| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, New York University Medical Center, 560 First Ave.,
Ste. HW 202, New York, NY 10016.
2 Department of Urology, New York University Medical Center, New York, NY.
Received November 9, 2007;
accepted after revision February 11, 2008.
Address correspondence to E. M. Hecht
(hechte01{at}med.nyu.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-nine women (mean age, 57 years) with suspected pelvic floor dysfunction underwent MRI using both a sagittal true FISP sequence, acquired continuously during rest alternating with the Valsalva maneuver, and a sagittal HASTE sequence, acquired sequentially at rest and at maximal strain. Data sets were evaluated in random order by two radiologists in consensus using the pubococcygeal line (PCL) as a reference. Measurement of prolapse was based on a numeric grading system indicating severity as follows: no prolapse, 0; mild, 1; moderate, 2; or severe, 3. A comparison between sequences on a per-patient basis was performed using a Wilcoxon's analysis with p < 0.05 considered significant.
RESULTS. Overall, 66.1% (39/59) of patients had more severe prolapse
(
1°) based on dynamic true FISP images, with 28.8% (17/59) of the
cases of prolapse seen exclusively on true FISP images. Only 20.3% (12/59) of
patients had greater degrees of prolapse on HASTE images than on true FISP
images, with 10.2% (6/59) of the cases seen exclusively on HASTE images. A
statistically significant increase in the severity of cystoceles (p
< 0.01) and urethral hypermobility (p < 0.01)—with a
trend toward more severe urethroceles (p < 0.07), vaginal prolapse
(p < 0.09), and rectal descent (p < 0.06)—was
shown on true FISP images.
CONCLUSION. Overall, greater degrees of organ prolapse in all three compartments were found with a dynamic true FISP sequence compared with a sequential HASTE sequence. Near real-time continuous imaging with a dynamic true FISP sequence should be included in MR protocols to evaluate pelvic floor dysfunction in addition to dynamic multiplanar HASTE sequences.
Keywords: MRI • pelvic floor dysfunction • pelvic organ prolapse • women's imaging
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Several imaging techniques may be used to supplement physical examination findings in patients with suspected pelvic floor dysfunction. Traditionally, fluoroscopic methods, including cystourethrography and evacuation proctography, have formed the cornerstone of the imaging evaluation. However, these techniques require ionizing radiation, are uncomfortable and time consuming for patients, and can assess only one compartment at a time [5]. Sonography is operator dependent, has poor soft-tissue contrast, has a limited field of view, and can be difficult to perform dynamically for the evaluation of changes with maneuvers such as the Valsalva maneuver.
Dynamic MRI offers a noninvasive alternative with no ionizing radiation and has several advantages for the evaluation of pelvic floor disorders including a large field of view, soft-tissue contrast superior to that of sonography, direct multiplanar capability, and high temporal resolution. Initially, T1 gradient-echo sequences were used, but acquisition times were relatively long at 6–12 seconds and soft-tissue contrast was poor [8, 9]. With advances in MR technology, including faster and stronger gradients, HASTE imaging was introduced allowing more rapid acquisition time (0.5–2.5 seconds per image) with excellent soft-tissue contrast. However, the HASTE approach requires 1–2 seconds between acquisitions to allow T1 recovery; therefore, real-time imaging is not possible [10–12]. Steady-state free precession gradient-echo imaging provides an alternative T2-like imaging contrast (T2- and T1-weighted imaging) with robust signal and a rapid acquisition time of less than 1 second, thereby permitting near real-time continuous imaging [13].
The purpose of this study was to compare two MR sequences, true FISP and HASTE, for the evaluation of pelvic floor dysfunction. To our knowledge, no study comparing these two sequences in the same patient population has previously been performed.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI
All patients underwent pelvic floor MRI examinations using a 1.5-T magnet
(Vision, Symphony, or Avanto, Siemens Medical Solutions) and a torso
phased-array coil. In accordance with the routine protocol at our institution,
patients did not undergo bowel preparation and no intraluminal contrast
material was administered. Each patient was asked to empty her bladder before
entering the magnet, which typically resulted in the bladder being half full
at the time of image acquisition. Patients were informed by the technologist
that they would be asked to "bear down or strain like you are having a
bowel movement" and then to relax. This was practiced by the patient
before imaging, and instructions were given by the technologist before
performing each individual sequence. If the technologist did not observe any
change in the pelvic floor during or after image acquisition due to suspected
suboptimal patient effort, the sequence was repeated.
Initially, axial, sagittal, and coronal T2-weighted turbo spin-echo (TSE) images were acquired at rest with imaging planes orthogonal to the orientation of the uterus. These images were acquired to evaluate for coexisting pelvic abnormalities and localize anatomic structures for optimal positioning during the dynamic sequences; however, these images were not evaluated as part of this study. Subsequently, based on the static T2-weighted TSE images, optimal midline sagittal or sagittal oblique planes were selected for the true FISP and HASTE sequences to include the pubic symphysis, urethra, vagina, rectum, and coccyx in the field of view accounting for anatomic asymmetry in individual patients. Patients underwent dynamic imaging for both sequences both at rest and at maximal strain.
The midline sagittal or sagittal oblique dynamic true FISP sequence was performed as a continuous acquisition with the patient alternating every 5 seconds between rest and maximal strain. The sequence parameters were a TR/TE of 3.9/1.9; refocusing flip angle, 70°; matrix, 256; rectangular field of view optimized to the patient's body habitus, 300–350 mm; slice thickness, 8 mm; bandwidth, 673 Hz per pixel; acquisition time per measure, 0.6 second; 90 consecutive measures; and acquisition time, 54 seconds.
The sagittal HASTE sequence was performed using five slice positions, one
midline and four parasagittal, with seven measurements per slice position
alternating between rest and maximal strain (total = 35 slices including 7
measurements at 5 slice positions). The following parameters were used:
TR/TEeff, 
/64; refocusing flip angle, 140°; matrix, 256;
rectangular field of view, 300 mm; acquisition time, 2.5 seconds per slice;
slice thickness, 5 mm; and bandwidth, 781 Hz per pixel.
In accordance with our standard departmental protocol, the HASTE sequences were acquired before the true FISP sequences. This order was used to standardize the examination for technologists in terms of workflow so that the order of acquisitions was not randomized prospectively. However, in 10 cases due to technical inadequacy of the initial HASTE scans, the HASTE acquisition was repeated after the true FISP acquisition, and the HASTE images from the second acquisition were used in the analysis. A statistical analysis of the importance of sequence order was performed.
Image Analysis
Images from each data set—HASTE or true FISP—were reviewed
retrospectively by two experienced radiologists, both with more than 4 years'
experience, in consensus on a PACS workstation; the reviewers were blinded to
patient identity. Images were evaluated in random order with respect to HASTE
or true FISP. Images were evaluated during two reading sessions separated by 3
weeks so that no individual patient's two data sets were reviewed during the
same reading session. For each technique, the optimal midline sagittal
sequence and image that revealed the greatest degree of organ prolapse were
selected for analysis. If the sagittal HASTE or true FISP sequence was
repeated, the interpreting radiologists reviewed both data sets and used the
data set that showed the greatest degree of prolapse. The order of acquisition
of the true FISP and HASTE sequences was recorded during a later session
separate from the image interpretation.
In addition, during a separate reading session more than 6 months after the first two reading sessions, one author recorded the time interval between acquisition of the HASTE and true FISP sequences and performed bidimensional bladder and cross-product measurements on the midline rest images of both sequences in a similar fashion. To ensure comparable measurements, the reviewer placed images side by side on a PACS station for comparison and to perform measurements.
Grading of Prolapse
The presence and degree of pelvic organ prolapse at maximal strain were
graded using the pubococcygeal line (PCL) as the reference standard
[7,
9,
13–15].
The PCL is defined as a line extending from the inferior margin of the pubic
symphysis to the last joint of the coccyx. Grading of cystocele; urethrocele;
vaginal, cervical, and uterine body prolapse; and rectal descent was based on
the distance of a perpendicular line drawn from the PCL to the inferior margin
of the organ of interest as follows: negative, < 1 cm; mild, 1 to < 2
cm; moderate, 2–4 cm; or severe, > 4 cm. A rectocele was defined as a
bulging of the anterior rectal wall greater than 2 cm from a line drawn
parallel to the center of the anal canal and was graded as mild, 2–4 cm;
moderate, > 4–6 cm; or severe, > 6 cm. There is no standard
grading system in the literature for urethral hypermobility; therefore, we
derived our grading system from a study in which the investigators found that
rotation of the urethra of > 30° from its resting axis during straining
indicates urethral hypermobility
[13]. The angle between a
centerline drawn parallel to the long axis of the patient's urethra at rest
and a centerline drawn parallel to the long axis of the urethra at maximal
strain were measured.
The severity of hypermobility was graded as follows: mild, < 45°; moderate, 45–90°; or severe, > 90°. The levator plate angle was also measured for both sequences by measuring the angle between the PCL and levator plate, which represents the posterior confluence of the iliococcygeal muscles anterior to the coccyx. The normal orientation of the levator plate is parallel to the PCL. A levator plate angle of > 20° was considered significiant pelvic laxity. Additional findings such as peritoneocele or enterocele were also noted. All measurements were recorded in centimeters or degrees.
Statistical Analysis
For analysis, organ prolapse measurements, as defined earlier, were given
numeric grades to indicate severity using 0–3 as follows: 0, no
prolapse; 1, mild prolapse; 2, moderate prolapse; 3, or severe prolapse. The
degrees of organ prolapse were compared between sequences on a per-patient
basis using a Wilcoxon's matched-pairs signed rank test. The number of
patients with an increase of at least one grade of prolapse based on either
sequence was tabulated. A Mann-Whitney test was performed to determine whether
any differences in the results between sequences could be attributed to the
order in which the sequences were acquired.
Three potentially key factors—patient age, bladder volume change over
time, and time interval between sequence acquisitions—can lead to over-
or underestimation of organ prolapse and thereby affect the results of a
comparison between the two sequences. The impact of these factors was assessed
by first defining cutoff values for the numeric factors. For patient age, the
cutoff was set to 60 years: 31 patients were < 60 years and 28, 60
years. For bladder volume change, the cutoff was set to a 20% increase in
bladder volume: 30 patients showed a < 20% volume increase over time
between sequences and 29 patients showed a 20% bladder volume increase
over time between sequences. For the time interval between sagittal HASTE and
true FISP sequences, a cutoff was set at 10 minutes: 34 patients were imaged
with < 10 minutes between sequences and 25 with 10 minutes between
sequences. These cutoff values were chosen to divide the population relatively
evenly.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
Only 20.3% (12/59) of patients had greater degrees of prolapse shown on the HASTE sequence. Among these 12 patients, 8.3% (1/12) showed anterior or 25.0% (3/12) middle compartment prolapse only, and 66.7% (8/12) had multicompartment prolapse (anterior and middle [n = 4], anterior and posterior [n = 3], or middle and posterior [n = 1] compartments) (Figs. 3A and 3B). Overall, 10.2% of patients (6/59) had compartmental prolapse shown on the HASTE sequence alone (Figs. 3A and 3B). The degree of prolapse in most of these cases was mild and involved the anterior and middle compartments only. These cases included mild urethral hypermobility (n = 1), mild cervical prolapse (n = 3), mild uterine prolapse (n = 1), and moderate uterine prolapse (n = 1). The two enteroceles that were depicted on the true FISP sequence were not depicted on the HASTE imaging data sets.
|
|
A statistically significant increase in the severity of cystoceles (p < 0.01) and urethral hypermobility (p < 0.01) (Figs. 4A and 4B), with a trend toward higher-grade urethroceles (p < 0.07), vaginal prolapse (p < 0.093), and rectal descent (p < 0.064), was shown on the true FISP sequence (Table 2). The levator plate angle (range, 0–101°) was greater on true FISP images in 13.6% (8/59) of patients.
|
|
|
Impact of Sequence Order
In most patients (n = 49), the true FISP sequence was performed
after the HASTE sequence in accordance with departmental protocol, as
discussed earlier. However, in the remaining 10 patients the order of
sequences was changed due to repetition of the HASTE sequence after the true
FISP sequence because the initial HASTE acquisitions were considered
inadequate and needed to be repeated. Although not part of the study design,
this second HASTE sequence permitted a more thorough evaluation of order and
sequence effects. The order in which the sequences were performed was not
observed to have a significant effect (p > 0.27) on any measures
derived using either sequence except urethral hypermobility and levator plate
angle, each of which is discussed in detail later in this article.
In terms of urethral hypermobility, the degree of hypermobility recorded for true FISP and HASTE images was significantly lower (p < 0.01) when either sequence was performed second. Nonetheless, with regard to performing the true FISP sequence or the HASTE sequence before or after the other, there was no statistical impact on the observation that urethral hypermobility was greater on true FISP images than on HASTE images (p = 0.03) irrespective of the order of sequence acquisition. When the true FISP sequence was performed first, the mean hypermobility was 0.16 units higher on true FISP than HASTE. This difference was not statistically significant (p = 0.64) from the corresponding difference when true FISP was performed second: The mean hypermobility was 0.10 units higher on true FISP than HASTE.
The mean levator plate angle determined using true FISP (mean = 39.2°) was significantly higher (p = 0.002) than that determined using HASTE (mean = 33.3°) when the true FISP sequence was performed first, but the levator plate angles derived using the two sequences were quite similar (mean = 20.9° for FISP, 21.8° for HASTE; p for difference = 0.27) when true FISP was performed last. Thus, the order in which the sequences were performed had a larger impact on the levator plate angle measures determined using true FISP than on those determined using HASTE.
There was a significant (p < 0.001) but weak (r = 0.4) correlation between change in bladder volume and time interval between acquisitions of the HASTE and true FISP sequences. Overall, the mean fractional increase in bladder volume between acquisition of the sequence performed earlier and the one performed later was 29% with an SD of 38%, but individual variability was high in this parameter. Using a cutoff of a 20% increase in bladder volume between sagittal HASTE and true FISP sequences, there was no significant difference in the incidence of finding greater degrees of cystocele on true FISP compared with HASTE (chi-square test: p = 0.34 for cystocele, p = 0.49 for urethrocele, and p = 0.29 for urethral hypermobility).
The mean time interval between sequences was 10 minutes 10 seconds with an SD of 5 minutes 12 seconds. Using 10 minutes as a cutoff, there was no significant difference in the incidence of finding greater degrees of cystocele on true FISP compared with HASTE (chi-square test: p = 0.91 for cystocele, p = 0.42 for urethrocele, and p = 0.41 for urethral hypermobility). The mean patient age was 57 years with an SD of 15 years. Using an age cutoff of 60 years, there was no significant difference in the incidence of finding greater degrees of cystocele on true FISP compared with HASTE (chi-square test: p = 0.54 for cystocele, p = 0.56 for urethrocele, and p = 0.82 for urethral hypermobility).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Overall, our results show that imaging using the true FISP sequence identified more instances of organ prolapse and greater degrees of prolapse than imaging using the HASTE sequence. This difference was greatest in the anterior compartment, with a statistically significant increase in the degree of bladder prolapse and urethral hypermobility. In addition, in 28.8% of patients, prolapse was seen only on the true FISP images. Of the 19 cases of prolapse seen exclusively on true FISP images, 18 were mild in degree. Of the six cases of prolapse seen exclusively on the HASTE sequence, five were mild in degree and one was moderate, and these cases predominately involved the middle compartment.
In any MR protocol evaluating organ prolapse, the patient needs to achieve maximal straining effort to ensure diagnostic accuracy. We postulate that the near real-time imaging that can be achieved with a true FISP sequence most likely accounts for the differences between the sequences that we observed in this study. Although the HASTE sequence offers superior contrast resolution and has a high temporal resolution, its temporal resolution is inherently limited because a brief pause is needed between acquisitions to permit T1 recovery. The true FISP sequence is robust in signal because of its combination of T2 and T1 weighting; however, T2 tissue contrast is relatively reduced on true FISP images compared with HASTE images because of the introduction of T1 weighting. Unlike HASTE, there is no need to pause for T1 recovery when using the true FISP sequence. Therefore, true FISP imaging offers faster acquisition times that permit continuous image acquisition with the patient alternating between resting and straining, allowing patients to achieve a greater degree of effort during straining. During the static sequential acquisitions of the HASTE sequence, patients may not be able to sustain maximal strain for the entire duration of scanning, and there is a margin for error such that there is potential for mismatch between timing of the acquisition and the actual point at which the patient achieved maximal strain.
Although our data indicate that the true FISP sequence reveals greater degrees of prolapse, there are recognized limitations of our study. This study was retrospective, and in most of the patients, the true FISP sequence was performed after the HASTE sequence. Although this protocol could introduce a learning bias such that the effort of straining was greater during the true FISP sequence because the patient had "learned" how to more effectively strain during acquisition of the HASTE sequence, our results indicate that the order of sequence acquisition did not impact the results. The readers could not be blinded to the sequence being evaluated because each is readily recognizable by an experienced reader, and this may have also introduced some bias. In addition, we limited our assessment to evaluation of midline sagittal images because this imaging plane is generally accepted as the most important one for the diagnosis of pelvic floor disorders.
We did not evaluate potential advantages of the HASTE sequence in terms of improved spatial resolution relative to the true FISP sequence. When identifying and grading prolapse, it is important to identify certain anatomic landmarks, as described earlier, which may be facilitated by the improved soft-tissue contrast of the HASTE sequence. Interestingly, the small number of cases in which the HASTE sequence revealed prolapse not seen on true FISP images involved the middle compartment.
The degree of strain achieved, fatigue, and effects of bladder volume are important to consider when determining the extent of organ prolapse, and we acknowledge that they are potentially confounding effects given that the sagittal HASTE and true FISP sequence acquisitions were not randomized in this retrospective study. A distended bladder, for example, can inhibit the degree of strain and descent of prolapse such that it is generally recommended that the bladder be emptied before dynamic pelvic floor MRI examination [16–18].
Patient effort, compliance, and fatigue may also have contributed to differences in organ prolapse detected between sequences. For example, over time patients may have become tired and failed to strain sufficiently; on the other hand, after repetitive strain maneuvers, the pelvic floor may have grown weaker, leading to greater degrees of organ prolapse on the sequence performed at a later time. In this study, the time interval between the HASTE and true FISP sequences varied, with the true FISP sequence being more frequently performed after the HASTE sequence; thus, over time increased bladder volume could have led to false-negatives on the true FISP images compared with the HASTE images. However, dividing the population approximately equally according to, first, degree of bladder volume increase; second, temporal separation between sequences, as a surrogate of fatigue; or, third, patient age, as a putative index of compliance, failed to show statistically significant differences in the relative advantage of true FISP over HASTE. Overall, greater degrees of prolapse were found on true FISP images compared with HASTE images, particularly in the anterior compartment.
We did occasionally observe, however, that some patients were shown to have greater degrees of prolapse on the HASTE sequence than on the true FISP sequence. In fact, 10.2% of the cases of prolapse were diagnosed exclusively on HASTE. To some degree, the strain effort likely contributed to the random differences among the two different sequences and may suggest why a small subset of cases showed greater degrees of prolapse on the HASTE sequence. Given the inferior soft-tissue contrast on true FISP images, the lack of clear distinction between the cervix, vagina, and uterus may possibly limit accurate diagnosis on true FISP images. However, on review of the six cases in which the HASTE sequence revealed prolapse not seen on the true FISP sequence, this was not the case. The degree of strain was more likely the contributing factor to discordance in those specific cases.
In dynamic pelvic floor imaging, one attempts to approach physiologic conditions to reproduce a patient's symptoms and provoke pelvic floor dysfunction. Some investigators argue that upright positioning in the magnet might even be preferable to supine positioning [19], but upright magnets are of limited field strength and traditionally pelvic floor muscle function and strength are measured with the patient in the supine position. Although HASTE imaging is feasible at a low field strength, true FISP imaging would not be achievable at a magnetic field strength of less than 1.0 T, and the signal-to-noise ratio of high-field-strength imaging in a closed system is preferable for optimal image quality.
Although investigators of previous studies in the literature have justified the use of dynamic MRI for the evaluation of pelvic organ prolapse using both of these sequences, we recognize that a potential limitation of this study is the lack of clinical correlation. Among the 47 patients with a discrepancy between diagnosis or grade of prolapse based on true FISP or HASTE, only 26 had charts available for review. Charts were reviewed for physical examination findings, but many records were insufficient to compare with MRI findings. Because of the limited clinical data and the retrospective nature of the study, it was not feasible to assess the impact of our findings on changes in patient management and outcome.
To assess the impact of dynamic MRI of the pelvic floor using one protocol or another, a long-term prospective study is required. This, however, was not the goal of this study. Our purpose was to compare two commonly used imaging sequences for dynamic pelvic floor imaging in the same patient population to determine whether there was any difference in the degree of prolapse observed. Further prospective studies are warranted to determine the overall clinical impact of such imaging techniques.
In conclusion, we found greater degrees of organ prolapse in all three compartments with a dynamic true FISP sequence than with a sequential HASTE sequence. Near real-time continuous imaging with dynamic true FISP in the midline sagittal plane requires less than 2 minutes of examination time and should be included in MR protocols to evaluate pelvic floor dysfunction in addition to dynamic multiplanar HASTE sequences.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
