8: Static and dynamic MR images in 30-year-old patient with anal incontinence and POP but notSUI. (a, b) Static axial T2-weighted turbo spin-echo images (5000/132) show (a) normal urethral (U) support-ing structures and (b) normal level I fascial support (arrows). (c) Static coronal balanced fast-field echo image(9/4) of anal sphincter shows no gross abnormality. EAS � external anal sphincter, PR � puborectalis mus-cle. (d) Dynamic sagittal balanced fast-field echo image (5/1.6) at maximum straining shows excessive levatorplate sagging. The levator plate angle is 88.2°, indicating marked levator muscle weakness (arrow, bladderneck descent). P � peritoneocele, UD � uterine descent. Anterior repair of the puborectalis muscle and theperitoneocele and postanal repair of the levator hiatus were performed, without repair of the bladder neck. Anormal anal sphincter was confirmed during surgery. At the 1-year follow-up examination, this patient did notreport the appearance of masked SUI.



urethra and its supporting structures inwomen with stress urinary incontinence: MRimaging using an endovaginal coil. AJR Am JRoentgenol 2003;180:1037–1044.

7. Huddleston HT, Dunnihoo DR, HuddlestonPM, Meyers PC. Magnetic resonance imag-ing of defects in DeLancey’s vaginal supportlevels I, II, and III. Am J Obstet Gynecol1995;172:1778–1784.

8. Beets-Tan RG, Morren GL, Beets GL, et al.Measurement of anal sphincter muscles: en-doanal US, endoanal MR imaging, or phased-array MR imaging? a study with healthy vol-unteers. Radiology 2001;220:81–89.

9. Laghi A, Iafrate F, Paolantonio P, et al. Mag-netic resonance imaging of the anal canalusing high resolution sequences and phasedarray coil: visualization of anal sphinctercomplex. Radiol Med (Torino) 2002;103:353–359.

10. Yang A, Mostwin JL, Rosenshein NB,Zerhouni EA. Pelvic floor descent in women:dynamic evaluation with fast MR imagingand cinematic display. Radiology 1991;179:25–33.

11. Comiter CV, Vasavada SP, Barbaric ZL,Gousse AE, Raz S. Grading pelvic prolapseand pelvic floor relaxation using dynamicmagnetic resonance imaging. Urology 1999;54:454–457.

12. Goh V, Halligan S, Kaplan G, Healy JC,Bartram CI. Dynamic MRI of the pelvic floorin asymptomatic subjects. AJR Am J Roent-genol 2000;174:661–666.

13. Kelvin FM, Maglinte DD, Hale DS, BensonJT. Female pelvic organ prolapse: a compar-ison of triphasic dynamic MR imaging andtriphasic fluoroscopic cystocolpoproctogra-phy. AJR Am J Roentgenol 2000;174:81–88.

14. Fielding JR. Practical MR imaging of femalepelvic floor weakness. RadioGraphics 2002;22:295–304.

15. Kelvin FM, Pannu HK. Dynamic cystoproc-tography: fluoroscopic and MR techniquesfor evaluating pelvic organ prolapse. In: Bar-tram CI, DeLancey JO, Halligan S, KelvinFM, Stoker J, eds. Imaging pelvic floor dis-orders. New York, NY: Springer, 2003;51–68.

16. Pannu HK, Kaufman HS, Cundiff GW, et al.

Dynamic MR imaging of pelvic organprolapse: spectrum of abnormalities. Radio-Graphics 2000;20:1567–1582.

17. Mengert WF. Mechanics of uterine supportand position. Am J Obstet Gynecol 1936;31:775–782.

18. DeLancey JO. Anatomic aspects of vaginaleversion after hysterectomy. Am J ObstetGynecol 1992;166(6 pt 1):1717–1724.

19. DeLancey JO. Anatomy and biomechanics ofgenital prolapse. Clin Obstet Gynecol 1993;36:897–909.

20. DeLancey JO. Functional anatomy of the pel-vic floor. In: Bartram CI, DeLancey JO, Hal-ligan S, Kelvin FM, Stoker J, eds. Imagingpelvic floor disorders. New York, NY:Springer, 2003; 27–38.

21. Jeffcoate TN, Roberts H. Observation onstress incontinence of urine. Am J ObstetGynecol 1952;64:721–738.

22. Enhorning G. Simultaneous recording of in-travesical and intra-urethral pressure: astudy on urethral closure in normal andstress incontinent women. Acta Chir ScandSuppl 1961;276:1–69.

23. Mostwin JL, Yang A, Sanders R, Genadry R.Radiography, sonography, and magnetic reso-nance imaging for stress urinary incontinence:contributions, uses, and limitations. Urol ClinNorth Am 1995;22:539–549.

24. DeLancey JO. Structural support of the urethraas it relates to stress urinary incontinence: thehammock hypothesis. Am J Obstet Gynecol1994;170:1713–1720.

25. DeLancey JO. Anatomy and physiology ofurinary incontinence. Clin Obstet Gynecol1990;33:298–307.

26. Olsen AL, Smith VJ, Bergstrom JO, CollingJC, Clark AL. Epidemiology of surgicallymanaged pelvic organ prolapse and urinaryincontinence. Obstet Gynecol 1997;89:501–506.

27. DeLancey JO. The hidden epidemic of pelvicfloor dysfunction: achievable goals for im-proved prevention and treatment. Am J Ob-stet Gynecol 2005;192:1488–1495.

28. Petros PP. Appendix 1. Patient question-naires and other diagnostic resource tools.In: The female pelvic floor: function, dys-

function and management according to theintegral theory. New York, NY: Springer,2004; 194–197.

29. Goodrich MA, Webb MI, King BF, et al.Magnetic resonance imaging of pelvic floorrelaxation: dynamic analysis and evaluationof patients before and after surgical repair.Obstet Gynecol 1993;82:883–891.

30. Pescatori M, Anastasio G, Bottini C,Mentasti A. New grading system and scoringfor anal incontinence: evaluation of 335 pa-tients. Dis Colon Rectum 1992;35:482–487.

31. Theofrastous JP, Swift SE. The clinical eval-uation of pelvic floor dysfunction. Obstet Gy-necol Clin North Am 1998;25:783–804.

32. El Sayed RF, Morsy MM, El-Mashed SM,Abdel-Azim MS. Anatomy of the urethralsupporting ligaments defined by dissection,histology, and MRI of female cadavers andMRI of healthy nulliparous women. AJRAm J Roentgenol 2007;189:1145–1157.

33. Rociu E, Stoker J, Zwamborn AW, LamerisJS. Endoanal MR imaging of the anal sphinc-ter in fecal incontinence. RadioGraphics1999;19(spec no):S171–S177.

34. Bergman A, Elia G. Three surgical proceduresfor genuine stress urinary incontinence: 5-yearfollow-up of a prospective randomized study.Am J Obstet Gynecol 1995;173:66–71.

35. Horbach N. Choosing the appropriate sur-gery for genuine stress urinary incontinence.Oper Tech Gynecol Surg 1997;2:1–4.

36. Terra MP, Beets-Tan RH, van Der Hulst VP,et al. Anal sphincter defects in patients withfecal incontinence: endoanal versus externalphased-array MR imaging. Radiology 2005;236:886–895.

37. El Sayed RF, Fielding JR, El Mashed S,Morsy MM, Abd El Azim MS. Preoperativeand postoperative magnetic resonance imag-ing of female pelvic floor dysfunction: corre-lation with clinical findings. J Womens Imag-ing 2005;7:163–180.

38. Kaufman HS, Buller JL, Thompson JR, et al.Dynamic pelvic magnetic resonance imagingand cystocolpoproctography alter surgicalmanagement of pelvic floor disorders. DisColon Rectum 2001;44:1575–1584.

39. Altringer WE, Saclarides TJ, DominguezJM, Brubaker LT, Smith CS. Four-contrastdefecography: pelvic “floor-oscopy.” Dis Co-lon Rectum 1995;38:695–699.



genitourinary imaging pelvic floor dysfunction · 2010. 12. 12. · pelvic floor dysfunction:...

Documents