| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, Duke University Medical Center, Box 3808, Rm. 2526 Blue Zone, S., Durham, NC 27710.
Received June 19, 2000;
accepted after revision October 6, 2000.
Address correspondence to M. A. Kliewer.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. In 358 fetuses, the anterior iliac separation, iliac length, and iliac angle were prospectively measured on antenatal sonography. All measurements were obtained at two axial levels (superior and inferior). The gestational age of the fetus was recorded. The anterior iliac separation was normalized by iliac length, and coefficients of variation were calculated for all measurements. The effects of axial level and gestational age were assessed in a linear regression model. The diagnostic use of the anterior iliac separation relative to that of the iliac angle was assessed in a comparison of 24 fetuses with Down syndrome and 247 non-Down syndrome fetuses.
RESULTS. The anterior iliac separation was less variable than the iliac angle at both superior and inferior levels. There were statistically significant effects for gestational age and axial level on both the anterior iliac separation and the iliac angle, but there was no significant effect for either factor when the anterior iliac separation was normalized by the iliac length. Comparing Down and non-Down syndrome fetuses, we found that the normalized anterior iliac separation had discriminating power similar to the iliac angle.
CONCLUSION. The linear measurement of the anterior iliac separation has diagnostic properties similar to the iliac angle and is subject to less measurement variability. This simpler measurement may be particularly useful when normalized by the iliac length.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
An alternative measurement for the iliac angle that is simpler, more reproducible, and stable would be welcome, if this index has equal or greater diagnostic efficiency for the detection of the dysmorphic features characteristic of the Down syndrome pelvis. In the axial image, the geometry of the fetal pelvis can be abstracted as an isosceles triangle with the iliac wings being the two equal sides (Fig. 1A,1B,1C). The side of the triangle opposite the iliac angle corresponds to the anterior separation of the iliac wings. The anterior iliac separation is, therefore, a simple linear measurement that theoretically contains the same information as the iliac angle itself and can be easily measured by simply positioning two electric calipers on the anterior ends of the iliac bones during the study. To our knowledge, this linear correlate to the iliac angle, however, has not been previously described or studied for prenatal diagnosis.
|
|
|
In an effort to test whether the linear measurement of the anterior iliac separation could serve as a simpler and more elegant surrogate for the iliac angle, we prospectively measured both these indexes in a large consecutive sample of healthy fetuses during prenatal sonography. In addition, we retrospectively compared the anterior iliac separation with the iliac angle in a group of fetuses with Down syndrome to assess the relative diagnostic efficacy of the two indexes. The goals of this study, then, were to characterize the intrinsic sampling variabilities of both the anterior iliac separation and the iliac angle and to compare the diagnostic quality and content of the two measurements in the antenatal diagnosis of Down syndrome.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The examinations were performed with standard sonography equipment (Acuson 128 and XP scanners; Acuson, Mountain View, CA) with 2.5-, 3.5-, or 5-MHz linear, sector, or curved linear electronically focused transducers. The transducer was chosen and positioned to optimize depiction of the fetal pelvis in the axial plane.
The pelvic profile was imaged at a superior level in the upper half of the iliac wing and at an inferior level in the lower half of the iliac wing. At both the superior and inferior axial levels, measurements were made of the length of the iliac wings, the linear separation of the anterior margins of the iliac wings, and the iliac angle (Fig. 1A,1B,1C). The iliac angle was defined as the angle formed by the convergence of lines drawn along the posterolateral aspect of the right and left wings of the ilium [1]. Linear measurements of the lengths of both iliac wings and the separation of the anterior iliac wing margins were made with electronic calipers by 10 sonographers who performed the obstetric studies. For iliac-length measurements, calipers were placed at the anterior and posterior limits of the bone (Fig. 1B). For the anterior iliac-separation measurements, calipers were placed at the lateral margins of the anterior ends of the ilium to locate two of the vertices of an idealized isosceles triangle (Figs. 1A and 1C). The iliac angle was measured by a single reviewer using a handheld goniometer on hard-copy images. A single measurement of each parameter was made.
Descriptive statistics were generated for the iliac angle, anterior iliac separation, and iliac length. In addition, the anterior iliac separation was divided by the length of the right iliac wing to produce a normalized index that could be relatively independent of growth changes over gestation. Means and standard deviations were calculated for six gestational age ranges. Coefficients of variation were calculated by dividing the standard deviation by the mean and then multiplying the result by 100. The correlation between anterior iliac separation and the iliac angle was characterized with Pearson's product-moment correlation coefficient. A linear regression model was constructed to account for the effects of axial level and best estimated gestational age on the anterior iliac separation, the iliac length, and normalized anterior iliac separation measurements. Test results were considered statistically significant at p values of 0.05 or less.
Comparison of Indexes in Fetuses with Trisomy 21
To compare the diagnostic efficacies of the anterior iliac separation and
the iliac angle, the records and results of all amniocenteses performed in our
laboratory from 1991 to 1997 were examined for fetuses identified with trisomy
21. All genetic amniocenteses were performed between 15 and 25 menstrual
weeks, determined with sonography performed at the time of the procedure.
A total of 28 fetuses with trisomy 21 were identified, and sonograms in these pregnancies were obtained. Images of the fetal pelvis, an area routinely documented on obstetric sonography in our laboratory, were selected. Axial images of the iliac wings were found in 24 of the 28 studies. All fetuses had been studied between 15 and 25 menstrual weeks. In these 24 fetuses, the iliac angle was measured with a handheld goniometer, and the anterior iliac separation and iliac length were measured by a single reviewer with a handheld caliper. To generate an age-matched comparison group, all fetuses in the 15-25 weeks gestational age range were selected from the prospective database. The non-Down syndrome group comprised 247 fetuses.
For the subsets of Down and non-Down syndrome fetuses, descriptive statistics were calculated for the iliac angle, anterior iliac separation, iliac length, and anterior iliac separation normalized by iliac length. Linear regression models were developed to test the effects of axial level and best estimated gestational age on anterior iliac separation, normalized anterior iliac separation, and iliac angle for both the non-Down and Down syndrome groups. We then compared the four indexes as measures of discrimination for Down syndrome, using approximate Student's t tests; the mean differences between the Down and non-Down syndrome groups were divided by the estimated standard errors of the difference to provide a ratio of the mean morphometric difference to the background variability of the index (similar to a signal-to-noise ratio). Mann-Whitney estimates of the area under the receiver operating characteristic curve were generated for the normalized anterior iliac separation and the iliac angle, and these areas were compared for statistically significant differences. A table of sensitivity and specificity values was produced from the receiver operating characteristic analysis for the normalized anterior iliac separation and iliac angle parameters. Using the conditional probabilities from the sensitivity and specificity estimates and assuming a prevalence of Down syndrome in the general population of one in 800 [9], we calculated the odds ratios and likelihood ratios at specific values of these parameters. Test results were considered statistically significant at p values of 0.05 or less.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The changes in the anterior iliac separation and the iliac angle over gestation are displayed in Table 1. The anterior iliac separation increases with gestational age as the pelvis grows and develops. The iliac angle tends to decrease slightly. The correlation coefficient for the anterior iliac separation and the iliac angle was 0.47 (p < 0.001). The anterior iliac separation was less variable than the iliac angle at both the superior and inferior levels. At the superior level, the coefficients of variation ranged from 13.0% to 16.0% for the anterior iliac separation and from 18% to 23% for the iliac angle. Both the anterior iliac separation and the iliac angle measurements were smaller and more variable at inferior axial levels. The anterior iliac separation at the superior level showed the least variability of the two indexes measured at the two levels.
|
The effects of best estimated gestational age and axial level were statistically significant for the measurements of the anterior iliac separation and the iliac angle. After we adjusted for the effects of best estimated gestational age, the anterior iliac separation was an estimated 5.5 mm shorter at the inferior level than at the superior level (p = 0.0001). After adjusting for axial level, the anterior iliac separation increases by 1.8 mm per week over gestation (p = 0.0001). Similarly, the iliac angle is 15.7° less at the inferior level compared with the superior level (after adjusting for best estimated gestational age) (p < 0.001) and tends to decrease by an estimated 0.37° per week (after adjusting for axial level, p < 0.001).
When the anterior iliac-separation measurement is divided by the iliac-length measurement, the resulting ratio is a normalized variable that is not significantly affected by axial level (p = 0.41) or best estimated gestational age (p = 0.13).
Comparison of Indexes in Fetuses with Trisomy 21
The means and standard deviations for the anterior iliac separation and
iliac angle are given for the Down and non-Down syndrome groups in
Table 2. The anterior iliac
separation was an average 2.3 mm shorter in Down syndrome fetuses, and the
iliac angle was an average 10° greater in Down syndrome fetuses. When
these mean differences between the Down and non-Down syndrome groups were
divided by the estimated standard errors of the difference, we found that the
resulting t statistic was greatest for the normalized anterior iliac
separation and the iliac angle indexes (3.4 and 3.1, respectively) and was
least for the anterior iliac separation and iliac-length indexes (1.8 and 0.1,
respectively). This conclusion indicates that the mean differences between the
Down and non-Down syndrome groups for the normalized anterior iliac separation
and the iliac angle are more than three times the background variability for
these indexes. Therefore, these indexes are the most promising measures of
discrimination and have similar discriminatory value.
|
Mann-Whitney estimates of the areas under the receiver operating characteristic curve were 0.74 for the normalized anterior iliac-separation measure and 0.67 for the iliac angle. There was no statistically significant difference in these areas (p = 0.75). Sensitivities and specificities for five levels of the normalized anterior iliac separation index and for six levels of the iliac angle index are provided in Table 3. Using an estimated prevalence for trisomy 21 in one of 800 live births, we calculated the odds ratio and the likelihood of a Down syndrome birth for each level of the two indexes. A normalized anterior iliac-separation value of 2.0 produces a likelihood ratio of one in 400, which is comparable to the risk level (one in 250) generally considered to indicate the need for amniocentesis [10].
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The geometry of the fetal pelvis indicates that the iliac angle and the
anterior iliac separation should contain the same morphometric information. In
an idealized form, the fetal pelvis is an isosceles triangle (Fig.
1A,1B,1C).
The mathematic relationship between the iliac length and the anterior iliac
separation can be expressed mathematically with a sine function:
 |
 |
Although our study indicates that the normalized anterior iliac separation has discriminatory power similar to the iliac angle, the comparative analysis of the Down and non-Down syndrome fetuses is limited by the fact that the measurements of the fetuses with Down syndrome were made retrospectively on images not specifically generated for morphometry of the fetal pelvis. As a result, the measurements of length and angle were likely made at various axial levels. An earlier study by our group has shown that iliac length and angle measurements are strongly influenced by the gestational age of the fetus, axial level of measurement, and the orientation of the fetal spine relative to the transducer [7]. These effects may partly explain the paradoxical finding that the iliac angle is wider than normal in Down syndrome fetuses, yet the anterior iliac separation tends to be shorter than normal (Table 3). Moreover, we did not examine the potential effects of intraobserver and interobserver variability in this study, and these factors may be significant when small measurements are made in the clinical setting. A larger prospective comparison of Down and non-Down syndrome cases, therefore, would be necessary to fully examine the clinical effectiveness of these parameters for distinguishing the two populations.
Nonetheless, our preliminary receiver operating characteristic analysis suggests what levels might be most clinically useful for discriminating in Down and non-Down syndrome fetuses in future prospective studies. Specifically, a level of 2.0 for the normalized anterior iliac separation yields a sensitivity of 46% and a specificity of 77% and a likelihood of one in 400 that a fetus would have Down syndrome, assuming a prevalence of Down syndrome in the general population of one in 800 fetuses [9]. This likelihood is similar to the generally accepted risk of amniocentesis [10]. For screening populations with a higher prevalence of Down syndrome fetuses, a higher sensitivity level may be desirable.
In summary, this study shows that the anterior iliac separation is strongly correlated with the iliac angle and may provide an alternative method of distinguishing fetuses with trisomy 21 from fetuses without trisomy 21. The particular value of the anterior iliac separation is that it represents a simple linear measure that can be easily obtained with electronic calipers on the monitor. This linear index has similar diagnostic properties to the iliac angle and is less variable. Normalizing this index to the iliac length provides the additional benefit that the linear measurement is no longer strongly dependent on best estimated gestational age or axial level. The role of fetal pelvic morphometry in the diagnosis of Down syndrome remains unresolved, though it appears that it has the potential to serve as an adjunctive and supportive piece of evidence in a larger context of findings used to diagnose Down syndrome in fetuses.
Acknowledgments
We thank Susan Murray for assistance with manuscript preparation.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
is measured at convergence
of lines drawn tangent to posterolateral margins of right and left wings of
ilia. Anterior iliac separation is chord of iliac angle. Therefore, these two
measurements are intimately related, both conceptually and mathematically.