Dynamic MRI of the Pelvic Floor in Stress Urinary Incontinence: Technique, Normative Data in Continent
Women, and Comparison with Physical Examination
Alan H. Stolpen1, Marc A. Hodroff2, Melody A. Denson2, Lizann Bolinger1, Bridget Zimmerman3, Bruce P. Brown1, Karl
University of Iowa Hospitals and Clinics, Departments of 1Radiology, 2Urology, and 3Biostatistics, Iowa City, Iowa United States;
Introduction Each of the three MRI measurements was recalculated as the
Pelvic floor prolapse and stress urinary incontinence are significant difference between the contracted and straining states (eg, bladder
and costly health problems. The Pelvic Organ Prolapse Quantification descent = bladder height at contraction - bladder height at straining).
(POPQ) system offers a standardized and fairly reproducible method The Table shows the correlation of MRI measurements with POPQ
for evaluating prolapse on physical examination. However, physical staging (values are Mean +/- SEM). The correlation between POPQ
examination may provide an incomplete description of pelvic floor Stage 0 versus Stages I or II and MRI was significant for bladder
abnormalities. Cystourethrography, ultrasound, and (most recently) descent (p=0.010). Also significant was the correlation between
MRI are becoming important tools for assessing pelvic floor prolapse POPQ Stage 0 versus Stage I and levator angle (p<0.0001).
and stress urinary incontinence. Here we describe a new, dynamic There was good interobserver agreement on bladder height
MRI technique for quantifying pelvic floor motion. The MRI measurements, but only fair to poor agreement on levator angle and
technique was applied to a group of continent female volunteers in posterior urethrovesicle angle measurements. For the eight patients
order to learn about its operating characteristics as well as the range of who were restudied after a 3-4 month interval, there was good
"normal" values in continent females. MRI measurements were agreement between the bladder height and levator angle measurements,
correlated with POPQ scores. but poor agreement for posterior urethrovesicle angle measurements.
Fifty-two women, ages 19-67, were recruited for this IRB-approved Our results in continent females suggest that dynamic MRI offers a
study. The main inclusion criterion was urinary continence. Exclusion promising method for quantifying pelvic floor motion in stress urinary
criteria included symptoms of urinary urgency or pelvic floor prolapse incontinence. To our knowledge, this is the first study to correlate
or prior corrective surgery for these symptoms or use of POPQ staging with MRI measurements. Interestingly, subjects with
anticholinergic agents. A history and physical examination were considerable pelvic floor laxity on POPQ and MRI can remain
performed. Pertinent historical data included the number of vaginal continent and asymptomatic. Clearly, imaging alone is insufficient for
and cesarean deliveries, birth weights, prior surgeries, and current managing patients with pelvic floor dysfunction. A limitation of our
medications. An experienced urologist performed a standardized study was an inability to ensure adequate and reproducible contraction
pelvic examination and quantified the degree of pelvic floor prolapse and straining efforts during dynamic MRI. Rectal manometry offers a
on a 0-4 scale (0=no prolapse, 4=gross prolapse) using the POPQ potential solution to this problem. Another limitation was the large
system. Subjects practiced the straining (ie, Valsalva) and contraction interobserver variability for levator angle and posterior urethrovesicle
(ie, Kegel) maneuvers in preparation for the MRI study. angle measurements. Some of this variability may reflect differences
Prior to the MRI, the urinary bladder was catheterized and emptied. in reader experience. A formalized training session for readers and
After exchanging for a 5F catheter, the bladder was filled with 200 cc standardized rules for performing the three measurements may reduce
of 10 mM gadodiamide in normal saline. The vagina and rectum were this variability. A third limitation relates to use of three scalar
each opacified with 8 cc of 12 mM gadodiamide in Trimosam jelly. parameters to describe complex, three-dimensional motions of the
The 5F catheter was flushed with 2-3 cc of gadodiamide-doped jelly to pelvic floor and organs. A more sophisticated engineering approach
increase the conspicuity of the urethra on T1-weighted MR images. may provide a more complete description of these motions.
MRI of the pelvis was performed on a 1.5T GE Signa LX CV/i
scanner equipped with a phased array pelvic multicoil. Imaging
included a localizer, a sagittal T1 spin echo sequence to identify bony
landmarks, a sagittal T2 3D fast spin echo sequence to identify pelvic
floor muscles, and a 20 sec breath-hold T1 sagittal 3D fast spoiled
gradient echo (FSPGR) sequence. The 3D FSPGR sequence was
repeated three times: at rest, during contraction, and during straining.
MRI data was transferred to a GE Advantage Windows workstation
for analysis. A reference line was drawn from the pubic symphysis to
the sacrococcygeal junction. For each set of 3D FSPGR images (ie, at
rest, during contraction, and during straining), the following
measurements were obtained from a midline sagittal section: (1) the
height of the bladder neck above the reference line; (2) the posterior
urethrovesicle angle; and (3) the posterior levator plate angle relative
to the reference line. To study interobserver variability, the studies
were analyzed independently by a urologist and a radiologist. To
study data reproducibility, eight randomly selected subjects from the
original 52 were restudied after a 3-4 month interval. Data were
analyzed using one-way ANOVA, Fisher´s Exact Test, and Sagittal MRI during contraction (left) and straining (right)
multinomial logistic regression.
Results 1. Bland DR, et al. Am J Obstet Gynecol. 181:1324, 1999.
Forty percent of the 52 subjects were parous and 60% were 2. Klute C, et al. J Urol. 143:563, 1990.
nulliparous. Of the parous subjects, 67% had either vaginal deliveries 3. Fielding JR, et al. AJR. 171:1607, 1998.
or cesarean section, whereas 33% had both. 4. Comiter CV, et al. Urology. 54:454, 1999.
On physical examination 56% of the subjects were POPQ Stage 0, 5. Goh V, et al. AJR. 174:661, 2000.
27% were Stage I, and 17% were Stage II. The only item in the POPQ Stage versus MRI Measurements
medical history that correlated with POPQ staging was the number of Bladder Levator Posterior UV
vaginal deliveries: women who had one vaginal delivery had an descent (mm) Angle (deg) angle (deg)
increased risk for being POPQ Stage I (p=0.018, odds ratio 7.43), and
Stage 0 7.2 (1.2) -8.7 (1.8) -8.3 (1.4)
women who had more than one vaginal delivery had an increased risk
for being POPQ Stage II (p=0.013, odds ratio 15.63). Stage I 14.4 (2.0) -20.4 (2.8) -15.1 (4.0)
Stage II 19.9 (3.0) -18.7 (5.4) -12.6 (4.3)
Proc. Intl. Soc. Mag. Reson. Med 9 (2001) 586
Proc. Intl. Soc. Mag. Reson. Med 9 (2001) 586